Tumi Bjornsson Ph.D

Regional scale modelling of the lower River Murray

wetlands

A model for the assessment of nutrient retention of floodplain

wetlands pre- and post-management

Dipl. Ökol. Kjartan Tumi Björnsson B.Sc.

Thesis Submitted for the Degree of

Doctor of Philosophy

June 2007

School of Earth and Environmental Sciences

Regional Scale Modelling of the lower River Murray wetlands

II


I

Table of Contents

Table of Contents ....................................................................................................... I

List of Figures and Tables .........................................................................................V

Figures ..................................................................................................................V

Tables .................................................................................................................. IX

Declaration .............................................................................................................. XI

Acknowledgements ................................................................................................ XII

Abstract ................................................................................................................ XIV

1 Background ....................................................................................................... 1

1.1 Introduction ............................................................................................... 1

1.1.1 Wetland processes .............................................................................. 5

1.1.2 Spatial relationships of wetlands to transport processes .................... 10

1.2 Degradation of floodplain wetlands .......................................................... 12

1.2.1 Eutrophication of aquatic environments ............................................ 14

1.2.2 Alternate stable states and permanent inundation impacts on wetlands

17

1.2.3 Irrigation drainage and constructed wetlands .................................... 19

1.3 Restoration of degraded floodplain wetlands ............................................ 21

1.3.1 Management strategies for restoration .............................................. 21

1.4 Predictive modelling of wetland processes and services; current state and

potential alteration due to management ................................................................ 23

1.4.1 Complexity and feasibility of modelling ........................................... 25

1.4.2 Qualitative and quantitative assessment of model accuracy and generic

applicability ..................................................................................................... 27

1.4.3 Validation ........................................................................................ 30

1.4.4 Modelling role in environmental decision-making ............................ 31


II

2 Aims and objectives ......................................................................................... 39

3 Materials and Methods..................................................................................... 41

3.1 Model Description ................................................................................... 41

3.1.1 Design Considerations ...................................................................... 41

3.1.2 WETMOD 1 .................................................................................... 45

3.1.3 WETMOD 2 .................................................................................... 54

3.2 Data: Model Driving Variables ................................................................ 63

3.2.1 “Exemplar” Wetland Sites ............................................................... 65

3.2.2 Wetland Data ................................................................................... 71

3.2.3 River Data ........................................................................................ 80

3.3 Data Handling .......................................................................................... 85

3.3.1 Model Calibration ............................................................................ 88

3.3.2 Validation Procedure ........................................................................ 88

3.4 Wetland Management .............................................................................. 89

3.4.1 Options ............................................................................................ 89

3.4.2 Management scenarios for cumulative assessment ............................ 92

4 Validation of the model WETMOD 2 and Discussion ...................................... 97

4.1 Fitting and Validation based on calibrated (“exemplar”) wetlands ............ 97

4.1.1 Implication for irrigation affected wetland representation ................118

4.1.2 Implication for wetland representation.............................................120

4.2 Validation based on non-calibrated wetland data .....................................125

4.3 Evaluating model performance ................................................................136

4.3.1 Generic nature and structural restrictions of model ..........................136

4.3.2 Relevance of project objectives .......................................................137

4.4 Chapter summary and Implication for the first hypothesis .......................139

5 Simulation results of potential management scenarios and Discussion ............140


III

5.1.1 Implications for Management ..........................................................157

5.2 Chapter summary and Implications for the second hypothesis .................161

6 Results of the cumulative assessment of management scenarios, visualisation and

discussion ...............................................................................................................163

6.1 Cumulative assessment: category 3 wetlands...........................................164

6.2 Cumulative assessment: category 4 wetlands...........................................185

6.3 Implications of cumulative impact of multiple wetland management .......192

6.4 Chapter summary and Implications for the third hypothesis ....................198

7 Summary, Context and Discussion ..................................................................200

7.1 Assessment methodology ........................................................................201

7.2 Current capabilities .................................................................................202

8 Conclusion & Future Work .............................................................................209

9 References ......................................................................................................216

Glossary .................................................................................................................232

Appendix A: WETMOD differential equations .......................................................234

$Macrophytes .....................................................................................................235

$Phytoplankton ..................................................................................................237

$Nutrients ...........................................................................................................242

$NutrientExchange .............................................................................................246

$Wetland&RiverFlowExchange .........................................................................251

$SpatialRelevantTimeSeries ...............................................................................252

$RiverNutrients ..................................................................................................252

$WetlandsTimeseriesUpdateMeasuredValues .....................................................252

$WetlandTimeseriesUpdate ................................................................................252

$RiverTimeseries4WetlandUpdateTimeseries .....................................................252

$PotentialContributionToRiver ...........................................................................252


IV

Appendix B: Driving Variables ..............................................................................253

Appendix C: Key to wetland numbers ....................................................................263

Appendix D: Cumulative Management Scenarios ...................................................266

Appendix E: WETMOD 2 Code .............................................................................291


V

List of Figures and Tables

Figures

Figure 1: Wetland exchange modelling ...................................................................... 3

Figure 2: Cumulative assessment of wetland processes .............................................. 3

Figure 3: Study Area ................................................................................................. 5

Figure 4: Driving Variables, State Variables and Major Interactions in WETMOD 1 46

Figure 5: Macrophyte Module ................................................................................. 49

Figure 6: Plankton Module ...................................................................................... 51

Figure 7: Nutrient Module ....................................................................................... 53

Figure 8: WETMOD 2 Structure and Data Flow ...................................................... 56

Figure 9: Volume Exchange Module ....................................................................... 58

Figure 10: External Nutrient Module ....................................................................... 60

Figure 11: Outflow Module ..................................................................................... 61

Figure 12: “Exemplar” Wetlands & River Monitoring Sites .................................... 65

Figure 13: Paiwalla & Sunnyside wetlands .............................................................. 68

Figure 14: Lock 6 and Pilby Creek wetlands ............................................................ 69

Figure 15: Reedy Creek wetland .............................................................................. 70

Figure 16: Wetlands (Categories 1 to 5) Driving Variables Turbidity, Water

Temperature & Solar Radiation (see also in Appendix B) ................................ 73

Figure 17: Sunnyside Irrigation Drainage PO4-P, NO3-N, Phytoplankton and

Estimated Flow Volume (see also in Appendix B) ........................................... 79

Figure 18: River Murray Nutrient & Phytoplankton Time Series as well as River Flow

Volume (see also in Appendix B) .................................................................... 84

Figure 19: Wetlands (Categories 1 to 5) Monitored Nutrients and Phytoplankton .... 87

Figure 20: Wetland exchange modelling .................................................................. 92

Figure 21: Cumulative assessment of wetland processes .......................................... 96

Figure 22: Percentage Deviation based estimate of flow exchange: Reedy Creek

wetland ............................................................................................................ 98

Figure 23: Validation of simulation results for Paiwalla wetland of PO4-P, and NO3-N

for both conditions with and without water exchange ......................................101


VI

Figure 24: Validation of simulation results for Paiwalla wetland of Macrophyte

Biomass, Zooplankton and Phytoplankton for both conditions with and without

water exchange ...............................................................................................102

Figure 25: Validation of simulation results for Sunnyside wetland of PO4-P, and NO3-

N for both conditions with and without water exchange ..................................105

Figure 26: Validation of simulation results for Sunnyside wetland of Macrophyte


water exchange ...............................................................................................106

Figure 27: Validation of simulation results for Lock 6 wetland of PO4-P, and NO3-N

for both conditions with and without water exchange ......................................108

Figure 28: Validation of simulation results for Lock 6 wetland of Macrophyte


water exchange ...............................................................................................109

Figure 29: Validation of simulation results for Reedy Creek wetland of PO4-P, and

NO3-N for both conditions with and without water exchange ..........................112

Figure 30: Validation of simulation results for Reedy Creek wetland of Macrophyte


water exchange ...............................................................................................113

Figure 31: Validation of simulation results for Pilby Creek wetland of PO4-P, and

NO3-N for both conditions with and without water exchange ..........................116

Figure 32: Validation of simulation results for Pilby Creek wetland of Macrophyte


water exchange ...............................................................................................117

Figure 33: Sunnyside monitoring area ....................................................................119

Figure 34: Validation of simulation results for Lock 6 wetland PO4-P and NO3-N,

using non-calibrated wetland data ...................................................................128

Figure 35: Validation of simulation results for Lock 6 wetland Macrophyte Biomass,

Zooplankton and Phytoplankton biomass, using non-calibrated wetland data ..129

Figure 36: Validation of simulation results for Reedy Creek wetland PO4-P and NO3-

N, using non-calibrated wetland data ..............................................................131

Figure 37: Validation of simulation results for Reedy Creek wetland Macrophyte

Biomass, Zooplankton and Phytoplankton biomass, using non-calibrated wetland

data .................................................................................................................132


VII

Figure 38: Validation of simulation results for Pilby Creek wetland PO4-P and NO3-

N, using non-calibrated wetland data ..............................................................134

Figure 39: Validation of simulation results for Pilby Creek wetland Macrophyte

Biomass, Zooplankton and Phytoplankton biomass, using non-calibrated wetland

data .................................................................................................................135

Figure 40: Lock 6 impacts on Nutrient concentration due to Turbidity reduction ....143

Figure 41: Lock 6 impacts on Macrophyte, Zooplankton & Phytoplankton due to

Turbidity reduction .........................................................................................144

Figure 42: Reedy Creek wetland impacts on Nutrient concentration due to irrigation

drainage reduction ..........................................................................................147

Figure 43: Reedy Creek wetland impacts on Macrophyte, Zooplankton &

Phytoplankton due to irrigation drainage reduction .........................................148

Figure 44: Reedy Creek wetland impacts on Nutrient concentration due to irrigation

drainage reduction and 75% turbidity reduction ..............................................152


Phytoplankton due to irrigation drainage reduction and 75% turbidity reduction

.......................................................................................................................153

Figure 46: Reedy Creek wetland impacts on Nutrient concentration due to 95 %

irrigation drainage reduction at 25, 50 and 75% turbidity reduction ................154


Phytoplankton due to 95% irrigation drainage reduction at 25, 50 and 75%

turbidity reduction ..........................................................................................155

Figure 48: Reedy Creek wetland PO4-P % reduction in outflow .............................156

Figure 49: Reedy Creek wetland NO3-N % reduction in outflow ............................156

Figure 50: Reedy Creek wetland Phytoplankton % reduction in outflow .................157

Figure 51: Cumulative retention- category 3 wetlands ............................................165

Figure 52: PO4-P Concentration Trends ..................................................................168

Figure 53: Macrophyte Biomass Growth Trends .....................................................169

Figure 54: Phytoplankton Biomass Growth Trends .................................................170

Figure 55: Zooplankton Biomass Growth Trends ....................................................171

Figure 56: NO3-N Concentration Trends.................................................................172

Figure 57: Macrophyte Biomass (size of sphere, kg/m3) plotted against Wetland

Volume and Wetland Depth ............................................................................175

Figure 58: Macrophyte Biomass vs. Wetland Depth ...............................................175


VIII

Figure 59: Average Macrophyte Biomass (size of sphere, kg/m3) plotted against

Average Wetland Volume and Wetland Depth ................................................176

Figure 60: Macrophyte Biomass vs. Wetland Volume ............................................176

Figure 61: Average Macrophyte Biomass (size of sphere) Plotted against Average

Wetland Volume and Wetland Depth range 1 – 2 m ........................................178

Figure 62: Average PO4-P (size of sphere) Plotted against Average Wetland Volume

and Wetland Depth range 1 – 2 m ...................................................................178

Figure 63: Average PO4-P vs. Macrophyte Biomass at Wetland Depth range 1 – 2 m

.......................................................................................................................179

Figure 64: Average NO3-N (size of sphere) Plotted against Average Wetland Volume

and Wetland Depth range 1 – 2 m ...................................................................179

Figure 65: Average NO3-N vs. Macrophyte Biomass at Wetland Depth range 1 – 2 m

.......................................................................................................................180

Figure 66: Comparison of Macrophyte, Phytoplankton and Zooplankton Biomass for

each category 3 wetland (Key to wetland numbers adapted from (Jensen et al.

1996), see list in Table 18 in Appendix C) ......................................................181

Figure 67: Nutrient uptake for full year wet vs. uptake for summer wet/winter dry .184

Figure 68: Cumulative loading to category 4 wetlands ............................................186

Figure 69: Macrophyte Growth Trends ...................................................................187

Figure 70: Phytoplankton Growth Trends ...............................................................188

Figure 71: Zooplankton Growth Trends ..................................................................189

Figure 72: PO4-P Trends.........................................................................................190

Figure 73: NO3-N Trends .......................................................................................191

Figure 74: Data - Model Driving Variables; From Figure 9 in section 2.3 ...............254



Figure 77: Time Series Irrigation Drainage ; From Figure 10 section 2.3.1 .............257

Figure 78: Time Series Irrigation Drainage; From Figure 10 section 2.3.1 ..............258

Figure 79: Time Series Irrigation Drainage ; From Figure 10 in section 2.3.1 .........259

Figure 80: River Data; From Figure 11 in section 2.3.2 ..........................................260




IX

Tables

Table 1: Data Sources, Type & Monitoring Frequency ............................................ 64

Table 2: Wetland Morphology ................................................................................. 76

Table 3: Calibration of inflow data for the 5-wetland categories .............................. 99

Table 4: Non calibrated validation of inflow data for 3 wetland categories .............126

Table 5: Assessment summary of wetlands realistic simulation ..............................139

Table 6: Lock 6 wetland Percentage Outflow Reduction .........................................142

Table 7: Reedy Creek wetland Percentage Inflow reduction vs. Percentage Outflow

Reduction .......................................................................................................149

Table 8: Assessment summary of wetlands management scenarios .........................162

Table 9: Impact, of category 3 wetland‟s management, on river load per annum .....192

Table 10: Impact, of category 3 wetland‟s (depth range shallow <1m) management,

on river load per annum ..................................................................................194

Table 11: Impact, of category 3 wetland‟s (depth range medium 1-2m) management,

on river load per annum ..................................................................................194

Table 12: Impact, of category 3 wetland‟s (depth range deep >2m) management, on

river load per annum .......................................................................................194

Table 13: Impact, of Lock 6 wetland management, on river load per annum ...........195

Table 14: Impact, of Lock 6 wetland management, summer wet winter dry, on river

load per annum ...............................................................................................195

Table 15: Impact, of category 4 wetland‟s management, on river load per annum ...196

Table 16: Impact, of Reedy Creek wetland management, on river load per annum ..197

Table 17: Initial values ...........................................................................................234

Table 18: Wetlands simulated as category 3 wetlands .............................................263

Table 19: Wetlands simulated as category 4 wetlands .............................................265

Table 20: Change in PO4-P wetland loading and percentage outflow due to

management; category 3 wetland scenarios .....................................................267

Table 21: Change in NO3-N wetland loading and percentage outflow due to


Table 22: Change in Phytoplankton wetland loading and percentage outflow due to


Table 23: PO4-P comparison between Full year wet versus Summer wet Winter dry

for three selected wetlands; category 3 wetland scenarios ...............................285


X

Table 24: NO3-N comparison between Full year wet versus Summer wet Winter dry

for three selected wetlands; category 3 wetland scenarios ...............................286

Table 25: Phytoplankton comparison between Full year wet versus Summer wet

Winter dry for three selected wetlands; category 3 wetland scenarios ..............287

Table 26: Change in PO4-P wetland loading and percentage in and outflow due to


Table 27: Change in NO3-N wetland loading and percentage in and outflow due to


Table 28: Change in Phytoplankton wetland loading and percentage in and outflow

due to management; category 4 wetland scenarios ..........................................290


XI

Declaration

I declare that this thesis is my own work and to the best of my knowledge and belief,

contains no material used for the award of another degree, or published or written by

another person(s), except where appropriately referenced in the text. To the best of

my knowledge and belief, this thesis contains no material previously published or

written by another person, except where due reference has been made in the text.

I consent to a copy of my thesis, when deposited in the university Library, being made

available for loan or photocopying.

Kjartan Tumi Bjornsson

Date 2007


XII

Acknowledgements

This project used monitoring data from several sources. River flow data, which is

collected at all locks, was obtained from the Murray Darling Basin Commission

(MDBC). This flow data, which was included in the model, was collected at Locks 1

through to 8 (Figure 12 on page 65). The River Murray nutrient data was provided by

the Department of Environment and Heritage of South Australia (DEH). This nutrient

data was a collection of data originally sourced from the South Australian

Environmental Protection Authority (EPA), the MDBC, and the South Australian

Department of Water (SA Water). The river nutrient data monitoring points are at

Lock 5, Mannum, and Murray Bridge. For simplicity in this report, all river data is

referred to consistently as MDB river data. However, the contributions by the MDBC,

DEH, EPA and SA water are gratefully acknowledged, as without their support this

project would not have been possible.

Planning SA provided GIS data covering the wetlands (the South Australian Wetlands

Atlas (Jensen et al. 1996)), Locks, and the River Murray. Wetland Care Australia

provided the Wetlands Management Study report 1998 ((Nichols 1998)), which was

used in obtaining wetland depth information. Solar radiation was obtained from the

Bureau of Meteorology (BOM). Bartsch (1997) Marsh (1997) Wen (2002a) Wielen

(nd) have collected a substantial quantity of water quality data for some wetlands of

the lower River Murray, as well as irrigation drainage into affected wetlands and

some river data at a site close to the wetlands for the same monitoring dates. Table 1

on page 64 describes the source and frequency of data collection. I thank them for

their contribution of data.

To my supervisors, starting with my principal supervisor Friedrich Recknagel, I first

thank for the initiative to develop this project. I would also like to thank him for his

role in the supervision of the project and especially input in the thesis structure and

making sure I finished. Bertram Ostendorf I thank for taking on a larger role in this

project than anticipated, the input was invaluable. Megan Lewis I thank for the

support and input particularly during the early stage of the project when I needed the

assistance the most.


XIII

Lydia Cetin I thank for the input through the Honours work. The effort invested in the

WETMOD contributed greatly to my work.

Mardi van der Wielen I thank for the many discussions and „lessons‟ on the lower

River Murray wetland particulars.

I would particularly like to thank Leslie Jackowski for his reviewing of my thesis.

You outperformed your role as a good friend. I thank you for helping me through a

very difficult writing stage and for giving me the encouragement I needed.

I would also like to thank Bjorn Björnsson, Magnus Björnsson and Jason Bobbin for

their role in reviewing sections of the thesis.

I would also like to thank the research group at the University of Maryland.

Particularly Thomas Maxwell and Roelof Boumans for their assistance with the SME

(Spatial Modeling Environment) during the early part of my project, although the

project diverted from this course I thoroughly enjoyed the learning experience.

Without the finance provided by the SPIRT grant and the River Murray Catchment

Water Management Board this project would never have existed. I am grateful for this

financial assistance.

Last and by far not least I would like to thank Georgina Tate, for showing me what

true patience and support is. The importance you have played can never be measured

nor expressed adequately. It is now my turn to give you the same during your

speciality training.


XIV

Abstract

Most of the lower River Murray and its floodplain wetlands are impacted upon by

degradation caused by river regulation. Increasingly the restoration of these

ecosystems and the river water quality has become a high priority for federal and state

governments and associated departments and agencies. Public concern is adding to the

pressures on these departments and agencies to restore these ecosystems and to

sustainably maintain the river water quality.

The long term monitoring of floodplain wetlands has been limited, compounding the

difficulties faced by managers and decision makers on assessing the potential

outcome of restoration options. The role of this project in the broad scheme of

restoration/rehabilitation is to contribute to the construction of a model capable of

increasing managers and decision makers understanding, and build consensus of

potential outcomes of management option. This model was to use available data.

The developed model, based on WETMOD developed by Cetin (2001), simulates

wetland internal nutrient processes, phytoplankton, zooplankton and macrophyte

biomass as well as the interaction (nutrient and phytoplankton exchange) between

wetlands and the river. The model further simulates the potential impact management

options have on the wetlands, and their nutrient retention capacity, and therefore their

impact on the river nutrient load.

Due to the limitation of data, wetlands were considered in categories for which data

was available. Of these two had sufficient data to develop, calibrate and validate the

model. Management scenarios for these two wetlands were developed. These

scenarios included, the impact of returning a degraded wetland in a turbid state to a

rehabilitated clear state, and the impact the removal of nutrient from irrigation

drainage inflows has on wetland nutrient retention, and consequent input to the river.

Scenarios of the cumulative impact of the management of multiple wetlands were

developed based on using these two wetlands, for which adequate data was available,

as “exemplar” wetlands, i.e. data from these wetlands were substituted for other

similar wetlands (those identified as belonging to the same category). The model

scenarios of these multiple wetlands provide some insight into the potential response


XV

management may have on individual wetlands, the cumulative impact on river

nutrient load and how wetland morphology may relate to management considerations.

The model is restricted by data availability and consequently the outputs. Further,

some limitations identified during the development of the model need to be addressed

before it can be applied for management purposes. However, the model and methods

provide a guide by which monitoring efforts can assist in developing future modelling

assessments and gain a greater insight not only at the monitoring site but also on a

landscape scale.


1

1 Background

1.1 Introduction

Wetlands are increasingly becoming valued and used for some of the functions or

services they provide. Costanza et al. (1997) prepared a study on the value of the

world‟s different ecosystem services, wetland services scoring the highest of all

terrestrial and aquatic ecosystems. The services and functions offered by wetlands can

be broadly divided into 3 categories (Anonymous 1995; Morris 1991; Scheffer 1998).

The first of these is hydrologic or flood amelioration, where wetlands can act in aid of

short-term surface water storage, long-term surface water storage, or the maintenance

of a high water table. The second is the preservation of flora and fauna habitat and

associated food webs, through the maintenance of characteristic plant communities

and characteristic energy flow. The third is biochemical or nutrient and sediment

uptake, where wetlands can be involved in the transformation or the cycling of

elements, the retention or the removal of dissolved substances, and the accumulation

of inorganic sediments.

Not all functions of wetlands are regarded as an asset; the value of a wetland function

is usually only then recognised when useful or required services have been identified.

However, a wetland function that presently does not have a recognised value may

obtain one in the future. For example, the value of maintaining water quality by a

small wetland may not be recognised until it is acquiring a relative greater percentage

of representation in the area, or if it is close to a drinking water source (Anonymous

1995). Maintenance and restoration of wetlands and associated aquatic environments

should therefore have a high priority for sustainable development.

Whereas the flood amelioration and preservation of habitat and biodiversity has seen

ongoing recognition, the nutrient uptake and sediment uptake has started gaining a

greater significance than previously was the case due to the loss of wetland function.

This project focused primarily on the nutrient uptake aspect of wetland function,

although the potential management interventions simulated aimed at rehabilitating

degraded wetlands are also expected to contribute to wetland biodiversity and habitat

availability rehabilitation.


2

From an anthropocentric standpoint, there are a number of reasons for improving the

management of wetland function and the resources or services they provide.

Freshwater habitats have a very important role in sustaining human activities

(Burbridge 1994). The natural functions of wetlands produce a range of resources,

which affect the economic and social welfare of a diverse range of people. With the

degradation of wetlands these resources are being severely and adversely affected

(Burbridge 1994).

One justification for reversing the trend of degradation of wetlands is that the sum of

the services provided by the functioning of wetlands, which include economic and

social values, is of a greater value than can be gained from degraded or converted

wetland use (Burbridge 1994; Costanza et al. 1997; Pimm 1997). Furthermore, the

function of a number of small wetlands may not be recognised until their cumulative

capacity is fully understood. For example, swamp reclamation or flood amelioration

can also lead to wetland reduction or even destruction; with a decrease in overall

wetland area, reduction in average size, total numbers, linkage and density, the

cumulative function of wetlands will decline (Anonymous 1995; Johnston et al. 1990;

Preston et al. 1988). Therefore, the functions of wetlands, which include the uptake

and storage of nutrients and sediment retention, will have an impact on a landscape

scale through the improvement of water quality.

The primary driving force of nutrient exchange, being the flow of nutrient into and

out of a wetland, is through the water flow between the wetland and the river. The

model developed in this project used a nutrient balance simulation within a wetland to

calculate this exchange rate, thereby elucidating a significant unknown for wetland

management. The process, by which the model assesses the exchange rate, is

simplified in Figure 1.


3

Figure 1: Wetland exchange modelling

To understand the impacts that wetland functions have within a catchment and the

implications of management of wetlands (or of reduction of wetlands, i.e. continued

wetland loss or degradation), an evaluation of the cumulative impact of the functions

of multiple wetlands is required. Landscape scale modelling of the wetland processes

and associated functions, such as nutrient retention by healthy wetlands or lack

thereof in degraded wetlands, would contribute to knowledge and understanding and

therefore provide information for decision making. A model that can be applied

generically across multiple wetlands can be used to assess the cumulative nutrient

retention estimate on a landscape scale; Figure 2 represents the use of a model in such

a scenario.

Figure 2: Cumulative assessment of wetland processes

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

Wetland 1 Wetland 2 Wetland 3 Wetland n

+ + + River Load => Change in

River Load

Wetland process modelling

Nuteirnt load from river

Nutrient retention becomes a factor of exchange volume, river concentration and wetland concentration

calculated using the wetland process model. Impact on river calculated using this output and the river nutrient

load.

River flow volume and river nutrient load

Fraction of river flow volume (f)

Nutrient load from wetlands


4

The riverine ecology system is still inadequately understood (Young et al. 2000),

complicating the issue of aquatic modelling. However, even with limited

understanding and data resources, it is possible to develop an aquatic model to test

hypotheses of wetland function and management; and to improve general

understanding. To identify the processes required within a wetland model and be

aware of the interactions these processes have both within the wetland as well as

externally and appreciate some of the issues affecting water quality it is necessary to

examine some of the wetland characteristics in detail. The complex interactions

between sedimentation, re-suspension, turbidity, eutrophication, primary producers

and consumers are to varying extent considered in the model developed during this

project, and are therefore briefly discussed below in reference to the study area.

This project focuses on the floodplain wetlands of the lower River Murray, the South

Australian section of the Murray-Darling Basin in Australia (see Figure 3). The

catchment area is approximately 1 million km2 or approximately one seventh of

Australia (Hills 1974; Walker 1985; Walker et al. 1994). The headwaters comprise of

only 500 km of the 2560 km of the river (Mackay et al. 1990; Roberts et al. 1991;

Walker 1985), which has a total floodplain area of approximately 10,000 km2

(Roberts et al. 1991). The approximately 2,000 km of river floodplain section has a

very shallow gradient with a drop of mere centimetres over distances of kilometres

(Mackay et al. 1990; Walker 1985). The average annual runoff is approximately

11,000 GL but can vary from 2,500 GL in a dry year to 40,000 GL in a wet year

(Mackay et al. 1990; Walker 1985).


5

Figure 3: Study Area

1.1.1 Wetland processes

Wetlands are complex ecosystems with numerous interactions which link to separate

aquatic systems (river, creeks, drainage flow paths etc.), terrestrial systems such as the

surrounding riparian zone and atmosphere. The complex interactions such as between

primary producers, consumers, predators and their feedback loops; as well as the

multiple sources and losses of nutrient and energy, can make full accounting an

impossible task in wetland assessment and therefore modelling seem an impossibility.


6

Of the interacting facets within a wetland some can however be focused on to obtain

an understanding of the function of the wetland. The ones that are seen as the major

processes or facets within a wetland are discussed below.

Sedimentation

Any wetland processes that act to decrease waterborne sediment and nutrient

concentrations are considered to benefit water quality (Johnston 1991). Sedimentation

and sediment re-suspension are processes that operate continually in wetlands and can

have an impact on the nutrient availability and wetland turbidity. Increasing

sedimentation and decreasing sediment resuspension would, through their impact on

improving water quality, be seen as part of rehabilitation. That is, wetland turbidity

and consequent nutrient availability affect the state of wetlands and the primary

producer (phytoplankton and macrophyte) composition. In a sequence of events, the

state of primary producers, along with turbidity and nutrients, compound the impacts

on water quality within wetlands, i.e. self regulating processes.

Turbidity

Turbidity in a wetland can effectively shade out the incoming light, thereby

minimising the underwater light availability. Walker and Hillman (1982) have found

that even in eutrophic waters of the River Murray high turbidity can restrict primary

productivity. The high turbidity is therefore an important factor controlling plant

growth in River Murray wetlands (Walker et al. 1982). The reduction of turbidity

particularly within wetlands is consequently seen as a major management focus. The

Secchi depth of water bodies (an indication of turbidity) is increased both through an

increase in suspended matter and the high nutrient flux from the sediment, which also

stimulate the algal production (Soendergaard et al. 1992).

Nutrients

Dissolved and particulate inorganic nutrients such as phosphorus, nitrogen and silica

are a natural part of the water content in rivers. In excess, these substances become

pollutants and contribute to growth of phytoplankton and other aquatic plants

(Shafron et al. 1990). Laboratory studies have shown that the release of phosphorus

can be increased 20-30 times in a resuspended sediment compared to that of an


7

undisturbed sample (Soendergaard et al. 1992). Such increased phosphorus levels can

lead to eutrophication of wetland water.

Phytoplankton

The increased growth of phytoplankton caused by eutrophication contributes to an

increase in turbidity and a decrease in water transparency of the water column,

limiting light penetration and therefore submerged macrophyte growth.

Phytoplankton, detritus and resuspended inorganic sediment particles thereby

contribute to lake turbidity (Scheffer 1998). The algal blooms tending to increase

turbidity of the water column both through their presence and because macrophytes

are effectively shaded out, their contribution to sedimentation thereby is lost. Binding

of sediment through compaction and/or minimisation of resuspension is therefore an

important management objective.

Macrophytes

Macrophytes are not only a part of the primary productive activity of wetlands but

also contribute to its self regulated maintenance. For example, established

macrophytes have been said to function as biological engineers as they act as

buffering systems in wetlands and have a large role in maintaining a clear state (Sand-

Jensen 1998; Stephen et al. 1998). Some of the ‟engineering characteristics or

mechanisms‟ include the reduction in flow velocity, the stabilisation of the sediment

and the provision of habitats for micro-organisms, invertebrates and fish (Carpenter et

al. 1997a; Sand-Jensen 1998).

Biota such as macrophytes contribute to the long-term storage of nutrients, with some

residual accumulating in newly formed soils (Graneli et al. 1988; Kadlec 1997).

Further, macrophytes can become permanent sinks of phosphorus through the burial

of plant litter (Graneli et al. 1988). Some benefits of macrophytes include habitat

provision for zooplankton which feed on phytoplankton (Baldry 2000; Stephen et al.

1998; Timms et al. 1984), uptake of nutrients (Chen et al. 1988), including luxury

uptake and enriched denitrification (Meijer et al. 1994; Stephen et al. 1998). Reduced

chlorophyll-a (i.e. phytoplankton) has been found to occur close to macrophyte

growth. This has been associated to the presence of zooplankton which find a refuge

within the macrophyte growth (Stephen et al. 1998). The role of macrophytes in


8

supporting zooplankton and therefore control of phytoplankton can therefore be a

significant aspect in wetland management.

Macrophytes also reduce water movement (turbulence) and therefore reduce re-

suspension and increase sedimentation; they can also shade benthic algae and

phytoplankton (Mitchell 1989; Sand-Jensen et al. 1988; Stephen et al. 1998). Sand-

Jensen and Mebus (1996) showed a steep reduction in flow velocity within dense

macrophyte growth. The lower energy environment above the sediment within the

macrophyte patches leads to a retention of fine sediment and organic matter, carbon,

nitrogen and phosphorus (Chambers et al. 1994; Sand-Jensen 1998; Sand-Jensen et al.

1992; Sand-Jensen et al. 1996). Effectively the sedimentation within macrophyte beds

reduces the transportation of nitrogen, phosphorus and other particles downstream

(Sand-Jensen 1998). Therefore, a healthy wetland with a large macrophyte biomass

should self propel a reduction in turbidity and nutrient retention. Due to the many and

diverse mechanisms provided by the macrophytes they are recognised as a key step in

restoring wetlands (Meijer et al. 1994; Stephen et al. 1998).

Macrophytes obtain phosphorus from the surrounding water and the substrate, with

minimal release found in actively growing macrophytes (Graneli et al. 1988).

However, decaying macrophytes can account for a substantial contribution of

phosphorus to the open water (Graneli et al. 1988). The growth and decay of

macrophytes will therefore have an impact on the phosphorus balance of an aquatic

system.

Macrophytes affect nutrient levels in wetlands in more ways than just uptake and

sedimentation. For example, phosphorus release may also be reduced through

oxidation of the sediment (Stephen et al. 1998). Macrophytes readily take up soluble

nitrogen from recycling processes (Stephen et al. 1998). Macrophytes also serve as a

bottom up control mechanism of nitrogen both through uptake and denitrification

(Carpenter et al. 1997a; Stephen et al. 1998). Macrophytes also influence the nitrogen

cycle by increasing water residence time and therefore enhancing the denitrification

cycle. This can be up to 3 times otherwise expected due to the organic enrichment

among rooted macrophytes (Sand-Jensen 1998). Effectively, macrophyte biomass

contributes to nitrification and denitrification within shallow water bodies and

therefore plays a significant role in the nitrogen budget (Caffrey et al. 1992).


9

River flow

Seasonal changes in nutrient and turbidity levels are influenced by river flow

behaviour. As a result of decreased flow and increased nutrient availability the

impounding of water (e.g. instillation of the locks in South Australia) will possibly

favour the growth of phytoplankton leading to algal blooms (Shiel et al. 1982; Walker

1979). However, despite the eutrophic conditions there may be some limiting of algal

production due to the turbid waters of the lower River Murray (Walker 1985). The

turbid conditions may however not be limiting to Anabaena as it is able to control its

buoyancy thereby increasing its light harvesting potential (Baker et al. 2000). Nutrient

control to manage algal blooms would in this case be a significant management

achievement.

Zooplankton

Nutrient availability (N and P) may determine the potential algal biomass production,

however zooplankton grazing can have a large role in determining the biomass

balance (Stephen et al. 1998). Of the zooplankton in the lower River Murray, the most

common are indicative of eutrophic conditions, some of which are influenced by

temperature changes, turbidity and salinity (Shiel et al. 1982), reflecting the state of

the system. The zooplankton grazing rates can be correlated positively to water

temperature and have a negative impact on phytoplankton biomass, i.e. chlorophyll-a

(Kobayashi et al. 1996; Schwoerbel 1993). Studies by Griffin et al. (2001) showed

that zooplankton grazing had a significant impact on phytoplankton biomass. They

found zooplankton biomass peaks follow that of the phytoplankton biomass peaks,

which is a typical Lotka-Voltera predator-prey cycle (Griffin et al. 2001). The degree

to which zooplankton impacts on phytoplankton biomass is dependent on the

zooplankton species as well as the species and size of phytoplankton (Schwoerbel

1993). Zooplankton are therefore an important constituent within wetlands playing a

role in stabilising phytoplankton growth. They are therefore a significant aspect to

consider as part of management.


10

1.1.2 Spatial relationships of wetlands to transport processes

Flow and flood regulation

The seasonal distribution of flow has been changed by flow regulation of the River

Murray. The winter flows have decreased as surplus water is taken into storage, and

the summer flows have increased as irrigation demands are met (Walker 1979). There

has also been significant flood amelioration; that is, through water retention of some

of the surplus water followed by controlled release, the severity and incidence of

flooding has been reduced significantly (Walker 1979).

Flood regulation drastically affects aquatic environments, including the reduction of

interaction between all but the closest wetlands to the river (Walker 1979; Walker et

al. 1993). This renders wetlands, which are not adjacent to the river, dry due to the

lack of periodic flooding thereby virtually eliminating these wetlands. The wetlands

closer to the river remain for the most part permanently inundated with associated

consequences (e.g. lack of sediment compaction and lack of macrophyte

regeneration). Flow regulation has consequently been widely recognised as a major

contributor to river and floodplain wetland degradation (Arthington et al. 2003; Bunn

et al. 2002; Walker 1979; Walker 1985).

Nutrient retention and Exchange capacity

The natural retention of nutrients in wetlands occurs by cumulative fluxes into storage

compartments of the wetland ecosystem. These compartments include the soil,

vegetation and plant litter (Johnston 1991). Through their retention of nutrient,

wetlands act as sinks of waterborne nutrients and thereby act to improve the water

quality (Johnston 1991). The impact that a wetland sink or storage compartment has

on the water quality depends on both the rate of nutrient uptake and the retention time

(turnover rate) (Johnston 1991; Kadlec et al. 2001). The flow of water through a

wetland therefore controls the nutrient transport into the wetland as well the nutrient

transport out of the system, see Figure 1. The nutrient retention of the wetland is

significantly determined by the water residence time that is controlled by the flow

speed, wetland size and linkage to the river.

The proximity of a wetland to the river as well as the wetland shape, size, depth and

volume can have a substantial impact on the effectiveness of the function of the


11

wetland in the landscape; this impact can be influenced by the exchange capacity as

well as residence time with in the wetland. The exchange capacity can be impacted on

by channel volume, shape and length or by such factors as the location of the wetland

in the landscape. The location of the wetland in the landscape relating back to

variables such as wind direction, which in the case of the lower River Murray plays a

significant role in the flow direction and flow rate of the river (Webster et al. 1997).

The depth, area and volume of the wetland itself will also impact on the exchange of

water between the wetland and the river, for instance wind can push the water in a

large shallow wetland away from the connection channel; or evaporative processes

can be influenced by the volume and surface area of a wetland.

The transport of material in and out of wetlands is primarily a function of water flow

(Johnston 1991). That is, the exchange rate has an impact on the exchange of nutrients

and transport of salinity between wetlands and the river. These aspects must therefore

be taken into consideration for wetland management, however obtaining exchange

data can be prohibitive due to cost or the complexity of environmental factors

mentioned. Consequently, any method of obtaining an estimate of exchange between

the wetland and the river will be a valuable tool in wetland management and a

significant addition to budgeting aspects of nutrient and salinity impacts to the river.

This estimation of the transport of material by river exchange has the potential of

being the most significant external influence acting upon a wetland and a wetlands

impact on the river, and its estimation is currently a significant data gap.

Despite the size, shape and position of wetlands in the landscape having a potentially

large influence on the functioning of wetlands, these parameters are infrequently

measured. These wetland properties can however be radically changed by human

influence (Preston et al. 1988). Therefore, an understanding of how the wetland

properties influence wetland functioning on a landscape scale is relevant to restoration

and management decision-making. Scenario analysis of different wetland properties

may therefore assist in increasing this understanding.


12

1.2 Degradation of floodplain wetlands

Water quality is a key indicator of river and wetland health, and of wetland

functioning. Maintenance of good water quality helps to prevent further degradation

of wetland and riverine ecosystems. River and wetland water quality need to be

maintained in the interest of primary industry as well as water supply for urban

environments.

The River Murray basin accounts for a large part of Australia's agricultural

production. The demands of settlement and land use have placed considerable

pressure on the river system, resulting in a decline in biodiversity and aquatic habitats

and therefore altered the structure and function of river and wetland ecosystems. As a

result the water quality of the lower River Murray, which covers an approximately

650 km stretch of the river in South Australia, has drastically diminished.

The River Murray is often viewed as the lifeblood of South Australia, the driest state

on the driest continent, and water quality is a significant issue for its inhabitants. The

River Murray is a significant water source for South Australia. The city of Adelaide

derives between 55% and 90% of its water from the River Murray, and other South

Australian towns, including those of the ``Iron Triangle'' (Whyalla, Port Augusta and

Port Pirie), receive up to 90% of their water supply from this source (Jacobs 1990).

Agricultural areas along the River Murray use it as a primary water source for crop

irrigation, as there is very little rainfall in these areas. Other uses of the River Murray

within SA includes tourism (camping, fishing, house boats and other cruises) and

commercial fishing.

Wetlands perform important services and functions for river water quality, such as

accumulating nutrients and trapping sediments (Anonymous 1995; Johnston 1991;

Mitsch et al. 2000). Wetlands also act as habitats for a wide range of flora and fauna

(Boon et al. 1997; Recknagel et al. 1997). It is therefore imperative to restore and/or

maintain the structure and functions of wetlands, such as nutrient retention.

Of the wetlands along the River Murray few, if any, can be considered to be pristine

environments (Walker 1979). Due to the present regulation of the flow regime of the

River Murray, the development of new wetlands (billabongs is reduced significantly

(Walker 1979). Therefore, it is becoming increasingly important to preserve, maintain


13

and manage the remaining diversity, and significant areas of flood-plain habitats

(Walker 1979). It has become increasingly recognised that rivers and wetlands are

legitimate users of water (Arthington et al. 2003; Naiman et al. 2002), with

government departments, such as the South Australian Department of Water, Land

and Biodiversity Conservation (DWLBC), recognising their role in preserving and

restoring ecological processes and ecosystems. Legislation for the protection of

aquatic ecosystems, such as the Water Resources Act 1997 and as amended by the

Natural Resources Management Act 2004, shows the progress towards the recognition

of the importance of aquatic ecosystems such as wetlands.

As it is, there have been biological changes to the lower River Murray and its

floodplain wetlands due to the introduction of the locks (Walker 1985). Effectively

the river has been replaced by a series of cascading pools, which due to their

difference from the normal river flow encourage a change in biodiversity such as

plant community composition towards exotic species. These fish and plant species

being more accustomed to permanent inundation and slow flowing pools (Pressey

1987; Walker 1985). Along the River Murray there are more than 100 different

storages (Walker 1985), the lower River Murray wetlands are therefore to a large

extent now either permanently inundated, or above pool level left dryer for a longer

period than before (Pressey 1987). The river regulation has affected the riparian

vegetation by disrupting regeneration and affecting the mature period (Roberts et al.

1991). Due to the lack of periodic flooding black box (Eucalyptus largiflorens)

communities are showing a reduction in numbers, the river red gum (Eucalyptus

camaldulensis) is not regenerating in significant numbers and in many areas there is a

significant dieback due to drowning (Walker 1985).

Some of the problems contributing to water quality degradation in the River Murray

are associated with changes in catchment condition due to land use in the Murray

Darling Basin over the past 100 years. The increased nutrient load in riverine water

has led to an increase in algal growth and conversely a decrease in water quality.

River and wetland management therefore has an important role in preserving a very

significant resource for Australia.


14

1.2.1 Eutrophication of aquatic environments

A full understanding of wetland eutrophication is still in its infancy (Keenan & Lowe

2001). However, research has shown two alternate stable states exist for shallow

water bodies; that of the turbid phytoplankton-dominated state and the clear water

macrophyte-dominated state (Blindow et al. 1993; Boon et al. 1997; Scheffer 1998;

Scheffer et al. 1993; Stephen et al. 1998). A wetland in one state will tend to remain

so due to a number of buffer mechanisms (Boon et al. 1997; Moss 1990; Stephen et

al. 1998), but with an increase in nutrient loading to a system, a wetland may change

from a clear to a turbid state (Boon et al. 1997; Scheffer 1998). A reverse change can

be difficult to obtain, however changes in water level and the removal of a part of the

fish stock have been used as successful restoration approaches in returning wetlands

to a clear state from a turbid one (Scheffer 1998).

In the River Murray catchment, agricultural development (land clearing, irrigation and

pasture management) has caused substantial increases in the river sediment load

(Walker 1979). There is also an increase in the organic, and nutrient load to the river

brought on through agricultural practices and as a consequence of loss of buffering

activity of the cleared vegetation (Lijklema 1994). This, combined with the turbidity

and sediment deposition downstream, affect the water quality and habitat suitability of

the river and its wetlands. Through leaching of nutrients from fertilised and irrigated

surrounding farmland, some wetlands of the River Murray floodplain have become

eutrophic.

This eutrophication, combined with turbid waters and degraded systems (e.g. by

permanent inundation, or the presence of exotic species such as carp), has turned the

wetlands into a turbid, algal dominated state where phytoplankton out-competes

macrophytes, leading to algal blooms (Scheffer 1998). Some of the nutrients

increased in the river are phosphorus and nitrogen. Both are essential nutrients for

plant growth, but in excessive amounts they can reduce water quality through

eutrophication, algal blooms, decreased light penetration and loss of dissolved oxygen

of the water body (Marsden 1989). Eutrophication can also contribute to the reduction

of macrophytes due to the shading impact of increased phytoplankton and can force a

wetland into a turbid state (Asaeda et al. 2001; Graneli et al. 1988; Scheffer 1998).

Therefore, increased eutrophication can alter the species composition of a wetland

(Johnston 1991) and therefore change the function of a wetland.


15

In most wetlands of the lower River Murray, phosphorus and nitrogen concentrations

exceed the limit of what is considered critical for eutrophication, reflected in the high

chlorophyll-a concentrations in the water columns (up to 256 mg/l) (Boon et al. 1997;

Goonan et al. 1992)). Boon et al. (1997) concluded that nutrient enrichment poses a

significant threat to the ecological integrity of wetlands throughout Australia.

Management of nutrients in the landscape can therefore have an impact on a large

range of ecosystems. Using wetlands or at least managing wetlands to fulfil the

function of nutrient retention can thereby be a strong tool to their own preservation.

The sources of nitrogen to wetlands include external inflow as well as fixation of

gaseous N2 that is converted into organic nitrogen (Johnston 1991). The removal of

nitrogen from wetlands however often occurs through a process called denitrification

where nitrogen is released into the atmosphere (Bowden 1987; Kadlec et al. 2001;

Mitsch et al. 2000; Morris 1991; Reddy et al. 1989; Scheffer 1998; Schindler 1977).

Therefore, the concentration of nitrogen or NO3 is in a continual state of flux

depending on the rate of nitrification and denitrification.

Unlike nitrogen, phosphorus cannot escape a wetland system. It therefore remains in

the system and is recycled. Phosphorus uptake in wetlands is regulated by physical,

(e.g. sedimentation), and biological mechanisms, (e.g. uptake and release by

vegetation) (Kadlec et al. 2001; Schindler 1977). The measurement of phosphorus

within the wetland is therefore more stable and indicative of availability.

Nitrogen and phosphorus can play a significant role in the eutrophication of wetlands

(Reddy et al. 1995). Phosphorus is the major limiting nutrient to nitrogen fixing algae

such as Anabaena (Schindler 1977) whereas increased nitrogen concentration can

contribute to a shift in species composition within wetlands (Morris 1991; Schindler

1977). They are also both the most likely nutrients to limit primary productivity

within wetlands (Baker et al. 2000; Beardall et al. 2001; Hecky et al. 1988; Morris

1991; Oliver 1993; Schindler 1977; Walker 1979; Walker et al. 1982).

Total phosphorus includes crystalline, occluded, adsorbed, particulate organic, soluble

organic and soluble inorganic phosphorus. However, not all of this phosphorus found

in a water body is biologically available. The biologically available phosphorus

includes the soluble reactive phosphorus (entirely biologically available), soluble

unreactive phosphorus (available through enzymatic hydrolysis) and labile


16

phosphorus (available through desorption) (Holtan et al. 1988). The main

anthropogenic input of phosphorus is through fertilisers and detergents (Holtan et al.

1988). In a system such as the lower River Murray, the phosphorus source to a

wetland can be almost exclusively through fertiliser or irrigation drainage runoff,

whereas sediments can act as sinks of phosphorus (Holtan et al. 1988). The

phosphorus, which is found in the sediment, is to a great extent sorbed to soil particles

or as part of organic matter (Holtan et al. 1988) reducing its availability. This

biologically unavailable phosphorus, found in the sediment, can be released through

various mechanisms such as turbulence, animal activity (bioturbation), and plant

growth (Scheffer 1998), thereby becoming biologically available. Sediment can

therefore in circumstances contribute to the maintenance of eutrophication in a water

body where inflow may have been reduced (Lennox 1984; Lijklema 1994; Nürnberg

1984; Nürnberg 1998; Olila et al. 1995; Recknagel et al. 1995). Recknagel et al.

(1995) found through simulations of Lake Mueggelsee that the best management for

the reduction of eutrophication was in fact sediment dredging. In a different approach

as in the case of lower River Murray wetlands, sediment compaction (drying of

wetland beds) to minimise resuspension and consequent release of bound nutrients is

a management strategy currently employed.

The internal phosphorus loading of the sediment is a significant factor in internal

loadings once external loading has been reduced. However, van der Molen (1994)

found that including this in the model of phosphorus concentration did not

significantly improve the predictive capacity of their model for shallow lakes which

experienced high external phosphorus loadings. In shallow lakes, where the external

loading was reduced, the sediment water exchange of phosphorus became significant

in estimating the variability experienced. This indicated that the external source was

the significant contributor of phosphorus concentration. Their hypothesis is that in

shallow lakes with significant external phosphorus loadings the sediment water

interaction is held in an equilibrium (van der Molen et al. 1994). It can be assumed

that this is also the case in wetlands, phosphorus simulation in eutrophic lower River

Murray wetlands can therefore concentrate on the external and suspended phosphorus

and disregard the impact of current sediment released phosphorus.


17

1.2.2 Alternate stable states and permanent inundation impacts on

wetlands

To maintain natural ecological integrity rivers, floodplains and their wetlands need

their natural flow regime in its full spatial and temporal variability (Arthington et al.

2003; Bunn et al. 2002; Poff et al. 1997). The wetlands, as part of their ecological

function, provide resilience mechanisms by which extreme events are buffered. Some

of these have been discussed above, such as phosphorus and nitrogen loads, the role

of macrophytes, plankton (phytoplankton and zooplankton) and flow regime. All of

these complex interactions, many of which have not been described, act to provide the

wetlands with a certain resilience mechanism. However, with the destruction of these

resilience mechanisms new resilience mechanisms develop to adapt to the new state

of the ecosystem (Carpenter et al. 1997a; Ludwig et al. 1997; Scheffer et al. 1993).

The change of wetlands from one buffered state to another is due to the resilience

being overcome by an extreme event. Such a change will transform an aquatic

ecosystem, such as a wetland, from one stable state to another (Carpenter et al. 1997a;

Carpenter et al. 1997b). The change can be driven by a complex interaction of

eutrophication, loss of macrophytes, water regime and turbidity, or by extreme events

for any of these (Carpenter et al. 1997a; Carpenter et al. 1997b; Scheffer et al. 1993).

The two states can be seen as alternate stable states of clear and turbid (Scheffer et al.

1993). Through river regulation many of the lower River Murray wetlands have

degraded to the turbid state reducing the function of the wetland in the landscape.

Returning a wetland to a clear state, once it has switched to a turbid one, can be more

complex than reversing the cause (Scheffer et al. 1993). For example, eutrophication

contributes to changing a shallow aquatic ecosystem such as a wetland from a clear

stable state to a turbid one. Reducing the nutrients may however not bring the wetland

back to a clear state due to the resilience of the alternate turbid stable state, which acts

through the buffering release of nutrients from the sediment or resuspension as winds

are not reduced by the otherwise present macrophytes (Scheffer et al. 1993).

Management of these wetlands could lie in the forceful change from one state to

another, such as the reduction of turbidity. This would induce macrophyte growth

through increased light availability that then reinvigorate the resilience of the clear

stable state (Scheffer et al. 1993). Scheffer et. al. (1993) suggest the reduction of


18

turbidity through either the management of fish stock such as carp or the management

of water levels to induce macrophyte growth.

Australian wetlands do not need constant inundation, and in fact their constant

inundation is detrimental. Drying and refilling of wetlands are natural processes in

Australian wetlands to which the flora and fauna are adapted and dependent (Pressey

1990). Permanent inundation reduces the growth and regeneration potential of

ephemeral vegetation common to River Murray floodplains (Nielsen et al. 1997), and

the lack of periodic flooding, due to river regulation, may contribute to the lack of

regeneration of terrestrial vegetation.

This permanent inundation of wetlands resulting from the regulation of river flow has

favoured invasion by the common carp (Cyprinus carpio). The feeding habits of carp

are thought to be a potential contributor to wetland turbidity further limiting

macrophyte growth. Carp together with the lack of drying cycles in the floodplain

wetlands are therefore believed to have contributed to the demise of wetland

macrophytes (Blindow et al. 1993; Pressey; van der Wielen 2001; Walker et al.

1993).

Further macrophyte loss is due to a lack of dry periods in wetlands. The lack of drying

cycles reduces sediment compaction leading to easier re-suspension and increased

wetland turbidity (McComb et al. 1997). Through increased turbidity macrophytes

can be shaded out causing their dieback. Their regeneration cycle, which is dependent

on dry spells, is also interrupted through the permanent inundation. The lack of

competition for underwater light due to the loss of macrophytes, as well as the loss of

nutrient buffering actions of macrophytes, stimulates phytoplankton growth and

increases the potential for future algal blooms (Carpenter et al. 1997a; Recknagel et

al. nd). Although the introduction of drying cycles is partly expected to reduce

wetland turbidity other couses for water turbidity also exists for the River Murray and

have an impact on the potential reduction of turbdidty possible in a wetland. Darling

River water for example, which is known to be turbid and sodic soils, widespread in

Australia (Rengasany et al. 1991), contribute to maintaining tubidity within wetlands.

In 2000 the Department of Water, Land and Biodiversity Conservation introduced a

flood event to a section of the lower River Murray floodplain (DWLBC 2004;

Siebentritt et al. 2004). A study on the impacts of flooding on riparian plants of the


19

River Murray showed an increase in flood dependent species, the reduction in flood

intolerant species but no change in aquatic species (associated with an impoverished

seed bank) (Siebentritt et al. 2004). The recommendation of this study was future

repeat flooding to increase the aquatic species seed bank and enhance their

regeneration. This flooding in 2000 also seemed to be effective in reducing exotic

species numbers. This study confirms the hypothesis of the impact river regulation

has had on riparian and aquatic species, with the reduction in aquatic species seed

bank of the lower River Murray wetlands. The study also shows one method of

influencing and improving species regeneration, i.e. flooding.

Nielsen and Chick (1997) conducted a study on sixteen artificial billabongs on the

River Murray floodplain. Their findings were that the longer a billabong remained

flooded the less diverse the plant communities became. The permanent flooding in

their study did not allow ephemeral or terrestrial species to grow, whereas in

billabongs where extended periods of drying followed by spring flooding was

introduced more diverse plant growth including terrestrial taxa were seen as a

consequence. This shows that should wetlands in the lower River Murray floodplain

have a natural water regime a more diverse plant community should become evident.

As a wetland management strategy the alteration of wetland inundation through the

introduction of dry periods and consequent re-flooding should stimulate responses to

species regeneration in turn returning a wetland to a stable clear state. In the lower

River Murray this management response may however be reduced when the main

water source is from the more turbid Darling River.

1.2.3 Irrigation drainage and constructed wetlands

Both constructed wetlands and natural wetlands can be used to improve water quality

(Keenan et al. 2001). Braskerud (2002) found that constructed wetlands placed at first

order streams removed between 21% and 44% of the phosphorus inflow. Constructed

wetlands, in a study by Burgoon (2001), were found to remove from 50% to 99% of

the nitrate inflow load. In a study of constructed wetlands in Flanders Belgium, the

nutrient removal efficiencies ranged from 31% to 65% for nitrogen and 26% to 70%

for phosphorus (Rousseau et al. 2004). Whereas Schulz et al. (2004) found

constructed wetlands for the treatment of aquaculture runoff were able to remove 41%

to 53% of phosphorus and 19% to 30% of nitrogen. Lüderitz et al. (2002), who


20

studied the effectiveness of constructed wetlands on sewage treatment, found removal

rates of phosphorus to be between 27% and 97% depending on the constructed

wetland design and management. Stormwater treatment, in Australia, with constructed

wetlands were found to remove up to approximately 80% of phosphorus (Bavor et al.

2001). This shows that the reason for using constructed wetlands in the removal of the

dissolved nutrients phosphorus and nitrogen can be diverse, the effectiveness of the

constructed wetlands also varying significantly. The optimal design parameters of the

constructed wetlands and the retention time required for nutrient removal depend on

the nutrients being removed (Bavor et al. 2001; Hunter et al. 2001; Rousseau et al.

2004). However, all studies showed that wetlands can be used for nutrient removal.

One of the major contributors of nutrients to wetlands is irrigation drainage from

agriculture. Constructed and natural wetlands are capable of absorbing phosphorus

and can be used for phosphorus load reduction (Kadlec 1997). How they impact on a

river system (i.e. capacity of wetlands to remove nutrients from the system) depends

on their location in the landscape (Crumpton 2001). In identifying the best landscape

position of restored wetlands Crumpton (2001) found that where wetlands are placed

to intercept a higher load of nutrients there is an increased retention capacity.

Studies by Wen and Recknagel (2002) and Wen (2002a) at a wetland in the lower

River Murray show that constructed wetlands can reduce wetland nutrient inflow

from irrigation drainage by up to 90% (Wen 2002a; Wen et al. 2002). Therefore, in

cases where the main wetland degradation impact comes from „reclaimed swamp‟ or

dairy pasture irrigation drainage outflow, the eutrophication source can be reduced

substantially. Consequently, where possible the interception of irrigation drainage and

treatment prior to its flow into wetlands could contribute considerably to the reduction

of nutrient inflow loads into wetlands.


21

1.3 Restoration of degraded floodplain wetlands

1.3.1 Management strategies for restoration

Following wetland restoration, through the re-introduction of drying cycles and carp

restriction during re-wetting of wetland degraded by permanent inundation,

Recknagel et al. (nd) observed the recovery of wetland habitats and the improvement

of water quality. By introducing drying periods or partial draw down, the germination

and growth of macrophytes are stimulated allowing for a return of macrophytes in a

reflooded wetland. Although initial conditions following re-wetting show increased

nutrient availability and therefore algal growth in the wetlands, macrophytes once

established out compete the algal community for nutrients (Recknagel et al. nd).

Where possible, such as in constructed wetlands, the harvesting of macrophytes can

partially remove the nutrients from the system (Hunter et al. 2001).

The main benefit of the drying of a wetland is the consolidation of the wetland

sediments, which reduces re-suspension, minimising turbidity and release of nutrients

from the sediment (McComb et al. 1997; Recknagel et al. 2000; van der Wielen

2001). Therefore, re-introducing a dry period to a wetland can have the impact of

switching a wetland from a turbid stable state to a clear stable state (Scheffer 1998;

Scheffer et al. 1993) as discussed above. Consequently, the reintroduction of dry

phases has been recommended as a management strategy to improve or restore

wetlands of the Murray-Darling Basing (Scholz et al. 2002).

Equipping wetland inlets with grills will prevent large carp from entering the re-

flooded wetland. It is assumed that this will protect macrophytes from being uprooted

by carp, as well as reducing the re-suspension of sediment expected as a consequence

of their feeding behaviour.

As a summary of the above discussed issues of wetland degradation, the potential

management strategies therefore available to improve water quality of degraded

wetlands, which have been considered in this project, are as follows:

1. The reintroduction of drying and wetting cycle‟s thereby reducing turbidity of

wetlands. Through this measure, the function provided by emergent and

submerged macrophytes can be reinstated revitalising a degraded wetland

(Recknagel et al.; Recknagel et al. 1997; Recknagel et al. 2000; Scholz et al.


22

2002; van der Wielen 2001). Drying consolidates the sediment and therefore

reduces the quantity of suspended solids in the water column. The re-emerging

macrophytes act to improve water quality by nutrient uptake, reduce flow

speed increasing sedimentation (Sand-Jensen 1998) and by out competing

phytoplankton for nutrient (Recknagel et al. nd). Experiments have shown that

water quality in wetlands managed in this manner can improve (Recknagel et

al.; Recknagel et al. 1997; Recknagel et al. 2000; van der Wielen 2001). There

are two possible mechanisms for introducing dry periods; these are through

the construction of regulators at individual wetlands or to implement it at a

broader scale through the change in water retention and therefore river height

at individual locks.

2. Irrigation drainage nutrient reduction through constructed wetlands. In

wetlands where external point nutrient sources such as irrigation runoff

contribute to the wetland nutrient load, there is an opportunity for the

construction of artificial wetlands where macrophyte harvesting can be used to

reduce nutrient loads. An example of this management strategy, in the lower

River Murray, was a research project situated between the Reedy Creek

wetland and the Basby farm near Murray Bridge (Wen 2002a). Here an

experimental pond system was constructed to eliminate inorganic phosphorous

from agricultural drainage by native water plants (Wen 2002a). This research

may lead to the design of constructed wetlands for the treatment of agricultural

drainage water before it enters floodplain wetlands.


23

1.4 Predictive modelling of wetland processes and services;

current state and potential alteration due to management

Real environmental systems are complex and it is therefore extremely difficult to

measure the parameters with accuracy (Parsons et al. 1995). The predictive ability of

water quality models is seriously limited by the difficulty in identifying complex

environmental processes and defining these within parameters (McIntosh 2003;

McIntyre et al. 2003; McIntyre et al.; Reckhow 1994). One method to overcome this

is to invest in monitoring and field based research. However, this quantitative

understanding and data are difficult and expensive to obtain. Ecologists and

resource/land managers cannot always employ traditional quantitative simulation

because of financial and temporal constraints (McIntosh et al. 2003), and therefore

need to use alternative approaches such as modelling.

Clearly substantial and complex data are required in order to assess and understand

the processes within a wetland, the interactions of these processes within the wetland,

and processes having influences upon wetlands, let alone assessing the implication of

potential management strategies. Assessing such a substantial and complex data set is

therefore outside the capacity of an individual. To facilitate the understanding of

processes operating on such a large scale computer models can be created to assist in

evaluating the wetland processes. This enables an assessment of management

scenarios as well as the testing of hypotheses of wetland function (Caswell 1988;

Goodall 1972; McIntosh 2003; McIntosh et al. 2003; Oreskes et al. 1994; Rykiel

1996; Wallach et al. 1998). As a consequence of the complexity of assessing such a

vast and complex data, there has been an increasing use of simulation models in the

study of aquatic and other ecological systems over the past couple of decades (Elliott

et al. 2000; Oreskes et al. 1994; Wallach et al. 1998).

There are two strategies for the management of degraded wetlands considered for this

modelling work, the choice being dependent on the reason underlying the

degradation. For wetlands where the main degradation is the inflow of nutrients

constructed wetlands would be considered. These constructed wetlands would

eliminate nutrients by absorption to nutrient poor sediments and nutrient uptake by

macrophytes. Simulation of the management of these wetlands would help determine

the impact of successful nutrient removal on the wetland and its exchange rate of


24

nutrients with the river. Where permanent inundation is the primary cause of wetland

degradation, a model could help determine the impact of the introduction of drying

and wetting cycle, on internal nutrient dynamics and wetland nutrient uptake. Both of

the simulations would provide assistance in decision support by providing an estimate

of:

overall wetland recovery and restoration

wetland specific responses to restoration management, and

degree of response required from either nutrient removal (constructed

wetlands) or from turbidity reduction (sediment compaction)

Modelling can be useful in understanding ecosystem processes and predicting

intervention outcomes. There is an inherent value in the analysis of past and present in

setting goals and objectives for the future (Thomann 1998). That is, modelling can be

used as a tool in predicting the ecological consequences of restoration plans or

management scenarios (Costanza et al. 1998; DeAngelis et al. 1998), which are vital

factors in environmental decision-making. Recognition of model validity, and

transparency to stakeholders, increases understanding and contributes to informed

dialogue, thereby enhancing decision making by consensus (Thomann 1998). A

model can also be helpful in a situation where non-linear mechanisms cause

unexpected patterns that cannot be grasped intuitively (Scheffer et al. 2000), or where

systems are too complex or cumbersome for human interpretation alone. For example,

computer models can be used to predict wetland response to environmental change

(Sklar et al. 1993).

A model is not expected to achieve exact predictions of ecosystem function, but its

development provides a tool for an approximation of outcomes. After all, modelling

often involves stressed systems with a view to return them to a natural state (Beck

1997). Not all potential impacts can be modelled successfully following intervention,

as there is always some lack of knowledge. However, modelling can help minimise

(but not eliminate) the variability of potential outcomes (Beck 1997).

Simulation models have become widespread and are playing an increasingly

important role to assist in the decision making process (Griffin et al. 2001; McIntyre

et al.; Sullivan 1997). With the fragmentation and lack of wetland specific data and

knowledge in the lower River Murray region, it is difficult for managers and other


25

stakeholders to make decisions in the management of wetland restoration. This is

particularly true in assessing the highest return of investment (cost-benefit analysis).

Managers ideally desire models capable of quantitative predictions of restoration

scenarios, taking into account hydrology and ecological processes of aquatic

environments (Arthington et al. 2003). However such models are not yet available for

Australian conditions as their development is prohibitively expensive, particularly

models that take into account catchment or regional scale (Arthington et al. 2003). A

simplified model capable of regional scale scenarios may however be able to answer

some of the questions posed by managers.

1.4.1 Complexity and feasibility of modelling

There are two important factors which will dictate the complexity of any model of an

ecological system. The first is the purpose of the model, dependent on the aims of the

potential user, e.g. flexibility may be an important issue. The second factor is the

feasibility of a model. This can be dependent on the understanding or knowledge

available for a system. That is, to what extent can the system be explained within a

modelling framework based on the current knowledge (McIntosh et al. 2003;

Reckhow 1994; Young et al. 2000). Furthermore, the incorporation of too many

factors into a model can obscure the action of some processes and render the model

mathematically inflexible (Caswell 1988). Caswell (1988) even suggested omitting

important factors to avoid obscuring the focus of the model. De Wit and Pebesma

(2001) compare four models of increasing complexity to assess the value of complex

models versus simple models. Their conclusions are that the complexity of models

may not improve the modelled results if the data quality is restrictive.

A model does not have to be extremely complex as good data for model development

may be all that is required to produce a simulation that will answer questions (Gibbs

et al. 1994). Taking this further, the simplest possible model, which can accurately

predict an observed phenomenon, provides a valuable contribution to ecological

knowledge (Caswell 1988). It provides a starting point on which there is a possibility

to build on observations and develop new theories (Caswell 1988). Whereas

unnecessarily complex models may lack the flexibility that may be required and may

contain inherent flaws (Wood 2001).


26

The choice between simple and complex models is affected by knowledge and data

availability. Young et al. (2000) found aquatic ecology to be complex and dynamic

with a multitude of interactions. However limited data, such as for the lower River

Murray wetlands, and limited detailed knowledge and understanding, of aquatic

ecology (Keenan et al. 2001; Young et al. 2000) argues for a simple model structure

(de Wit et al. 2001; Li et al. 2002; Li et al. 2003; Reckhow 1994). Reckhow (1994)

claims that limited data and knowledge are incompatible with high detail and large

models. The lack of detailed understanding of each process required to develop a

holistic quantitative model of an aquatic ecosystem restricted the modelling by Young

et al. (2000) to a few parameters. Young et al. (2000) therefore adopted a simplistic

modelling approach. The degree of knowledge is therefore an important determinant

of the level of complexity allowable within the model to achieve a meaningful and

accurate scenario (Wood 2001).

Wetlands are variable ecological systems and can be complex to model. This is due to

their morphology, susceptibility to sporadic external influences such as wind,

temperature, river flow (directional change is a possibility in the case of lower River

Murray wetlands (Webster et al. 1997)) and a multitude of complex dynamics and

interactions that cannot be monitored and studied without disturbing (and therefore

influencing) the system. Modelling wetlands can therefore become a complex venture

often hampered by the lack of detailed monitored data as well as rapid and sporadic

change in condition such as water availability, weather etc.

Despite the lack of knowledge, many complex descriptions of wetland behaviour and

nutrient cycling have been developed for modelling purposes. However, the

complexity complicates and often defies calibration and validation. Through this

complexity, the ecosystem can behave counter-intuitively despite individual

components being well understood. Therefore, wetland models are often kept simple,

with well-understood parameters and processes assigned a defined value (Kadlec et

al. 2001). Additionally, simple models are easier for non-modellers such as water

quality managers and the general public to understand (Murray 2001), which

contributes to consensus building.

Young et al. (2000) began with a simple model, intending to extend model

complexity in the future. The key to their model development was keeping the degree

of complexity consistent with the current level of understanding. Therefore, the model


27

can be developed as the understanding develops. This premise was also used in

WETMOD development (Cetin and Recknagel, pers. Com). Therefore, WETMOD

can be further developed with increasing understanding and data availability, and is

therefore a basis from which to conduct further research for lower River Murray

wetlands. Another example of simple wetland nutrient retention models, simple due to

due to limited data and knowledge, are described by Li, Xiao et al. (2003) and Li et

al. (2002) and discussed in relation WETMOD below. Different ways how the

accuracy of models that simulate complex systems can be assessed are examined

below.

1.4.2 Qualitative and quantitative assessment of model accuracy

and generic applicability

As the output requirements for models can vary depending on their intended

application and purpose; further differentiated by data availability by which to run

scenarios, many opinions on the need for quantitative vs. qualitative modelling output

have developed. Different methods of assessment of model performance have

therefore been developed. Judgmental terms such as excellent, good, fair, and poor are

useful because they can invite, rather than discourage, contextual definition (Oreskes

et al.). It is not uncommon for water quality models to have a small amount of data

available for model development, leaving even less for model evaluation and testing.

In this situation, rigorous testing and assessment of model predictions is rare and has

little meaning (Reckhow 1994). Water quality model calibration should compensate

to some degree for errors arising from model limitations (spatial averaging, model

structure errors and numerical dispersion) (McIntyre et al.).

Data restriction to modelling

Due to the limit in data availability “exemplar” data have been used to develop

predictive models. The model output along with continued monitoring can then be

used for adaptive management relating model outcomes with real occurrences (Young

et al. 2000). It must however be understood by both the model developer and future

users that the level of assumptions regarding the use of “exemplar” data will affect

the modelling accuracy in a quantitative way (Beck 1997; Wood 2001). In using

assumptions within models some otherwise unsolvable process given the current data

availability or knowledge can be resolved.


28

McIntosh (2003) states that there is no reason why relationships between abiotic

quantities such as soil and nitrogen, or between biotic and abiotic quantities such as

vegetation biomass and soil or water, cannot be modelled imprecisely if such an

approach is required by the level of available knowledge/data or the model purpose.

The model output in such a case should however not be expected to be quantitatively

accurate. However, despite a lack of quantitative accuracy, qualitative results can be

used as a guide in future monitoring, research needs and further model development.

The argument may be that qualitative models outputs have an intrinsic uncertainty due

to the imprecision of the outcomes. In fact stakeholders and managers are often aware

of model uncertainty, however they do not see this as detrimental to the value of

models in decision support (Andersson 2004). That is, the role of the model may be

such that the only output possible is a qualitative one due to data limitations and

therefore inherent assumptions. However, such a qualitative output can be informative

and therefore assist in decision-making, even if this decision is of the necessity of

further research. A qualitative model therefore fulfils its function where inadequate

data is available for quantitative predictions.

Assessing accuracy

Many studies have discussed imprecise and qualitative modelling techniques

(Dambacher et al. 2003; Guerrin et al. 2001; McIntosh 2003; McIntosh et al. 2003;

Parsons et al. 1995; Wood 2001). An understanding of, and rigorous comparison to

monitored data, can be used to assess a models qualitative and quantitative accuracy.

Wetland management decision support models are not necessarily dependent on

optimal statistical accuracy, and may fulfil their role when assessed as qualitative

models (McIntosh et al. 2003), despite their quantitative output and need therefore not

be validated as rigorously as purpose built quantitative models. Some models such as

WETMOD (Cetin 2001) are developed as quantitative models that provide qualitative

outcomes.

The comparison of model output, to establish modelling accuracy, can be performed

in a qualitative or a quantitative manner, the decision of the methodology being

dependent on circumstances (Wood 2001). The qualitative approach assumes that if

the composition and structure of the ecosystem is known, it can be encompassed in

models qualitatively (Dambacher et al. 2003). Users of such a model must however


29

understand that qualitative predictions are only a relative benchmark of expected

system behaviour (Dambacher et al. 2003). In fact, Andersson (2004) found that

stakeholders and decision makers were more interested in relative changes over long

periods than prediction of exact time-series. Andersson (2004) found that the use of a

simple qualitative model to assess future environmental conditions depending on

management strategies was effective in stimulating a three-way communication

between model developers, stakeholders and decision makers, the stakeholder

however were cautious about regional information. The modelled results were found

to provide a common base for understanding the impact of management. Andersson

(2004) also found that qualitative information based on a generic environment was an

effective model output for stakeholders and that quantitative information could be

seen as confusing.

A descriptive analysis of ecological model output compared to monitored data is a

valid method of assessment (Wood 2001). The assumptions required in the

development of ecological models also cause a mismatch between model output and

monitored data, therefore reducing the potential of statistical accuracy (Wood 2001).

Dambacher et al. (2003) suggest that an over-emphasis on precision in ecological

research is neither necessary nor essential for mathematical rigor. They argue that an

emphasis on statistics and precision may detract from a valuable qualitative

understanding of the system. Another view is that the pursuit of optimal quantitative

mathematical programs is not necessarily the primary concern of modellers. Beck

(1997) argues that rather than asking what numerical optimisation can do for us, we

should be asking how we can use our understanding of a system to successively

improve numerical optimisation. That is to say, an obsessive pursuit of optimal

numerical precision is not necessarily the role of a modeller. The successive

development of models or model parameters should be seen as a step forward in the

modelling process. The identification of good candidate models or equations may

assist in directing research, leading to the discovery of better future potential models

(Beck 1997).

Generic applicability

McIntosh et al. (2003) present the view that flexible and cost effective models are

more beneficial than one-off models, which perform very well for one ecosystem


30

only. In an extension of this concept, the applicability of a model as a testing tool in

wetlands where minimal or virtually no data exist, can expand the potential of

management understanding and thus advance the decision making process as well as

guide future research needs. Goodall (1972) and Rykiel (1996) make a distinction

between testing the adequacy of a model‟s predictions for a particular ecosystem, and

generalisation of its applicability to a range of ecosystems. A model might be

applicable and accurate for one particular ecosystem, however it may not be generic

and therefore applicable to other ecosystems (Rykiel 1996). This is clearly important

in determining whether predictions from a given model can be generally applied to

decision-making and management strategies in other ecosystems. WETMOD (Cetin

2001) sacrifices some quantitative accuracy for a few wetlands, in favour of

flexibility, applicability and cost. The point should be taken, as stated by Rykiel

(1996), that a model may accurately simulate the qualitative behaviour of the system

without the quantitative accuracy. In the development of WETMOD it was decided

that the qualitative assessment was the more appropriate assessment approach, so as

to maintain the generic applicability required in the region (Cetin 2001).

1.4.3 Validation

The necessity of validation is an issue that is the subject of considerable controversy

in the literature. To increase the understanding of the approach used for a given

project, and to potentially minimise conflict, a modeller should clearly state what the

validation criteria are in the context of the model. Modellers should also state any

restrictions and limitations of the application of the models. For the modeller to fulfil

this obligation, the purpose of the model, the criteria the model must meet to be

declared acceptable for use, and the context in which the model is intended to operate,

must be specified (Rykiel 1996).

Rykiel (1996) discusses that models should be judged on usefulness rather than

validity. However, model validation is required regardless of whether a model is

expected to produce quantitative or qualitative outputs. Model validation is also

important for end-user acceptance in the decision-making processes (Power 1993;

Rykiel 1996). Mayer and Butler (1993) relate validation to the potential application

and users of the model, where the validation is a comparison of model prediction to

real world monitoring, to determine whether the model is suitable for its intended


31

purpose (Mayer et al. 1993; Rykiel 1996). Rykiel (1996) states that a valid model is

one whose scientific or conceptual content is acceptable for its purpose. According to

Goodall (1972), validation is testing to determine the degree of agreement between a

model and the real system, that is, how good is the prediction, not whether it should

be accepted or rejected (Goodall 1972; Rykiel 1996). Caswell (1988) argues against

the case of validation being the decisive part of a successful model. His view is of the

role of a model in expanding understanding and contributing to knowledge in a

similar vein to experiments contributing to empirical problems.

Validation procedures range from general qualitative tests to highly restrictive

quantitative tests (Rykiel 1996). Rykiel (1996), Oreskes et al. (1994) and Tsang

(1991) all examine and discuss different validation concepts. Although a detailed

discussion of this topic is beyond the scope of this thesis, it is necessary to state how

the term “validation” is understood in the context of this project. WETMOD is

assessed on the basis of Rykiel‟s (1996) description of validation, where validation is

a test to confirm that the model is acceptable for its intended use and meets its

specified performance requirement (Rykiel 1996). For pragmatic purposes “a model

only needs to be good enough to accomplish the goals of the task to which it is

applied” (Rykiel 1996)

1.4.4 Modelling role in environmental decision-making

Scale

The study of ecological function and the management of natural resources have often

been at a local scale, even though the ecological processes within wetlands, streams,

and rivers occur at a larger (catchment) scale. One of the reasons for this local scale

approach has been an inability to manage and analyse large and complex data sets.

However, there has been a gradual recognition that management must be handled at

large spatial scales to obtain meaningful results (Crumpton 2001; Fitz et al. 1996;

Johnson et al. 1997). Fortunately, technology, spatial data, and software tools have

advanced to such an extent that landscape-scale studies are now feasible (Johnson et

al. 1997). As discussed in the previous chapter, to fully understand management

implications and evaluate options the full impacts of restoration of wetland functions

will need to be assessed on a landscape scale rather than at an ecosystem scale.


32

Individual wetlands, through the food web, provision of habitat and flood mitigation,

have an impact on surrounding wetlands, on the surrounding ecosystems and local

land use (Bedford et al. 1988). Without consideration of wetland processes at

watershed, landscape, and ecosystem scales, the most effective management strategies

cannot be assessed (Lemly 1997). That is, the spatial modelling of ecosystems is

necessary to develop a description of past behaviour, or to predict impacts caused by

alternative management strategies (Mitasova et al. 1998; Sklar et al. 1993), and their

impacts beyond their boundaries.

The benefit of landscape models is the ability to use them for the prediction of

management impacts on wetlands, without actual alteration or potential destruction

(Sklar et al. 1993). Spatial variation is important in assessing the response of a system

to excessive nutrient loads and the impact on the system (Murray 2001). Specifically,

landscape models can be used to study ecological principles, evaluate cumulative

impacts, mitigate environmental alterations, and prevent large-scale anthropogenic

mistakes from degrading wetland functions (Sklar et al. 1993). Models can also be

used to predict “missing” data that can further be used in management decisions (such

as flow exchange). Part of the strength of landscape models is the integration of

disciplines due to their ability to handle large amounts of data and information, and

provide output that is simple to convey (Boumans et al. 2001). Perhaps the major

advantage of landscape models in catchment management is their comprehensive and

systematic integration of knowledge and data for a specified region (Voinov et al.

1999a). Thereby a model user can be forced to view and interpret data normally not

considered.

Cumulative impacts

As mentioned above, environmental impacts have often been assessed in the past on

the local scale, and have not considered the broader scale impacts (Bedford et al.

1988). However, in a cumulative approach, the different external activities that impact

upon a study area are considered. Therefore, on a landscape scale cumulative impacts

from processes or activities external to the project area may become apparent that

otherwise were not apparent using a local scale approach.

The cumulative impacts can result from individually minor but collectively significant

actions taking place over a period of time (Preston et al. 1988). When assessing


33

cumulative impact, the impacts caused by external activities and projects set the

assessment boundaries i.e. the landscape scale (Bedford et al. 1988; Preston et al.

1988). Therefore, the area considered in cumulative assessment can expand from the

wetland scale to catchment or regions. Only by allowing all external activities and

processes that affect a wetland to determine the project boundary, can cumulative

impacts be monitored or measured (Bedford et al. 1988).

The ultimate aim of a cumulative impact assessment is to evaluate the impacts that

may result from change. These impacts include the physical, chemical, and biological

changes to an environment (Abbruzzese et al. 1997). The cumulative impact of

nutrient uptake due to management, whether improved or degraded, falls within the

scope of impact assessment of potential landscape scale wetland management

application. Accordingly, the cumulative impact observed due to the simulation of

multiple wetland management scenarios can be viewed as a cumulative impact

assessment of the proposed management strategies.

Models role in estimation of nutrient retention

Simulating nutrient flux within a river environment using models taking into account

pollution sources through to river outlets should be able to assist managers to target

intervention options for nutrient load reduction (de Wit 2001). There are models of

various complexity which attempt to provide this capability such as PolFlow by de

Wit (2001), which is based on physical laws and is embedded in a GIS (geographical

information system). As well as a model by Crumpton (2001) who attempt to identify

the position in the landscape of wetland restoration sites for optimal nutrient

(Nitrogen) removal. Peijl et al.(1999) developed a model that was able to describe the

carbon, nitrogen and phosphorus dynamics and interactions in riverine wetlands, and

Muhammetoglu et al. (1997) developed a dynamic three dimensional water quality

model for macrophyte dominated shallow lakes. An example of a simple spatial

wetland model which simulates nutrient retention of wetlands is described by Li et al.

(2002) and Li et al. (2003).

These models all try to simulate the nutrient retention capacity of wetlands and relate

this back to the landscape scale, e.g. downstream nutrient load. The model by

Crumpton (2001) attempts to direct management for an optimal return on investment,


34

i.e. by addressing the question of where in the landscape investment in a wetland

would deliver the greatest return as far as nutrient retention is concerned.

The PolFlow model was designed as a complex model. In following investigations de

Wit and Pebesma (2001) found that under given circumstances, where the available

data quality is the limiting factor in model development, simple models may in fact

provide as accurate model results as complex models. Crumpton (2001), who study

wetland restoration on a catchment scale, found through the application of their model

that the location of restored wetlands was decisive in the ability of the wetland to

effectively remove Nitrogen loads of the system. They showed that the interception of

nutrients by the wetland should be a focus by managers in deciding on wetland

restoration sites. Wetland managers lack the information to make any such decisions

for lower River Murray wetlands. Landscape scale assessment through modelling can

be used in a similar manner to Crumpton (2001), with due consideration given to the

availability of data for the lower River Murray catchment.

Peijl et al. (2000b), who investigated the importance of landscape geochemical flows

using a dynamic model, that simulated carbon, nitrogen and phosphorus cycling in

riverine wetlands, show an example of a wetland model that did not manage to predict

the field experiment. However this model did contribute to their understanding of the

system (Peijl et al. 2000a). This shows that a model can be counted as successful

simply based on the improvement of knowledge or understanding.

Muhammetoglu et al. (1997) developed a dynamic three-dimensional water quality

model for macrophyte dominated shallow lakes. The model simulates the interactions

between macrophytes and water quality parameters. The parameters, which are

considered, include dissolved oxygen (DO), organic nitrogen, ammonia, nitrate,

organic phosphorus, orthophosphate, biochemical oxygen demand, phytoplankton and

the sediment. The model has been successful in prediction compared to measured

values and can be used as a eutrophication management tool (Muhammetoglu et al.

1997).

The model developed and described by Li et al. (2002) and Li et al. (2003) is an

example of a simple wetland model which simulates nutrient retention of wetlands

and relates this to a landscape scale. Their stance is similar to that of research needs

identified for the lower River Murray wetlands model; in that the data availability and


35

knowledge of the system being modelled was limited, the model was therefore

impacted upon by a number of assumptions. As a consequence they opted for a simple

model. Their model outcomes are in some instances contrary to those anticipated. But

they point out that the trends displayed by the model are useful in guiding land use

planning (Li et al. 2002). Due to the simple structure of their model Li et al. (2002)

and Li et al. (2003) claim that it is applicable to other areas and therefore not location

specific. The model output, from which management recommendations are made are

only indicative of a trend (Li et al. 2003). This shows that in circumstances where

limited data is available, model scenarios of wetland nutrient retention can be used for

land use and other environmental management decision-making.

Sediment compaction

Sediment resuspension accounts for a large part of wetland turbidity influenced by

climatic factors. To study the impact within one wetland a model could be made to

account for wind direction and speed, and macrophytes role in sediment resuspension.

An example of a project which accounts for these influences is a study performed by

Hamilton and Mitchell (1996). The objective of the study performed by Hamilton and

Mitchell (1996) in shallow New Zealand lakes was to derive relationships between

suspended sediment concentrations and the physical forces caused by wave action,

and to quantify the factors responsible for the differences between a number of lakes.

They were successful in obtaining statistical evidence of the stabilisation mechanism

of sediments and the inhibition of resuspension posed by macrophytes. One of the

major causes degradation of the wetlands of the lower River Murray is their turbidity

as discussed previously. Climatic data on a regional scale could be an option to be

included in future monitoring studies of select wetlands thereby providing a

representative case for the region. Wetland specific issues such as vegetation cover

would also have to be included, complicating a landscape model. In the meantime a

case exists for the development of a model capable of simulating the impact turbidity

has on wetland ecosystem process and thereby assist managers in evaluating wetland

rehabilitation needed to achieve a set turbidity reduction target.

River Murray models

The Murray-Darling Basin Commission (MDBC) has been using computer models for

more than thirty years for water resources planning, development of operating rules,


36

development of salinity and drainage strategies and forecasting of flow and salinity

(Close 1996). The history of mathematical modelling for the MDBC to evaluate

management options dates back to 1902 (Close 1986). From 1965 a computer water

supply model was being used. Since then flow and salinity models have been created

and their interactions improved (Close 1986). In 1996 a model called BigMod was

taken up by the MDBC, which replaced the older models and had the role of salt

routing prediction in planning studies, short term flow and salinity forecasts,

calculating solute loads based on historical data and modelling daily flow variations

(Close 1996). Its role is to estimate electrical conductivity (EC) and track parcels of

water throughout the system so that salinity load can be defined anywhere in a reach

(Close 1996). The use of models by the MDBC has been a successful venture. Due to

the complexity of the Murray Darling Basin, and therefore the difficulty to qualify the

impacts of changes to the system and the impossibility of quantification without

modelling, the developed models have been extremely useful to the MDBC to aid in

management decisions (Close 1986).

The Flood Inundation Model (FIM) is based on historical flood inundation extent

extracted from satellite imagery, known flow at the border, flood levels and lock

levels (Overton 2000; Overton 2005). The FIM takes into consideration backwater

curves. It provides managers of lower River Murray assets, such as wetlands and

floodplains, with a tool to simulate potential inundation areas by changing lock levels

at given flows across the SA border (Overton 2000; Overton 2005). The model output

is for example used for assisting wetland management by simulating their inundation

at given flow levels and relating this back to a potential hydrological regime. The FIM

however identifies neither the flow paths connecting the wetlands and the river nor

the turnover rate (water volume exchange) within wetlands.

A salinity model was developed for The Department of Water, Land and Biodiversity

Conservation (DWLBC) to account for salinity impacts of wetlands on the lower

River Murray, i.e. salinity accounting (Murdoch et al. 2004; RMCWMB 2002). The

use of the model was intended provide a generic daily salt water balance as a

consequence of wetland hydrology regimes .This model Salinity Impacts of Wetland

Manipulation (SIWM) is a generic model relying on “exemplar” data and qualitative

outcomes for generating quantitative assessments. The hydrology estimations within

SIWM were taken from BigMod which propagates inaccuracies based on BigMod


37

assumptions. Consequently SIWM output quality is degraded (Murdoch et al. 2004).

Although a novel approach the accuracy of the model output was not adequate for the

purposes intended, i.e. salinity accounting for individual wetlands. It was therefore

withdrawn from use by DWLBC. A replacement model is currently being developed

by DWLBC (Croucher 2005). Further models for the lower River Murray are a

number of groundwater models. These models simulate groundwater sources and

impact on floodplain and river salinity and have been combined to make up a

Floodplain Risk Methodology (FRM). The FRM is a collection of models used to

assess the impact of groundwater on floodplain vegetation and the impacts periodic

flooding and weir manipulation would have (Holland et al. 2005).

As discussed, to fully understand management implications and evaluate options the

full impacts of restoration strategies on wetland functions will need to be assessed on

a landscape scale rather than at an ecosystem scale. This having been identified as a

research need for the wetland processes of the lower River Murray, a wetland

ecosystem model called WETMOD 1, initially developed by Cetin (2001), was

identified as a first step from which landscape scale could be built on. The aim of

WETMOD 1 was to simulate macrophytes, phytoplankton, zooplankton and nutrients

in the open water of wetlands. Cetin (2001) based the structure of WETMOD 1 on the

Patuxent Landscape Model (PLM) (see Maxwell et al. 1997; Voinov et al. 1999a;

Voinov et al. 1999b), and the lake ecosystem model SALMO (Simulation by means

of an Analytical Lake Model) (see Benndorf et al. 1982; Recknagel et al. 1982).

The PLM is a complex landscape scale model of aquatic ecosystems including

wetlands where ecosystem processes are simulated. This model allows simulations of

a catchment using detailed morphological data (digital elevation models DEM‟s) of

the catchment and wetlands as well as time series of point source nutrient inflow. The

model simulates an entire catchment using raster GIS data as the main driving

variables where the model is run for each cell at each time step, propagating the

results to the next cell for the next time step (Boumans 2001; Voinov et al. 1999b).

The complexity of the PLM particularly the requirement of a detailed DEM prevents

it from being applied in the lower River Murray system. An added complication for

the lower River Murray system for a linear model is the non linear nature of flows in

the lower River Murray. This can be through the reversing of flows “upstream” by

wind and the bypassing of the main channel through rapidly flowing anabranches


38

such as those of the Chowilla floodplain. Therefore, cell to cell modelling such as the

PLM methodology would not be easy to adapt or implement. Part of the PLM could

however be adapted to the lower River Murray system. Particularly when adapted

with equations such as from time series dependent models such as SALMO.

SALMO was designed for the management of lake ecosystems, based on state

variables phytoplankton, zooplankton and orthophosphate time series data. The

SALMO model allowed for management simulations of nutrient cycles within lakes

and the consequences of different management strategies for the control of

eutrophication in lake and reservoir ecosystems (Benndorf et al. 1982). Using select

equations from both as well as further literature Cetin (2001) was able to develop a

generic model (WETMOD 1) for simulation of internal nutrient dynamics. The

WETMOD 2 model described in the remaining chapters, built on WETMOD 1, is a

contribution to the simulation of the lower River Murray system to aid informed

decision making research and management of the lower River Murray wetlands.


39

2 Aims and objectives

The main aim of this work is to develop a model, which facilitates the analysis of

management options for informed selection of wetlands requiring restoration, with the

aim of re-establishing wetland landscape function through optimal means. To fulfil

this aim the following objectives must be met:

I. Adapt a generic wetland process model for the lower River Murray floodplain

wetlands and improve the resolution of the spatial influences acting upon a

wetland

II. Evaluate these spatially relevant impacts on wetland nutrient uptake

III. Appraise the potential river nutrient-load buffer capacity of wetlands both pre-

and post-management, on a landscape scale.

It is generally expected that restoring wetlands will reinstate their river-nutrient

buffering capacity, consequently improving water quality in rivers by reducing

nutrients otherwise available to support algal growth. The model is expected to

deliver an estimate of the potential nutrient reduction in the lower River Murray

following management intervention.

This project focuses on the lower River Murray wetlands and relies on previous work

done in that area. Some of the research in the Lower Murray area has focused data

collection and survey work, and has been summarised in the Wetlands Atlas of the

South Australian Murray Valley by Jensen et al. (1996). Other projects in the Murray

Darling Basin have been compiled and catalogued by the Murray Darling Basin

Commission (Kirk 1998). However, the work this project mostly depends on are

projects in the lower River Murray that have had objectives of producing solutions for

particular problems. These past projects include for example the creation of weirs at

individual wetlands for the introduction of drying cycles, and the construction of

wetlands for nutrient removal from agricultural drainage water (prior to being

released into the system). Recent baseline surveys (SKM 2004; SKM 2006) have

added to the information available on the condition of individual wetlands for the

purpose of wetland management. This data provides a simplified snapshot of the

current condition of a few wetlands. However, a key lack of data, which impacts on a


40

number of research projects (such as fish habitat) and management decisions, is the

exchange of water (turnover rate) between wetlands and the river.

Hypotheses:

I. A simplified generic wetland model can be used to realistically simulate

multiple and different wetlands qualitatively. It is the premise of this project

that a simple model will, with the available data, produce results that are

sufficiently accurate so as to aid in decision-making (see 1.4.1).

II. A simplified generic wetland model can be used to answer “what if” questions

for landscape scale scenarios, and

III. A simplified generic wetland model can be used to assess the cumulative

impact of managing multiple wetlands.

This project adopts a generic wetland process model WETMOD 1 to account for

wetland local external influences. These influences include improvement of the

resolution of spatial influences such as river nutrient content, river flow volume and

where appropriate external irrigation drainage inflow, which act upon a wetland. The

model will evaluate the impact these influences have on wetland uptake of PO4-P,

NO3-N, and production of phytoplankton, as well as how uptake can change at

different locations. To be able to apply the model at different locations, despite

restricted data availability, a wetland classification system incorporating the use of

monitored data from intensely studied wetlands as regional-scale exemplars will be

adopted. Therefore, the model will be applicable on a regional (landscape) scale

providing qualitative understanding of the cumulative impact of wetland management.


41

3 Materials and Methods

3.1 Model Description

For application of the model in decision-making, managers and stakeholders need to

understand the models purpose as understood by the developer, any assumptions

made during its development and the model structure (Bart 1995). The following

chapter describes the assumptions of wetland behaviour as well as the model

structure. The model used for the basis of extended development is WETMOD 1

developed by Cetin (2001), which was based mainly on literature data.

3.1.1 Design Considerations

Problems current for wetland management is the acute lack of awareness of impact of

management on a regional scale. Given that wetlands will have a varying nutrient

retention capacity depending on the turnover rate, i.e. longer turnover rate will allow

for more nutrients to be absorbed, finding an optimum turnover rate to maximise the

nutrient retention capacity of wetlands could be a management aim for river nutrient

reduction. To assess the impact of wetland management for nutrient reduction in a

river it is necessary to assess the capacity of a wetland to retain nutrients individually

and cumulatively at the landscape scale. Therefore, multiple variables come into play

to assess the capacity of wetlands to retain nutrients on a landscape scale.

The first step of assessing individual wetland nutrient retention was addressed in part

by WETMOD 1 (Cetin 2001). The limiting factors are, as is often the case, the acute

lack of sufficient data when the model is to be applied to a landscape scale. The

wetland model WETMOD 1 has the ability to simulate the general internal dynamics

of a wetland with minimal monitored driving variables, therefore allowing the model

to be applicable at sites with minimal data. With site-specific data on water exchange,

nutrient through flow and wetland morphology, introduced during the development of

WETMOD 2, the modelling of wetland dynamics becomes more specific for an

individual wetland although using landscape scale available data. None the less, with

the limited resources invested in the monitoring of wetlands, only very few can

reliably be simulated. To overcome the restriction hindering the testing of


42

management strategies for wetlands and assessing the potential cumulative impact

two options remained:

1. The substantial investment of resources in monitoring of wetlands, prior to the

extensive development of a model. Such a model would potentially be capable

of the simulation of each of the monitored wetlands in detail, thereby

providing managers with a robust and comprehensive decision making tool.

This option has the drawback of the investment of substantial resources, loss

of valuable time in monitoring and model development. The largest drawback

being that the model would still be restricted to the monitored wetlands.

2. The development of a modelling tool capable of assisting in the development

of understanding of potential management decisions, which would be based on

current available data and knowledge. The restriction on the complexity of

such a model would be ensuring its applicability to all wetlands within the

catchment area based on the current data availability. Therefore, as such a

model would need to rely on available data some of the accuracy of model

results would be dependent on the range of data quality and quantity.

Going with the second option, a developed generic model which allows the

assessment on a landscape scale of wetland function and cumulative impact, the

simplification of wetland into wetland classes becomes necessary, to such an extent

that no wetland is seen as unique nor all wetlands as equal (Bedford et al. 1988). If

this simplification is not introduced, the data required for a landscape scale

assessment becomes insurmountable.

There are multitudes of ways to classify wetlands. The system that is chosen is

dependent on the purpose of the classification, the time available, the data and the

knowledge available, as well as the preconception of the classifier, which will affect

any wetland classification. In a general sense, there are 2 approaches, one through

geomorphology and the other through the hydrological relationship of the wetland to

the river (Bedford et al. 1988; Pressey 1990).

As an example of a classification procedure Strager et al. (2000) used a landscape

based approach to classify wetlands and riparian areas based on habitat requirements

of amphibians and reptiles. This classification also included forested and non-forested

groupings (Strager et al. 2000).


43

The classification used in this project is partially driven by the limited data

availability for both geomorphology and hydrological relationship between the

wetland and river. The approach was therefore a very simplified hydrological

connectivity classification, which will be discussed in more detail below.

The description of the model is broken down into two segments, WETMOD 1 and

WETMOD 2. The first description, WETMOD 1, relates to the model sections

developed by Cetin (2001) that relate to internal nutrient dynamics. The second

section, WETMOD 2, relates to the redesign of the model to account for external

influences acting upon a wetland. The methodology for the application of WETMOD

2 to assess cumulative impact of wetland management is discussed at the end of this

chapter.

The macrophyte biomass module is described in the macrophyte sector below. The

phytoplankton and zooplankton biomass module are described together as part of the

plankton sector, and the PO4-P and NO3-N module is described as part of the nutrients

sector. The “Fitted River exchange and Irrigation Drainage Inflow” was, due to its

complexity, split into separate modules within WETMOD 2, which are described as

Flow Exchange Sector and External Nutrient Source Sector. Both of these relate to the

significant addition to the model where internal nutrient dynamics are related to

external and therefore landscape scale impacts such as river nutrient load. The output

of both of these sectors contributes significantly to management considerations on a

landscape scale. The sources of differential equations are described in Appendix A.

The descriptions of the macrophyte, plankton and nutrient sectors have been adapted

from Cetin (2001).

Units of input data (conversions are performed within the model, descriptions of

which can be found in section 3.3);

MDBC river data;

o Nitrate + Nitrite as N: mg/L

o Filterable Reactive Phosphorus as P: mg/L

Reedy Creek river data;

o Nitrate as NO3-N: mg/L

o Soluble Reactive Phosphorous as PO4-P: mg/L


44

Miscellaneous;

o Turbidity: NTU

o Temperature: C

o Secchi depth: metres

o Chlorophyll-a: μg/L

o Solar Radiation: MJm2

o Wetland Volume: cubic metres (m3)

Units of output data;

o Nutrients (PO4-P and NO3-N): g/L

o Phytoplankton biomass: cm3/m

3

o Zooplankton biomass: cm3/m

3

o Macrophyte biomass: kg/m3

Principal Model Assumptions (for simplification of model design)

The following assumptions were made at the commencement of the project to

compensate for a general lack of data, and data quality for the lower River Murray.

These were needed as part of the simple model design strategy employed.

1. As the considered wetlands are permanently inundated wetlands, it is assumed

that as a result of lock management, where locks are maintained at a constant

level, all wetlands included in potential management scenarios have a constant

volume as well as a permanent connection with the lower River Murray.

Consequently, there is a bi-directional and permanent exchange of water and

nutrient with the river, the exchange volumes (in- and out-flow) being equal.

2. There were no data on exchange flow or channel morphology; therefore it was

assumed that the exchange volume was solely dependent on the river flow

volume.

3. It was assumed the wetland is homogeneously mixed for each modelling time

step. Simulated wetland nutrient data would therefore represent the

concentration throughout the wetland.


45

4. South Australia is a dry state, and there generally are no significant catchment

areas for individual lower River Murray wetlands. It was therefore assumed,

that there would be only low or insignificant nutrient inflow though

precipitation runoff for most wetlands. The exception is Reedy Creek wetland,

and therefore by extrapolation, all category 4 wetlands (wetland classification

is described below).

5. For management simulation purposes, it was assumed that the introduction of

dry periods to wetlands would compact the sediment and reduce the turbidity

within the wetland during the next wetting event. The inherent assumption is

that the turbidity is caused by suspended sediment and is not significantly

contributed to by phytoplankton. However, phytoplankton will increase the

turbidity in proportion to its own growth. A future user of the model must

therefore take this into account when assessing model output. The

management scenarios also assume that the turbidity is independent of the

potential inflow of suspended sediment from the river as the river turbidity

fluctuates dependent on the water source of upper River Murray versus the

Darling River. A modeller must therefore take into account the realistic

potential reduction in turbidity for a given wetland dependent on external

sources as well as its internal dynamics, i.e. resuspension and sedimentation.

6. For management simulation purposes it was assumed that all same category

wetlands resemble each other in exchange volume. In an operational

application local knowledge of the exchange volume for simulated wetlands

would assist in improving potential modelling output.

3.1.2 WETMOD 1

The WETMOD 1 model (Cetin 2001; Cetin et al. 2001) is a generic wetland

ecosystem model. WETMOD 1 simulates internal wetland nutrient dynamics, i.e. the

growth of macrophytes, phytoplankton and zooplankton through mass balance

equations (Figure 4). WETMOD 1 simulates internal wetland nutrient processes using

water temperature, turbidity, Secchi depth and solar radiation as driving variables

(model time-series input). Phosphorus as PO4-P, nitrogen as NO3-N, macrophytes,

phytoplankton and zooplankton are state variables (model output). This section,

represented as WETMOD 1, was rigorously adapted into the WETMOD 2


46

environment to account for advancement in data and addressing model limitations.

The description in this thesis is therefore of the model sections, WETMOD 1, as they

are found in the WETMOD 2 model. Any generic wetland ecosystem model such as

WETMOD 1 could be adapted in a similar fashion to account for the river, wetland

interaction and cumulative impact on a landscape scale.

The Model WETMOD 1 was developed and implemented by means of the modelling

developmental software STELLA (2000). STELLA provides an intuitive user

interface for domain experts with little modelling experience. Models developed

within STELLA are, due to its rigid structure, transparent.

Figure 4: Driving Variables, State Variables and Major Interactions in WETMOD 1

Nutrient contribution to the wetland occurs through sediment release, surface runoff,

irrigation drainage and river inflow. The data used being either sourced from literature

or approximate values obtained from expert recommendation. Nutrient loss occurs

through uptake by macrophytes and phytoplankton as well as sedimentation and

wetland outflow, this data being mainly calibrated or based on expert

recommendation. Zooplankton increase is through growth, and reduction is through

mortality. Macrophyte and phytoplankton increase is through growth (phytoplankton

inflow being introduced in WETMOD 2). Macrophyte and phytoplankton biomass

loss is through respiration and mortality, phytoplankton additionally through

sedimentation, zooplankton grazing and outflow.

Zooplankton

M acrophytes

Phytoplankton

N O 3-N

PO 4-P

IN PU TS W ETLA N D D A TA

Turbidity

W ater Tem perature

Solar R adiation

Secchi D epth

W etland V olum e

N utrient Inflow (From

L iterature)

Sedim entation/O utflowD riving V ariables

M odel Interactions


47

The model is divided into individual modules where related process equations are

grouped together. The descriptions below are of the individual modules as they appear

in WETMOD 2.

Macrophyte Sector

The macrophytes are simulated within WETMOD 1, and represented by their

photosynthetic biomass in the open water (Cetin 2001). To obtain a simple model

structure, all submerged macrophyte species found within the wetland were

aggregated and represented as macrophyte biomass. Emergent macrophytes play an

important role in the lower River Murray ecosystem, not only through their nutrient

retention but also as habitat and sediment traps. WETMOD 1 however did not

consider emergent macrophytes in the model and consequently neither will

WETMOD 2.

The increase in macrophyte biomass within the model is controlled by the

productivity of the photosynthetic biomass („Mac Gross PP‟ Figure 5). The loss of

macrophyte biomass is through mortality and respiration of photosynthetic biomass

(„Mac mortality‟ and „Mac respiration‟).

The growth of macrophyte (photosynthetic) biomass is influenced by the growth rate,

turbidity and nutrients, underwater light and water temperature. In Australian waters

turbidity can be a controlling factor in macrophyte growth (Roberts et al. 1986), with

the growth rate reaching a maximum when the turbidity is below 70 NTU (Shiel et al.

1982). Therefore, the reduction in turbidity is seen as a major aim of wetland

management and is consequently the main focus of management scenarios of the

model.

Productivity of the photosynthetic biomass („Mac Gross PP‟), i.e. macrophyte

biomass growth, is contributed to by the macrophyte production coefficient („mac

prod cf‟), Gross Primary Production rate for the total plant biomass („Mac GPP‟) and

can be limited by turbidity if it surpasses the 70 NTU threshold. The production

coefficient („mac prod cf‟) is calculated from the availability of nutrients, underwater

light and water temperature (see Appendix A for equations).

The underwater light coefficient calculation is based on the Beer-Lambert Law for

light attenuation, where the data required is Secchi depth and solar radiation. Solar

radiation input is MJm2/day. The equation used in WETMOD 1, which was obtained


48

from literature, demands units in Jcm2/day, therefore, in WETMOD 2 MJm

2/day is

multiplied by 100 to convert to Jcm2/day. Where Secchi depth data are missing or of

very poor quality, the Secchi depth is calculated based on the turbidity within the

wetland. The equation for calculating the Secchi depth from the wetland turbidity is

discussed in detail in section 3.2.1. Therefore, the Secchi depth data source can be

either monitored, calculated from the turbidity or fixed manually.

The water temperature is one of the driving variables of WETMOD (1 and 2). The

macrophyte temperature coefficient („mac temp cf‟) is based on the optimum water

temperature for macrophyte growth (Boumans 2001). The macrophyte nutrient

coefficient („mac nut cf‟) is based on the Michaelis-Menten expression, where the

nutrient uptake is dependent on the concentration of the nutrient in the water and the

nutrient half saturation constant („mac Ks N‟ and „mac Ks P‟, see Appendix A) for

uptake.


49

Figure 5: Macrophyte Module

Plankton Sector

The plankton sector, seen in Figure 6, (labelled as phytoplankton module) comprises

both the phytoplankton and zooplankton equations within the model. Both

phytoplankton and zooplankton have, for the sake of model simplicity, been

aggregated into their respective state variable, i.e. phytoplankton biomass and

zooplankton biomass.

The phytoplankton biomass can have two sources of input. One is the wetland growth

of phytoplankton expressed as the phytoplankton gross primary productivity („pht


50

Gross PP‟), which is dependent on the phytoplankton production coefficient („pht

prod cf‟), and limited by the maximum biomass of phytoplankton („pht max‟), i.e. the

carrying capacity which is calibrated for each wetland. The phytoplankton production

coefficient („pht prod cf‟) is obtained in a similar manner to the macrophyte sector.

The second input source of phytoplankton is from external sources such as the river or

irrigation drainage inflow („Phytoplankton In‟), the load is fitted with the exchange

estimate; see External Nutrient Source sector description below.

The phytoplankton biomass reduction is through five sources, mortality („pht

mortality‟), respiration („pht respiration‟), sedimentation („pht sedimentation‟)

outflow („Phytoplankton Out‟) and zooplankton uptake („Pht grazing‟). The

phytoplankton respiration is dependent on water temperature and the temperature

limitation coefficient („pht temp cf‟). The phytoplankton temperature coefficient

equation is adapted from Hamilton and Schadlow (1997), which relates the growth to

a constant multiplier of the water temperature.

Within wetlands there is a net increase in effective sedimentation as a consequence of

increasing turbidity. That is, due to the increase in suspended particles available there

is a net increase in sedimentation as when compared with a clear wetland. The

sedimentation is controlled through a calibrated sedimentation rate, which is altered

by the turbidity of the wetland. In WETMOD the change in sedimentation rate is

controlled through a calibrated turbidity threshold at 95 NTU. Mortality of

phytoplankton is controlled through a set mortality rate and the respiration through a

respiration rate. The outflow is dependent on the fitted (estimated) exchange rate of

the wetland, described in Flow Exchange and External Nutrient Source Sectors. The

zooplankton uptake is dependent on the zooplankton growth rate controlled through

the grazing rate of phytoplankton by zooplankton („Pht grazing‟).

The zooplankton equations are adapted exclusively from SALMO (Simulation by

means of an Analytical Lake Model) (Recknagel et al. 1982). The zooplankton

outflow is simplified with the zooplankton mortality, controlled through a calibrated

mortality rate for each wetland, accounting for all sources of zooplankton biomass

reduction. The phytoplankton biomass growth is controlled through the phytoplankton

grazing equation („Pht grazing‟), which is the grazing of phytoplankton by

zooplankton. Affecting the grazing, and therefore zooplankton growth, is the

zooplankton respiration rate and the zooplankton growth rate. The zooplankton


51

growth rate is a function of both the macrophyte biomass and the phytoplankton

grazing rate. The phytoplankton grazing rate is a function of the day length and the

water temperature. The macrophyte biomass has an influence on the zooplankton

growth rate due the assumption that it provides a shelter for zooplankton (Asaeda et

al. 1997). Therefore, if the macrophyte biomass is low, the zooplankton biomass will

reduce. The zooplankton respiration rate is controlled by the phytoplankton grazing

rate and the water temperature.

Figure 6: Plankton Module

Nutrients Sector

Both of the nutrient equations consist of similar inflows and outflows, Figure 7. As

discussed in section 1.1.2 and 1.2 the main contributors of nutrient inflow to wetlands


52

are external sources such as the river or irrigation drainage inflow. As with

phytoplankton, the inflow rate is determined by the fitted rate for the particular

wetland, which is described in Flow Exchange and External Nutrient Source Sectors.

Other inflows include „P loading‟ and „N loading‟ respectively, as well as „P

sediment‟ and „N sediment‟ release. Nitrate flux is also potentially affected by

nitrification and denitrification. However, due to insufficiencies in data, the sediment

dynamics could not be modelled within WETMOD. The nutrient dynamics of the

wetland are for the open water only, with sedimentation rates calibrated to adjust for

missing complexity. This has simplified the model, but future research may need to

invest in expanding this section of the model despite increasing complexity, as the

present simplification does account for some model limitation.

The outflows include „P soil coprecipitation‟, or the sedimentation of PO4-P and NO3-

N „N soil coprecipitation‟, P or N uptake and nutrient outflow as per the fitted

exchange estimate, described in Flow Exchange and External Nutrient Source Sectors.

The sedimentation rate accounts for the coprecipitation of nutrients, (which is the

sorption of nutrients to suspended soil particles that then precipitate to the wetland

floor). The coprecipitation is more pronounced at high turbidity due to the high

availability of suspended soil particles, and can account for significant nutrient uptake

by wetlands. The model assumes a calibrated sedimentation rate (calibrated for

wetland categories) for both PO4-P and NO3-N of 50% at turbidity levels 70 NTU or

above (the 70 NTU being a calibrated estimate that acts as a threshold), and 10%

below 70 NTU (for Lock 6 wetland) or 50% vs. 20% (for Reedy Creek wetland),

wetland classification is discussed below.

The uptake of nutrients by macrophytes and phytoplankton was adapted from the

PLM (Patuxent Landscape model) (Boumans 2001). The uptake is dependent on

nutrient to carbon ratio and the net primary productivity of both macrophyte and

phytoplankton biomass.


53

Figure 7: Nutrient Module


54

3.1.3 WETMOD 2

External influences were not well constructed, or represented, in WETMOD 1,

therefore to study the impact of respective external influences for differing wetlands

these need to be added to the model. WETMOD 2 is a modification of the original

WETMOD 1 model to suit the requirements of this project. As accounting for external

influences is essential in regional scale scenarios, a modified version of WETMOD

had to be developed, i.e. WETMOD 2. The capacity of the model to simulate potential

impacts of restoration scenarios for the different wetlands would, through improved

local relevant data, also be enhanced. The first two aims in the modifications of

WETMOD 2 playing a part in fulfilling the third. These aims are listed below.

I. Overcome shortcomings in knowledge due to limited data and incomplete

system understanding.

II. Address processes requiring further development, which were identified at the

beginning of the study. These included river and wetland water exchange,

nutrient exchange, and irrigation drainage data influence, and

III. Adapt and test the application of the model on a regional scale; i.e. develop a

cumulative assessment of potential management impacts of multiple wetlands

on the river nutrient load.

One challenge in modifying WETMOD to simulate landscape scale scenarios was to

account for external influences acting on wetland water quality. This involved the

further development of estimates of the inflow and outflow of nutrient to the

wetlands. To accomplish this, the following sources were included in the model:

I. Irrigation (Reedy Creek wetland and Sunnyside wetland)

II. River Exchange modelling (Lower Murray River flow dependent)

During these key WETMOD modifications the following two principles were

maintained:

I. Model transferability (Model Generic Nature)

II. Model Expectations (Reliable prediction of system dynamics, i.e. trends)

The model overview, WETMOD 2, and the data flow between modules are presented

in Figure 8. The sections initially sourced from WETMOD 1 are represented as the


55

three green modules where internal wetland processes are simulated. The newer

modules, which encompass the major modifications of WETMOD 2, include wetland

data updates (yellow, rigorous reconstruction and update of the data base), new

wetland specific morphology data (white) and wetland external nutrient sources

(blue).


56

Figure 8: WETMOD 2 Structure and Data Flow

W ETM O D Structure and D ata F low

M acrophyte Sector

W etland B iom ass Load

M acrophytes

Plankton Sector

W etland B iom ass Load

Zooplankton

Phytoplankton

D ata

Solar R adiation

W etland

T im es-Series

W etland Tim e Series

T urbidity, Secchi &

T em perature

R iver T im e-Series

River Tim e Series at Lock 5, M annum

and M urray Bridge

PO4-P , N O

3-N &

Phytoplankton

Locks 1-8

Flow V olum e

Fitted R iver

exchange and

Irrigation D rainage

Inflow

D rainage Inflow

River N utrient Exchange:

PO4-P , N O

3-N ,

Phytoplankton Inflow

Irrigation D rainage

T im e-Series

D rainage Inflow

PO 4-P , N O 3-N

Phytoplankton

D rainage V olum e

M orphological

D ata

W etland A rea

W etland D epth

N utrients Sector

W etland C oncentration

PO4-P & N O

3-N

M acrophyte

B iom ass

PO4-P & N O

3-N W etland V olum e

PO4-P & N O

3-N

PO4-P & N O

3-N Inflow

(including Irrigation D rainage)Solar R adiation

T urbidity, Secchi &

T em perature

PO4-P , N O

3-N , Phytoplankton &

D rainage V olum e

PO4-P , N O

3-N ,

Phytoplankton

&

Flow V olum e

Phytoplankton

O utflow

PO4-P & N O

3-N

O utflow

D ata Flow

Feedback on V olum e

(Fitted for each w etland)

External

Sources

W ET M O D 2

W etland Internal

D ynam ics

W ET M O D 1

W etland Specific

D ata

W ET M O D 1

W etland Specific

D ata

W ET M O D 2

Phytoplankton

Inflow

(including Irrigation

D rainage)


57

Flow Exchange Sector

As discussed in section 1.1.2 the transport of material in and out of wetlands is

primarily a function of water flow (Johnston 1991). Therefore, the transport of

material through water flow has the potential of being the most significant external

influence acting upon a wetland. The process of fitting the exchange volume to a

particular wetland based on Equation 6 which is discussed in detail in section 3.3.

Essentially Equation 6 relates the exchange volume to the wetland nutrient

concentration, river nutrient concentration and the river flow rate. This sector shows

the adjustment of exchange volume estimation (Percentage of River Flow regarded as

exchange) based on the river flow ML per day (converted to appropriate units as

required within the model) (Figure 9). As the wetland is assumed to maintain a

constant volume, any irrigation drainage inflow into the wetland is included in the

respective wetland outflow volume.


58

Figure 9: Volume Exchange Module


59

External Nutrient Source Sector

This sector encompasses the Nutrient Exchange module, as seen in Figure 10, where

both the river exchange and the irrigation drainage inflow are introduced. That is, the

inflow load of the nutrients and phytoplankton can be from two sources. The first is

the river and the second, when applicable, the irrigation drainage inflow. The

calculation of the individual loads is discussed in section 3.4 and calculated as per

Equation 9 (irrigation drainage load to the wetland) and Equation 10 (nutrient load to

the wetland from the river). Both are described in section 3.4.1. The sum of both loads

is fed into the relevant modules described in the Plankton Sector and Nutrients Sector.

Within the nutrient exchange sector the irrigation drainage concentration and volume

are selected and adjusted based on the wetland being simulated. Time-series for both

irrigation affected wetlands Sunnyside and Reedy Creek wetland (described in section

3.2.1) are selected if either is being simulated; the option of testing for irrigation

drainage affecting Paiwalla wetland was included in the model as it was also

potentially impacted by irrigation drainage. The irrigation flow volume is manually

set for Sunnyside as accurate volume data were not available, see section 3.2.1. Reedy

Creek wetland irrigation flow was fixed at a set volume. The calculated irrigation

drainage load („PDrainLoad‟, „NDrainLoad‟ and „Chla DrainLoad (Reedy or

Sunnyside)‟, see Figure 10) is distributed for each of the wetlands („P Drain Water

Inflow‟, „N Drain Water Inflow‟ or „(Reedy or Sunnyside) Chla divided into

wetland‟) as per the seasonal flow pattern („Seasonal Flow Pattern SunnyORReedy‟)

described in section 3.2.1. The methodology of conversion of Chlorophyll-a to

phytoplankton biomass is discussed in section 3.3, and performed within the model in

„Phytoplankton Inflow cm3m3‟.

The outflow module, Figure 11, is where the outflow of PO4-P, NO3-N and

phytoplankton are calculated based on the fitted exchange volumes from the Flow

Exchange Sector, expressed in terms of Equation 11 or Equation 12 (both equations

for estimating the nutrient load from a wetland to the river, Equation 12 taking into

account irrigation drainage, see section 3.4.1). These outflow concentrations are then

fed back to the relevant modules described in Plankton and Nutrient Sectors.


60

Figure 10: External Nutrient Module


61

Figure 11: Outflow Module

Miscellaneous Sectors

Other modules (Sectors) within the model contain data handling such as data source

selection (including driving variables based on wetland category), wetland volume

calculation, and appropriate solar radiation and river data selection. In certain

circumstances data conversions between units are handled within these modules e.g.

river data conversion.

Model accuracy is tested using MS Excel based on the evaluation criterion D

described in section 3.3.2 (statistical estimation of the accuracy of model output in

comparison to monitored data). Excel was also used to calculate the retention capacity

of the wetland, the potential impact that this retention capacity has on the river

nutrient load, the potential impact of management scenarios (Equation 13, Equation

14, Equation 16 and Equation 17), and the cumulative impact of multiple wetland

management. Equation 13 and Equation 14 both relate to the change in nutrient


62

outflow from a wetland to the river. They are both described in section 3.4.1.

Equation 16 and Equation 17 calculate the change in river nutrient load following

wetland management and its percentage change in river nutrient load respectively.

Both Equation 16 and Equation 17 are used to calculate the impact the management of

a wetland or multiple wetlands has on the river nutrient load (they are both described

in section 3.4.2).


63

3.2 Data: Model Driving Variables

WETMOD 1 was developed to study the impacts of internal wetland nutrient

dynamics. At the time of its development, most of the external nutrient inflow data

had not yet been collated, nor was all the wetland data available. Literature and other

“exemplar” data were used to supplement the datasets used as the driving variables of

the model. A working model was therefore developed which could be improved

through the introduction of appropriate real data.

As this (WETMOD 2) project was to focus on regional scale applications of wetland

models and to develop scenarios to study the potential impact of management,

WETMOD 1 was deemed to be an acceptable basis from which to continue

development. By using WETMOD 1, the time otherwise invested in redeveloping an

internal wetland nutrient dynamic model was saved. However, rigorous data pre-

processing was required to adapt WETMOD 1 to both the regional data set

requirements and the updated data set. This describes the data set used and it‟s pre-

processing for WETMOD 2.

This project used several different monitoring databases as summarised in Table 1.

Processes and conditions within wetlands have been monitored in several studies

focusing on select wetlands. These studies occurred between 1997 and 2001 for

periods ranging from 9 months to 2 years. Data were collected approximately every

two weeks (Bartsch 1997; Marsh 1997; van der Wielen nd; Wen 2002a; Wen 2002b).

These case studies are the source for all type-specific wetland properties as well as

most abiotic and biotic time-series that are used as the main driving variables in the

model, i.e. “exemplar” data. The monitored wetland locations, used in the modelling,

were from open water sampling in the centre of the wetland. Error Bars in the data

graphs displayed below were calculated from all monitoring sites within a particular

wetland.

The exchange of nutrients between the river and wetlands depends on the river flow

and river nutrient load. River flow data and water quality data (nutrient load) are

collected at all locks and were obtained from the Murray Darling Basin Commission

(MDBC) and the South Australian Department of Environment and Heritage (DEH).

The flow data included in the model were collected at Locks 1 through to 8 (Figure

12), therefore for the model the most appropriate river data can be chosen for a given


64

scenario. The main climatic driving variable is solar radiation data obtained from

BOM (Bureau of Meteorology).

To apply the model to all wetlands along the river, location specific data have been

incorporated. These include wetland size, depth, influence of irrigation drainage and

connection to the river; and were obtained from Planning SA and Wetland Care

Australia. From this morphological data the wetland could be assigned to categories,

depending on hydrology and irrigation drainage influence.

Table 1: Data Sources, Type & Monitoring Frequency

Data Type Monitoring Frequency

Data Included Source

Wetland, Drainage Inflow & River (water quality)

Fortnightly NO3, PO4, Turbidity, Temperature, Chl-a & Secchi depth

University of Adelaide

River Monitoring

Weekly Temperature & Turbidity

DEH

Fortnightly Chl-a, DEH

Monthly PO4 & NO3 DEH

River Flow Volume Daily Water Flow MDBC

Solar Radiation Daily Solar Radiation BOM

Wetlands Management Study Report

N/a Wetland Depth Wetland Care Australia


65

3.2.1 “Exemplar” Wetland Sites

The study area, which contains the modelled wetlands, covers a length of the River

Murray of just over 600 km from the South Australian and Victorian border to the

entry of the river into Lake Alexandrina (Figure 12).

Figure 12: “Exemplar” Wetlands & River Monitoring Sites

The overall purpose of the model is to simulate as many wetlands as possible along

the lower River Murray in order to identify management strategies that may

potentially improve wetland state and function. A range of different wetlands have


66

been studied in the past. Of these, selected wetlands that best represent the range of

wetlands, based on hydrological connections, act as “exemplars” of driving variable

data time-series used in management simulations. The assumption is that if physically

similar wetlands respond in the way “exemplar” wetlands do it will be possible to

expand the model application and simulate the cumulative impact of multiple

management intervention. The wetlands for which data was available and serve as

“exemplar” data sources are Paiwalla, Sunnyside, Lock 6, Reedy Creek and Pilby

Creek wetlands. Their locations within the lower River Murray catchment are shown

in Figure 12.

The classification of wetlands is based on the hydrological relationship of the

wetlands to the river. This has been divided into 2 basic types, 1. through-flow

wetlands and 2. dead-end river connections. Through-flow wetlands occur where river

water can flow through the wetland, either as the wetland has one complete side open

to the river or the wetland has two distinct flow channels acting as water inflow and

outflow channels. Dead-end river connected wetlands occur where the river water

flows in and out at the only available flow channel in the wetland (i.e. one channel

only connects the wetland to the river). Both of these have furthermore been divided

into the following two categories, permanent inundation (with carp presence) and

permanent inundation (with-out carp presence) as well as being influenced by

agricultural drainage. In the fifth category are the managed wetlands, which are

controlled through drying and wetting cycles. These managed wetlands could be made

of either through flow or dead end wetlands, in both cases carp restriction would be

built into the wetland control barriers to restrict large carp from entering during re-

flooding of the wetland and potentially disturbing the sediment. This study addressed

five categories of wetlands as follows:

Category 1, Through flow, permanent inundation (Paiwalla wetland)

Category 2, Through flow, permanent inundation & irrigation drainage

(Sunnyside wetland)

Category 3, Dead end, permanent inundation (Lock 6 wetland)

Category 4, Dead end, permanent inundation & irrigation drainage (Reedy

Creek wetland)


67

Category 5, Managed - Dry periods & carp restriction (Pilby Creek wetland, in

this case a dead end wetland)

Category 1: Through flow wetlands permanent inundation and no irrigation drainage

(Paiwalla wetland)

Paiwalla and Sunnyside wetlands are situated approximately 14 km North of Murray

Bridge in the lower River Murray region. Prior to swamp reclamation, the two

wetlands were a part of the same riparian wetland. Sellicks swamp was reclaimed in

1967 (Bartsch 1997) and was until recently used as irrigated dairy pasture. Due to the

nature of the reclamation, the retired pasture area was situated lower than the average

pool height of the River Murray (Philcox 1997). Both water seepage from the river

and irrigation, necessitated the construction of drainage channels within the reclaimed

area to remove excess water and prevent water logging. The collected irrigation

drainage water was pumped into the Sunnyside wetland.

Paiwalla wetland is situated directly north or upstream of Sellicks swamp (Figure 13).

For the purpose of this study, as in Bartsch (1997), it was assumed that Paiwalla

wetland was not influenced by irrigation drainage discharge. This assumption was

justified by Paiwalla being upstream of Sellicks swamp and did not receive direct

irrigation drainage through active pumping. Paiwalla acts as an “exemplar” of

category 1 wetlands; permanently inundated through flow wetlands with no irrigation

drainage.

Category 2: Through flow wetlands with irrigation drainage (Sunnyside wetland)

Sunnyside is south of and downstream from Sellicks swamp (Figure 13). Like

Paiwalla wetland, Sunnyside was considered to be a through flow wetland, the main

difference between the two wetlands being the influence of Sellicks swamp irrigation

drainage outlet that flowed directly into the northeast corner of Sunnyside. Sunnyside

was used in the study as an “exemplar” for category 2 wetlands; through flow

permanently inundated wetlands with irrigation drainage.


68

Figure 13: Paiwalla & Sunnyside wetlands

Category 3: Dead end wetlands with no irrigation drainage (Lock 6 wetland)

Lock 6 wetland (Figure 14) is a dead end wetland situated immediately upstream of

Lock 6 in the Riverland region of the River Murray. Due to the controlled and

constantly maintained volume of Lock 6, the wetland is permanently inundated. As

with all unmanaged wetlands directly connected with the lower River Murray, there is

carp presence potentially contributing to resuspension of sediment and therefore

wetland turbidity. There is no irrigation drainage directly affecting this wetland.

Permanent inundation and high turbidity levels have led to a reduction in macrophyte

growth and therefore nutrient uptake. Lock 6 wetland is therefore, considered to be in

a degraded state, with an increased possibility of blue green algae growth (Blindow et

al. 1993; Boon et al. 1997; Scheffer 1998; Scheffer et al. 1993; Stephen et al. 1998),

see section 1.2.

Lock 6 wetland was used in the modelling project as an “exemplar” for category 3

wetlands; dead end wetlands with no irrigation drainage. It served, in the modelling of

potential management strategies (described in section 3.4), as a prime example of a


69

wetland that has the potential of being improved through management. The

management considered in modelling scenarios was the construction of a wetland

weir, as found in neighbouring Pilby Creek wetland, for the introduction of dry

periods.

Figure 14: Lock 6 and Pilby Creek wetlands

Category 4: Dead end wetlands with irrigation drainage (Reedy Creek wetland)

Reedy Creek wetland (Figure 15) is permanently inundated and situated

approximately 6 km south of Mannum in the lower River Murray region. It is

influenced by irrigation drainage runoff from surrounding agricultural areas. Intensive

monitoring of this wetland over a 2-year period provided data for wetland internal

nutrient dynamics, river nutrient load and irrigation drainage from Basby farm. It was

used in the project as an “exemplar” for category 4 wetlands; dead end permanently

inundated wetlands with irrigation drainage. The management strategies employed in

simulations for this wetland (described in section 3.4) are based on nutrient reduction

of irrigation drainage using constructed wetlands.


70

Figure 15: Reedy Creek wetland

Category 5: Dead end wetlands, managed through implementation of dry periods

followed by large carp exclusion and no irrigation drainage (Pilby Creek wetland)

Pilby Creek wetland is found directly north of Lock 6 wetland (Figure 14). A minor

through flow creek “Pilby creek” feeds into the wetland at the northern end. As this

creek feeds in and out at one point of the wetland only, Pilby Creek wetland is

considered to be a dead end wetland with no through flow (any wetland managed

wetland is considered to fall within this category for the purpose of this project

although it is recognised that through flow wetlands can also be managed). There is

no irrigation drainage considered to influence Pilby Creek wetland. The introduction

of a control structure and the consequent management with dry periods has dried and

compacted the sediment and returned the wetland to a clear stable state (discussed in

section 1.2.2). A further advantage of the management has been the exclusion of large

carp though screening off of the inflow channel. The potential re-suspension of

sediment by bottom feeding carp has therefore been reduced.


71

Pilby Creek wetland was used in the model to simulate an ideal target condition

wetland, which is considered to be in a natural, clear, non-degraded, stable state. Pilby

creek was used as an “exemplar” for category 5 wetlands; dead end wetlands

managed through implementation of dry periods with carp restriction and no irrigation

drainage.

3.2.2 Wetland Data

The data presented in this chapter were used in developing the model as well as

serving as data “exemplars” for each wetland category. The main driving variables of

the model are turbidity, water temperature, solar radiation, Secchi depth and the

morphological data; wetland volume and surface area. Spatially relevant driving

variables include external sources of the nutrients Nitrate (as NO3-N), Soluble

Reactive Phosphorous (as PO4-P) and phytoplankton, the external sources being river

exchange; and if applicable irrigation drainage. Additional monitoring time-series of

wetlands, not used in WETMOD 2 development, were used for validation and

confirmation. The validation data were prepared in the same manner as the driving

variable data as described below.

Time Series of Wetland Physical Condition

One of the key driving variables is wetland turbidity, which affects PO4-P and NO3-N

sedimentation and re-suspension, as well as macrophyte and phytoplankton growth.

The turbidity time-series are provided in Figure 16A, D and G. Most of the wetland

data was monitored in 1997 however, Reedy Creek wetland (category 4) was

monitored between 20/10/1999 and 16/09/2001, and represents the most complete and

reliable study in the database.

Wetland water temperature data can be seen in Figure 16B, E and H. This driving

variable affects zooplankton and phytoplankton growth, grazing and mortality, and

macrophyte growth.

The Secchi depth is another driving variable required for the modelling of macrophyte

growth. Secchi depth was not monitored constantly for either category 1 & 2

(Paiwalla & Sunnyside) wetlands, but assumed to be constant at 0.7 metres due to the

wetland depth. In Reedy Creek wetland, a turbid wetland, the Secchi depth was

assumed to be constant at 0.2 metres. The Secchi depth for Pilby Creek wetland,


72

being in a stable clear state where the bottom could be observed, was assumed to be at

a constant 1.8 metres. The Secchi depth for Loch 6 was considered to be variable and

was therefore calculated from turbidity data. Equation 1 was used to calculate Secchi

depth from turbidity data and was derived from the power regression of Secchi data

versus turbidity data from van der Wielen‟s time-series (van der Wielen nd), where

the only reliable monitoring of both had been undertaken. The R2 of the power

regression was 0.7748.

Equation 1: -0.5675^4355.2 TurbiditySecchi


73

Figure 16: Wetlands (Categories 1 to 5) Driving Variables Turbidity, Water Temperature & Solar Radiation (see also in Appendix B)

T urbidity

-5 0

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

NT

U

Wate r T e mpe rature

0

5

1 0

1 5

2 0

2 5

3 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

de

g C

Solar R adiation R e e dy C re e k We tland

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

MJ

pe

r s

qu

are

me

ter

T urbidity

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

1 4 0

1 6 0

1 8 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

NT

U


0

5

1 0

1 5

2 0

2 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

de

g C

T urbidity

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

NT

U


0

5

1 0

1 5

2 0

2 5

3 0

3 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

de

g C

Solar R adiation Paiwalla & Sunnyside We tlands

0

5

1 0

1 5

2 0

2 5

3 0

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

MJ

pe

r s

qu

are

me

ter

Solar R adiation P ilby C re e k & Lock 6 We tlands

0

5

1 0

1 5

2 0

2 5

3 0

3 5

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

MJ

pe

r s

qu

are

me

ter

A

B

C

D

E

F

G

H

I

P aiwalla W etland 1997 S unnyside W etland 1997 Reedy C reek W etland 2000-2001Lock 6 wetland 1997 P ilby C reek W etland 1997


74

Climatic Time Series Solar Radiation

The solar radiation data used in WETMOD 1 were obtained from literature (Bowles et

al. 1979; Cetin 2001). This literature data were adequate in the early development of

the model. However, the source area of the radiation is somewhat remote from the

lower River Murray and did not provide the model with reliable daily values. A CD

containing solar radiation data was obtained from the Bureau of Meteorology (Forgan

2001).

Solar radiation time-series were obtained from BOM solar data, as derived from the

processing of Japanese Geostationary Meteorological Satellite (GMS) imagery. The

data is essentially exposure data from Meteorological Satellite Imagery collected

daily. Data for any given location is obtained for the pixel encompassing the given

area and is not interpolated. The resolution of each pixel is between 6x6 to 24x24 km

(Forgan 2001). The BOM model calculated the surface insolation (solar radiation)

from the measured upward solar radiation measured by the Visible and Infrared Spin

Scan Radiometer (VISSR) taking into account atmospheric influences such as the

absorption by water vapour and ozone, cloud reflection and absorption. Effectively

the solar radiation is modelled for hourly images from which a daily total is derived.

For a detailed account of the model used to calculate the solar radiation refer to

(Weymouth et al. 1994).

Figure 16C, F and I show the solar radiation used as driving variables in the model.

Solar radiation is used in the model to calculate macrophyte and phytoplankton

productivity. Unfortunately, no data were available for the period between February

1994 and July 1997, which is the period that Paiwalla, Sunnyside, Pilby and Lock 6

wetlands were monitored. However, South Australia is a very dry State with minimal

cloud cover; therefore the seasonal pattern of the solar radiation for 1998 is similar to

what would be expected for 1997. The intensity of the solar radiation, which impacts

on macrophyte and phytoplankton biomass growth, follows such a seasonal pattern. It

was found during simulation test runs of WETMOD 2 that slight variation in the solar

radiation time-series does not have a noticeable impact on the simulation output. As

the use of 1998 solar radiation pattern is assumed to have minimal impact on the

modelling accuracy, the Solar Radiation for 1998 is used in WETMOD 2 for the

simulation of Paiwalla, Sunnyside, Pilby and Lock 6 wetlands. The solar radiation


75

data were available for the period where Reedy Creek wetland was monitored, and

was used accordingly. The solar radiation at two locations, one at either end of the

study area, was adopted into the model. Simulation of either solar radiation positions

did not alter the modelled output significantly. Solar radiation from the northern end

of the study area was therefore used for all modelling scenarios, as this contained the

least amount of missing daily values and therefore represented the most complete

seasonal range of solar radiation.

Wetland Morphology - Spatial Data

In WETMOD 1, subjective estimated wetland volume was used in each wetland

simulation. One modification for WETMOD 2 was to use a more correct wetland

volume during wetland scenario modelling. The wetland volumes for all wetlands that

can be potentially simulated by WETMOD 2 were obtained or estimated using the

wetland surface area multiplied by the wetland depth.

The GIS data covering the wetlands and the lower River Murray and Locks were used

for a number of data extractions. These GIS data reflect the wetlands as shown in the

“Wetlands Atlas of the South Australian Murray Valley” (Jensen et al. 1996). The

data extracted, related to wetland morphology (surface area, depth and river

connection), as well as the geographical position of the wetland in relation to the

river. The wetlands data sets, “Locks”, and “lower River Murray”, were also used in

determination of regional scale scenarios.

Wetland volume was used in the model to calculate nutrient concentration as well as

the nutrient and water exchange capacity of the wetland. Therefore, relatively

accurate wetland volume estimation was required. As no DEM‟s were available the

surface area in conjunction with the wetland depth provided the necessary wetland

volume estimation. The surface areas of the wetlands were obtained from the digitised

version of the SA Wetland Atlas. The “Wetlands Management Study report” (Nichols

1998) surveyed many of the lower River Murray wetlands, and contains some data

relating to average wetland depth. Many wetlands in the lower River Murray are

regular in depth (Recknagel nd; van der Wielen nd), it therefore seemed justified,

given the lack of better data, to assume each wetland to be a basin of uniform depth

and the “Wetlands Management Study report” (Nichols 1998) depth data used in the

model.


76

Although many of the wetlands described in the “Wetlands Management Study

report” (Nichols 1998) had an estimate of their average depth, some did not have

quantitative data for depth, and were simply referred to as shallow, deep, or unknown.

Given that the model needs quantitative data for depth, assumptions had to be made

regarding descriptive terms. For all the wetlands described as:

shallow a depth of 0.3 metres was used in the model,

deep a depth of 2 metres was used, and

“unknown” an average value of 0.92 metres was used.

This last figure was calculated from the actual wetland depths presented in the

“Wetlands Management Study report” (Nichols 1998).

The wetland volume was calculated using Equation 2, the results, for “exemplar”

category wetlands only, are presented in Table 2. The wetland volume was used in the

nutrient sector of the model (section 3.1.2) and the nutrient exchange sector (section

3.1.2).

Equation 2: DepthWetlandAreaSurfaceWetlandVolumeWetland ____

Table 2: Wetland Morphology

Wetland Category

Wetland Name Area Hectares

Depth m Volume m3

1 Paiwalla Wetland 49.009 0.7 343061

2 Sunnyside Wetland 27.309 0.7 191160

3 Lock 6 Wetland 17.92 0.92 164860

4 Reedy Creek Wetland 98.633 0.8 789064

5 Pilby Creek Wetland 11.991 1.8 215843

Time Series Irrigation Drainage

A number of the wetlands under consideration are influenced by irrigation drainage

runoff to varying degrees. As irrigation drainage can be a source of high nutrient

loads, this may have a significant impact on wetland nutrient content and must

therefore be taken into consideration in wetland modelling. In WETMOD 1, a

constant nutrient load contribution from irrigation drainage flow was assumed for all

irrigation affected wetlands. The extended model includes time-series data for

irrigation drainage nutrient contribution, and therefore nutrient and flow variations

within the irrigation drainage.


77

“Wetlands Management Study report” (Nichols 1998) was again used to identify

features of wetlands. In this instance where Nichols (1998) identified wetlands subject

to 10% irrigation inflow or more (where 10% of inflow into a wetland can be assumed

to be from irrigation areas) were considered as irrigation impacted wetlands during

modelling scenarios.

Two of the wetlands (Sunnyside and Reedy Creek wetlands) considered in the

development of the wetland category structure, have irrigation drainage inflow. For

both of these wetlands, monitoring of the drainage inflow was included during the

wetland-monitoring project. These data were used in simulation modelling of these

wetlands and their respective categories.

The pump supplying the irrigation drainage to Sunnyside wetland was not observed at

every monitoring date. The pumping of irrigation into Sunnyside wetland may have

occurred either intermittently or daily. In either case, the volume pumped will have

varied with requirements. In a situation of intermittent pumping, it is not possible to

retrospectively estimate when pumping occurred, nor the nutrient concentration of the

drainage water. In the absence of better data constant daily pumping was assumed

based on the agricultural need to prevent water logging of reclaimed dairy pasture and

the raising of water tables that can cause damage to pasture growth (Harrison 1994).

The data shown in Figure 17 provides the model with additional input of NO3-N &

PO4-P and phytoplankton to Sunnyside wetland, received as irrigation drainage. The

inflow amount into the wetland can be set at a constant volume, the units being in

litres per day.

The supply of irrigation drainage to Reedy Creek wetland was monitored at one inlet.

The flow volume at this inlet was not monitored and an annual rate of 600 ML was

estimated for this inlet into Reedy creek wetland (Wen 2002b). The inflow amount is

controlled by an estimate where the volume distribution pattern is based on the

relative average monthly precipitation, the distribution pattern having a mean of one

over a one-year period. Therefore, the monthly drainage pattern resembles that of the

average precipitation pattern. The irrigation drainage flow pattern for Reedy Creek

wetland was adopted to account for the estimated load of 600 ML per annum. The

irrigation and drainage multiplication factor chosen during modelling, in the case of

Reedy Creek wetland, is therefore a direct multiplication of estimated nutrient inflow

loads. The Reedy Creek wetland base irrigation rate of 600 ML per annum is included


78

once irrigation drainage flow simulation is selected for the wetland category. Figure

17D shows the irrigation and drainage inflow pattern developed and Figure 17E, F

and G the additional input of NO3-N & PO4-P and phytoplankton loads supplied as

part of the irrigation and precipitation drainage.

Surface Runoff Data

As the lower River Murray flows through a predominantly arid landscape water

contribution through precipitation does not account for a significant nutrient or water

source for most of the wetlands, the exception being Reedy Creek wetland. Therefore,

to maintain the generic nature of this model site-specific surface flow would

unnecessarily complicate the model with no significant advantage to modelling

scenarios. Precipitation and consequent surface flows were ignored for most wetlands

in this generic model, with the exception of Reedy Creek wetland that had a separate

contributing minor catchment.

Annual average rainfall in the east Adelaide hills was used to provide the seasonal

precipitation pattern. This was believed to be the most appropriate source of a rainfall

pattern as the east Adelaide hills is the source of surface runoff flowing into Reedy

Creek sub-catchment, see Figure 17D.


79

Figure 17: Sunnyside Irrigation Drainage PO4-P, NO3-N, Phytoplankton and Estimated Flow Volume (see also in Appendix B)

Se aso n a l D ra in ag e Pa tte rn R e e d y C re e k Su b catch m e n t

0 .0 0

0 .2 0

0 .4 0

0 .6 0

0 .8 0

1 .0 0

1 .2 0

1 .4 0

1 .6 0

1 .8 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

Re

lati

ve

Ra

te P

er

Mo

nth

-

-

D ra in ag e Ph yto p lan kto n

0

0 .0 5

0 .1

0 .1 5

0 .2

0 .2 5

0 .3

0 .3 5

0 .4

0 .4 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

-

D rainage N O 3-N

0

0 .2

0 .4

0 .6

0 .8

1

1 .2

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

-

D rainage PO 4-P

0

0 .5

1

1 .5

2

2 .5

3

3 .5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

-

A

B

C

D

D rainage Phytoplankton

-2 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

1 4 0

1 6 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

--

D rainage N O 3-N

-0 .4

-0 .2

0

0 .2

0 .4

0 .6

0 .8

1

1 .2

1 .4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-

Drainage PO4-P

-1

0

1

2

3

4

5

6

7

8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-

E

F

G

S unnyside W etland Reedy C reek W etland


80

3.2.3 River Data

External sources, such as river exchange, precipitation, and irrigation drainage, impact

upon wetlands. The most important of these for most of the considered wetlands is

river exchange. Although the river flow data are limited to the Lock locations, using

this data for relatively long stretches of the river is more appropriate than using

models of river flow and inundation. Around Mildura the river fall is less than 5cm

per kilometre and near the sea is as little as 1.6 cm per kilometre (Mackay et al.

1990). Therefore, due to the shallow gradient of the river as it flows in its course

through South Australia (Walker 1985) with alternating flow direction based on wind

direction, the development of a rudimentary flow model becomes difficult and would

add a complexity and inaccuracy that would compound in the generic ecosystem

model WETMOD 2.

In the past, series of aerial photographs and satellite images have been used to

develop a flood inundation model (FIM) (Overton 2000; Overton 2005). The data

required which is water exchange between a wetland and the river, is dependent on

river flow and could not be extracted from FIM for individual wetlands. The

development of an estimation of the exchange volume between the wetlands and the

river was achieved using WETMOD 2 in combination with river flow and nutrient

load. The methodology of estimating the exchange volume, between the river and

individual wetlands, is described below and is a major output of the model.

River Flow Volume

As all the wetlands considered in this model are permanently inundated and have a

constant connection with the river, it was assumed that the controlling factor for

nutrient exchange between the river and the wetlands is river flow volume. The river

flow is monitored at each river lock and this data is presented in Table 1. The flow

volume of the River Murray has been monitored daily since the construction of the

Locks in the late 1920‟s (Lock 6 being completed in 1930), the relevant time-series

for this project was obtained from the Murray Darling Basin Commission (MDBC).

This provides an important source of information that can be related to the connection

between wetlands and the river and consequently exchange of nutrients.


81

On occasional days where river flow data were unavailable, a linear interpolation

between the monitored dates was performed. However, for a fortnight in December

2000, a number of the locks failed to monitor the flow volume due to particularly high

flow levels during a flood. Fortunately, the locks at the beginning and end of the river

stretch under consideration recorded the flow through their location. Regression

equations, based on the correlation of flow during simultaneously monitored dates in

the weeks preceding the flood, were used to estimate the missing flow data based on

the nearest lock with monitored flow volume. The R2 values for these regression

equations ranged from 0.95 to 0.99. To corroborate these estimated flow volumes, the

data were compared to the flow levels monitored at an independent lock. Through this

methodology, it was possible to reconstruct a probable flow volume pattern during the

fortnight of high flow event through all the relevant locks. Figure 18 shows the river

flow pattern and volume used in the modelling.

River Water Quality

The River Murray is the major nutrient source or deposit area for wetlands within the

study area. The river nutrient data, monitored simultaneously with wetland data,

provides a more accurate representation of the modelled situation and also a more

accurate comparison of wetland vs. river nutrient load than lock monitored data. This

is due to river nutrient data monitored simultaneously with wetland data was a direct

indication of the river nutrient load at that time and not at a location further from the

site such as at locks. In those wetlands where the river and wetland were monitored

concurrently (Reedy Creek, Sunnyside and Paiwalla wetlands) only the Reedy Creek

river nutrient data were comprehensive enough to be considered for the wetland

modelling. Consequently, for Reedy Creek wetland it is possible to simulate the

wetland with either the MDBC data or the concurrent monitored river data.

Other sources of river data were required for the remaining wetlands. River data from

the same time period as wetland monitoring that could be included in the model were

acquired from the sources listed in Table 1. The river data quality, obtained from

DEH and MDBC see Table 1, were suitable for use in the model, see Figure 18. As

with the wetland data, all the river data were extrapolated linearly to obtain daily

values.


82

River concentrations of both NO3-N and PO4-P were generally higher than the

concentration within the wetlands (Figure 18 and Figure 19), exceptions occurred

where wetlands had a high-modelled river exchange volume. This suggests that where

there is an inflow of nutrient from the river to the wetland, the river will act as a

source of both NO3-N and PO4-P to the wetland. If the wetland processes manage to

take up the nutrients in macrophyte and phytoplankton growth, and these are retained

within the wetland, the water outflow from the wetland into the river would contain

lower nutrient concentrations. The wetland would therefore act as a nutrient sink. For

wetlands with higher concentrations of nutrients than the river, the wetlands may act

as point sources of nutrients to the river.

Figure 18A and E contain the MDBC river time-series of PO4-P and Figure 18I

contains the Reedy Creek river time-series of PO4-P. River Filterable Reactive

Phosphorus as PO4-P monitoring was discontinued early in the wetlands study time

period, whereas Filterable Reactive Phosphorus as P was continued. As WETMOD

requires phosphorus as PO4-P, a linear regression was calculated between Filterable

Reactive Phosphorus as P and Filterable Reactive Phosphorus as PO4-P, from a time

when both were monitored. Equation 3 was used to convert the monitored Filterable

Reactive Phosphorus as P to PO4-P. The R2 for Equation 3 was 0.9988.

Equation 3: 0004.00575.34 PPPO

The Reedy Creek river monitored Filterable Reactive Phosphorus time-series (PO4-P)

could be used in the model without any conversion.

Figure 18B and F contain the MDBC and Figure 18J the Reedy Creek time-series of

river Nitrate as NO3-N. As with PO4-P, river Nitrate as NO3-N monitoring was

discontinued early in the wetlands study time period. As the model input required is

Nitrate as NO3-N, a linear regression to obtain estimated NO3-N was calculated using

Equation 4 to convert from Nitrate as N to Nitrate as NO3-N. The R2 for the linear

regression was 0.9998.

Equation 4: 0044.0412.43 NNNO

The Reedy Creek river monitored Nitrate time-series (NO3-N) could be used in the

model without any conversion.


83

Chlorophyll-a is used as a surrogate for phytoplankton. The conversion between the

two is given in Equation 5 below where C is Chlorophyll-a in μg/L and P is

phytoplankton in cm3/m3 (Recknagel nd).

Equation 5: 5.2

CP

The MDBC could not supply river Chlorophyll-a, Figure 18C and G for the entire

study area, as monitoring ceased in early 1998 for some locations, therefore

Chlorophyll-a was not available at all monitoring locations. Only the Reedy Creek

project monitored river Chlorophyll-a concurrently and comprehensively for the

entire study period, see Figure 18K. However, as phytoplankton exchange plays an

important role in the wetland modelling it was opted to use data from further

downstream rather than none at all. Therefore, for all other wetlands the MDBC river

Chlorophyll-a time-series monitored at Murray Bridge was used in the model for all

river to wetland inflow. The remoteness of Murray Bridge from Pilby and Lock 6

wetlands must be taken into consideration when assessing the model simulation

performance for these wetlands. The modelling of Category 4 wetlands used the

Chlorophyll-a time-series obtained from the Reedy Creek wetland data. This approach

was far from optimal. However, as the data was not central in the development of the

model and it could be ignored during validation it was deemed acceptable during this

stage of the modelling process. Through future river Chlorophyll-a monitoring this

discrepancy could be remedied.


84

Figure 18: River Murray Nutrient & Phytoplankton Time Series as well as River Flow Volume (see also in Appendix B)

River Flow

0

5000

10000

15000

20000

25000

30000

35000

40000

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

ML

/day

-

River Phytoplankton

0

2

4

6

8

10

12

14

16

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

-

River Phytoplankton

0

1

2

3

4

5

6

7

8

9

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

-

River Flow

0

2000

4000

6000

8000

10000

12000

14000

16000

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

ML

/day

-

River Flow

0

2000

4000

6000

8000

10000

12000

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

ML

/day

-

River Phytoplankton

0

1

2

3

4

5

6

7

8

9

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

-

River NO3-N

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-

River PO4-P

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

-

River PO4-P

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

-

River PO4-P

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-

River NO3-N

0

0.5

1

1.5

2

2.5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

-

River NO3-N

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

-

A

B

C

E

F

G

I

J

K

D H L

Reedy C reek W etlandLock 6 & P ilby C reek W etlandsP a iwa lla & S unnyside W etlands


85

3.3 Data Handling

Calculating River and Wetland Exchange

One of the major attributes of WETMOD 2 is its ability to calculate exchange rate

(turnover) of water and nutrients between the wetlands and the river. The nutrient

exchange between the river and the wetland is calculated for each time-step in the

model. The net outflow of nutrient from the wetland is subtracted from the net inflow

of nutrient. The equation for the bi-directional exchange between the wetland and the

river t

N R[mg/day] (Nutrient Retention) can be expressed as per Equation 6 with CR

and CW denoting concentrations of nutrients in the river and wetland respectively, and

ƒ being a fraction of river flow rate R [L/day], see Figure 1.

Equation 6: RfCCt

NWR

R

)(

The factor f quantifies in a simple way, how the wetland is connected to the river. It

summarises the complex morphology of linkage of wetlands and the river through

channels, topographic conditions and distance.

The factor f is varied for each modelling scenario, and the model performance with

respect to PO4-P and NO3-N is tested. The best performing scenario is chosen to

represent the optimum exchange volume for a given wetland. An example of the

exchange volume estimation is provided in section 4.1. The methodology for the

assessment of model performance is discussed in section 3.3.1.

Based on the modelled exchange volume it is possible to estimate the wetland water

turnover rate where the turnover rate (τ [1/day]) relates to the factor f, R and Vw as per

Equation 7, Vw being the wetland volume.

Equation 7: WV

Rf

The turnover rate gives a secondary method to assess the potential accuracy of the rate

of exchange expected for a given wetland, see section 3.4.2. As mentioned in section

1.1.2 the potential nutrient uptake of wetlands is related to the turnover rate, i.e. the

retention time.


86

Model Expected Simulation Output (monitored data)

As mentioned in section 3.1 the model simulates the PO4-P and NO3-N concentration

in a wetland and the phytoplankton, macrophyte and zooplankton biomasses. The

wetlands used as “exemplars” were monitored for the outputs PO4-P, NO3-N and

phytoplankton. These output data were used to test, develop, and calibrate the model,

and to adjust the exchange volume and nutrient inflow to achieve a best fit. Neither

zooplankton nor macrophyte biomass were used in the first instance as these data

were unavailable for comparison with model outputs and thus could not be used to

assess the model. Any discussion and conclusions made based on macrophyte and

zooplankton modelled biomass is limited by this lack of data and may not necessarily

reflect what may occur in a natural setting. Validation of the model continued as

discussed in section 4.2.

The monitored data for the different wetlands representing the categories, i.e.

providing the “exemplar” data, are presented in Figure 19, which an ideal model

would simulate accurately. For Paiwalla and Sunnyside wetlands Figure 19A, B and C

represents the monitored PO4-P, NO3-N concentration and phytoplankton biomass

respectively. Figure 19D, E and F represent the Lock 6 and Pilby Creek wetlands

monitored PO4-P, NO3-N concentration and phytoplankton biomass respectively.

Figure 19G, H and I the Reedy Creek wetland monitored PO4-P, NO3-N concentration

and phytoplankton biomass respectively. At least three monitoring sites were used for

each of the wetlands, usually one close to the inlet (or the river), one in the littoral

zone of the wetland, and one in the open water of the wetland. The model however

uses the driving variables from the open water monitoring site of the wetland. The

monitored data used to test and validate the model were also derived from the open

water location. To represent the variability of the wetlands and therefore the potential

variability of the modelling outcome, the Standard Error was calculated for each

sampling date and is displayed along with monitored concentrations in Figure 19.

Only data from one sampling location in Reedy Creek was obtained (i.e. only one

measurement per monitoring date), the Standard Error for the entire monitoring period

had been calculated based on all sampling dates Figure 19G, H and I.


87

Figure 19: Wetlands (Categories 1 to 5) Monitored Nutrients and Phytoplankton

PO 4-P

-0 .5

0

0 .5

1

1 .5

2

2 .5

3

3 .5

4

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

PO4-P

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

NO3-N

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

PO4-P

-0.05

0

0.05

0.1

0.15

0.2

Feb-9

7

Mar-

97

Apr-

97

May-9

7

Jun-9

7

Jul-97

Aug-9

7

Sep-9

7

mg

/L

NO3-N

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Feb-9

7

Mar-

97

Apr-

97

May-9

7

Jun-9

7

Jul-97

Aug-9

7

Sep-9

7

mg

/L

NO3-N

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

A

B

D

E

G

H

C F I

P aiwalla W etland 1997 S unnyside W etland 1997 Reedy C reek W etland 2000-2001Lock 6 wetland 1997 P ilby C reek W etland 1997

Phytoplankton

0

2

4

6

8

1 0

1 2

1 4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

Phytoplankton

0

2

4

6

8

10

12

14

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

Phytoplankton

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

C F I


88

3.3.1 Model Calibration

As WETMOD 1 was substantially adapted and the driving variable database rebuilt

with updated data for WETMOD 2, the model calibrations needed re-evaluation. The

model was run based on its original calibrations and the optimal exchange rate

established. Once the optimal exchange rate had been estimated the model output was

assessed to identifying discrepancies such as unexpected trends. The model

parameters identified to be adversely affecting the model output were recalibrated to

account for the new data set. Many parameters calibrated in the original WETMOD 1

model were unaltered with only the following parameters being recalibrated.

Turbidity sedimentation threshold for phytoplankton was recalibrated from 70

NTU to 95 NTU. The ones for PO4-P and NO3-N were unaltered at 70 NTU.

The sedimentation rate for phytoplankton (pht sed) was recalibrated

Zooplankton mortality rate (ZooMortRate) was recalibrated

The maximum phytoplankton growth rate (Phyt max) was recalibrated

Once the model had been recalibrated the exchange rate between the wetlands and the

river was reconfirmed and readjusted as appropriate.

3.3.2 Validation Procedure

It was found during the initial validation procedure that squared error estimates over-

represented errors at peaks in the model output. This was seen as an inaccurate

representation of a generic model where short term peak fluctuations can not be

modelled. Therefore, an evaluation criterion where the average linear deviation from

the measured values as a fraction of the average observed values was used and is

referred to as D (Equation 8). The index D is derived as per Equation 8 with M being

the modelled and E monitored PO4-P, NO3-N concentrations or phytoplankton

biomass at the monitoring dates.

Equation 8: E

EMABSD

Any reduction in D was considered to be an improvement in performance of a model

scenario, however some improvements had a greater impact than others and should be

emphasised. The following descriptive grades of improvement were adopted to better


89

convey the importance of each improvement. Improvements of 10% or greater were

regarded as noteworthy, improvements of 20% or greater to be considerable and 30%

or above to be a significant improvement to the modelling performance.

When assessing modelling performance the PO4-P D was valued prior to NO3-N D as

PO4-P does not escape to or return from a gaseous phase, like NO3-N does in a

wetland environment, and is therefore more constant in the system, see section 1.2. In

scenarios where PO4-P D optimum performance could not be achieved due to data or

modelling peculiarities, NO3-N optimum D performance was strived for. As

mentioned in section 3.2.3 the Chlorophyll-a data, used to calculate phytoplankton

biomass, were sourced from Murray Bridge. Due to this limit of location specific

concentrations of phytoplankton, and the methodology for calculating the

phytoplankton from Chlorophyll-a concentration, phytoplankton was never used to

assess the model performance.

3.4 Wetland Management

3.4.1 Options

In the application of the model there were two management strategies simulated by

WETMOD for the wetlands of the lower River Murray, turbidity reduction and

irrigation drainage reduction. Scenarios were developed for potential turbidity

reduction management for both Lock 6 wetland and Reedy Creek wetland. Scenarios

of the second management strategy were developed for only the Reedy Creek

wetland; however, it was applied both with and without the management strategy of

turbidity reduction. The management strategies have two different approaches to

nutrient reduction within a wetland, therefore potentially reducing both nutrient and

phytoplankton outflow from the wetland.

Strategy; Turbidity reduction - Construction of wetland flow control structures

and grids for introduction of wetland dry periods and consequent carp

restriction.

The presumed wetland response expected is sediment compaction as a consequence of

a wetland drying, see section 1.3. Through grids being constructed at the wetland flow

inlet large carp re-colonisation would be avoided minimising any bioturbation impact

i.e. sediment re-suspension and therefore turbidity. Management simulations were


90

performed for 0% reduction in turbidity, 25%, 50% and 75% (100% reduction in

turbidity regarded as unattainable).

Secchi depth increases with the reduction of turbidity; therefore the Secchi depth was

altered appropriately for each assumed turbidity reduction scenario. In Lock 6 wetland

management scenarios, where turbidity was reduced by:

o 25% the Secchi depth was set at 0.3 metres,

o 50% the Secchi depth was set at 0.6 metres, and

o 75% the Secchi depth was set at the wetland depth of 0.9 metres.

Strategy; Irrigation drainage nutrient reduction - Constructed wetlands for

nutrient removal.

Nutrient normally entering the wetland through irrigation drainage would be diverted

into constructed wetlands, where macrophytes would assist in nutrient uptake.

Theoretically the harvesting of the macrophytes would remove the nutrients

permanently from the system. The effective removal of nutrients can be variable, as

discussed in the introduction, see section 1.3. Therefore, variable nutrient removal

successes were modelled with scenarios representing 0% nutrient reduction, 25%,

50% and 75%. An example of “fully” restored wetlands with 85%, 90% and 95%

nutrient reduction was also simulated.

To examine the impact of a two pronged management strategy a combination of both

management interventions was simulated for Reedy Creek wetland. It was assumed

that in the period prior to simulation the wetland had been dry and therefore resulted

in turbidity reduction. The scenarios of the Reedy Creek twin management strategies

were simulated for twelve months with no allowance made for a second dry period.

Simulations were made for 25, 50, 75, 85, 90 and 95% irrigation drainage load

reduction (nutrient reduction scenarios). High irrigation nutrient reductions were

performed to display the hypothetical impacts of a nearly fully restored wetland. For

each of the nutrient reduction scenarios, scenarios of 25, 50 and 75% turbidity

reduction were also simulated. In Reedy Creek wetland simulations, the wetland

Secchi depth during the turbidity reduction scenarios were adjusted to 0.2, 0.3, 0.6

metres and the maximum wetland depth of 0.8 metres used in the 75% turbidity

reduction scenario.


91

Assessment of Management Scenario Impact

To assess the management scenario impact on nutrient retention capacity of a wetland

a comparison between the change in inflow and outflow was made, see Figure 20. The

percentage Reduction of Inflow (%RI) is calculated as per Equation 13 where ID is

the total Irrigation Drainage load (calculated using Equation 9). IC denotes the

concentration of irrigation drainage nutrient and I the Irrigation Drainage flow in

litres/day. ∆ID is the change in total Irrigation Drainage load after management and

RF the total River Inflow load, RF is calculated as per Equation 10. Equation 14

calculates the percentage Reduction in Outflow (%RO), where OF is total Outflow

load (calculated as per Equation 11), and ∆OF is the change in total Outflow load post

management. The OF for Reedy Creek wetland and category 4 wetlands is calculated

by Equation 12 to account for the additional irrigation flow volume exiting the

wetland. As in Equation 6, CR and CW denote concentrations of nutrients in the river

and wetland respectively and ƒ represents a fraction of the river flow rate R.

Equation 9: ICID I

Equation 10: RfCRF R

Equation 11: RfCOF W

Equation 12: ))(( IRfCOF W

Equation 13: 100/100% RFIDRFIDRI

Equation 14: 100100% OFOFRO

The %RO Equation 14 above is therefore, the change in outflow due to management

when compared to the status quo (no management). With a positive %RO there is a

net improvement of the nutrient or phytoplankton retention of the wetland due to

management. The %RI Equation 13 only applies to Reedy Creek and category 4

wetlands and represents the effective change in wetland nutrient inflow due to nutrient

reduction scenario as compared with the status quo.

The impact of water loss through other means, specifically evaporation, has not been

included in the mass balance equations. The current method of evaporation estimation

is itself inaccurate and would have added further complications, to model calibration

and validation, than is acceptable at such an early stage of the model development.

This is an aspect that can in future be included in the model when full monitoring (of


92

at least one wetland) including all water sources, sinks (including evaporation) and

nutrient balance becomes available to effectively calibrate and validate the model.

Figure 20: Wetland exchange modelling

3.4.2 Management scenarios for cumulative assessment

Wetland candidates for simulations

There are more than a thousand individual wetlands in the lower River Murray,

ranging from small, temporary wetlands to large and more permanent examples.

However, of this multitude of wetlands, only 250 individual wetlands or groups of

closely related wetlands (complexes) are identified in the „Wetlands Atlas of the

South Australian Murray Valley‟ (Jensen et al. 1996). For the purposes of this project,

the 250 identified wetlands were perused with the intent of consideration for

management. In the cumulative assessment of management scenarios two wetland

categories were considered, these being category 3 (wetlands resembling Lock 6

wetland) and category 4 (wetlands resembling Reedy Creek wetland). Identified

wetlands were assigned to a particular category, depending on their similarities to the

Wetland process modelling

Nutrient concentration in

river (CR) X exchange

volume (f)

Nutrient retention (t

N R

) becomes a factor of exchange volume (f & R), river concentration (CR) and wetland

concentration (CW) calculated using the wetland process model. Irrigation inflow is considered where appropriate

using CI and I. Change in (t

N R

) due to management is assessed for different scenarios (influenced by the

change in (CW)).

River Flow volume (R) (River Nutrient load (LR) = R X CR)

Fraction of river flow volume (f)

Nutrient concentration in

wetland (CW) X

(exchange volume (f) +

irrigation flow volume

(I))

Nutrient concentration

from Irrigation runnoff

(CI) X exchange volume

(I)


93

Lock 6 and Reedy Creek ”exemplars”, with each category having a defined

management strategy.

In the lower River Murray 54 of 250 wetlands (wetland groups) were identified as

being similar to Lock 6 wetland and therefore classified as category 3 wetlands.

Including Lock 6 wetland, 35 were found to be over 0.6 metres depth, the minimum

depth of wetlands found to be effectively simulated by WETMOD 2. These 35

wetlands and wetland groups, make up a total of 57 individual lagoons that can be

simulated within WETMOD 2 (a list of these wetlands is provided in Table 18 in

Appendix C). The method for Secchi depth adjustment in cumulative wetland

management scenarios was handled in the same way as for Lock 6 wetland

simulations discussed in section 3.4.1.

Due to the nature of Reedy Creek wetland, more stringent restrictions had to be placed

on the wetlands that could be regarded as potential category 4 wetlands. If wetlands

less than half the volume of Reedy Creek were simulated using the exchange volume

found for Reedy Creek then the average volume exchanged per day would exceed the

total wetland volume. When the exchange volume exceeds the wetland volume the

nutrient retention time within the wetland is reduced below that of the model time-

step. WETMOD 2 has not been developed nor calibrated for such a continual high

exchange volume. WETMOD 2 was therefore restricted to simulation of wetlands

where the average exchange volume is below that of the wetland volume.

Consequently, due to the high river exchange volume estimated for Reedy Creek

wetland, category 4 modelled wetlands are restricted to those with a volume greater

than half the volume of Reedy Creek wetland.

A further restriction, for wetlands to be considered for management scenarios of

category 4 wetlands, was based on the irrigation flow volume. Reedy Creek wetland

was estimated to receive a high volume of irrigation drainage flow. Therefore,

wetlands that were deemed to receive only a low volume of irrigation drainage flow

were also excluded from management consideration. Therefore, 7 of the 250 wetlands

(wetland groups), including Reedy Creek wetland, were identified as being category 4

wetlands for which WETMOD 2 had the potential capacity to reliably simulate (a list

of these wetland is provided in Table 19 in Appendix C). This did not include the

potential irrigation drainage nutrient concentration that these wetlands may receive, as


94

this information was unavailable, Reedy Creek irrigation data was therefore used as

the driving variables.

Exchange volume

During the simulation of wetland management scenarios using “exemplar”-driving

variables, the wetland volume is changed to reflect the wetland that is being

simulated. However, the exchange volume between the wetlands and the river was

maintained at the same percentage of river volume as was estimated for the

“exemplar” wetland (i.e. the wetland which provided the driving variable data). For

category 3 wetlands the river exchange was maintained at 0.1% of the river flow

volume per day, this being the volume fitted for Lock 6 wetland. For category 4

wetlands the fitted exchange rate for Reedy Creek wetland of 3.5% of the river flow

volume per day was used. This fitted volume for each of the two categories was

maintained based on the assumption that all category wetlands resemble each other

unless specific data is available. Consequently future improvement of the model could

be achieved with a proper estimate of individual wetland water exchange with the

river, thereby providing improved wetland scenario accuracy.

For each category wetland scenario the driving variables for the river data are sourced

from the nearest upstream monitoring location, the exception being Reedy Creek

wetland which has its own monitored nutrient river data set. Therefore, the flow

volume was adjusted below each successive lock and the river nutrient data was

adjusted to each individual nutrient monitoring locations. The behaviour of wetlands

of a particular category was expected to be similar, particularly where the only major

difference between the wetlands is the morphology.

Implication of the change in nutrient retention capacity on river nutrient load

Through the management of both category 3 and category 4 wetlands, a cumulative

impact on the river nutrient load would become evident. Although the modelling

accuracy of category wetlands allows only a qualitative understanding of the trends

expected due to wetland management and not quantitative accuracy, the model results

will, for this section, be assumed to be quantitatively accurate. The rationale is two

fold.


95

First, although the results are not quantitatively accurate the assessment of the

quantitative output helps to develop a qualitative trend analysis of the

cumulative impact of management.

Second, although this model, due to the poor data quality, is of low

quantitative accuracy the methodology of assessing the cumulative impact

could be applied in the same manner should the model quantitative

performance improve through future data improvement.

However, this assumption is made in order to understand and discuss the potential

cumulative impact on nutrient loads within the river, and should only be seen as a

trend analysis.

To understand the cumulative impact that the management of multiple wetlands

would have on the river nutrient load, the change in wetland nutrient retention

capacity was compared to the river load. To this purpose the initial river nutrient load

( RL ) was required see Equation 15, see Figure 21.

Equation 15: RCL RR

The initial river load is calculated from the first available monitoring locations post

inflow into South Australia, i.e. the flow volume data is obtained from Lock 6

whereas the river nutrient concentration is obtained from Lock 5. The calculation of

the river nutrient load based on the earliest available monitoring locations was chosen

so that the river data would not reflect the status quo impacts of the wetland that are

simulated, i.e. wetland impacts would otherwise be counted both status quo and as per

management scenario.

The wetland nutrient retention calculation is similar to Equation 6 (see Box) where the

retention in the wetland is calculated per day. Equation 16 needs to calculate the sum

over the modelled period for each of the management scenarios. The status quo (i.e.

no wetland management) subtracted from the nutrient retention in the wetland as per a

management scenario, gives the change in nutrient retention ( RN ) due to

management. Where, RN is the change in wetland retention due to management and

is calculated as per Equation 16 where sqt

N R

is the nutrient retention at the status

quo scenario and mst

N R

the nutrient retention at the respective management

scenario.


96

Equation 16 sqt

Nms

t

NN

RR

R

The RN was used to calculate the change in river load where the % River Load

removed due to the wetland management (%RL) is calculated as per Equation 17, see

Figure 21.

Equation 17: R

R

L

NRL%

Equation 16 and Equation 17 are used to calculate the impact of a single wetland on

the river nutrient load as well as the cumulative impact the management of multiple

wetlands would have on the river nutrient load, see Figure 21.

Figure 21: Cumulative assessment of wetland processes

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

W

etl

an

d

pr

oc

es

s

m

od

ell

in

g

Wetland 1 Wetland 2 Wetland 3 Wetland n

RN 1 RN 2 RN 3 RN n + + +

RL

RL

RL

)/ RL

(

=> %RL

Equation 6: RfCCt

NWR

R

)(


97

4 Validation of the model WETMOD 2 and Discussion

During the development of WETMOD 1, neither river flow nor nutrient load data

were available. To varying extents the wetlands are reliant upon the exchange of

water and nutrient with the river. The addition of river flow and nutrient load as well

as the exchange volume between the wetlands and the river is therefore a significant

improvement of the WETMOD 2 model. This chapter will test the first hypothesis of

whether “a simplified generic wetland model can be used to realistically simulate

multiple and different wetlands qualitatively”.

4.1 Fitting and Validation based on calibrated (“exemplar”)

wetlands

The results presented in this chapter show the validation steps used for WETMOD 2

using data from the five different wetlands. The validation of WETMOD 2 is based on

D (Percentage Deviation of modelled time-series from monitored time-series) for

PO4-P, NO3-N and phytoplankton, and is represented in Table 3.

River water quality is influenced by adjacent wetlands. The water exchange estimate

is a step in the process of developing a model capable of simulating management

strategies for wetlands of the lower River Murray and their impact on nutrient load in

the river. WETMOD 2 was used to find the water exchange between wetlands, where

there is a lack of channel morphology data and no measured wetland water turnover.

The added spatial driving variables for WETMOD 2 are used to account for local

variations and inflow into a wetland, particularly to reflect bi-directional water and

nutrient exchange between the River Murray and the wetlands (see section 3.3). This

was based on a combination of the river flow volume and the wetland specific budget

of PO4-P or NO3-N simulated by WETMOD 2. Through this methodology it is

possible to obtain the turnover volume of water in a wetland using nutrient modelling

output (Bjornsson et al. 2003).

The optimal river exchange estimate was determined by WETMOD 2 based on the

best percentage deviation (D) (see box). Given the availability of accurate daily river

flow data as well as fortnightly nutrient data, it was possible to estimate the flow of

nutrients carried by the lower River Murray. This provided accurate data for the

estimation of the most significant external nutrient source, i.e. the river. Combined


98

with successive calculations of wetland internal nutrient load by WETMOD 2, the

wetland simulation results improved until the optimum exchange was attained. Once

the external load was increased past the optimum the wetland simulations degraded,

see Figure 22.

As discussed in section 3.3.1, PO4-P was in most cases used as the primary indicator

of model D as PO4-P is the most reliably modelled and monitored nutrient within the

system (once PO4-P enters a wetland it is not diminished through a gaseous state). The

flow exchange between a wetland and the River Murray was mostly estimated based

on the model percentage deviation (D) calculation of PO4-P, with NO3-N only used

for Lock 6 wetland. Figure 22 presents an example of the selection of D for Reedy

Creek wetland. In this example the PO4-P shows the best fit at a river exchange of

3.5% of the daily river flow volume.

Figure 22: Percentage Deviation based estimate of flow exchange: Reedy Creek wetland

Estim ation o f R iv er and W etland Exchange Vo lum es: R eedy C ree k w etland

0

10

20

30

40

50

60

70

80

90

100

1 1 .5 2 2 .5 3 3 .5 4 4 .5 5 5 .5

% of R iv er F low Exchaged per D ay

% D

ev

iati

on

(%

D )

% D e viation PO 4-P

% D e viation N O 3-N

% D e viation Ph ytoplankton

A verag e W etland Volum e % E xchang ed per D a y

The lower the D the closer the fit of modelled data to monitored data.


99

Table 3: Calibration of inflow data for the 5-wetland categories

Wetland Category

Wetland Name

Wetland External Input Variables

Modelled

PO4-P

D

Modelled

NO3-N

D

Modelled

Phytoplankton

D

1

Paiwalla Wetland

NO River Exchange & NO Irrigation Drainage

92 71 74

0.7% River Flow/day Exchange NO Irrigation

74 73 28

2

Sunnyside Wetland


71 69 75

NO River Exchange & 500L Irrigation Drainage

71 69 75

0.06% River Flow/day Exchange NO Irrigation

70 58 64

0.06% River Flow/day Exchange 500L Irrigation Drainage

70 58 64

3

Lock 6 Wetland

NO River Exchange 55 54 44

0.1% River Flow/day Exchange 81 34 49

4

Reedy Creek Wetland


94 101 67

NO River Exchange & 3500 X Irrigation Drainage

87 100 58

3.5% River Flow/day Exchange & NO Irrigation Drainage

61 72 49

3.5% River Flow/day Exchange & 3500 X Irrigation Drainage

56 71 40

5 Pilby Creek Wetland



Wetland modelled results are presented and discussed in the sections below. Each

wetland is assessed independently, and some comparisons are made.

Category 1: Through flow wetlands with carp presence and no irrigation drainage

(Paiwalla wetland)

Paiwalla wetland is situated upstream of Sunnyside wetland (see section 3.2.1), with

an area of reclaimed „swamp‟ situated between them, which was used as dairy pasture

prior to 1997 (refer to map in chapter 2). The runoff from this pasture was pumped

into Sunnyside wetland and thereby transported nutrients from the irrigation drainage

into Sunnyside wetland. In contrast there was no direct input of nutrient from the

dairy pasture into Paiwalla wetland (Bartsch 1997). Paiwalla wetland was therefore


100

chosen to represent through flow wetlands with possible carp presence and no

irrigation.

The comparison between modelled and monitored concentrations of PO4-P is seen in

Figure 23A and Figure 23B for NO3-N; macrophytes, zooplankton and phytoplankton

are represented in Figure 24A, B and C respectively. Each graph of Figure 23 and

Figure 24 includes results for the scenarios “no flow exchange” and “optimum flow

exchange”. For monitored data in Figure 23 and Figure 24, error bars represent the

standard error for measurements made on that date.

As seen in Table 3 Paiwalla modelling results for PO4-P and phytoplankton improved

due to the consideration of river exchange, phytoplankton result being significant.

Figure 23A reflects this improvement. The NO3-N D shown in Table 3 does not show

an improvement, however the graph in Figure 23B indicates a distinctive change in

the model output due to river exchange. The NO3-N variability, range and seasonality

are realistically reflected by the river exchange scenario. It is therefore concluded that

the model validation improved with regard to qualitative trends even though the

quantitative accuracy is not optimal. There is a major improvement in the modelling

results for phytoplankton following the introduction of river exchange. The modelled

D for phytoplankton (Table 3) is the best result of all output from modelled wetlands

and scenarios; this modelling performance is also being displayed in Figure 24C

where the modelled phytoplankton corresponds well with the trends of the monitored

phytoplankton. There is some early macrophyte biomass growth in Paiwalla wetland

however; there is a rapid decline due to increasing turbidity, see Figure 24A.

Phytoplankton growth, as seen in Figure 24C, increases as expected following

diminished macrophyte competition. As the solar radiation and wetland water

temperature increase in spring, the growth of phytoplankton increases accordingly

(Figure 24C). Zooplankton biomass increases in response to the growth of

phytoplankton. This is due to the phytoplankton serving the zooplankton as a food

source (Figure 24B) following the typical Lotka-Voltera predator-prey cycle as

discussed in the introduction.

Through flow wetlands are highly variable due to the close link to the river and are

therefore difficult to model with a simplistic model such as WETMOD. Although the

modelling results for this category of wetlands were not as good as expected there was

an improvement in the model output for Paiwalla wetland due to the introduction of


101

river exchange. It shows the potential of simplistic models to assess the exchange

volume of water and nutrients between riparian wetlands and the river.

Figure 23: Validation of simulation results for Paiwalla wetland of PO4-P, and NO3-N for both

conditions with and without water exchange

NO 3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L -

PO4-P

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80F

eb

-97

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L -

A

B

Mode lled C oncentration

No R ive r E xchange No Irrigation


0 .7% R ive r E xchange

Monito red D ates O nly

C oncentration in W e tland


102

Figure 24: Validation of simulation results for Paiwalla wetland of Macrophyte Biomass,

Zooplankton and Phytoplankton for both conditions with and without water exchange

M acro p h yte B io m ass

0

1

2

3

4

5

6

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

kg

/m3

-

P h y to p lan k to n

0

1

2

3

4

5

6

7

8

9

10

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3 -

Zo o p lan k to n

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

- -

A

B

C


No R ive r E xchange No Irrigation






103

Category 2: Through flow wetlands with carp presence and irrigation drainage

(Sunnyside wetland)

Figure 25A portrays the PO4-P and Figure 25B the NO3-N monitored and modelled

concentrations for Sunnyside wetland. Figure 26A, B and C depicts macrophyte,

zooplankton and phytoplankton monitored and modelled concentrations respectively.

The monitored concentrations of PO4-P, NO3-N and phytoplankton in the wetland,

and of PO4-P and NO3-N concentrations in the irrigation drainage, are represented in

Figure 25A and B and Figure 26C. For each monitored concentration, error bars

represent the standard error for measurements. Each graph includes results of

scenarios where no river flow exchange and no irrigation drainage were considered.

Another trendline in each of the graphs includes river flow exchange estimated at a

modelled best-fit D (Table 3), according to monitored wetland nutrient concentration.

This scenario was re-run with irrigation drainage included. To estimate the impact of

irrigation drainage on the wetland simulation Sunnyside wetland was also simulated

with only irrigation drainage influencing the scenario results and no river exchange.

The response of the scenario where irrigation drainage inflow was the only outside

nutrient source was minimal and effectively covers the simulation where no outside

nutrient source was considered (Figure 25 and Figure 26).

Sunnyside wetland is an “exemplar” for the category 2 wetlands considered in the

modelling project, which are wetlands having river water through flow and are

directly affected by irrigation drainage. Simulation results demonstrated that an

improvement in D of only 0.01 was evident when a realistic volume of 500L of

irrigation drainage per day was included in a scenario. In order to clarify the reason

for this result, we must look at both assumptions made at the start of the modelling

project as well as the monitoring design; this is discussed in section 4.1.1.

A better scenario of a wetland with irrigation drainage inflow in this wetland category

is not possible due to the limited data available. However, the small response of the

model to scenarios with drainage nutrient and the success of modelling Reedy Creek

wetland with its irrigation drainage (described below in category 4 wetlands), indicate

the possibility of a more successful modelling scenario when adequate data for this

wetland category become available.


104

The modelling of PO4-P (Figure 25A) does not pick up the early high wetland

concentration monitored, neither with nor without the river exchange and irrigation

drainage. However, with the introduction of river exchange there is a slight

improvement in the trend modelled, as can be seen in the results between the months

of May to June in Figure 25A. The improvement in the modelling trend of NO3-N due

to the introduction of river exchange can similarly be seen in Figure 25B. The D for

PO4-P, NO3-N and phytoplankton (Table 3) does improve with the introduction of

river exchange, with a noteworthy improvement for NO3-N and phytoplankton,

however this improvement is not great. As mentioned, better data is required to

successfully model this wetland.

There was a longer growth period of macrophytes in Sunnyside wetland than in

Paiwalla wetland simulations (Figure 26A and Figure 24A respectively). Again, this

can be attributed to the turbidity of the wetlands. The delayed increase in turbidity in

Sunnyside wetland extended the growth period for the macrophytes. The growth in

zooplankton and its high concentration (Figure 26B) is probably due to the shelter

provided by macrophytes (Figure 26A), the first zooplankton growth phase followed

by the increased food source phytoplankton in the second growth phase (Figure 26C).

The combination of simulated nutrient competition by macrophytes and grazing by

zooplankton restrict the initial growth of the phytoplankton (Figure 26). The major

growth phase of phytoplankton simulated occurs from May to July corresponding well

with the monitored trend (Figure 26C).


105

Figure 25: Validation of simulation results for Sunnyside wetland of PO4-P, and NO3-N for both


For both Figure 25 and Figure 26 the grey line (modelled concentration (PO4-P or NO3-N) with 0.06% river

exchange and no irrigation) falls behind the green line (modelled concentration (PO4-P or NO3-N) with 0.06% river

exchange and 500L irrigation drainage inflow). The blue line (modelled concentration (PO4-P or NO3-N) with no river exchange and no irrigation) falls behind the pink line (modelled concentration (PO4-P or NO3-N) with no river exchange but with 500L irrigation drainage).

PO4-P

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L -

NO 3-N

0.00

0.50

1.00

1.50

2.00

2.50

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L -

Mo d e lle d C o nce ntratio n

No R ive r E xchang e


0 % R ive r E xchang e 5 0 0 L I rrig atio n


0 .0 6 % R ive r E xchang e


0 .0 6 % R ive r 5 0 0 L I rrig atio n

Mo nito re d D ate s O nly

C o nce ntratio n in W e tland


C o nce ntratio n in I rrig atio n D rain

A

B


106

Figure 26: Validation of simulation results for Sunnyside wetland of Macrophyte Biomass,


M a cro p hyte B io m ass

0

5

1 0

1 5

2 0

2 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

kg

/m3

Phytoplankton

0

1

2

3

4

5

6

7

8

9

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

--

Zooplankton

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

--

A

B

C

Mo de lle d C o nce ntra tion

No R ive r E xc hang e

Mod elle d C o nce ntra tion

0 % R ive r E xc hang e 5 00 L I rrigation

Mo de lle d C onc entratio n

0 .06 % R ive r E xchang e

Mo de lle d C o nce ntration

0 .0 6 % R ive r 5 00 L I rrig ation


C o nce ntration in W e tland


107

Category 3: Dead end wetlands with carp presence and no irrigation drainage (Lock

6 wetland)

Figure 27 and Figure 28 depict the modelled output of Lock 6 wetland for PO4-P,

NO3-N, macrophytes, zooplankton and phytoplankton respectively. The error bars

represent the standard error for the monitoring at that particular date based on three

separate measurements.

Lock 6 wetland is a permanently inundated wetland situated adjacent to Lock 6 of the

River Murray. It is a wetland classified as a “dead end” wetland. This wetland‟s

hypothetical management strategy was drying and compacting the sediment.

Therefore, it was assumed for the modelling that following a re-flooding event the

sediment re-suspension and wetland turbidity would be reduced (see section 3.4).

As there is no irrigation drainage flowing directly into Lock 6 wetland, only the river

exchange volume was considered as an external influence upon this wetland. It was

expected that all output parameters would have an improved response. It is possible

that the high PO4-P level modelled in the wetland was overestimated due to relatively

high river concentrations. However, the trend was clearly modelled correctly when

compared to monitored concentrations (Figure 27A) despite the D indicating a worse

fit (Table 3). This discrepancy is also reflected in the modelling result of the

phytoplankton (Figure 28C). This shows that although the D is a good method of

finding the best-fit scenario during modelling, it is by no means a perfect method and

model results should be analysed with an understanding of the expected trends. The

modelling performance of NO3-N was improved considerably by the introduction of

river exchange, as seen in Table 3 and Figure 27B, and is the best modelling response

of NO3-N for all wetlands and scenarios.

Due to the high turbidity levels of Lock 6 wetland, the modelled macrophyte growth

is inhibited showing that the original estimate of the potential macrophyte biomass

used in the modelling scenario was probably overestimated (Figure 28A). It can be

assumed that the high turbidity levels limited underwater light for macrophyte growth.

However, due to the high nutrient levels within the wetland (Figure 27), and the lack

of competition provided by the macrophytes, the phytoplankton were able to grow

effectively (Figure 28C), reaching a peak biomass prior to the onset of winter. The

lack of the spring growth phase can be attributed to the large volume of Lock 6


108

wetland effectively buffering early rise in water temperatures. It must be remembered

that the river chlorophyll-a used in calculating the river phytoplankton, which is

consequently used in representing the exchange rate inflow into the wetland, was not

available for this part of the river. As phytoplankton has a significant role in this

model, its part in the wetland simulations could not be ignored. The zooplankton

growth (Figure 28B) in Lock 6 wetland follows the phytoplankton growth as

expected, and declines during the winter months.

Figure 27: Validation of simulation results for Lock 6 wetland of PO4-P, and NO3-N for both


NO 3 -N

0 .00

0 .05

0 .10

0 .15

0 .20

0 .25

0 .30

0 .35

0 .40

0 .45

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L -

P O 4-P

0 .00

0 .02

0 .04

0 .06

0 .08

0 .10

0 .12

0 .14

0 .16

0 .18

0 .20

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L -

A

B

Mo d elled C oncentratio n

0R iver


0 .1 % R ive r

Mo nito red D ates O nly C o nce ntration in

W e tland


109

Figure 28: Validation of simulation results for Lock 6 wetland of Macrophyte Biomass,


M acrophyte Biomass

0

1

2

3

4

5

6

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

kg

/m3

-

Phytoplankton

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

20.00

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

--

Zooplankton

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

-=

A

B

C

Mo d elled C oncentratio n

0R iver


0 .1 % R ive r

Mo nito red D ates O nly C o nce ntration in

W e tland


110

Category 4: Dead end wetlands with carp presence and irrigation drainage (Reedy

Creek wetland)

The Reedy creek wetland data set monitored by Wen (2002a) includes time-series for

the water quality of the wetland, the River Murray and the irrigation drainage

originating from the adjacent Basby farm. A period of 12 months with high internal

wetland nutrient variability (1st Jun 2000 to 31

st May 2001) was chosen from the data

set, to represent the condition of Reedy Creek wetland. Figure 29A & B and Figure

30A, B & C contain the simulated results for PO4-P, NO3-N, macrophytes,

zooplankton and phytoplankton respectively for Reedy Creek wetland. The monitored

concentrations for PO4-P, NO3-N, and phytoplankton are displayed in Figure 29A, B

and Figure 30C; the error bars represent the mean error for the entire monitoring

period of 20th

October 1999 to 16th September 2001.

A limitation of the drainage inflow time-series is that it was obtained from one source,

that being a small drainage inflow from Basby farm. The catchment area of Reedy

creek is 315 km2, whereas Basby farm covers an area of 85ha (0.85 km

2) (Wen

2002a). The Reedy Creek catchment area results in significant natural flows and

nutrient loadings to Reedy Creek wetland in response to precipitation. Unfortunately,

no monitoring data existed of the nutrient inflow from Reedy Creek, as this was not

required for the project responsible for the monitoring. Its contribution was therefore

approximated by higher surface runoff and irrigation drainage into Reedy Creek

wetland than was monitored at the one source; inflow from Reedy Creek catchment is

known to grow to a substantial amount following rains in the region (Frears 2006).

Accordingly it was assumed that the expected seasonal precipitation (described in

section 2.3.1) would have reflected the relative seasonal flow pattern over the

modelling timeframe. The monitored drainage source would have reflected the

average concentration of nutrients per unit volume expected from surrounding farms

contributing to the Reedy Creek. In order to determine the most appropriate flow,

multiple scenarios were run each with an increasing multiplication of the irrigation

volume entering the wetland. The best fit was chosen depending on the deviation of

modelled values from the monitored values D (Table 3). As with previous wetlands

the best values D for the river exchange was separately modelled.


111

As seen in Figure 29A & B and Figure 30C there was a significant improvement in

the modelling results of both PO4-P and NO3-N, and a considerable improvement on

the modelling of phytoplankton. The PO4-P results in Reedy Creek wetland improved

clearly through the introduction of the irrigation drainage inflow; however the D

(Table 3) shows the river exchange flow to have the greater impact. As can be seen in

Figure 29A this result is skewed by a particularly good fit for a short period from

March to the end of May. The combination of both river exchange flow and irrigation

drainage not only produced the best D for both PO4-P and NO3-N, but also showed a

better fit when the trend is observed as seen in Figure 29A. The Reedy Creek PO4-P

modelling shows the most significant improvement of PO4-P simulation when

compared with the other modelled wetlands. NO3-N is influenced by both the river

flow exchange and the irrigation drainage inflow to produce a significant

improvement in model fit D (Table 3). The phytoplankton modelling of Reedy Creek

wetland shows a considerable improvement in D through the introduction of river

exchange and drainage.

Some of the extreme events in PO4-P and NO3-N concentrations from October to

December (Figure 29A) were not realistically simulated by the model, although the

trend is clearly visible. A limitation of the generic nature of the model WETMOD2

may be that short lived and extreme events cannot be successfully simulated.

Reedy Creek wetland is in a turbid state with minor macrophyte growth (section

3.2.1). The macrophyte growth curve shown in Figure 30A is a result of the high

turbidity, which limits underwater light for growth. The zooplankton, lacking the

shelter assumed to be provided by macrophytes, are reliant on the phytoplankton as

their food source. The zooplankton growth, seen in Figure 30B, closely follows the

phytoplankton growth seen in Figure 30C. As seen in Figure 30C, a combination of

both river exchange and irrigation drainage inflow was required for phytoplankton to

resemble the monitored and therefore expected concentrations. This further

strengthens the validation of the model, showing that one external influence such as

the river exchange is not enough to drive the simulation for a wetland such as Reedy

Creek wetland. But rather the combinations of external influences such as the river

flow exchange and irrigation drainage are required to successfully and

comprehensively simulate the Reedy Creek wetland.


112

Figure 29: Validation of simulation results for Reedy Creek wetland of PO4-P, and NO3-N for

both conditions with and without water exchange

PO4-P

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Jun

-00

Jul-

00

Au

g-0

0

Se

p-0

0

Oct

-00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L -

-1

0

1

2

3

4

5

6

7

8

mg

/L D

rain

ag

e o

nly

-

---

NO3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Jun-0

0

Jul-00

Aug-0

0

Sep-0

0

Oct-

00

Nov-

00

Dec-0

0

Jan-0

1

Feb-0

1

Mar-

01

Apr-

01

May-0

1

mg

/L --

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

mg

/L D

rain

ag

e o

nly

-

---

A

B


No R ive r E xc hang e No I rrig atio n

Mod elled C o nce ntra tio n

0% R ive r E xc hang e

35 00 X m onito red I rrig atio n In f low


3 .5 % R iver E xchange 0 I rrigatio n Inf lo w


3 .5 % R iver E xchange

3 50 0 X m onito re d I rriga tio n Inf lo w



Mo nito re d D ate s Only

C oncentratio n in I rrigatio n D rain


113

Figure 30: Validation of simulation results for Reedy Creek wetland of Macrophyte Biomass,


Phytoplankton

0

2

4

6

8

10

12

14

16

Ju

n-0

0

Ju

l-00

Au

g-0

0

Sep

-00

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Mar-

01

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

--

-20

0

20

40

60

80

100

120

140

160

cm

3/m

3 D

rain

ag

e o

nly

----

Zooplankton

0

0.5

1

1.5

2

2.5

3

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

--

M acrophyte Biomass

0

0.02

0.04

0.06

0.08

0.1

0.12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

--

A

B

C


No R ive r E xc hang e No I rrig atio n

Mod elled C o nce ntra tio n

0% R ive r E xc hang e

35 00 X m onito red I rrig atio n In f low


3 .5 % R iver E xchange 0 I rrigatio n Inf lo w


3 .5 % R iver E xchange

3 50 0 X m onito re d I rriga tio n Inf lo w



Mo nito re d D ate s Only

C oncentratio n in I rrigatio n D rain


114

Category 5: Dead end wetlands managed through implementation of dry periods with

carp restriction and no irrigation drainage (Pilby Creek wetland)

Figure 31 represents simulation results for PO4-P and NO3-N concentrations in Pilby

Creek wetland and Figure 32 the simulation results for macrophyte, zooplankton, and

phytoplankton biomass within the wetland. The error bars represent the standard error,

of three separate measurements, of the monitored concentration for each monitoring

date.

Pilby Creek wetland is a dead end wetland adjacent to Lock 6 wetland (Category 3).

Pilby Creek wetland is managed by artificial drying and wetting cycles resulting in

sediment compaction. Restriction on the presence of large bottom-feeding fish such as

carp, which are believed to stir up wetland sediment, is also believed to have

contributed to reduced turbidity. The case study for Pilby Creek wetland was included

in the modelling project to test the model validity for a restored wetland.

Although Pilby Creek wetland is not directly connected to the river, as well as being a

dead end wetland, an exchange of water and nutrient with the river was assumed. The

justification for this assumption is the possibility of an exchange through Pilby creek,

which flows through at one end of the wetland (see Figure 14). The possible nutrient

load change during the exchange through an intermediary creek should be taken into

consideration when assessing the modelling success of this wetland. The model

results support the assumption of water exchange through Pilby creek, as the model

scenario D improves with the introduction of river flow exchange (Table 3). The D

shows a considerable improvement for the PO4-P modelling (Table 3). The peak

concentration of PO4-P simulated by the river exchange scenario (Figure 31A) was

due to both a high peak in river flow and high river PO4-P concentration (see section

3.2.3). This nutrient peak did not reach the wetland during the monitoring period as

indicated by the internal wetland nutrient monitoring (Figure 31A), which may be due

to the lag time of nutrient flow to Pilby Creek wetland from the River Murray. The

NO3-N curve is lower than expected during late February until April. However, with

the exception of an extreme event at the end of April the curve does show a similar

trend to that of monitored concentrations (Figure 31B), which is not as apparent in the

simulation without the river exchange. The improvement in NO3-N simulation is also

reflected by the D value (Table 3).


115

Following a drying period of two months in 1997, of Pilby Creek wetland, that was

long enough to compact the sediments the high macrophyte growth seen in Figure

32A was a result of low turbidity as expected in a managed wetland within a short

time after re-flooding. The macrophyte biomass decreased over the winter months

with low water temperatures but increased during spring. Monitoring ceased at the

beginning of October.

The observed phytoplankton growth in Figure 32C showed a rapid growth phase prior

to the macrophyte growth, directly following wetland re-flooding. In this instance, the

phytoplankton took advantage of the lack of competition as well as the high nutrient

availability. Once competition set in with the growth of macrophytes, there was a

reduction in the phytoplankton biomass. The phytoplankton biomass growth was

thereby restricted until the decreasing macrophyte biomass in winter when

phytoplankton again took advantage of less nutrient competition and increased its

biomass. The phytoplankton had a faster response time in growth than macrophytes at

the onset of the warmer period of spring. As with Lock 6 wetland, the river

phytoplankton was derived from river chlorophyll-a levels monitored further

downstream.

The zooplankton growth can be linked to the provision of a nourishment source, the

phytoplankton growth, and possibly to a lesser extent the assumed provision of a

shelter from predators by macrophytes (Figure 32). The lowest number of

zooplankton occurred when there was a combination of both low phytoplankton and

low macrophyte biomass. The lack of phytoplankton as a food source explains the

reduction in zooplankton observed despite the potential supply of shelter provided by

the macrophytes. The secondary growth phase of zooplankton corresponded to the

secondary growth phase of the phytoplankton. During the spring growth phase of

phytoplankton the zooplankton follows suit, again possibly as a consequence of

shelter provided by the increase in macrophyte growth. The modelled growth

behaviour of the macrophytes, phytoplankton and zooplankton described follows

expectations of a wetland in the Pilby Creek wetland category (category 5). It must

however be remembered that no data was available to validate model output for

zooplankton and macrophyte biomass.

It is interesting to note that the growth of phytoplankton in a category 5 wetland

(Pilby creek) was less than in a category 3 wetland (Lock 6). This can be attributed to


116

the competition between the macrophytes and the phytoplankton in Pilby Creek

wetland, which is virtually absent in Lock 6 wetland. However, Pilby Creek wetland

shows a relatively greater zooplankton growth than Lock 6 wetland when compared to

the phytoplankton availability in each of the wetlands. The cause of the relatively

larger zooplankton growth in Pilby Creek wetland may be as a consequence of added

shelter opportunity within Pilby Creek wetland assumed to be provided by the

macrophytes. The only discrepancy in the modelling of Pilby Creek wetland is the

very late spike in PO4-P levels described earlier, attributed to river flow and river

nutrient concentration.

Figure 31: Validation of simulation results for Pilby Creek wetland of PO4-P, and NO3-N for both


NO3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

-

PO4-P

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

--

A

B


No R ive r E xchange

Mode lled C o ncentratio n

0 .3 3% R ive r E xchange

Mo nito re d D ates O nly

C once ntration in W e tland


117

Figure 32: Validation of simulation results for Pilby Creek wetland of Macrophyte Biomass,


Phytoplankton

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

--

Zooplankton

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

--

M acrophyte Biomass

0.00

20.00

40.00

60.00

80.00

100.00

120.00

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

kg

/m3

--

A

B

C





Mo nito red D ate s O nly



118

4.1.1 Implication for irrigation affected wetland representation

Considering the generic nature of the model and its structural restrictions and how this

interacts with potential quantitative modelling performance, the qualitative modelling

performance, the time and data available for model development and most

importantly the project goals, the model displays the potential of a developed tool

with purpose designed monitoring scenarios. The following discussion aims to

represent the performance of the model in a dispassionate approach, focusing on

where it has succeeded in fulfilling its objective and is at a stage where it can be

applied to answer wetland specific management questions and therefore fulfilling the

project aims.


Creek wetland)

The modelling results from Reedy Creek wetland are an example of a successful

simulation of a wetland that is affected by irrigation drainage. Both the quality of the

trend as well as the statistical comparison improved with the introduction of irrigation

drainage. The methodology of estimating the inflow volumes from the Reedy Creek

catchment can at this stage not be confirmed as no monitoring of the sub-catchment

inflow was performed concurrent with the wetland-monitoring project. However,

although the inflow volume used in the model may be debateable, the methodology of

the model derived optimum level gives future modellers the option to adjust the

scenarios as this data becomes available. Any consequent monitoring could

potentially refute or confirm the range of estimated nutrient and volume inflow. It is

therefore not regarded as a high priority at this stage to invest expense and time in the

improvement of the Reedy Creek wetland modelling scenarios. The validation of the

macrophyte and zooplankton modelling output may however increase the confidence

in the model. Future monitoring could assist in this regard by providing adequate data

for model validation.

Category 2: Through flow wetlands with carp presence and irrigation drainage

(Sunnyside wetland)

Modelling scenarios of Sunnyside wetland improved with the introduction of river

exchange. However, the inflow of monitored irrigation drainage did little to improve


119

the scenario performance. The monitoring of Sunnyside wetland was not designed

with this project in mind. Bartsch (1997) designed her monitoring project with the

sole intention of comparing the two wetlands Paiwalla and Sunnyside; and therefore

study the impacts irrigation drainage has had on Sunnyside wetland. Minor effort was

therefore made to assess internal wetland dynamics by that project. Due to Bartsch‟s

(1997) project aims, and particularly the need to assess the impact of irrigation

drainage into Sunnyside wetlands, most of the monitoring sites were located at one

end of the wetland and close to the drainage outlet. The monitoring of nutrients was

made mainly in, what can be recognised in aerial photos as, a channel through the

macrophyte growth leading from the irrigation drainage outlet to the river (see Figure

33).

Figure 33: Sunnyside monitoring area

One of the central assumptions made in this modelling project is that all wetlands are

homogeneously mixed from one time step to the next. However, partly due to the

sporadic point source inflow of nutrients (irrigation drainage) the concentrations of

nutrients are highly variable within the wetland. Further, the nature of Sunnyside

wetland, macrophyte growth within the wetland and the channel through the


120

macrophytes from the irrigation drainage source to the river (see Figure 33), hampers

the mixing of water and nutrients within the wetland. Sunnyside wetland, due to its

highly variable nature, can therefore not be considered as homogeneously mixed.

Monitoring within and close to the channel will therefore represent the concentration

of nutrients entering the wetland. If this concentration is then assumed to encompass

the entire wetland, the true concentration and particularly the inflow of irrigation

drainage will be over represented. The monitoring of irrigation flow did not include

volume. It is therefore not possible to estimate the true impact the irrigation drainage

has on the concentration of nutrient within the wetland.

Another issue impacting on the use of the model to estimate a realistic irrigation

inflow scenario may be due to the drainage inflow being sporadic (infrequent and

short-lived), despite our assumptions of daily pumping. However, the methodology

available to estimate the best-fit scenario shows a slight change in model fit as the

drainage inflow only affects a minimum number of monitored dates. That does not

mean the model proves there to be an insignificant detrimental impact on the wetland

due to drainage, but rather that the infrequent nature and the unknown exact drainage

volume for each particular pumping date complicates the modelling estimate. The

cumulative impact of the drainage inflow on the wetland would however still persist

as suggested by the slight change in model fit.

4.1.2 Implication for wetland representation

Comparison of wetlands Paiwalla and Sunnyside

For both Paiwalla wetland and Sunnyside wetland, being similar wetlands and in close

proximity, the modelling scenarios performed well enough to allow a comparison.

Sunnyside wetland scenario underestimated PO4-P considerably. The D (Table 3) is

due to the low variability of PO4-P concentration in the wetland. Further, from early

April to the end of the simulation period the PO4-P trend was simulated very well

(Figure 25A). The Paiwalla wetland scenario PO4-P D (Table 3) showed a somewhat

worse fit than the Sunnyside wetland scenario, although the concentrations within

Paiwalla wetland scenario are for the most part closer to the monitored (Figure 23A).

In the Paiwalla wetland scenario the model fails to mimic the PO4-P trend as well as it

does in the Sunnyside wetland simulation (Figure 23A and Figure 25A). The most


121

likely reason for under prediction in Sunnyside wetland was discussed in section

4.1.1. The Paiwalla wetland scenario PO4-P prediction, although with a worse fit than

anticipated, did improve considerably with the introduction of river exchange.

The Sunnyside wetland NO3-N simulation performance was good with a low D (Table

3) and with a trend, displayed in the time-series, showing a very good fit (Figure

25B). The Paiwalla wetland D (Table 3) of NO3-N simulation, although not poor, is

an indication to the potential failings of D, if used alone, in assessing a comparative

modelling output. This statement is made as, although this is not evidenced in the D,

the trend or rather the time-series fit (Figure 23B) shows a great improvement with

the introduction of river exchange.

The Paiwalla wetland D (Table 3) for phytoplankton improves dramatically and can

also be seen in the time-series display (Figure 24C) and provides a strong argument

for the validity of WETMOD 2 phytoplankton simulation capacity. The Sunnyside

wetland phytoplankton simulation also improves with the introduction of river

exchange both as represented by the D (Table 3) and by the visual trend assessment

(Figure 26C).

The macrophyte biomass increase within the Paiwalla wetland scenarios is low, with a

rapid decline following the initial growth phase (Figure 24A). The cause of the

decline is related to the turbidity level within the wetland limiting underwater light

penetration. This monitored wetland turbidity does not increase in the Sunnyside

wetland scenario until two weeks later, therefore allowing for a longer macrophyte

growth phase. The later increase in turbidity in Sunnyside wetland is assumed to be as

a result of the higher macrophyte levels within Sunnyside wetland that act both to

settle out the turbidity and to reduce sediment re-suspension. The lower exchange

volume of Sunnyside wetland (Table 3) is also assumed to be as a result of the

macrophyte growth, whereby the water flow through Sunnyside wetland in

comparison to Paiwalla wetland is reduced. The significance of the difference in the

macrophyte growth phase between the two wetlands is reflected in the phytoplankton

time-series. Where, as a consequence of competition for nutrients and light the

Sunnyside wetland scenario shows a very small summer phytoplankton growth phase

compared with Paiwalla (Figure 26C vs. Figure 24C). Another consequence of the

higher macrophyte biomass content within the Sunnyside wetland scenario is the

habitat availability assumed to be provided to zooplankton represented by the higher


122

summer zooplankton biomass compared with the winter biomass (Figure 26B). In

contrast in Paiwalla wetland, with low macrophyte biomass, the zooplankton growth

(Figure 24B) mimics the phytoplankton growth (Figure 24C) more closely. The

macrophyte and zooplankton model output assessments are limited however by the

lack of validation data.

Comparison of wetlands Lock 6 and Pilby

Lock 6 wetland and Pilby Creek wetland are located geographically close. Prior to the

management of Pilby Creek wetland they were both in a similar degraded state.

Unfortunately no monitoring of Pilby Creek wetland was undertaken prior to

management so no direct comparison can be made at this time of simulations of a

particular wetland in a degraded and in a restored state.

For Lock 6 wetland, both the PO4-P and phytoplankton D (Table 3) increased once

river exchange was introduced. However, a visual assessment of the time-series trend

(Figure 27A and Figure 28C) showed a marginal improvement in both cases. The

improvement in the Lock 6 NO3-N simulation performance was exceptionally good

both visually (Figure 27B) and according to D (Table 3 reducing by a full 20%),

supporting the claim that the simulation of Lock 6 was successful. This discrepancy in

PO4-P and phytoplankton results was not seen in the Pilby Creek wetland scenarios,

where there was an improvement in both D (Table 3) and the visual assessment

(Figure 31A and Figure 32C). Both wetlands showed the assumed expected

macrophyte biomass growth trends (Figure 28A and Figure 32A). In the Lock 6

wetland scenario there was a rapid decline from initial macrophyte biomass and in

Pilby Creek wetland there was a substantial macrophyte biomass increase post re-

wetting followed by an expected winter reduction. The phytoplankton biomass

growth, in both wetlands (Figure 28C and Figure 32C), responded appropriately to the

level of competition expected in respect to the macrophyte biomass present (Figure

28A and Figure 32A). In the Lock 6 wetland scenario low macrophyte competition

caused phytoplankton biomass to reach high levels, only matched by Reedy Creek

wetland, which can be viewed as another wetland with high nutrients concentrations

(Figure 29) and low macrophyte competition (Figure 30A). The phytoplankton

biomass in the Pilby Creek wetland scenario matched the time-series trend expected,

with a growth phase both prior to and directly following the macrophyte growth phase


123

(Figure 32A and C). Zooplankton in the Pilby Creek wetland scenario responded to

the shelter availability assumed to be afforded by macrophytes. However, the

zooplankton in Pilby Creek wetland (Figure 32B), despite being relatively more

abundant when compared to the phytoplankton availability in both wetlands (Figure

32 and Figure 28), were restricted by the low food source of phytoplankton (Figure

32C). Whereas in the Lock 6 wetland scenario there was a large zooplankton biomass

increase (Figure 28B) due to the ample nutrient source the phytoplankton biomass

(Figure 28C). The ample phytoplankton biomass therefore minimised the otherwise

negative impact of the lack of habitat normally provided by macrophytes (Figure

28A).

The good scenario trend results provided by the model in the case of Lock 6 wetland

and Pilby Creek wetland confirms the applicability of WETMOD 2 to wetlands in

both extreme stable states (turbid and clear). The model can therefore be applied with

confidence to category wetlands belonging to either Lock 6 wetland or Pilby Creek

wetland (category 3 and 5). This confidence being both placed in the representation of

the realistic trend of wetland nutrient concentration as well as in the impact respective

external nutrient sources have upon the wetlands. However, as stated for Paiwalla and

Sunnyside wetlands the macrophyte and zooplankton model output assessments are

limited by the lack of validation data.

Comparison of wetlands Sunnyside and Reedy Creek

The main difference between the two wetlands is the data quality and quantity. Reedy

Creek wetland has more comprehensive data so is more suitable for modelling

purposes. The Reedy Creek wetland simulation succeeds where the Sunnyside

simulation struggles. Results from Reedy Creek wetland simulations provide the

strongest argument for the validity of WETMOD 2.

For the Reedy Creek wetland scenarios, as can be seen by the D in Table 3 and the

wetland time-series data in Figure 29 and Figure 30, there are obvious improvements

in the model output both with the introduction of river exchange, as well as the

introduction of irrigation drainage nutrient inflow. There were significant

improvements in the overall modelling performance at Reedy Creek wetland for NO3-

N and PO4-P modelling (Figure 29) as well as considerable improvement in

phytoplankton modelling performance (Figure 30C). Visual assessment of Figure 29


124

and Figure 30 shows the model to simulate the Reedy Creek nutrient and

phytoplankton time-series trend satisfactorily. The Reedy Creek wetland macrophyte

simulated biomass is low due to the high turbidity and low Secchi depth, the

zooplankton therefore mimicking only the growth of its food source the

phytoplankton.

WETMOD 2 shows great success with the notable performance in simulating Reedy

Creek wetland. The results of Reedy Creek wetland simulations support the argument

that the model is capable of simulating wetlands with both river and irrigation

drainage as external nutrient sources. Therefore, the reasoning that the poor data

quality for Sunnyside wetland affects its simulation performance is justified based on

the successful Reedy Creek wetland scenarios.


125

4.2 Validation based on non-calibrated wetland data

WETMOD 2 has a generic nature; through the use of wetland categories and its

simplicity it is applicable to wetlands and timescales other that where it was

developed. In order to verify the model applicability at different timescales and

wetlands, the model must show itself to be accurate outside of the data range where it

was developed. Therefore, to rigorously test the model, it should be fitted to one set of

data, while checking for agreement with independent data (Goodall 1972; Tsang

1991; Wood 2001). Extra validation therefore, not only serves the validation of the

model for the monitored wetlands, but also supports the argument of the models

generic applicability. If the model is capable of accurately simulating a separate set of

data than used in the calibration, the acceptance of the qualitative simulations for

category wetlands where no time-series are available should be strengthened.

For the purpose of rigorous validation, some of the data for Reedy Creek wetland

(category 4 wetland) were withheld during the model calibration stage. This extra data

stems from the same source project and covers the seven months prior to the data used

in the model calibration stage (the data used in the model calibration stage spanned

one year, see Box).

Following the monitoring project that provided data for the modelling of Lock 6 and

Pilby Creek wetlands, another project monitored the same wetlands. The data from

this second monitoring study, performed by van der Wielen (nd), was kept separate

from the data used in the model development. It is therefore also possible to validate

the developed WETMOD 2 on the data not used in 3 of the 5 category wetlands.

The method used by van der Wielen in assessing the NO3-N concentration was a

colorimetric method (Cadmium Reduction Method) (van der Wielen nd). Colorimetric

methods require an optically clear sample as the turbidity of a sample can conflict

with the colorimetric measurement (APHA et al. 1992). After discussions with van

The time period chosen, for Reedy Creek wetland data, in the model development

stage was due to a two significant factors;

1. It was a highly variable year therefore providing the model with complex

data and dynamics.

2. It was from the winter period of low growth to the next winter period (so it

encompassed an entire growth cycle)


126

der Wielen (nd), it was considered likely that the very turbid waters of the River

Murray wetlands compromised the monitored NO3-N values. As a consequence the

NO3-N measurements in the Pilby Creek and Lock 6 data set cannot be relied upon. In

this case the modelled PO4-P compared to the monitored PO4-P gives the best

estimation of model validity. The D for the modelled results of the validation data is

presented in Table 4 below. The individual results are discussed below.

Table 4: Non calibrated validation of inflow data for 3 wetland categories

Wetland Category

Wetland Name

Wetland External Input Variables

Modelled PO4-P

D

Modelled NO3-N

D

Modelled Phytoplankton

D

3 Lock 6 Wetland




4 Reedy Creek Wetland


95 71 55

NO River Exchange & 3500 X Irrigation Drainage Time Series

86 70 54

3.5% River Flow/day Exchange & NO Irrigation Drainage

80 52 35

3.5% River Flow/day Exchange & 3500 X Irrigation Drainage Time Series

74 52 35

5 Pilby Creek Wetland



Category 3: Dead end wetlands with carp presence and no irrigation drainage (Lock

6 wetland)

The simulated time-series for the non calibrated data validation of WETMOD 2 for

category 3 wetlands are presented in Figure 34 and Figure 35. The standard error, at

each monitoring date, is represented for PO4-P, NO3-N and phytoplankton biomass.

The wetland scenario did not initially perform as well as was expected. The D actually

degraded with the introduction of exchange (Table 4). The time-series graph in Figure

34 however, does show an improvement in the modelling trend after the introduction

of the river exchange. For this scenario the default exchange volume for Lock 6

wetland was kept at the same level as used during the model development stage.


127

However, during simulations using monitored data from three different locations

within Lock 6 (available in van der Wielen‟s data), it was discovered that the impact

of the river exchange diminishes as the distance of the monitoring location from the

river channel increases. It was therefore assumed to be reasonable to examine a

different exchange volume as the monitoring site locations within the wetland

differed. A reduced exchange rate at 0.05% of the river daily flow volume showed an

improved D (Table 4) and a well fitting time-series as can be seen in Figure 34A.

The Lock 6 D improvement for PO4-P is noteworthy despite the model not being

calibrated for this data. Therefore, the model was considered valid for the PO4-P

scenario within Lock 6 wetland. However, the NO3-N and phytoplankton D results

were poor. As discussed previously the NO3-N monitored data was to be considered

with scepticism and cannot be relied upon. Looking at the result in Figure 34B one

can however see a slight improvement in NO3-N estimation during October 1998.

Based on this scenario and due to the unreliable nature of the monitored NO3-N the

model can, for NO3-N wetland concentration simulation, neither be considered valid

nor invalid.

In Figure 35C the phytoplankton shows an improvement, despite the D results, during

the October 1998 to January 1999 modelled period. For this scenario the

phytoplankton modelling results show a significant overestimation for the modelled

period. However, due to the performance of the model with regard to phytoplankton,

both during model development and in the following validation scenarios at other

wetlands described below, the model should not yet be considered invalid. Future

model development should focus on addressing the phytoplankton discrepancy, which

may be as simple as the monitoring methodology, the conversion of chlorophyll-a to

phytoplankton or addressing the sediment impact on wetland water nutrient load. In

the mean time phytoplankton volume estimation from modelling scenarios should be

reviewed carefully before management decisions are made based on the model results.


128

Figure 34: Validation of simulation results for Lock 6 wetland PO4-P and NO3-N, using non-

calibrated wetland data

PO4-P

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oct-

98

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

Ma

r-9

9

Ap

r-9

9

Ma

y-9

9

Ju

n-9

9

Ju

l-9

9

Au

g-9

9

Se

p-9

9

Oct-

99

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

mg

/L -

NO3-N

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

Ma

r-9

9

Ap

r-9

9

Ma

y-9

9

Ju

n-9

9

Ju

l-9

9

Au

g-9

9

Se

p-9

9

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

mg

/L -

A

B

Mode lled P O 4-P m g /L

0R ive r


0 .1% R ive r

Monito re d D ate s O nly P O 4-P m g /L in

W e tland

Mo d e lle d P O 4 -P m g /L

0 .0 5 % R ive r


129

Figure 35: Validation of simulation results for Lock 6 wetland Macrophyte Biomass,

Zooplankton and Phytoplankton biomass, using non-calibrated wetland data


0R ive r


0 .1% R ive r

Monito red D ates O nly P O 4-P m g /L in

W e tland


0 .05% R ive r

Phytoplankton

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

Ma

r-9

9

Ap

r-9

9

Ma

y-9

9

Ju

n-9

9

Ju

l-9

9

Au

g-9

9

Se

p-9

9

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

cm

3/m

3

--

Z o o p la n k to n

0

0 .2

0 .4

0 .6

0 .8

1

1 .2

1 .4

1 .6

1 .8

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

Ma

r-9

9

Ap

r-9

9

Ma

y-9

9

Ju

n-9

9

Ju

l-9

9

Au

g-9

9

Se

p-9

9

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

cm

3/m

3

-=

M acrophy te Biom ass

0

1

2

3

4

5

6

No

v-9

7

De

c-9

7

Ja

n-9

8

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

Ma

r-9

9

Ap

r-9

9

Ma

y-9

9

Ju

n-9

9

Ju

l-9

9

Au

g-9

9

Se

p-9

9

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

kg

/m3

-

A

B

C


130


Creek wetland)

Reedy Creek wetland data provided the best data for model development. The data

from Reedy Creek wetland withheld during model development also provided the

most comprehensive and reliable data for extensive model validation based on non-

calibrated driving variables. Figure 36 and Figure 37 display the simulated output for

the non-calibrated data validation of category 4 wetlands. The standard error at each

monitoring data based on three separate measurements is included where appropriate.

Evaluating these scenario results for PO4-P simulation there is a significant

improvement in the D (Table 4) where both river exchange and irrigation drainage are

considered. Although the model was not calibrated for this time-series the results

show a satisfactory resemblance to the monitored PO4-P as seen in Figure 36A.

The NO3-N D (Table 4) during this time-series actually shows a better fit to the

monitored data than the original calibrated data time-series. This can be attributed to

the high variability in the development data series, which were partly chosen as a

consequence of this variability. The time-series seasonality and fit can be seen in

Figure 36B. The NO3-N modelling result is the only NO3-N data available with which

to verify the model outside of the data used in model development. The notable

improvement in the improvement of the D and the good fit shown in Figure 36B

provide a strong case for the validity of WETMOD 2 with regard to NO3-N

simulation. The phytoplankton D shows a significant improvement (Table 4),

although the seasonality is somewhat exaggerated as seen in Figure 37C.

The performance of WETMOD 2 for Reedy Creek wetland, with both data sets

calibrated and non-calibrated, demonstrates the performance that can be obtained

when adequate data is available. The model performance for Reedy Creek wetland is

the strongest argument in the favour of model validity. Therefore, the shortcoming of

the model in previous instances can to a large degree be attributed to data quality.

The Reedy Creek wetland results show that the availability of adequate quality data

improves the performance of the model. However, it is a generic modelling tool where

simple data sets can be used giving reasonable trends, thereby assisting potential

management decisions. The lack of quality data should in this case not necessarily

hinder scenario analysis however; the decision maker must understand that the quality


131

of the modelling output is very dependent on the quality of the data used as driving

variables.

Figure 36: Validation of simulation results for Reedy Creek wetland PO4-P and NO3-N, using

non-calibrated wetland data

N O3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Ap

r-0

0

Ma

y-0

0

mg

/L

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

mg

/L D

rain

ag

e o

nly

PO4-P

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Ap

r-0

0

Ma

y-0

0

mg

/L

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

mg

/L D

ra

ina

ge o

nly

A

B

Mod e lle d C once ntratio n

No R ive r E xchang e No I rrig atio n

Mod e lle d C o nce ntratio n

0% R iver E xc hange

35 0 0 X m o nito re d I rrig atio n Inf low

Mo d elle d C once ntratio n

3 .5% R iver E xc hange 0 I rrig atio n Inf lo w


3.5 % R ive r E xc hange

35 0 0 X m o nito red I rrig atio n Inf lo w

Monitore d D ate s O nly

C o nc entra tion in W e tland


C o nce ntratio n in I rrigation D rain


132

Figure 37: Validation of simulation results for Reedy Creek wetland Macrophyte Biomass,


M acro ph y te B io m a ss

0

0.02

0.04

0.06

0.08

0.1

0.12

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Ap

r-0

0

Ma

y-0

0

kg

/m3

--

Zoo plan kton

0

0.5

1

1.5

2

2.5

3

3.5

4

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Ap

r-0

0

Ma

y-0

0

cm

3/m

3

--

Phytoplankton

0

5

10

15

20

25

30

Oc

t-9

9

No

v-9

9

De

c-9

9

Ja

n-0

0

Fe

b-0

0

Ma

r-0

0

Ap

r-0

0

Ma

y-0

0

cm

3/m

3

--

-20

0

20

40

60

80

100

cm

3/m

3 D

rain

ag

e o

nly

----

A

B

C


No R ive r E xchang e No I rrig atio n

Mod e lle d C o nce ntratio n

0% R iver E xc hange

35 0 0 X m o nito re d I rrig atio n Inf low

Mo d elle d C once ntratio n

3 .5% R iver E xc hange 0 I rrig atio n Inf lo w


3.5 % R ive r E xc hange

35 0 0 X m o nito red I rrig atio n Inf lo w


C o nc entra tion in W e tland


C o nce ntratio n in I rrigation D rain


133

Category 5: Dead end wetlands managed through implementation of dry periods with

carp restriction and no irrigation drainage (Pilby Creek wetland)

The non-calibrated driving variable validation of WETMOD 2 based on category 5

wetlands is displayed in Figure 38 and Figure 39. The standard error of the monitored

data set is included where available.

Pilby Creek wetland PO4-P in this case has a noteworthy improvement in D.

Nevertheless, the improvement is best judged from the time-series in Figure 38A

where the concentrations in the early modelling period were very close to the

monitored concentrations. The September to February performance is exaggerated but

is at least showing a similar trend to the monitored data. The main discrepancy in the

PO4-P modelling is the lack of the late November early December peak.

Pilby Creek wetland validation data stems from the same source as the Lock 6

validation data. The NO3-N monitoring results are therefore, as in the case of Lock 6

wetland, to be considered suspect and therefore no model validation will be made

based on NO3-N model output for this data.

The phytoplankton biomass growth is greater than expected (see Figure 39C)

particularly the initial peak growth phase, which is due to the lack of macrophyte

competition. However, the model scenario does retain a low phytoplankton biomass

load as is expected of Pilby Creek wetland given the simulated macrophyte biomass.

The macrophyte biomass growth does show an increase; followed by a winter

decrease (see Figure 39A). The zooplankton biomass pattern as can be seen in Figure

39B follows both its food source pattern, i.e. phytoplankton, and assumed shelter

availability afforded by the macrophytes. The zooplankton does in this instance have a

more complex growth pattern than the phytoplankton due to the high shelter

availability provided by the macrophytes.

From the modelling results in this case as well as the two above, the model has shown

itself capable of simulating wetlands for which it has been calibrated, but with non-

calibrated data sets. Each of these wetlands is either in a different stable state, i.e.

clear vs. turbid, or has added external influences (Reedy Creek wetland irrigation

drainage inflow). This supports the argument that the model is generically applicable

to similar wetlands. Where data for these similar wetlands is non existent, the

accuracy WETMOD 2 trend development allows the use of “exemplar” data obtained


134

from the calibration wetlands, and consequently the development of qualitative

scenarios and hypothetical quantitative outcomes. The application of WETMOD 2 to

category wetlands in such a manner is explored in chapter 6.

Figure 38: Validation of simulation results for Pilby Creek wetland PO4-P and NO3-N, using non-

calibrated wetland data

NO3-N

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

mg

/L

PO4-P

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

mg

/L

--

A

B



Mode lled C o ncentratio n

0 .3 3% R ive r E xchange


C once ntration in W e tland


135

Figure 39: Validation of simulation results for Pilby Creek wetland Macrophyte Biomass,


M acrophy te B iom ass

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

kg

/m3

--

Phy toplankton

0.00

20.00

40.00

60.00

80.00

100.00

120.00

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

cm

3/m

3

--

Zooplankton

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

Oc

t-9

8

No

v-9

8

De

c-9

8

Ja

n-9

9

Fe

b-9

9

cm

3/m

3

--

A

B

C

Mo d e lle d C o nc e ntratio n

No R ive r E xc hang e

Mo d e lle d C o nc e ntratio n

0 .3 3 % R ive r E xc hang e


C o nc e ntratio n in W e tland


136

4.3 Evaluating model performance

4.3.1 Generic nature and structural restrictions of model

When wetland scenario results are evaluated and compared, WETMOD 2 performs

satisfactorily and as expected, even for wetlands with extreme conditions of turbid

and clear. The quantitative results may not reflect the accuracy expected of a

dedicated wetland model. However, as discussed in the introduction, the limiting

model structure, the lack of data availability and the models generic nature does not

allow for WETMOD 2 to be fitted to one wetland in particular. This allows the model

to be applied to a larger range of wetlands, even where verification may not be

possible, with confidence in the simulation results qualitative trend. Therefore, in

developing WETMOD 2, a compromise on quantitative accuracy was made in order

to be able to compare the relative conditions of wetlands, including impacts external

influences may have on the wetlands and/or wetlands with minimal or no time-series

data.

The data quality available for a given wetland has a direct impact on the accuracy of

WETMOD 2 to simulate internal nutrient dynamics, as seen for Sunnyside and Reedy

Creek wetlands. The potential to simulate management scenarios is directly linked

with model performance. Consequently, due to the lack of data quantity and

particularly quality for Sunnyside wetland management simulations for Sunnyside

wetland and therefore category 2 wetlands were not feasible. However, for both Lock

6 wetland and Reedy Creek wetland, scenarios of potential management strategies

were possible and are described and discussed in chapter 5. Using both Lock 6 and

Reedy Creek wetland data “exemplar” category wetlands could therefore also be

simulated, this is described and discussed in chapter 6.

In WETMOD 2 macrophyte growth is controlled to a large extent by light availability,

where the growth of macrophytes increases with a decrease in turbidity and therefore

increase in Secchi depth. This relates back to Secchi depth representation of

underwater light availability. The equation in the model assumes that at increased

Secchi depth there will be an increase in underwater light availability and therefore in

the macrophyte growth. A limitation discovered during model validation pertains to

the equation used. The equation shows a logarithmic growth curve with increasing

Secchi depth, which in itself is not regarded as inaccurate. However, the limitation of


137

this equation is its lack of consideration of the maximal depth of the wetland, i.e. there

is no correlation of the equation to the water depth and therefore the maximal light

penetration possible. Therefore, a shallow wetland with a depth less than 0.6m is not

simulated as having substantial macrophyte growth despite the underwater light being

fully available to macrophyte growth, represented by the Secchi depth effectively

penetrating to the wetland bottom. The wetlands, which were monitored and provide

the wetland time-series driving variables, are all of a depth where this restriction is not

of significant concern; and where appropriate the macrophyte growth is calibrated to

expected trends. This limitation impacts on the application of the model to very

shallow wetlands. As there currently is no model calibration data available or

sufficient data available as driving variables for very shallow wetlands this limitation

is currently not an issue. Future development of WETMOD should however take this

limitation into account and replace the current Secchi depth equation with a more

appropriate one.

4.3.2 Relevance of project objectives

The principal objective calls for the improvement of the resolution of spatial

influences acting upon wetlands. That is, to develop or adopt a generic wetland

process model to local external influences acting on a wetland. The purpose of the

objective is to improve the understanding of the respective spatial influences acting

upon a wetland, such as morphology and external nutrient sources, and how

management can impact on the nutrient retention capacity of wetlands at each spatial

location. The spatial differences considered in WETMOD 2 are any significant

external sources acting upon wetlands, including:

river nutrient load,

the presence or absence of agricultural drainage (irrigation drainage) with its

associated nutrient load contribution, and

in isolated cases the impact of precipitation on irrigation drainage nutrient

contribution. (South Australia is a very dry state with minimal precipitation.

Most wetlands in the South Australian stretch of the River Murray do not have

independent catchment areas. Precipitation is therefore in most cases not

relevant).


138

As the principal focus of WETMOD 2 development was the spatial context of the

wetlands, i.e. the individual external influences, it is important to discuss whether the

model behaves logically based on the anticipated impact of external influences as well

as the comparative differences of two wetlands. The validation of the model and the

comparison of wetlands, discussed above, have shown the successful improvement of

the model simulation output following the introduction of external influences. The

simulation outputs therefore enable the study of the local and assumed external impact

on wetland fulfilling this principal project objective.


139

4.4 Chapter summary and Implication for the first hypothesis

Based on the validation results presented above WETMOD 2 is considered capable of

simulating wetland seasonal nutrient flux for individual wetlands that are affected by

varying external influences. Further, WETMOD 2 is considered valid, based on the

wetlands that were used in model development. To improve this confidence further

model validation using separate wetland data was performed and is described in

section 4.2. From this the model can be applied to category wetlands with reasonable

confidence placed in the output trend.

The first hypothesis is that “a simplified generic wetland model can be used to

realistically simulate multiple and different wetlands qualitatively”. To address this

hypothesis the results of the different wetlands scenarios, developed as part of the

model calibration and model validation, were reviewed as to their realistic

representation of expected wetland nutrient and biomass growth trends. These

wetlands are listed in Table 5.

Table 5: Assessment summary of wetlands realistic simulation

Category Wetland Simulated realistically

Category 1 wetland Paiwalla YES

Category 2 wetland Sunnyside Limited

Category 3 wetland Lock 6 YES

Category 4 wetland Reedy Creek YES

Category 5 wetland Pilby Creek YES

Category 3 wetland (non

calibrated data)

Lock 6 YES


calibrated data)

Reedy Creek YES


calibrated data)

Pilby Creek YES

This shows that the model is capable of simulating different wetlands, for which

adequate data is available, realistically although not to the accuracy of individually

tailored wetland models. This argument is strengthened by the results of the model

validations based on non-calibrated wetland data.


140

5 Simulation results of potential management

scenarios and Discussion

This chapter attempts to address the second hypothesis of whether a simplified

generic wetland model can be used to answer “what if” questions. The two degraded

wetlands for which there is adequate data to simulate effectively, Lock 6 and Reedy

Creek wetlands, are used to test simulation effectiveness of potential management

strategies. In the management simulations of Lock 6 and Reedy Creek wetlands,

WETMOD 2 was used to explore potential management strategies. Both wetlands are

considered permanently inundated and one was additionally severely degraded due to

excessive nutrient inflow from irrigation drainage.

Lock 6 wetland

Table 6 displays the potential percentage reduction in the outflow of PO4-P, NO3-N

and phytoplankton biomass as a consequence of three different management

scenarios. Figure 40 and Figure 41 represent the impact, on nutrient concentration and

macrophyte, zooplankton and phytoplankton biomass respectively, due to potential

management of turbidity reduction through the introduction of wetland dry periods in

Lock 6 wetland. To illustrate the status quo, the monitored concentrations and

biomass of PO4-P, NO3-N and phytoplankton, and the standard error, are also

displayed.

The scenarios during the months of February, March and April reflect the anticipated

wetland response to turbidity management (see section 3.4) during the macrophyte

growth period. The sedimentation rate of PO4-P, NO3-N and phytoplankton during the

months of March and April was constant for all scenarios (see Box). During this

period, the reduction in both PO4-P and NO3-N wetland concentrations is therefore a

direct result of the reduction in turbidity and improved uptake of nutrients by the

wetland, Figure 40A and B. This is reflected in the increase in macrophyte and

zooplankton growth seen in Figure 41A and B. The nutrient reduction success during

this period improves with each increment of turbidity reduction management, as can

clearly be seen in Figure 40A and B. The improvement in wetland condition can also

be seen in the dramatic increase in macrophyte growth, first at the 50% turbidity

reduction management scenario, and then at the 75% turbidity reduction scenario, as


141

shown in Figure 41A. The 75% reduction in turbidity (Figure 41A) demonstrates a

healthy growth phase of macrophytes, which reduces in the cooler months. The

zooplankton population growth seen during March and April for the 75% turbidity

scenario is a consequence of the assumed improved habitat conditions provided by the

macrophytes. The reduction in phytoplankton during this time is a consequence of the

competition with macrophytes for underwater light. The initial growth spurt of

phytoplankton for the 75% turbidity reduction scenario during February, is caused by

the improved underwater light conditions, and reduced competition due to, an

expected, lag in macrophyte growth. The zooplankton growth during February also

shows a slight increase as a consequence of the improved nutrient source

(phytoplankton), followed by a slight reduction in its population during the transition

from phytoplankton to macrophyte dominant phase.

The 50% reduction in turbidity (Figure 41A) signifies the first real improvement in

macrophyte growth, with a corresponding wetland nutrient load reduction. However,

as expected, the macrophyte growth is not as pronounced as that of the 75% turbidity

reduction scenario. The 50% reduction in turbidity scenario also shows some of the

February increase in phytoplankton growth prior to macrophyte competition, as well

as a slight improvement in the zooplankton. The 25% turbidity reduction scenario

shows minimal improvement in macrophyte growth, reflected mainly in the slight

improvement in the uptake of nutrients (PO4-P and NO3-N) during March and April.

When the turbidity is below that of the sedimentation threshold of 70 NTU, there is a

reduction in sedimentation of both PO4-P and NO3-N. This is apparent during a short,

but clear, high turbidity event in February for the 0% turbidity reduction scenario

where the wetland concentration of both PO4-P and NO3-N show a sudden and

substantial reduction, as seen in Figure 40. This trough in PO4-P and NO3-N

concentration is due to a rise in turbidity above the sedimentation threshold of 70

NTU that, in the unmanaged scenario, causes a sudden increase in nutrient

sedimentation. More significantly, there is an early drop in nutrient concentration for

the 0% scenario at the beginning of May (Figure 40), which continues for the

reminder of the simulation period. The 25% simulation, where the turbidity was

reduced by 25%, has a similar but more drastic drop in nutrient concentration at the

end of May, followed by the 50% scenario at the end of June. The 75% scenario has

only a small drop in nutrient concentration for a relatively short period of time as in


142

this scenario the turbidity only surpasses the 70 NTU sedimentation threshold for a

short period of 7 days (28th

August 1997 to 3rd

September 1997).

In summary, due to the management simulation of turbidity reduction, the period of

time where the turbidity remained below that of the calibrated sedimentation threshold

steadily increased for each improved management scenario, i.e. each increase in

turbidity reduction. Accordingly, the sedimentation rate reduced for each increasing

turbidity reduction scenario until, at the 75% turbidity reduction scenario, only 7 days

remained where turbidity within the wetland exceeded 70 NTU. Consequently, Lock 6

wetland progressively lost modelled sink (adsorption) capacity for both PO4-P and

NO3-N with each increase in turbidity reduction. This does however not accurately

account for any resuspension of nutrient highlighting one discrepancy in a generic

model.

Phytoplankton was also affected by the sedimentation change, the phytoplankton

threshold being calibrated to 95 NTU. The phytoplankton maintained a longer growth

period both due to the reduction in turbidity with its inherent increased light

availability and a low sedimentation rate (Figure 41C, June onwards). This growth

period was extended with each improved management scenario, until the

phytoplankton sedimentation is absent in the 75% turbidity reduction scenario (Figure

41C). The augmented phytoplankton availability resulted in an increased

phytoplankton outflow from the wetland as can be seen in Table 6, with the 50%

turbidity reduction scenario showing the highest amount of phytoplankton in the

outflow. The increase in macrophytes and zooplankton in the 75% turbidity

simulation, during March and April, reduced the phytoplankton growth that can be

seen in the lower phytoplankton outflow in Table 6 as well as in Figure 41C. The

zooplankton growth increased as a consequence of, and proportionally to, the

extended phytoplankton growth, as can be seen in Figure 41B.

Table 6: Lock 6 wetland Percentage Outflow Reduction

25% Turbidity Reduction



PO4-P % Reduction in Outflow

-17.1 -34.1 -47.7

NO3-N % Reduction in Outflow

-17.2 -18.0 9.1


143

Phytoplankton % Reduction in Outflow

-3.0 -11.1 -5.6

Figure 40: Lock 6 impacts on Nutrient concentration due to Turbidity reduction

NO 3 -N

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

-

P O 4 -P

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

-

A

B

T urb id ity R e d uce d b y 0 % T urb id ity R e d uce d b y 2 5 % T urb id ity R e d uce d b y 5 0 %

T urb id ity R e d uce d b y 7 5 % Mo nito re d D ate s O nly



144

Figure 41: Lock 6 impacts on Macrophyte, Zooplankton & Phytoplankton due to Turbidity

reduction

P h y to p la n k to n

0 .0 0

5 .0 0

1 0 .0 0

1 5 .0 0

2 0 .0 0

2 5 .0 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

-

Zo o p la n k to n

0 .0 0

0 .5 0

1 .0 0

1 .5 0

2 .0 0

2 .5 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

-

M a c r o p h y te B io m a s s

0 .0 0

2 .0 0

4 .0 0

6 .0 0

8 .0 0

1 0 .0 0

1 2 .0 0

1 4 .0 0

1 6 .0 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

kg

/m3

-

A

B

C

T urb id ity R e d uce d b y 0 % T urb id ity R e d uce d b y 2 5 % T urb id ity R e d uce d b y 5 0 %

T urb id ity R e d uce d b y 7 5 % Mo nito re d D ate s O nly



145

Reedy Creek wetland

Figure 42A and B and Figure 43C show the management scenarios for Reedy Creek

wetland and the impact on wetland biomass and nutrient concentrations as a result of

the management scenarios of successful reduction of irrigation drainage. Table 7

shows the percentage reduction of the total inflow (irrigation drainage and river

concentrations) versus the percentage outflow reduction due to different management

scenarios for Reedy Creek wetland.

The Reedy Creek wetland is adjacent to dairy farms whose pasture areas are situated

on reclaimed swamps. The irrigation runoff from the dairy pastures is pumped from

the adjacent farms into the wetland. This irrigation drainage has heavily influenced

Reedy Creek wetland and caused substantial degradation. One potential management

strategy that can be applied to Reedy Creek wetland is the nutrient reduction of

irrigation drainage load through the use of constructed wetlands. Wen (2002a; 2002b),

who contributed his data to this project, conducted preliminary trials of constructed

wetlands on Basby farm, which is a dairy farm immediately adjacent to Reedy Creek

wetland. His findings were that PO4-P could potentially be reduced by 50% to 90%.

Based on his findings, three scenarios of management were performed. The

management scenarios represented increasing reductions of 25%, 50% and 75% of

PO4-P, NO3-N and phytoplankton irrigation drainage loads, the time-series of which

can be seen in Figure 42A and B and Figure 43C. The percentage reduction in the

wetland outflow concentration compared with the reduction in inflow concentration

can be seen in Table 7. In Table 7 the effective percentage of reduction of the total

nutrient inflow is labelled as %RI and is displayed for each of the irrigation drainage

reduction scenarios. The ensuing percentage reduction in outflow is labelled %RO

(see section 3.4.1). The management scenarios of increasing reductions in nutrient

inflow to the wetland are controlled through the irrigation nutrient reduction option of

the model.

Each of the PO4-P simulations, for increased nutrient removal capacity, shows the

same trend, and for a large time period a virtually identical wetland concentration.

However, in October and again in February, as seen in Figure 42A, the simulated

wetland PO4-P concentration shows a reduction as a result of the management. The

NO3-N reduction can also be seen in Table 7. The high phytoplankton biomass is due

to a high phytoplankton inflow load from the irrigation drainage. There is a major


146

reduction of phytoplankton during the January, February and March periods. The

phytoplankton reduction was successful, as clearly seen in Figure 43C, particularly

during the months of January, February and March. The reduction in percentage

inflow versus reduction in percentage outflow is most extreme for phytoplankton, as

seen in Table 7. This indicates that through a minor reduction in irrigation nutrient

inflow, there can be a substantial impact on the outflow concentration of nutrients and

phytoplankton from the wetland. Drop in phytoplankton growth phase 19th

to 27th

March is due to a spike in turbidity. The zooplankton growth follows the

phytoplankton concentration, with a similar reduction due to management.

The high turbidity of the wetland, which restricts the Secchi depth to an estimated

depth of 0.2 m, severely limits the macrophyte growth within the wetland. The

degradation of the wetland macrophyte concentration from the initial starting level

adopted for the model is a consequence of this macrophyte growth restriction.

Therefore, despite the positive impact that simulated management (irrigation nutrient

reduction) has on outflow nutrient reduction, the lack of macrophyte growth hampers

an increase in the nutrient retention capacity of the wetland. The impact that the

reduction of turbidity, as a second management strategy, may have on macrophyte

growth and therefore nutrient retention is examined below.


147

Figure 42: Reedy Creek wetland impacts on Nutrient concentration due to irrigation drainage

reduction

NO3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Jun

-00

Jul-

00

Au

g-0

0

Sep

-00

Oct

-00

No

v-0

0

De

c-0

0

Jan

-01

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

mg

/L D

rain

ag

e o

nly

PO4-P

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Jun

-00

Jul-

00

Au

g-0

0

Se

p-0

0

Oct

-00

No

v-0

0

Dec-0

0

Jan

-01

Fe

b-0

1

Mar-

01

Apr-

01

Ma

y-0

1

mg

/L

-1

0

1

2

3

4

5

6

7

8

mg

/L D

rain

ag

e o

nly

A

B

I rrig atio n D rainag e C o nce ntra tion

R e d uce d b y 0 %

Irriga tio n D ra inag e C o nce ntration

R e duce d b y 2 5%

Irrig atio n D rainag e C o nc e ntratio n

R ed uc ed b y 5 0 %

Irrig atio n D rainag e C o nce ntra tion

R e d uce d b y 7 5%


C o nce ntra tion in W e tland


C once ntratio n in I rrig atio n D rain


148

Figure 43: Reedy Creek wetland impacts on Macrophyte, Zooplankton & Phytoplankton due to

irrigation drainage reduction

Phytoplankton

0

2

4

6

8

10

12

14

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

Dec-0

0

Jan

-01

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

-20

0

20

40

60

80

100

120

140

160

cm

3/m

3 D

rain

ag

e o

nly

Zooplankton

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

M acrophyte Biomass

0

0.02

0.04

0.06

0.08

0.1

0.12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

A

B

C

I rrig atio n D ra inage C o nc entra tion

R ed uc ed b y 0%

Irrigation D rainag e C o nce ntratio n

R e duce d b y 2 5 %

Irrig atio n D ra inage C o nce ntratio n

R e d uce d b y 50 %

Irrig atio n D ra inage C o nc entra tion

R ed uc ed b y 75 %


C o nce ntratio n in W etland


C o nc entra tion in I rrigation D rain


149

Table 7: Reedy Creek wetland Percentage Inflow reduction vs. Percentage Outflow Reduction

0%

Nutrient Reduction

25% Nutrient Reduction



PO4-P %RI 0 0 0 0

%RO (Irrigation Reduction Only)

0 1.2 2.8 4

NO3-N %RI 0 0 0 0


0 0.7 1.4 2.1

Phytoplankton %RI 0 0 0 0


0 4.1 8.2 12.2

Reedy Creek wetland twin management strategies

For Reedy Creek wetland a combination of both management strategies was

simulated. This was in an attempt to assess the cumulative impact of intensive

management of one large and severely degraded wetland. The high irrigation drainage

reduction scenarios of 85, 90 and 95% were used to assess the impact of potential full

restoration of the wetland.

Figure 44 and Figure 45 represent the concentrations in the open water of Reedy

Creek wetland when the turbidity is modelled at 75% reduction and the nutrients are

reduced by 25%, 50% and 75% successively. Whereas, Figure 46 and Figure 47

represent the concentrations in the open water of Reedy Creek wetland when the

irrigation drainage nutrient reduction scenario is maintained at 95% and the turbidity

reduction scenarios are at 25, 50 and 75% respectively. Figure 44, Figure 45 and

Figure 46, Figure 47 are plotted separately to distinguish between the impacts of

various turbidity reduction scenarios at the best possible nutrient reduction scenario,

and the impact of the nutrient reduction scenario at the best turbidity reduction

scenario. Note, in Figure 44, Figure 45, Figure 46 and Figure 47 the monitored

irrigation drainage concentration and the monitored concentration in the wetland are

those monitored for Reedy Creek wetland. Figure 48, Figure 49 and Figure 50 show

the percentage reduction in outflow load from Reedy Creek wetland as a consequence

of the double management strategies. In these figures the results from all simulated

combinations are presented.


150

The reduction irrigation drainage inflow has also reduced the wetland outflow of

nutrients and phytoplankton. This can be seen in the gradual increase in “% Reduction

in wetland nutrient outflow compared to status quo” in Figure 48 for PO4-P, Figure 49

for NO3-N and Figure 50 for phytoplankton. In Figure 48 to Figure 50 the percentage

reduction in wetland nutrient outflow is related back to the status quo, i.e. without

management scenarios.

With the exception of NO3-N each increment of reduction in turbidity results in a

“drop in the percentage reduction” in wetland PO4-P outflow, and phytoplankton

outflow, see Figure 49, Figure 48 and Figure 50 respectively, i.e. PO4-P and

phytoplankton outflow increase. The reason for the comparatively increased nutrient

outflow, despite the management strategy of “reduction in irrigation” remaining the

same at 95%, is related back to the decrease in the sedimentation rate, i.e. the turbidity

simulated at below 70 NTU for nutrients, and 95 NTU for phytoplankton. The NO3-N

wetland retention, seen in Figure 49, as a general trend improves, however the NO3-N

retention reduces again for the 75% turbidity reduction as the turbidity passes below

the sedimentation threshold (discussed for Lock 6 wetland above). That is, the 75%

turbidity reduction scenario has a higher NO3-N outflow due to the loss of

sedimentation of nutrients (Figure 44B). This can be seen as an increase in NO3-N

wetland concentration, i.e. decreased nutrient retention, and is visible during

September in Figure 46. The scenarios with minimal turbidity reduction display a

higher NO3-N retention, attributed to higher sedimentation of NO3-N in more turbid

wetlands. This again raises the question whether the model needs an improvement to

account for sedimentation resuspension.

An improvement in PO4-P retention is observable in the irrigation nutrient scenarios;

however the turbidity reduction scenarios cause a steady drop in the PO4-P retention

(Figure 48). The turbidity reduction scenarios reduce the PO4-P sedimentation as they

do the NO3-N sedimentation. The difference between NO3-N and PO4-P is that the

PO4-P concentration is very low during the period that has such a great influence on

NO3-N, i.e. September see Figure 44A. Therefore, the variability in wetland

concentration for PO4-P during September becomes negligible.

There is an early low macrophyte biomass for turbidity reduction scenarios which is

not as apparent in nutrient reduction and status quo scenarios, which can be seen in

Figure 45A and Figure 47A for the periods July 2000 to January 2001. This fast drop


151

in macrophyte biomass for the turbidity reduction scenarios is a model artefact with

negligible repercussions. It is caused by the minimum fixed macrophyte gross primary

productivity being slightly higher than the calculated, when the turbidity is below the

70 NTU threshold and the macrophyte growth is restricted due to other causes. The

trend of macrophyte growth and its peak is due to the underwater light availability as

well as nutrient availability during the simulation period. This can be seen in Figure

45A when compared to Figure 44 where initially the underwater light for macrophyte

growth is limited. The macrophyte growth is again restricted, this time by NO3-N

limitation in late May 2001, see Figure 44B, which causes the rapid macrophyte

dieback seen in Figure 45A. The same limitations caused by underwater light and

NO3-N can be seen in Figure 47A for the scenario of 75% turbidity reduction and

95% nutrient reduction when compared to Figure 46A and B. However, in this

instance the macrophyte biomass is at its lowest for a 75% turbidity reduction

scenario, compare macrophytes at Figure 45A on page 153 and Figure 47A.

Effectively the higher macrophyte biomass growth is seen in the high turbidity

reduction scenario (75%) with an incremental increase in biomass with each

successive nutrient reduction scenario, seen in Figure 45A.

Phytoplankton retention improves with each irrigation drainage reduction (Figure 50).

However, with the decrease in turbidity (Figure 50) and the late start of the

macrophyte growth season discussed above, the phytoplankton has ample opportunity

to increase its biomass (Figure 50, Figure 45C and Figure 47C). Therefore, the

turbidity reduction scenarios actually contribute to the phytoplankton growth for

Reedy Creek twin management scenarios. As seen previously the zooplankton growth

trend, Figure 45B and Figure 47B, follows that of its food source the phytoplankton

seen in Figure 45C and Figure 47C.


152

Figure 44: Reedy Creek wetland impacts on Nutrient concentration due to irrigation drainage

reduction and 75% turbidity reduction

PO 4-P

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-1

0

1

2

3

4

5

6

7

8

mg

/L D

ra

ina

ge

on

ly

N O 3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-0 .2

0

0.2

0 .4

0 .6

0 .8

1

1.2

mg

/L D

rain

ag

e o

nly

A

B

I rrigation D rainage C o ncentration

R e duced by 0%

Irrigatio n D rainage C oncentration

R educed by 25%

Irrigatio n D rainage C o nce ntratio n

R educe d b y 5 0 %

Irrigation D rainage C o ncentration

R e duced by 75%




C once ntration in I rrig atio n D rain

S tatus Q uo (No Managem ent)


153


irrigation drainage reduction and 75% turbidity reduction


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

Zooplankton

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

Phy toplankton

0

2

4

6

8

10

12

14

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

-20

0

20

40

60

80

100

120

140

160

cm

3/m

3 D

rain

ag

e o

nly

A

B

C

I rrigation D rainage C oncentration

R educed by 0%

Irrigation D rainage C oncentration

R educed by 25%


R educed by 50%


R educed by 75%




C oncentration in I rrigation D rain

S tatus Q uo (No Managem ent)


154

Figure 46: Reedy Creek wetland impacts on Nutrient concentration due to 95 % irrigation

drainage reduction at 25, 50 and 75% turbidity reduction

PO 4-P

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-1

0

1

2

3

4

5

6

7

8

mg

/L D

ra

ina

ge

on

ly

N O 3-N

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

-0 .2

0

0.2

0 .4

0 .6

0 .8

1

1.2

mg

/L D

rain

ag

e o

nly

A

B

S tatus Q uo (No Managem ent) I rrigation D rainage C once ntration R educe d by 95 %

& T urb id ity R educed by 25%

Irrigation D rainag e C once ntration R ed uced by 95%

& T urb id ity R e d uced b y 50%

Irrigation D rainage C once ntration R educe d by 95 %



C oncentratio n in W e tland

Monito red D ates O nly C o nce ntratio n in I rrigatio n D rain


155


95% irrigation drainage reduction at 25, 50 and 75% turbidity reduction

Phy toplankton

0

2

4

6

8

10

12

14

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

-20

0

20

40

60

80

100

120

140

160

cm

3/m

3 D

rain

ag

e o

nly


0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

Zooplankton

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

A

B

C

S tatus Q uo (No Managem ent) I rrigation D rainage C oncentration R educed by 95%


Irrigation D rainage C oncentration R educed by 95%


Irrigation D rainage C oncentration R educed by 95%




Monito red D ates O nly C oncentration in I rrigation D rain


156

Figure 48: Reedy Creek wetland PO4-P % reduction in outflow

Figure 49: Reedy Creek wetland NO3-N % reduction in outflow

PO4-P Outflow at Reedy Creek

-6

-4

-2

0

2

4

6

25% Irrigation

Drainage

Reduction

50% Irrigation

Drainage

Reduction

75% Irrigation

Drainage

Reduction

85% Irrigation

Drainage

Reduction

90% Irrigation

Drainage

Reduction

95% Irrigation

Drainage

Reduction

Irrigation Drainage Reduction

% R

ed

uc

tio

n in

we

tla

nd

nu

trie

nt

ou

tflo

w c

om

pa

red

to

sta

tus

qu

o

Irrigation Drainage Reduction Only

Irrigation Drainage Reduction and 25% Turbidity Reduction



NO3-N Outflow at Reedy Creek

0

1

2

3

4

5

6

7

25% Irrigation

Drainage

Reduction

50% Irrigation

Drainage

Reduction

75% Irrigation

Drainage

Reduction

85% Irrigation

Drainage

Reduction

90% Irrigation

Drainage

Reduction

95% Irrigation

Drainage

Reduction


% R

ed

uc

tio

n in

we

tla

nd

nu

trie

nt

ou

tflo

w c

om

pa

red

to

sta

tus

qu

o






157

Figure 50: Reedy Creek wetland Phytoplankton % reduction in outflow

5.1.1 Implications for Management

Lock 6 wetland

The improvement in nutrient uptake during the macrophyte growth period, March and

April, shows that management scenarios, particularly the 75% turbidity reduction

scenario, are extremely successful in nutrient reduction. Scenarios of increasing

management success, represented by increased percentage of reduced turbidity,

demonstrate gradual improvement in nutrient retention, with 75% turbidity reduction

showing a drop in almost a third of wetland nutrient load. During the winter period,

where the poorest performance of managed wetlands can be seen, nutrient

sedimentation rate exceeds the nutrient uptake of macrophytes, phytoplankton and

zooplankton. As a result, the mass balance seen in Table 6 shows the turbid state to be

a more effective nutrient and phytoplankton sink. Although the macrophyte growth of

March and April indicated an improvement due to turbidity reduction the main

concern to wetland management was the dramatic reduction in the sedimentation of

PO4-P and NO3-N. This reduction of sedimentation of PO4-P and NO3-N was as a

direct consequence of the reduced turbidity, which is mainly apparent during the

periods of May through to late September. This does however not adequately consider

any potential resuspension of nutrient, which could be a future model enhancement.

The excess nutrient availability and lack of macrophyte competition in the cooler

Phytoplankton Outflow at Reedy Creek

-15

-10

-5

0

5

10

15

20

25% Irrigation

Drainage

Reduction

50% Irrigation

Drainage

Reduction

75% Irrigation

Drainage

Reduction

85% Irrigation

Drainage

Reduction

90% Irrigation

Drainage

Reduction

95% Irrigation

Drainage

Reduction


% R

ed

uc

tio

n in

we

tla

nd

nu

trie

nt

ou

tflo

w c

om

pa

red

to

sta

tus

qu

o






158

months led to an increased phytoplankton growth, and therefore a possible resultant

degradation of water quality for the river. However, when the 50% turbidity

management scenario is studied in detail, it responds to macrophyte growth and has

the lowest nutrient load during most of the winter period. This, along with the healthy

macrophyte growth of the 75% turbidity reduction scenario, indicates that the optimal

wetland state will be found in a balance between maintaining as high a sedimentation

rate as possible with some suspended sediment inflow and therefore slightly turbid

waters (effectively a sedimenting wetland).

Comparing the nutrient mass balance of the different management strategies shows

that the increasing macrophyte growth could not compete with the loss in nutrient

sedimentation in the management scenarios, the exception being the Lock 6 wetland

NO3-N retention in the 75% turbidity reduction simulation. This shows that the model

output may improve with some increased complexity, although this would need to be

weighed up against the loss in model applicability on a landscape scale.

The main reason for the PO4-P mass balance failing to show an improvement in the

mass balance, despite there being a very clear and significant PO4-P uptake during the

macrophyte growth phase, was a short-term high nutrient load in the inflow water

from the river. This inflow occurred in late September. During this month there were

high river PO4-P loads, which caused a large inflow load. The high turbidity of the 0%

and 25% scenarios contained the increased load through a high sedimentation rate, as

the turbidity levels were above the 70 NTU sedimentation threshold. Due to the

turbidity controlled sedimentation threshold, the 50% and 75% turbidity reduction

scenarios were unable to buffer this excess load, which is reflected in the increase in

phytoplankton growth during the final simulation week, seen in Figure 40. The 50%

and 75% turbidity reduction scenarios, having low turbidity and a low nutrient

sedimentation rate, have a seemingly greater wetland nutrient load, and hence there is

a higher outflow load of nutrient and phytoplankton during this period. This increased

nutrient load has an adverse impact on the nutrient mass balance, showing the 50%

and 75% turbidity reduction management scenarios to be ineffective in improving

wetland nutrient retention. However, the scenarios show that during the period with

increased macrophyte growth, see Figure 41, as predominantly seen with the 75%

turbidity reduction, the phytoplankton and particularly NO3-N outflow was reduced

(Table 6). Therefore, assessing the results for a season where the model assumes low


159

sedimentation of nutrients for all scenarios (i.e. all scenarios having the same turbidity

sedimentation) there is an obvious visual decrease in wetland nutrient load with each

reduced turbidity simulation. Increasing the complexity of the model through

introducing sediment resuspension and nutrient release may therefore not be

necessary.

The model in this case (Lock 6 wetland) can be used to assess the minimum turbidity

improvement required for the wetland to have a positive response to nutrient

retention. With this information, wetland managers can more confidently judge the

potential success rate of wetland restoration based on their expectation of turbidity

improvement. Another management option based on the Lock 6 wetland management

scenarios may lead wetland managers to inundate the wetland during the macrophyte

growth period only, and introduce wetland dry periods during the cooler winter

months where the nutrient removal may not be as successful or when macrophyte

health starts to deteriorate. This would then maximise the macrophyte driven nutrient

uptake of the wetland. In this case the model would have been used in optimising the

choice of wetland dry periods. This is examined in section 6.1.

Reedy Creek wetland

The management option of irrigation reduction through constructed wetlands shows

an improvement of wetland nutrient and phytoplankton retention. This model

simulation suggests a positive result on wetland nutrient and phytoplankton load, and

therefore outflow as a consequence of reducing irrigation drainage inflow into the

wetland. The outflow reduction was in each instance higher in percentage than the

percentage reduction of inflow, suggesting that a small change in irrigation drainage

inflow can have a substantial impact on the total exchange of nutrients between the

wetland and river. The impact this nutrient reduction has on river nutrient load is

discussed in section 6.3.

The model shows that Reedy Creek wetland itself, as a consequence of its high

turbidity and lack of macrophyte growth, is presently not capable of improving its

nutrient retention. Due to a lack of data, the effective turbidity reduction as a

consequence of a reduction in phytoplankton is not taken into account in the model.

Decision makers must therefore keep in mind the possibility that phytoplankton

reduction may also reduce turbidity and increase Secchi depth. This increase in Secchi


160

depth would allow macrophyte growth, which may further reduce wetland nutrient

load.

Reedy Creek wetland twin management strategies

Simulating twin management actions provides the opportunity to assess the

compounding impact one management strategy may have on the other. With an

effective net phytoplankton production, the management strategy of turbidity

reduction proved counter productive. Along with the ample nutrient availability in the

wetland, the primary cause of net phytoplankton production increase was the added

underwater light availability that was enough for phytoplankton growth but still

restricted macrophyte growth. This increased phytoplankton growth, became evident

during the high nutrient reduction scenario Figure 47C and Figure 46. Further, the

simulated loss of sedimentation of PO4-P and NO3-N resulted in the loss of nutrient

retention. Therefore, for a net increase in wetland nutrient and phytoplankton

retention, the nutrient reduction scenarios through constructed wetlands and without

the added management scenarios of turbidity reduction proved to be the more

effective management strategies.

Although this conclusion can be drawn at this stage from the twin management

strategies scenarios, Beck (1997) discusses the problems faced by modellers when

models are calibrated for stressed systems and may therefore have some difficulty in

simulating the system when returned to a natural state. WETMOD 2 was calibrated

for optimal wetland response for category 4 wetlands, i.e. for a degraded system,

largely influenced by irrigation drainage with no significant macrophyte growth.

Therefore, allowance must be made to question the accuracy of simulated macrophyte

biomass growth particularly as the model is compounding the potential errors of

assumptions for two management strategies, as in the case of twin management.

The confidence in the model output must rely on the assessment of expected trends

for the wetland as a consequence of twin management. Therefore, before deciding on

refraining from twin management of a wetland such as Reedy Creek wetland, the

question must be raised as to whether the simulated volume of macrophyte biomass

was realistic enough to truthfully represent the impact of macrophyte uptake of

nutrients. Although the conclusion drawn at this stage indicates that twin management

may be counterproductive, the results elicited help formulate new questions and


161

therefore focus further potential research. For example, further research could be

directed at discovering the true potential response of macrophyte growth trend in such

an instance, as well as to discover at what stage of nutrient reduction (through a

constructed wetland) would the introduction of wetland dry periods assist in

promoting macrophyte growth. As a start for example, monitoring would be required

to validate model macrophyte simulations.

5.2 Chapter summary and Implications for the second

hypothesis

The simulations of wetland management, based on the two wetlands presented, show

that WETMOD 2 can be applied to assess and better understand the impacts of

wetland management. The model was effectively applied in the management of

wetlands facing different degradation pressures. Both wetlands were degraded as a

consequence of permanent inundation, and one was additionally degraded due to

irrigation drainage inflow. WETMOD 2 could, as it is a generically applicable model,

be applied to other wetlands within these categories. The model developed

management scenarios that were successfully used to assess the impact of

management, see Table 8. Table 8 shows that “a simplified generic wetland model can

be used to answer what if questions”.


162

Table 8: Assessment summary of wetlands management scenarios

Category Wetland Management question Answer

Category 3

wetland

Lock 6 Can the model identify the

turbidity reduction that is

required for a positive response

if a turbid wetland is managed?

YES

Category 4

wetland

Reedy Creek Can the model simulate the

implications of the reduction of

irrigation drainage nutrient on a

wetland impacted on by

irrigation drainage?

YES

Category 4

wetland

Reedy Creek Can the model indicate the

impact of introducing two

management strategies to a

wetland such as Reedy Creek

wetland?

YES

(although

limited)

The use of deterministic differential models provides a platform with which some of

the complexity of wetland management can be organised and examined. Thereby the

model user is able to gain a better understanding of the impact of intervention options

such as different wetland management strategies or intensities, e.g. minimum turbidity

reduction required, and therefore answer “what if” questions. A modeller can

experiment with the model to study the impacts of minor alterations within a wetland

and therefore gain a larger understanding of the complexity of the ecosystem. By

using the model, decision makers can agree on which scenarios are to be run, assess

the output, and if necessary trace back the trigger variable to either gain a better

understanding or increase consensus. Whether modellers are also able to gain some

insight into the potential outcomes of multiple wetland management and therefore the

cumulative impact on the river nutrient load is discussed in the next chapter. This

would assist managers and decision makers in estimating what intensity of

management may be required for a desired regional response.


163

6 Results of the cumulative assessment of

management scenarios, visualisation and

discussion

As the model is generic, and therefore shown to be applicable to category wetlands for

which “exemplar”-driving variables are available, an assessment of the cumulative

impact of multiple managed wetlands was therefore possible. The cumulative impact

assessment allows the discovery of the potentially optimal management strategy, not

only for one wetland but multiple wetlands, and therefore the optimal strategy for

regional scale wetland management.

For a cumulative assessment of the impact wetland management would have within

regional scale management, scenarios were developed with WETMOD 2 for those

wetlands identified as belonging to category 3 and category 4 wetlands (“exemplar”

driving variables from Lock 6 wetland and Reedy Creek wetland respectively). The

management of 57 category 3, and 7 category 4 wetlands were simulated (see

methodology in section 3.4.2). The application of the model to category wetlands tests

the hypothesis of whether “a simplified generic wetland model can be used to assess

the cumulative impact of managing multiple same category wetlands”. This would

expand the applicability of the model to wetlands where limited data is available and

therefore the assessment of potential multiple wetland management on a regional

scale.

The category simulation output represents the estimation of the nutrient, plankton and

macrophyte trends within a wetland as a result of the differences between the

wetlands. These wetland differences are wetland volume, depth and location along the

river. The location along the river dictates river flow volume and river nutrient

concentration. However, there are important differences between wetlands, which

could not be considered in “category wetland” simulations. Future wetland simulation

modelling has the potential to upgrade category simulations with improved data for

the following, without substantial model alteration;

Specific exchange volume estimate for each simulated wetland (based on

future monitoring, digital elevation models and/or expert input)


164

Substrate composition, i.e. will the wetland sediment compact? (soils surveys

may have to determine the sediment compaction potential of individual

wetlands)

Specific irrigation volume, category 4 wetlands only (assumed to be equal to

Reedy Creek due to lack of data, results used to show feasibility of simulation

only, not result accuracy)

These limitations to scenario modelling were anticipated and, as this project did not

include on-site data collection, these limitations were deemed not to be a priority

concern. This is discussed further in the conclusions, in section 7.2.

6.1 Cumulative assessment: category 3 wetlands

The cumulative assessment of management of multiple wetlands (a list of all wetlands

selected for category 3 management simulations can be seen in Appendix C) shows a

trend towards the improvement in NO3-N retention, which can be seen in Figure 51.

As seen in Figure 51 the PO4-P retention does not show an improvement. However,

this could be due to the spike seen in the final week of simulation (Figure 52), which

relates back to the river load at that time where the river load inflow causes a spike in

modelling output, see section 3.2.3. The increase in phytoplankton outflow, seen in

Figure 51, is due reduced turbidity leading to the increased availability of underwater

light. Phytoplankton responds earlier to increased light availability than macrophytes,

see Figure 54. Consequently, there is a trend towards an increased phytoplankton

growth, particularly in the 50% turbidity reduction scenario, see Figure 51. The 75%

turbidity reduction scenario conversely shows a trend toward reducing phytoplankton

growth, see Figure 51. This reduction in phytoplankton, when compared to the 50%

turbidity reduction scenario, could be associated with increased macrophyte growth

seen in the 75% turbidity reduction scenario leading to competition with

phytoplankton for underwater light, see Figure 53.

For detailed output for category 3 wetlands cumulative assessment refer to Appendix

D Table 20 toTable 22 (PO4-P, NO3-N and phytoplankton biomass respectively). A

detailed change in retention for each wetland and each management scenario, as well

as the percentage change in the outflow concentration is shown. At the end of each

table there is a summary of the cumulative retention, as shown in Figure 51.


165

Figure 51: Cumulative retention- category 3 wetlands

Results of category 3 wetland management at the 50% (A) and 75% (B) turbidity

reduction scenarios, as compared with the average of the status quo for all 57

wetlands are shown in Figure 52, to Figure 56. As the cumulative impact and

particularly trends are of concern, not individual wetland responses, the results of

individual wetlands are shown in grey only. The average response is shown in green

and the median in red. The 50% turbidity reduction scenario was the first to show a

response to the management scenario. The 75% turbidity reduction scenario shows the

best simulated response to turbidity reduction, with healthy macrophyte growth. Thus

only the 50% and 75% turbidity reduction scenarios are shown.

For the 50% turbidity reduction scenario, the wetlands macrophyte growth vary from

no growth to healthy summer growth Figure 53A. A 50% reduction of turbidity

therefore leads to a response in the form of macrophyte growth. In the 75% turbidity

reduction scenario there is also a range in the successful growth of macrophytes of the

different wetlands (Figure 53B). For most wetlands the median shows a clear trend

towards summer growth (i.e. late summer immediately following inundation) slowing

down with temperature and light reduction in winter. In the 75% turbidity reduction

scenario some wetlands showed only minor macrophyte growth, such as wetland

Cumulative retention by Category 3 Wetlands

0

5000

10000

15000

20000

25000

30000

35000

40000

Status Quo

Turbidity

25% Turbidity Reduction50% Turbidity Reduction75% Turbidity Reduction

-400

-350

-300

-250

-200

-150

-100

-50

0

PO4-P Net Retention kg/annum NO3-N Net Retention kg/annum Phytoplankton Net Retention m3/annum


166

numbers S0115 (367), S0229 (978) and S0230 (47) (wetland numbers are as per

South Australian Wetlands Atlas (Jensen et al. 1996)). Other wetlands showed an

exceptional macrophyte growth such as wetland numbers S0174 (1036), S0203 (471)

and S0229 (84) (Figure 53B). These differences in macrophyte growth are related to

individual wetland morphology.

The clear trend towards summer growth phase with a winter dieback supports the

argument for managed winter dry periods with the aim of re-introducing sediment

compaction. Reflooding would lead to macrophyte germination and the summer wet

would maximise macrophyte growth and therefore nutrient retention. The

phytoplankton growth phase occurs in response to improved underwater light and the

lack of competition due to macrophyte dieback in winter. With the winter dry period

this would be minimised (Figure 54B). The net impact on a cumulative scale would be

nutrient retention by the wetlands. The winter dry/summer wet management strategy

is explored more below, with an example of three wetlands that are assumed to be

dried following the onset of macrophyte dieback (i.e. with the onset of winter and

therefore reduced modelled macrophyte growth).

Going back to the cumulative assessment of the 57 wetlands, some wetlands show a

trend towards a better macrophyte growth phase than others, such as wetland number

S0219 (996) that has a very short winter macrophyte dieback period. This wetland

shows a trend towards a long macrophyte growth period, a minimal phytoplankton

growth phase and positive nutrient retention. The main difference between these

wetlands is wetland morphology. The trends within wetlands based on wetland depth

and volume are discussed below.

The sudden reduction in phytoplankton biomass in the 50% turbidity reduction

scenario in mid July (Figure 54), stems from the increase in sedimentation due to

turbidity increasing past the sedimentation threshold as discussed previously. The

zooplankton biomass trend (Figure 55), follows suit due to its reduced food source. As

turbidity (NTU) never exceeds the sediment threshold in the 75% turbidity reduction

scenario there is no change in the rate of phytoplankton biomass sedimentation.

Consequently, phytoplankton biomass in the 75% turbidity reduction scenario (Figure

54) remains high through the winter period. However, the trend for phytoplankton

biomass outflow in the 75% turbidity reduction scenario is less than that of the 50%


167

turbidity reduction scenario, (Figure 51). The greater macrophyte growth in the 75%

turbidity reduction scenario accounts for this variation (as discussed above).

Based mainly on “category wetlands” morphological differences and the different

exchange and nutrient loads at the respective river locations, WETMOD 2 was

capable of simulating differences of biomass growth and nutrient retention within

these wetlands. The implication of multiple wetland management and the potential

cumulative impact on river nutrient load, through alteration of nutrient and

phytoplankton retention, is discussed below in section 6.3.


168

Figure 52: PO4-P Concentration Trends

PO4-P C o n ce n tr at io n in C ate g o r y 3 w e tlan d s at 50% T u r b id ity Re d u ctio n Sce n ar io

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

PO4-P C o n ce n tr at io n in C ate g o r y 3 w e tlan d s at 75% T u r b id ity Re d u ctio n Sce n ar io

0.00

0.10

0.20

0.30

0.40

0.50

0.60

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/LA

B

Ind ivid ual W e tland s A ve rag e Me d ian S tatus Q uo A ve rag e (no m anag e m e nt)


169

Figure 53: Macrophyte Biomass Growth Trends


170

Figure 54: Phytoplankton Biomass Growth Trends

Ph yto p lan k to n Bio m as s in C ate g o r y 3 w e tlan d s at 50% T u r b id ity Re d u ctio n Sce n ar io

0

2

4

6

8

10

12F

eb

-97

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

Phytoplank ton Biom as s in Cate gory 3 w e tlands at 75% Turbidity Re duction

Sce nar io

0

5

10

15

20

25

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

Note: First week where there is no competition with the macrophytes the

phytoplankton grows exponentially. Therefore to be able to view the effect the

management has on macrophyte and phytoplankton growth the first week of

phytoplankton growth has been removed.

A

B



171

Figure 55: Zooplankton Biomass Growth Trends

A

B


Z o o p lan k to n Bio m as s in C ate g o r y 3 w e tlan d s at 75% T u r b id ity Re d u ctio n Sce n ar io

0

5

10

15

20

25

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

Z o o p lan k to n Bio m as s in C ate g o r y 3 w e tlan d s at 50% T u r b id ity Re d u ctio n Sce n ar io

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8F

eb

-97

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3


172

Figure 56: NO3-N Concentration Trends

NO3-N C o n ce n tr at io n in C ate g o r y 3 w e tlan d s at 50% T u r b id ity Re d u ctio n Sce n ar io

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/L

NO3-N C o n ce n tr at io n in C ate g o r y 3 w e tlan d s at 75% T u r b id ity Re d u ctio n Sce n ar io

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

mg

/LA

B



173

Wetland size, volume and location

Category 3 wetlands scenarios include differences in wetland volume, and differences

in the monitoring location of river flow data and river nutrient data. The 75% turbidity

reduction scenario also includes wetland depth differences due to improved Secchi

depth (underwater light penetration was limited by wetland depth). Since in the 75%

turbidity reduction scenarios Secchi depth equals the actual wetland depth, these

scenarios would be best to compare with each other to understand the impact depth

has on wetland response to management.

When macrophyte biomass volume is compared against wetland size and wetland

depth (Figure 57), a trend towards greater macrophyte growth with an increasing

wetland depth is apparent. However, with a corresponding increase in wetland volume

macrophyte biomass reduces. This is however limited by the lack of validation data

for macrophyte biomass. These size assessments therefore are subject to this

significant model limitation. The assessments are however made to indicate the

potential use of the model once adequate validation has been undertaken.

First, there is a trend towards an increase in wetland depth leading to an increase in

macrophyte biomass (Figure 58). This goes back to the issue discussed earlier in the

model validation, section 4.3, where the underwater light availability, and therefore

macrophyte growth, is dependent on the Secchi depth (i.e. logarithmic increase in

macrophyte growth with increasing depth). This calculation is not taking into account

the maximal wetland depth and the amount of underwater light actually reaching the

wetland substrate nor the maximum growth depth of macrophytes (not currently an

acute issue).

Second, a wetland with the same depth but smaller surface area and therefore volume

seems to have more macrophyte growth. This would relate back to the amount of

nutrient entering the wetland, i.e. the model assumes the same fraction of river flow

volume is the exchange volume for both the larger and smaller wetland. A small

wetland therefore effectively has a greater turnover rate. A more accurate wetland

exchange volume for the wetland would improve the results in such an instance, again

highlighting the need for improved data on potential exchange volumes. With the

current modelling capacity, WETMOD 2 however poses the question whether

wetlands with a small volume would be more apt at nutrient uptake (retention) due to


174

the greater macrophyte growth within these wetlands compared to wetlands with a

larger volume?

In an attempt to address this question the results of cumulative wetland assessments

were investigated further. Figure 57 shows the relationship between macrophyte

biomass, wetland volume and depth. Indicating that greater macrophyte biomass is

related more to wetland depth than it is to wetland volume (Figure 57). This is

supported by Figure 58, which shows a slight increase in macrophyte biomass with

increasing wetland depth. The 57 category 3 wetlands were divided into three wetland

depth ranges shallow (<1 m), medium (1 to <2 m) and deep (>2 m). The average

depth of the shallow wetlands was 0.9 m. The average depth of the medium wetlands

was 1.3 m, and for the deep wetlands 2.1 m. Figure 59 shows the average macrophyte

biomass and average wetland volume for the wetland depth ranges. Figure 59

indicates that medium sized wetlands favour optimal macrophyte growth.

Figure 60 shows that for medium depth range wetlands, which have the largest

average macrophyte biomass, there is an exponential decline in macrophyte growth

with increasing wetland volume. These wetlands have a similar depth and the same

turbidity (same “exemplar” data source), therefore they have the same macrophyte

growth potential according to the modelled underwater light. Consequently, the major

difference between wetlands is volume. The cause of lower macrophyte growth in

greater volume wetlands can be correlated back to nutrient availability. That is, the

larger wetlands have a greater dilution of the inflow nutrient load within the water

body. It must however be remembered that the wetland has not been validated agains

macrophyte growth. These results are therefore only indicative based on the current

model capabilities.


175

Figure 57: Macrophyte Biomass (size of sphere, kg/m3) plotted against Wetland Volume and

Wetland Depth

Figure 58: Macrophyte Biomass vs. Wetland Depth

Macrophyte Biomass vs. Wetland Volume and Depth

1

10

100

1000

10000

100000

1000000

10000000

0 0.5 1 1.5 2 2.5 3

Wetland Depth m

Lo

g 1

0 (

Wetl

an

d V

olu

me m

3)

Macrophytes vs. Wetland Depth

1

10

100

1000

10000

100000

0 0.5 1 1.5 2 2.5

Wetland Depth m

Lo

g 1

0 (

Macro

ph

yte

s k

g/m

3)


176

Figure 59: Average Macrophyte Biomass (size of sphere, kg/m3) plotted against Average Wetland

Volume and Wetland Depth

Figure 60: Macrophyte Biomass vs. Wetland Volume

Same We tland D e pth R ange (1m - <2m), Incre asing We tland

Volume v 's M acrophyte B iomass

y = 2E + 08x-0.8526

R2 = 0.9113

0.0

5000.0

10000.0

15000.0

20000.0

25000.0

30000.0

0 500000 1000000 1500000 2000000 2500000

W e tla nd V olum e m 3

Ma

cro

ph

yte

Bio

ma

ss

kg

/m3

Average Cumulative Macrophyte Biomass vs. Average Wetland

Volume and Depth

0

200000

400000

600000

800000

1000000

1200000

1400000

1600000

1800000

2000000

0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3

Wetland Depth m

We

tla

nd

Vo

lum

e m

3


177

As discussed above, for maximum macrophyte biomass growth within wetlands,

WETMOD 2 indicates an optimal wetland volume and depth range. Figure 61 to

Figure 65 help to demonstrate the relationship between wetlands volume, macrophyte

biomass and nutrient dilution.

To establish which wetlands were producing the greatest biomass within the medium

wetland depth range (1 to 2 metre), macrophyte biomass was plotted against wetland

volume and depth (Figure 61). With the increase in volume macrophyte biomass

reduces significantly. If this is compared to Figure 62, where the average

concentration of PO4-P within the wetland for the simulation period is plotted instead

of macrophyte biomass, a similar pattern is produced. The pattern in Figure 62

indicates a lower average PO4-P load for the wetlands where the macrophyte biomass

is low. This suggests that PO4-P may be the limiting nutrient to macrophyte growth.

This is supported by Figure 63, which shows the macrophyte biomass vs. average

PO4-P load within these wetlands. No such dependency of macrophyte biomass on

NO3-N was seen in Figure 64 and Figure 65. Therefore, the optimal wetland volume

and depth discovered within WETMOD 2 simulations can within the confines of the

present model capability be related back to the PO4-P availability (volume relating to

dilution) and underwater light availability controlled by the wetland and Secchi depth.


178

Figure 61: Average Macrophyte Biomass (size of sphere) Plotted against Average Wetland

Volume and Wetland Depth range 1 – 2 m

Figure 62: Average PO4-P (size of sphere) Plotted against Average Wetland Volume and Wetland

Depth range 1 – 2 m

Macrophyte Biomass cm3/m3 vs. Wetland Volume and Wetland

Depth

-500000

0

500000

1000000

1500000

2000000

2500000

1 1.1 1.2 1.3 1.4 1.5 1.6

Wetland Depth m

We

tla

nd

Vo

lum

e m

3

Average PO4-P mg/L vs. Wetland Volume and Wetland Depth

-500000

0

500000

1000000

1500000

2000000

2500000

1 1.1 1.2 1.3 1.4 1.5 1.6

Wetland Depth m

We

tla

nd

Vo

lum

e m

3


179

Figure 63: Average PO4-P vs. Macrophyte Biomass at Wetland Depth range 1 – 2 m

Figure 64: Average NO3-N (size of sphere) Plotted against Average Wetland Volume and

Wetland Depth range 1 – 2 m

Average PO4-P Concentration vs. Macrophyte Biomass

y = 2E-06x + 0.018

R2 = 0.6904

y = -2E-10x2 + 6E-06x + 0.008

R2 = 0.921

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.0 5000.0 10000.0 15000.0 20000.0 25000.0

Macrophyte Biomass kg/m3

PO

4-P

mg

/L

Average NO3-N mg/L vs. Wetland Volume and Wetland Depth

-500000

0

500000

1000000

1500000

2000000

2500000

1 1.1 1.2 1.3 1.4 1.5 1.6

Wetland Depth m

We

tla

nd

Vo

lum

e m

3


180

Figure 65: Average NO3-N vs. Macrophyte Biomass at Wetland Depth range 1 – 2 m

Average NO3-N Concentration vs. Macrophyte Biomass

y = -1E-06x + 0.0613

R2 = 0.5466

y = -1E-10x2 + 2E-06x + 0.0547

R2 = 0.7452

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.0 5000.0 10000.0 15000.0 20000.0 25000.0

Macrophyte Biomass kg/m3

NO

3-N

mg

/L


181

As can be seen in Figure 66, zooplankton biomass trend follows macrophyte biomass

trend (the data has been ranked by macrophyte biomass (kg/m3)). This would indicate

that more so than the food source phytoplankton biomass, the assumed shelter

provided by macrophytes is very important for zooplankton (the wetland names

corresponding to the numbers used in Figure 66 can be found in Appendix D).

Nevertheless, despite a general increase in zooplankton biomass trend following the

macrophyte increase, zooplankton exhibits dependence on its food source

phytoplankton. This can be observed in wetland S0106 (645) where the phytoplankton

is relatively low, which is consequently reflected in the zooplankton.

Figure 66: Comparison of Macrophyte, Phytoplankton and Zooplankton Biomass for each

category 3 wetland (Key to wetland numbers adapted from (Jensen et al. 1996), see list in Table

18 in Appendix C)


182

Summer wet winter dry

Siebentritt (2003) describes a number of different water regimes in the restoration

options via flooding and draw down of water for the wetlands of the lower River

Murray. Each of these regimes is intended to illicit a diversity of vegetation and

habitat types. One of these is the use of the natural flow regime as suggested by Poff

et al. (1997), and which has been applied experimentally by the Department of Water,

Land and Biodiversity Conservation (DWLBC 2004; Siebentritt et al. 2004). Another

recommendation by Siebentritt (2003) is the implementation of restoration water

regimes to enhance a mosaic of vegetation structures within the lower River Murray

wetlands. Most current wetland management practices attempt to mimic the natural

flow regime and enhance macrophyte biomass. The model scenario discussed here

however, focuses on the minimisation of phytoplankton growth and suggests a return

to a more natural flow regime.

The natural (historical) flow pattern of the River Murray is minimal flow in March,

which increases slightly in April and May. In the upper reaches of the River Murray

catchment, where the majority of the water is sourced, the flow reduces in early

winter as freezing sets in, binding the precipitation in snow and ice. The major annual

flow occurs in spring due to snowmelt and continues into mid December due to

westerly influenced precipitation. The flow therefore achieves its height in spring and

slowly declines until it reaches a minimum in March (Burton 1974; Walker 1979;

Walker 1985). Due to the slow transport of water along the river the flow can be

delayed for 4 to 6 weeks until it reaches the lower River Murray wetlands (Mackay et

al. 1990).

Three wetlands were randomly selected to assess the impact, on wetland nutrient and

phytoplankton retention, of restricting the wet period to the macrophyte summer

growth period. Assuming that the wetlands are wet for the period of major

macrophyte growth only, a change in trend may be observed (Figure 67,

phytoplankton on secondary axis). Figure 67A shows a full year wet where the

retention is calculated as the average per day for the simulated time period. Figure

67B shows the results of summer wet/winter dry scenario; here the retention is

calculated from the average per day for the summer growth period of 88 days. The

PO4-P retention per day does not show a large improvement; however, there is a slight


183

improvement when comparing the status quo and the 75% turbidity reduction scenario

(detailed results can be seen in Appendix D Table 23 to Table 25). With summer wet

winter dry wetland management, there should be a large reduction in turbidity and

therefore increased macrophyte growth for this period. With macrophyte growth for

the entire wet period, PO4-P retention should be mainly through macrophytes rather

than phytoplankton. The scenarios in Figure 67 show the NO3-N retention per day to

improve, both when comparing the management scenarios “full year” and “summer

wet winter dry”, and as a response to increased turbidity reduction within each of the

different management scenarios. The reduction in phytoplankton growth is as a direct

consequence of the loss of its growth period, which would normally have occurred as

the macrophyte biomass reduced during the winter period. Therefore, the management

strategy of summer wet would assist in minimising phytoplankton growth. The

nutrient retention during the winter months would otherwise have been utilised for

phytoplankton growth that has now been limited. The cumulative trend shows that if

the aim of management was to minimise phytoplankton inflow into the river with a

maximum potential of nutrient retention through macrophyte growth then, the 75%

turbidity reduction scenario with summer wet winter dry management would be the

optimum management scenario as it produces less phytoplankton.

This scenario is limited by the monitoring period available. The modelled scenarios

are run for the time frame for which there is data available, which is in late summer.

The scenarios show that if the wetlands were to be flooded, i.e. the turbidity reduced,

at the time of year in which data was available, the macrophytes would be limited to

the available timeframe when water temperature is appropriate, and underwater light

and nutrients are available. However, the height of the natural flow regime of the

lower River Murray when wetlands would naturally have been inundated is

considerably earlier, i.e. during spring to early summer (Burton 1974). The results

provided here, although shifted in season, do show the impact of managing the flow

regime of the wetlands to optimise the use of macrophytes in nutrient removal and

reduction in phytoplankton. With full season (one year) data, scenarios could be

produced to obtain a more accurate assessment of the impact of mimicking the natural

hydrological regime in wetland management. In the mean time the scenarios

presented here give an indication of the impact the control of a wetland hydrological

regime may have on nutrient retention.


184

Figure 67: Nutrient uptake for full year wet vs. uptake for summer wet/winter dry

Full year

0

1

2

3

4

5

6

Status Quo

Turbidity

25% Turbidity Reduction 50% Turbidity Reduction 75% Turbidity Reduction

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0


S u m m e r W e t W inte r Dry

0

2

4

6

8

10

12

Status Quo

Turbidity

25% Turbidity Reduc tion 50% Turbidity Reduc tion 75% Turbidity Reduc tion

-0.07

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0


A

B


185

6.2 Cumulative assessment: category 4 wetlands

This section presents the results of category 4 wetland scenarios where 7 wetlands

were simulated and compared to status quo (a list of wetlands simulated can be seen

in Appendix C Table 19). Figure 68 shows the influence of the cumulative loading to

category 4 wetlands, where there is a steady increase in the PO4-P and phytoplankton

retention. NO3-N retention however is more variable. Due to the high turbidity of the

wetlands there is virtually no macrophyte growth (as discussed in section 5.1.1). The

phytoplankton shows some growth during the spring and summer months and the

zooplankton growth trend follows that of the phytoplankton (Figure 69 to Figure 71).

The concentrations PO4-P and NO3-N reduce slightly as evidenced by the slight

decrease in the wetland average (Figure 72 and Figure 73).

Of the five wetlands used in model development only Reedy Creek has adequate river

data, for its location, that is monitored on the same day as the wetland data, see (Wen

2002a; Wen 2002b). However, although Reedy Creek wetland data is used as an

“exemplar” for other wetlands of the same category, the river flow and nutrient load

for appropriate wetland locations must also be used (see Box) as in category 3

wetlands described above.

There is no significant role played by wetland internal nutrient dynamics. This is due

to the lack of macrophyte growth and therefore there being no change in the nutrient

uptake. The main impact of category 4 wetlands is therefore produced by the

reduction of irrigation drainage concentration. The results in Figure 68 to Figure 73

reflect the change of concentration within the open water of the various wetlands.

Detailed results for Figure 68 can be seen in Appendix D (Table 26 to Table 28). The

potential cumulative impact the management of the category 4 wetlands have on river

nutrient load is discussed in section 6.3.

The Reedy Creek monitored river nutrient data was compared to the available

river data (from river lock monitoring points) otherwise used in the model. The

scenarios that were based on the river data responded with relatively good results.

This is despite the model not being calibrated to this river data. Therefore, the use

of river data from the respective monitoring locations close to the simulated

wetlands was considered to improve the potential spatial accuracy of WETMOD.


186

Figure 68: Cumulative loading to category 4 wetlands

Cumulative retention by Category 4 Wetlands

0

10000

20000

30000

40000

50000

60000

70000

80000

Status Quo

Irrigation Drainage

25% Irrigation

Drainage Reduction

50% Irrigation

Drainage Reduction

75% Irrigation

Drainage Reduction

0

100

200

300

400

500

600

700

800

900

PO4-P Net Retention kg/annum NO3-N Net Retention kg/annum

Phytoplankton Net Retention m3/annum


187

Figure 69: Macrophyte Growth Trends

All fall below the red line showing that the irrigation drainage inflow has no impact on the macrophyte growth trends.

M acr o p h yte Bio m as s in C ate g o r y 4 w e tlan d s at 50% Ir r ig atio n Dr ain ag e Nu tr ie n t

Re d u ctio n Sce n ar io

0

0.02

0.04

0.06

0.08

0.1

0.12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

M acr o p h yte Bio m as s in C ate g o r y 4 w e tlan d s at 75% Ir r ig atio n Dr ain ag e Nu tr ie n t


0

0.02

0.04

0.06

0.08

0.1

0.12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

kg

/m3

A

B



188

Figure 70: Phytoplankton Growth Trends

Ph yto p lan k to n Bio m as s in C ate g o r y 4 w e tlan d s at 50% Ir r ig atio n Dr ain ag e Nu tr ie n t


0

2

4

6

8

10

12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

Ph yto p lan k to n Bio m as s in C ate g o r y 4 w e tlan d s at 75% Ir r ig atio n Dr ain ag e Nu tr ie n t


0

2

4

6

8

10

12

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

A

B



189

Figure 71: Zooplankton Growth Trends

Z o o p lan k to n Bio m as s in C ate g o r y 4 w e tlan d s at 50% Ir r ig atio n Dr ain ag e Nu tr ie n t


0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

Z o o p lan k to n Bio m as s in C ate g o r y 4 w e tlan d s at 75% Ir r ig atio n Dr ain ag e Nu tr ie n t


0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

A

B



190

Figure 72: PO4-P Trends

PO4-P C o n ce n tr at io n in C ate g o r y 4 w e tlan d s at 50% Ir r ig atio n Dr ain ag e Nu tr ie n t


-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8J

un

-00

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

PO4-P C o n ce n tr atio n in C ate g o r y 4 w e tlan d s at 75% Ir r ig atio n Dr ain ag e Nu tr ie n t


-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

A

B



191

Figure 73: NO3-N Trends

A

B

NO3-N C o n ce n tr at io n in C ate g o r y 4 w e tlan d s at 50% Ir r ig atio n Dr ain ag e Nu tr ie n t


-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4J

un

-00

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

NO3-N C o n ce n tr atio n in C ate g o r y 4 w e tlan d s at 75% Ir r ig atio n Dr ain ag e Nu tr ie n t


-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L



192

6.3 Implications of cumulative impact of multiple wetland

management

For the purposes discussed in the methodology, in this section the assumption is made

that the model is quantitatively accurate.

Category 3: Dead end wetlands with carp presence and no irrigation drainage

To assess and discuss the potential cumulative impact that the management of all

category 3 wetlands may have upon the river nutrient load the model quantitative

output is assumed to be relatively accurate. Therefore, evaluating these results, the

cumulative impact shows that there would be a net retention of NO3-N (Table 9).

However, the PO4-P inflow into the river may increase due to the loss of retention

through wetland sedimentation (the model does not fully take into account sediment

resuspension) and the phytoplankton load may also increase due to the increased

underwater light availability (Table 9).

Table 9: Impact, of category 3 wetland’s management, on river load per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction -803 6223 -61

% of River load removed through 75% Turbidity Reduction management -0.33 1.71 -1.58

The simulations of introducing dry periods as a management strategy for wetlands

need to be scrutinised further. WETMOD 2 uses a simplistic sedimentation and re-

suspension equation for PO4-P, NO3-N and phytoplankton. The wetland internal

nutrient concentrations are more dynamic than portrayed in the model. This is one of

the most significant limitations (i.e. the abrupt sedimentation threshold) of the model.

Although the model can be applied to more extensive wetlands due to its simplistic

construction and data prerequisites, it is acknowledged that for management purposes

a more accurate estimation of nutrient and phytoplankton retention by the wetlands

would be favourable. However, the model does provide a framework for expansion of

research to assist in the assessment of the cumulative impact of wetland management

on a regional scale.


193

Presently the model provides the opportunity of simulating the trends within a

wetland due to potential management strategies. These wetland simulations would

become more accurate with the present model (WETMOD 2) as more data, and

particularly comprehensive data, becomes available. As discussed above (chapter 6),

model accuracy could be improved if more local knowledge of particular wetlands

were applied in cumulative assessments (i.e. better turnover estimate) “exemplar”

driving variables could however still be used for category wetlands. Future work on

the extension of WETMOD 2 should focus on the inclusion of detailed water and

sediment interaction, particularly nutrient uptake, and the potential change that may

occur due to sediment compaction.

As discovered in section 6.1 (Figure 59 and Figure 60) there seems to be optimum

wetland morphology for macrophyte growth and therefore maximal nutrient and

phytoplankton retention. Wetlands were split into the depth categories shallow,

medium and deep (Table 10 to Table 12).

The cumulative impact scenarios made by WETMOD 2 for the shallow range of

wetlands (58% of wetlands), shows this range to be least effective at nutrient retention

(Table 10). In this shallow range of simulations, for the 75% turbidity reduction, there

is a net increase of 0.39% in the PO4-P river load and a full 1% of the phytoplankton

load. However, there is a decrease of NO3-N of 0.75%. In contrast, the medium and

deep wetlands (each 21% of wetlands) show retention for both PO4-P and NO3-N. Of

these two depth ranges, deep wetlands have a minimal impact on phytoplankton river

load with only a 0.06% increase. From these simulations the conclusion that can be

drawn is that the medium and deep wetlands on the whole have a greater impact on

nutrient retention than the shallow wetlands. Consequently, if only a small number of

wetlands were to be managed the medium and deep wetlands would potentially

provide the greatest cost benefit return.

Model application limitations

Prior to WETMOD 2 being used to make management decisions, beyond the

theoretical examination presented here, some restrictive issues must be addressed. The

reliability of the macrophyte growth representation in very shallow wetlands is

questionable. This issue was discussed in section 4.3. As the validation of Lock 6

wetland, which is within the range of shallow wetlands, confirmed the macrophyte


194

growth trend within this range the issue must be raised as to the accuracy of the

macrophyte growth trend of the deep wetland (for which the model was not

specifically calibrated). If the Secchi depth influence on macrophyte growth equation

were to be modified (i.e. to take into account maximum wetland depth) to better

reflect the situation in lower River Murray wetlands this result may change

considerably. Validation with monitored macrophyte data would however still be

required.

Table 10: Impact, of category 3 wetland’s (depth range shallow <1m) management, on river load

per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction -961 2741 -39

% of River load removed through 75% Turbidity Reduction management -0.39 0.75 -1.01

Table 11: Impact, of category 3 wetland’s (depth range medium 1-2m) management, on river

load per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction 91 1981 -20

% of River load removed through 75% Turbidity Reduction management 0.04 0.54 -0.51

Table 12: Impact, of category 3 wetland’s (depth range deep >2m) management, on river load

per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction 68 1501 -2.43


Looking at the impact of the management of Lock 6 wetland only, which is within the

shallow depth range, there is still a positive impact on the reduction of river nutrient

load, Table 13. There is a very small PO4-P uptake, which suggests that the retention

capacity of the wetland is improved through the turbidity reduction management,

although this is virtually negligible. Comparing Lock 6 wetland results in Table 13


195

with those produced when Lock 6 wetland is considered for summer wet winter dry

cycles in Table 14, Lock 6 wetland shows a slightly more promising retention

capacity. In this scenario Lock 6 wetland has a slightly greater effective PO4-P

retention and less phytoplankton contribution to the river.

Despite this very small improvement, an assessment of the simulation output of

multiple wetland management of category 3 wetlands can be used to gain insight into

the cumulative impact that might be obtained on the lower River Murray nutrient and

phytoplankton load. Some indication as to the wetlands that may be the most effective

at nutrient retention can also be deduced. Where qualitative scenario results can assist

in assessing a wetland in the lower River Murray, WETMOD 2 is a functional tool.

That is, WETMOD 2 can simulate category 3 wetlands for which limited data is

available. The modelling output reliability for these wetlands can be improved with

local knowledge of exchange volume, macrophyte growth trend and sediment

compaction potential. However, to reliably apply the WETMOD model and decide on

potential management scenarios to be applied, the model should still be developed

further, as discussed above.

Table 13: Impact, of Lock 6 wetland management, on river load per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction 0.44 24 -1.89


Table 14: Impact, of Lock 6 wetland management, summer wet winter dry, on river load per

annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 245604 364372 3880

Change from status quo at 75% Turbidity Reduction 1.79 17 -0.21


Category 4: Dead end wetlands with carp presence and irrigation drainage

As in category 3 wetlands, to assess the cumulative impact on river nutrient load, the

scenarios for category 4 wetlands are assumed to be quantitatively accurate. In


196

category 4 wetlands the nutrient and phytoplankton retention calculated includes the

irrigation drainage inflow reduction. Therefore, the improvement in category 4

wetland retention and its impact on river load includes the PO4-P, NO3-N and

phytoplankton assumed to be removed through constructed wetlands that would

otherwise have been flowing into the wetland as part of the irrigation drainage.

Table 15 shows the potential nutrient retention capacity of category 4 wetlands and

the impact on the river nutrient load. It must be remembered that due to the limited

data available on wetlands of the lower River Murray, which are affected by irrigation

drainage, the data available from Reedy Creek wetland was applied to wetlands within

this category as an “exemplar” data source. Despite these wetlands being within the

same category as Reedy Creek wetland, the irrigation drainage inflow would vary

more than is accounted for in these scenarios. However, although the irrigation

concentration and volumes would differ, some floodplain wetlands of the lower River

Murray are directly impacted by irrigation drainage, having very high nutrient loads.

Category 4 wetland cumulative assessment is hypothetical scenario testing intended to

examine the cumulative impact of management, and to assess the capacity of the

model to simulate category 4 wetlands.

Through the introduction of constructed wetlands, to reduce irrigation drainage

nutrients entering a wetland, a net retention of nutrients normally flowing into the

river is achieved. Table 15 shows the hypothetical cumulative retention if all category

4 wetlands are successfully managed. The model indicate that these 7 wetlands would,

in case of 75% irrigation drainage nutrient reduction, contribute a 2.68% reduction of

the river phytoplankton load, as well as a small reduction of PO4-P and NO3-N river

load.

Table 15: Impact, of category 4 wetland’s management, on river load per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 1376872 338228 12231

Change from status quo at 75% Irrigation Drainage Nutrient Reduction

5850 1205 328

% of River load removed through 75% Irrigation Drainage Nutrient Reduction management

0.42 0.36 2.68

The wetland retention portrayed in Table 16 is that of Reedy Creek wetland, whose

data quality and therefore modelling accuracy, both qualitatively and quantitatively,


197

was most comprehensive and accurate. Table 16 is therefore bound to be the most

accurate reflection qualitatively and quantitatively of the impact of wetland

management on wetland nutrient retention capacity.

Table 16: Impact, of Reedy Creek wetland management, on river load per annum

PO4-P kg/annum

NO3-N kg/annum

Phytoplankton m

3/annum

Load in River 1376872 338228 12231

Change from status quo at 75% Irrigation Drainage Nutrient Reduction

1052 163 48

% of River load removed through 75% Irrigation Drainage Nutrient Reduction management

0.08 0.05 0.40

Through the management of Reedy Creek wetlands, with the assumption of 75%

nutrient reduction capacity, the model indicates a small reduction of PO4-P, NO3-N

and phytoplankton to the river. Therefore, the model suggests that the management of

even one category 4 wetland may slightly reduce river nutrient load.


198

6.4 Chapter summary and Implications for the third

hypothesis

Although there is a limited availability of data for wetlands of the lower River Murray

modelling does allow for scenario development of multiple wetlands. The generic

nature of WETMOD 2 has therefore allowed its application to multiple wetlands

where only rudimentary morphological data is available. The model has thereby been

applied on a landscape scale. The modelling limitations have been described and

include the important point that the quantitative results can only be qualitatively

indicative of potential management outcomes.

Reviewing the data produced during cumulative assessment of multiple wetland

management the third hypotheses “A simplified generic wetland model can be used to

assess the cumulative impact of managing multiple same category wetlands” can be

addressed. The simulations above show that this is possible. The main outcomes from

the cumulative simulations are to find the optimum wetland morphology (for the best

return on investment), the hydrology season for optimum nutrient uptake, and the

impact of effective constructed wetlands in removing irrigation drainage load.

However, this ability is presently restricted as per the following;

The output is qualitative and not quantitative due to the nature of simplified

models and the use of “exemplar” data.

Due to the limitation in the simulation of turbidity reduction, i.e. turbidity

controlled sedimentation threshold of nutrients and phytoplankton biomass,

the management scenarios estimate comparisons of nutrient removal

efficiency may become biased towards a turbid wetland. This limitation is

solvable through further model development; therefore the methodology used

and described above remains applicable particularly when this limitation has

been addressed.

Thus, the output of the cumulative assessments of management of multiple category 3

wetlands is preliminary. However, the potential of using generic models for

cumulative assessments is substantiated by the methodology used as shown by its

application to category 4 wetlands and the application to the preliminary management

scenarios of the category 3 wetlands.


199

The nature of differential equations allows the conservative use of available mass.

Theoretically therefore the assessment of the mass balance should be possible; for the

wetlands where monitoring has taken place some indication of mass balance is

available, particularly for Reedy Creek wetland. However, due to the limitation of

data availability the current modelling effort should only be viewed as being capable

of estimating potential mass balance. That is, the qualitative information obtained

through the landscape scale scenarios allow for a simplistic understanding of the

cumulative impact of management of multiple same category wetlands on river

nutrient load.


200

7 Summary, Context and Discussion

The use of differential equations allows a deterministic approach where simulations or

scenario analysis are possible. The predictive modelling of wetlands contributes to

informed decisions on management strategies based on the data available. The

uncertainty or lack of knowledge and data does affect the quality of model predictions

(Wallach et al. 1998). However, this does not prevent management and decision-

making and is a part of ecological simulation modelling (Reckhow 1994; Wallach et

al. 1998). As long as the decision makers understand the limitations, they can still use

a model to assess scenarios within these model limitations. The model may, in fact,

provide decision makers with the only tool to experiment and increase understanding,

without which they could be limited in assessing potential management impacts. This

enhanced knowledge enables a better prediction of outcomes and therefore aids

decision making in regard to management. Further, with enhanced knowledge and

transparent assessment, consensus between stakeholders involved in the decision-

making is more readily achieved (Thomann 1998).

Management decisions for ecosystems may be made by many stakeholders, not all of

whom would fully understand the ecological implications of different intervention

options. Furthermore, experts in the field can hold opposing views on subjective

topics. Modelling can be seen as a structure to assist in regulating knowledge, data

and assumptions used for decision-making. Other experts can participate and

comment on the model as it is defined. Most decision support models have inherent

uncertainties of an acceptable magnitude (Reckhow 1994). It must however be made

clear where there is a lack of knowledge and/or other uncertainties, and how this has

been dealt with within the model. This information will reflect on how the model can

and should be used, and how much reliance can be placed on the modelling

predictions. Which detail is required and the appropriateness of assumptions is

dependent on the purpose of the model (Caswell 1988).

For the model developed in this project to be applicable on a regional scale, data

obtained from monitored wetlands was assumed to be appropriate for internal wetland

behaviour and relationships of similar wetlands. This assumption was used to

overcome the lack of knowledge and data for the lower River Murray floodplain

wetlands and simulate regional scale scenarios; and thereby obtain a cumulative


201

impact assessment of the management of multiple wetlands on river nutrient load.

Therefore, implicitly the understanding should be that the model output is of trend

behaviour and potential impacts on the river, both prior to and post management

scenarios.

7.1 Assessment methodology

The application of WETMOD 2 was designed for wetlands where minimal data, such

as morphology and spatial location, has been sourced. For these wetlands the driving

variables are borrowed from their associated “exemplar” wetland. Quantitative data

from parameters measured in wetlands were used in WETMOD 2 to act as

“exemplars” to provide qualitative outcomes in other wetland systems based on

wetland categories. Due to the assumptions made and described in section 3.1, as well

as the intended purpose or aims of the model, WETMOD 2 is maintained in a generic

form to be qualitatively applicable to wetlands where only basic morphological data

are available. Through this methodology a model was developed that is based not only

on scant data but is also applicable to wetlands with no time series data (the modelling

predictions of WETMOD 2 were therefore not assessed strictly in a quantitative

manner).

It is not possible to statistically assess the model outputs for these wetlands, as no data

with which to compare the output exist. There must therefore be general confidence in

the simulated time-series seasonal trend and approximate magnitude produced by

WETMOD 2, for the wetlands used in validation of the model. Otherwise, no

confidence will be placed in the scenarios produced for category wetlands, i.e. those

using “exemplar” driving variables. It can be said that the qualitative assessment of

such model scenarios may be a more significant assessment of the model

performance, than an improvement in statistical accuracy of individual wetlands (i.e.

optimisation of quantitative performance of the model). Modelling effort was

therefore directed at the development and improvement of spatial contributions to

wetland modelling, rather than focusing on the improvement of the individual wetland

process modelling. This approach is an extension of the view presented by McIntosh,

et al. (2003) that flexible and cost effective models are more beneficial than one off

models that perform very well for one ecosystem only.


202

To represent qualitative model performance the score D was used and served the

model development well and is used extensively in the model validation. Other

statistical options are discussed in (Mayer et al. 1993), however the statistical

accuracy of the model would not solely or adequately improve the confidence of users

when WETMOD 2 is used as a landscape decision support tool. When assessing the

performance of WETMOD 2, by comparing the modelled output with its monitored

counterpart, the model behaves qualitatively correctly and logically for each wetland

considered. This is reflected in the similarity of seasonality of the modelled response

and monitored concentrations. The seasonal response of the non-monitored wetland

parameters macrophyte and zooplankton in model scenarios was logical, supporting

model validity.

As discussed in the introduction (section 1.4.2), the qualitative difference in the

comparison of different wetlands is a legitimate model assessment methodology. The

purpose of the model determines its required precision. In the case of WETMOD 2,

the qualitative assessment of model results can in fact be the most appropriate

methodology when the model is applied outside its development envelope. Evidence

of this is in the discrepancy between visual assessment of validation results and the D

for some of the wetlands, see section 4.1. This would in fact particularly be the case

where WETMOD 2 is applied to category wetlands where data from “exemplar”

wetlands is used. Nevertheless, the statistical evaluation of the modelling accuracy is

a significant validation step required to assess the model performance. This can

however, only be undertaken for scenarios where the model is simulating actual

monitored data using monitored driving variables. Values of D are presented in Table

3 and Table 4 and discussed during the validation of the model.

7.2 Current capabilities

The results, described and discussed in chapters 4, 5 and 6, have shown the

applicability of the model at the present stage of development, its limitations and

identified areas requiring further research and model improvement. A summary of the

present capabilities of the model, in providing information which was previously not

available, include:


203

finding the exchange volume of water and therefore nutrient and

phytoplankton load between wetlands and the river (For wetlands with nutrient

date time series)

calculating the status quo nutrient retention of wetlands

developing estimates of the potential impact of management on the nutrient

retention of wetlands

estimating the impact wetlands may have on the river nutrient load due to

improved management

producing an estimate, based on qualitative output, of the cumulative impact

of multiple wetlands management on the river nutrient load, and

developing comparative studies of wetlands based on their morphological

differences, using the same driving variable time-series.

Further advantages of the model include:

presenting a framework from which to expand the model capabilities through

an improvement in the workings of the model (some of the model expansion

would not require a dramatic increase in model complexity)

presenting a framework from which to expand the model capabilities as data

availability increases

focusing future monitoring for improved assessment, and enhanced modelling

capabilities which may aid management decisions, and

posing questions where model limitations are encountered due to a lack of data

or knowledge

Currently the central problem for modelling wetlands of the lower River Murray is

data quality and quantity. Now that the model has been developed, future monitoring

can take its data prerequisites into account to alleviate this restriction, thereby the

model serves the purpose of focusing future research needs. Model limitations are

discussed below.


204

External influences and Landscape Scale

WETMOD 2 is capable of estimating the exchange volume between a wetland and the

river where wetland nutrient time series are available. Using the exchange volume the

model can further account for external influences acting upon, and therefore improve

the modelling of, wetland internal dynamics. Together with the exchange estimate and

the internal nutrient dynamics the probable outflow load, of nutrients and

phytoplankton biomass, can be estimated. Thereby, the model can be used to assess

the impact the wetland has on river nutrient load, and how this can be altered through

potential management strategies. A call for such a model for Australian wetlands was

made by McComb and Qiu (1997).

Due to the models simple structure and low driving variable demand it is generically

applicable to other wetlands within the region, which were not used in model

development. Thereby, the model can be used to estimate the status quo or the

potential impacts management may have on wetland nutrient and phytoplankton

retention and consequent river load, even if minimal data for the wetland is available.

WETMOD 2 simulates the qualitative behaviour without the quantitative accuracy. In

this case the qualitative behaviour of multiple floodplain wetlands (where

morphological data only is available), reflecting model potential as proposed by

Rykiel (1996). Specifically, although the data simulated for each category of wetlands

may not be quantitatively accurate, the trends are plausible. In a cumulative

assessment the simulated impact of multiple wetland management, is indicative of

potential results.

As discussed in chapter 6, cumulative assessment assists in focusing management

oriented research. The model allows the user to determine the implications of

assumptions made, i.e. whether they are valid or otherwise. The role of the modelling

tool is therefore (in part) to confront users with the implications of beliefs that they

may hold (Bart 1995). Therefore, the potential outcomes of modelling on a regional

scale, where minimal data are available, may assist managers in directing future

monitoring studies and thereby aid in eventual decision making. For example,

modelling outcomes of optimal wetland morphology are related to exchange rate for

nutrient retention.


205

Although the results presented can be used to the degree discussed in chapters 4, 5

and 6, it is stressed that the model is still in early development. Model improvements

and validation with specifically monitored data should be performed. Further research

is suggested in chapter 8.

The application of the model is at this stage still restricted to wetlands of category 1,

3, 4 and 5, as the data available for wetlands of category 2 were insufficient for proper

validation. Management strategies are available only for category 3 and 4 wetlands

however; model applicability can be enhanced as data becomes available.

Presently the model can be used to assess, qualitatively, the potential cumulative

impact of multiple wetland management. For example, the comparison of two

wetlands, for which there is limited data availability, is possible by developing

scenarios based on wetland categories and the morphological data available for these

wetlands. Future feedback when comparing model predictions with actual outcomes

will aid in identifying incorrect hypothesis, model inaccuracies and therefore

contribute to future improvement of the model and enhancing its performance and

applicability.

Limitations

There are four significant limitations to the model at this stage (in order of

significance), with the second and third being related.

The first is the abrupt sedimentation threshold (70 NTU for PO4-P and NO3-N

and 95 NTU for phytoplankton), which makes distinguishing change in

nutrient retention due to varying management scenarios difficult. More data on

sedimentation rate and resuspension would be helpful.

The second limitation is the inapplicability of the model to very shallow

wetlands. Although the wetlands, where data was available, were not

shallower than the 0.6 m, some wetlands of the lower River Murray are. An

update of the equation that considers macrophyte growth in relation to Secchi

and maximum depth of the wetland should improve this model aspect.

Currently the data available for wetland depth is only used to calculate

wetland volume within model simulation. The addition of wetland depth to


206

factor in the impact on macrophyte growth would improve the generic

modelling applicability and allow the simulation of shallow wetlands.

The third limitation relates to model output. The shallowness of modelled

wetlands (below 1 metre depth) may still be the best wetlands for

management, despite the model results. Asaeda et al. (2001) in fact found in

their modelling studies that, despite shallow wetlands having a higher

concentration of phytoplankton, macrophyte growth did increase due to more

favourable light conditions. In shallow wetlands macrophyte growth may be

expected throughout the wetland, causing increased sedimentation, increased

nutrient uptake and shading out of phytoplankton. The increased macrophyte

growth would also provide more shelter for zooplankton, which feed on

phytoplankton further reducing their numbers. Therefore, the equation that

was discussed in model validation section 4.3, and which shows a logarithmic

growth pattern with increasing wetland depth needs to be reviewed.

As stated in the methodology, zooplankton and macrophyte biomass data were

not available for model development, validation and calibration. The model

output and conclusions made are therefore limited by this lack of data and do

not necessarily accurately reflect what could occur in a real environment.

Despite these limitations to the methods applied, the WETMOD 2 simulation results

and assessment of potential cumulative impact of wetland management remain

applicable. WETMOD 2 is a work in progress, and this project mainly contributes to

the spatial factors of lower River Murray wetland modelling. The present assessment

of the model‟s capabilities has helped to identify future research requirements such as

the model structure (equation improvement/replacement), model expansion (sediment

water interaction), and data acquisition (wetland monitoring).

The use of river Chlorophyll-a levels from Murray Bridge as the driving variable for

all phytoplankton exchange (as discussed in section 3.2.3) led to Pilby Creek and

Lock 6 being the only wetlands that showed virtually no improvement in model

performance with regard to phytoplankton simulation (as is shown in section 4.1).

With additional monitoring of river Chlorophyll-a levels the accuracy of model

phytoplankton simulation should be improved.

Other avenues to improve model performance include:


207

measurement/establishment of exchange volume estimates for simulated

wetland (based on future monitoring and digital elevation models)

determining the sediment compaction potential of individual wetlands

measurement of irrigation volume (category 4 wetlands only), and

determination of evaporation impact on wetland nutrient retention balance

calculations.

In its present state the model can be used for some restricted management assessment.

This management focus would be on potential:

nutrient retention of wetlands

exchange volume and nutrient load

twin management (limited)

comparative studies of wetlands based on morphology (limited)

impacts on river nutrient loads (indication only), and

cumulative impact of multiple wetlands management.

Revisiting the Project Aims

Now the model capabilities have been assessed it is necessary to revisit the aims of

the project to assess whether the model extension has fulfilled the intended purposes.

Model extension aimed to:

I. overcome shortcomings in knowledge due to limited data and incomplete

system understanding

II. address processes requiring further development, which were identified at the

beginning of the study. These included river and wetland water exchange,

nutrient exchange, and irrigation drainage data influence, and

III. adapt and test the application of the model on a regional scale; i.e. develop a

cumulative assessment of potential management impacts of multiple wetlands

on the river nutrient load.

To fulfil the first aim the model first fulfilled the second, which is the extension of the

models capabilities. The model is now able to estimate water exchange, therefore


208

developing data for a previously unknown quantity for those wetlands where data is

available. This has led to the ability to estimate the nutrient retention capacity of

monitored wetlands and simulate potential change due to management. From this the

bi-directional nutrient exchange has been modelled. Based on a similar methodology

the irrigation drainage influence has also been accounted for, where relevant.

The third aim was fulfilled with the use of the different “wetland categories”, i.e.

using “exemplar” driving variable data. Thereby, qualitative estimates of the

cumulative impact of multiple wetlands on the river nutrient load could be developed,

as well as an assessment of the impact of management of these wetlands.


209

8 Conclusion & Future Work

This project set out to develop a model capable of simulating nutrient retention

capacity of the lower River Murray wetlands. The model was to be applied on a

regional scale encompassing wetlands for which limited data is available. In applying

the model, it was to assess the change in nutrient retention capacity of multiple

wetlands and the cumulative impact on the river following hypothetical management

interventions of these wetlands.

The application of the developed regional model WETMOD 2 is constrained by the

availability of comprehensive data of adequate quality and frequency in the lower

River Murray. However, the study does serve the purpose; demonstrate the provision

of a tool for examining the impact of management interventions on the broader scale.

The model also helps to purpose of focus future research, including purpose driven

monitoring and model improvement.

Hypotheses

The modelling has fulfilled most of the objectives and aims of the project, with the

assessment of model output and its limitations discussed in the respective results and

discussion chapters and summarised in section 7.2. These hypotheses were:

I. A simplified generic wetland model can be used to realistically simulate

multiple and different wetlands qualitatively.

Given adequate driving variables the model can simulate different

wetlands realistically, e.g. Lock 6 and Reedy creek using non-

calibration data see section 4.4.

II. A simplified generic wetland model can be used to answer “what if” questions,

and

Management simulations for selected degraded wetlands have been

successfully run, see section 5.2.

III. A simplified generic wetland model can be used to assess the cumulative

impact of managing multiple same category wetlands.


210

Limited qualitative assessments are possible. For category 3 wetlands

this assessment is preliminary due to (solvable) model limitations (see

section 6.4). For category 4 wetlands these same limitations restrict the

adequate modelling of two concurrent management strategies.

With the scenarios developed of the different wetlands, general understanding of the

system can be enhanced and the hypotheses tested with regard to alternate

management options and their required response. The differential equation based

deterministic model WETMOD 2 does provides a tool for hypothesis testing of

management effectiveness for wetland regeneration. WETMOD 2 is a tool that can be

used in the facilitation of understanding of the required management effort for

successful wetland restoration, i.e. percentage of turbidity reduction required for

macrophyte growth response and therefore wetland regeneration.

Understanding of the cumulative response of multiple wetland management is

enhanced by the model scenario output. Although the model output is qualitative it

does provide some assessment of cumulative impact. Further development of the

model would enhance this feature. Some understanding can also be obtained of the

general differences between wetlands (smaller versus larger, shallow versus deeper

etc.), although minimal data is available. While the model outcomes cannot be viewed

as quantitatively accurate (particularly in individual category wetland comparison),

the model outcomes do provide a point of reference from which further research can

be made. The model outcomes, in such a comparative use, are for general

understanding as well as an aid in facilitating consensus on the potential impact of

restoration options, assuming there is confidence in the model.

Future development of WETMOD

During the development, calibration, validation and application of the model, certain

limitations were discovered, as well as potential improvement identified for which

there was inadequate time to address. The following recommendations for future

model improvements are made (this list is not exhaustive as other improvements

could be made). Model improvements need to take into account the lack of data in the

region.

Underwater light and Secchi depth need to be fixed for very shallow wetlands.


211

This projects purpose was to use the previously developed wetland ecosystem process

model WETMOD 1 and extend this beyond theoretical wetland dynamics to include

spatially relevant data. The project therefore was not primarily concerned with

improving internal modelling dynamics. The prerequisite for this omission being that

limitations did not affect model verification, and that consequent management

simulation restrictions were identified. Where limitations were identified, future

improvements are suggested. This model restriction was therefore an issue that was

not only outside of the scope of this project, but also one for which there was not

sufficient data to address the problem. For future application of WETMOD 2 this

limitation must be taken into account, as very shallow wetlands will, with the present

model structure, not be simulated accurately. Therefore, this limitation is of a high

priority for future development of WETMOD 2.

River turbidity & temperature are not used in the model and are only included

as potentially relevant data for the future.

Both the river turbidity and temperature will impact on wetland ecosystems and

should therefore in an ideal model be included. Depending on the distance of the

wetlands from the river the full impact of river turbidity and temperature on wetlands

may be variable. Therefore, their inclusion in a model may add to its complexity. As

discussed previously the relative simplicity of WETMOD 2 should be maintained.

Given the implications added complexity has on the model generic applicability it

must therefore contribute substantially to model output. Testing of relative

improvement in model performance following increased complexity will need to be a

deciding factor as to its merit and ultimate acceptance (i.e. a sensitivity analysis).

Rather than relying on estimates of expected efficiencies of constructed

wetlands a separate module for which artificial wetlands can be individually

modelled should be added to WETMOD 2.

Although this would add complexities to the model this module would only need to be

operational in circumstances where the availability of data allows. Such a module

could be turned on in circumstances such as done for the external nutrient inflow

(irrigation drainage) in the Reedy Creek wetland example.

Include wetland soil substrate and therefore sediment re-suspension (turbidity)

potential in status quo (in permanently inundated wetlands) and as an


212

assessment of the potential success of management through the introduction of

dry periods.

Include sediment nutrient dynamics to more accurately account for sediment

nutrient source and sink.

Sedimentation of suspended particulate matter improves water quality by reducing

turbidity and suspended solids concentration. Any nutrients and contaminants adhered

to particulate matter are also deposited during sedimentation effectively removing

them from the water column thereby further improving the water quality (Johnston

1991; Oliver 1993; Walker et al. 1982). Sediment retention and reduction of turbidity

within multiple individual wetlands can have an important cumulative impact on

water quality at a catchment scale (Johnston 1991; Johnston et al. 1990). Despite

some sediment re-suspension, sedimentation is a long term and relatively irreversible

sink (Johnston 1991). This could therefore be included in sedimentation expansion of

the model to account for the nutrient impact of sedimentation and sediment

compaction/binding. However, some sediment nutrient source is still a possibility.

Modelling of sediment as a nutrient source is therefore necessary to accurately assess

the impact of sediment and water nutrient balance. Again a balance of model

complexity and generic applicability will need to be found.

The model is still in its infancy. When more spatial patterns are introduced more

complexities will develop within the model, making it more discriminate to individual

wetland characteristics. This can to some degree still be done whilst maintaining the

simplistic model structure. An example where this was accomplished is the inclusion

of spatial dependent wetland characteristics, wetland depth and volume (section 6.1

(Wetland size, volume and location)).

One of the next development stages could be to include soil substrate data. Sediment

properties are the deciding factor to changes due to drying and reflooding (McComb

et al. 1997), therefore the wetland substrate plays an important role in the

effectiveness of the reintroduction of wetland dry periods. The fieldwork would only

need to be conducted once, as the results would be conclusive and therefore not

constitute an ongoing expense. This would deliver a strong spatial criterion in

modelling of scenarios, so much so that a potential wetland may be found to be

entirely unsuitable for management through the introduction of dry periods.


213

Improve twin management simulations

Currently the twin management scenarios are effective in formulating further research

questions such as “is sedimentation (e.g. using clay to adsorb nutrients followed by

sedimentation) the best management strategy in a highly eutrophic system or will

constructed wetlands allow sufficient nutrient removal to facilitate wetland

rehabilitation?” Developing this capacity within the model may provide some

direction for further field based research.

Adoption of WETMOD 2 into Spatial Modelling Environment (SME)

modelling software

The initial attempt at using GIS (geographical information systems), with SME as a

platform, as a data source to the model was deemed as inappropriate in the case of the

development of a wetlands model for the lower River Murray. The sole reason for this

was the lack of GIS data, particularly a DEM (digital elevation model). The model

however was designed in a manner of keeping this option open should adequate GIS

data become available. The major advantages would be the simultaneous simulation

of all wetlands, thereby making cumulative assessments and/or comparisons between

wetlands that much easier. The recent baseline surveys of select River Murray

wetlands (SKM 2004; SKM 2006) have included relatively accurate DEM

developments, the accuracy of the DEM being between 0.25 and 0.5 meters.

Modifying WETMOD 2 for these wetlands may be possible in the future although this

would restrict the model to the monitored wetlands.

Development of an evaporation module

Evaporation as a water loss can be added using an evaporation spreadsheet complied

by DWLBC (Simpson 2003), the “Water Loss Calculator”. This was avoided early in

model development due to its own inherent inaccuracy that would have complicated

model development. The water exchange volume for a wetland was based on the

inflow estimation required to reach an optimal nutrient dynamic simulation. The

evaporation loss would reduce the simulated outflow from a wetland which was

previously assumed to be equal to the inflow. Therefore, a wetland could actually be

retaining higher loads of nutrient than so far simulated. Consequently the full

development of an evaporation module for WETMOD 2 may improve the assessment

of nutrient retention. As “Water Loss Calculator” is currently used by state


214

government agencies and wetland managers to calculate wetland evaporative water

loss, building this into the model would work in with current practice (despite its

inherent inaccuracy). The “Water Loss Calculator” is as generically applicable as

WETMOD 2 and would therefore not add to the model complexity.

Monitoring needs

Progress in model development to enhance results requires the availability of

validation data or improvement of and/or inclusion of new driving variable data (de

Wit et al. 2001). Although some of these would increase model complexity, the

relative improvement in model output may warrant their inclusion. Many would

therefore need to be considered. These new data could include:

all driving variables within a wetland;

o temperature

o turbidity

o Secchi depth

o PO4-P

o NO3-N

o phytoplankton

results of monitoring within wetlands for;

o dissolved oxygen

o zooplankton

o macrophyte biomass

o substrate (soil composition and compaction potential)

o ground truthing of through flow

information from monitoring external nutrient sources concurrent with the

wetland monitoring including concentration and volume, such as;

o irrigation drainage

o river

o groundwater


215

external climatic factors besides solar radiation, such as wind direction and

speed, shelter by surrounding vegetation (could contribute to resuspension

modelling and flow direction of water exchange).

All of these factors could impact on the division of the wetland categories. As an

example of a classification procedure Strager et al. (2000) used a landscape based

approach to classify wetlands and riparian areas based on habitat requirements of

amphibians and reptiles. This classification also included forested and non-forested

groupings as this had an impact on the wind reaching the wetlands (Strager et al.

2000). Borrowing this approach, forest cover mapping or obtaining a cover

representation from satellite imagery, might be used to differentiate classifications in

the Murray wetlands model in future work, particularly if wind and therefore sediment

resuspending equations are developed in the model.

The model developed by Muhammetoglu et al. (1997) is too complex to apply to the

lower River Murray wetlands given the lack of data, but it shows the work presently

underway to develop models of nutrient retention by wetlands. As such WETMOD 2

contributes to this research by providing an example of a simple generic model

applicable on a regional scale where very limited data are available. In the modelling

of complex environmental ecosystems, particularly where scant data is available,

simple models provide a basis with which to advance or focus management and future

research. The desire to increase complexity therefore needs to be carefully balanced

between improved model performance and applicability of the model on a landscape

scale.


216

9 References

Abbruzzese B, Leibowitz SG (1997) A Synoptic Approach for Assessing Cumulative

Impacts to Wetlands. Environmental Management 21, 457-475.

Andersson L (2004) Experiences of the Use of Riverine Nutrient Models on

Stakeholder Dialogues. Water Resources Development 20, 399-413.

Anonymous (1995) 'Wetlands: Characteristics and Boundaries.' (National Academy

Press: Washington, D.C.)

APHA, Greenberg AE, Clesceri LS, Eaton AD, Association AWW, (U.S.) WPCF

(1992) 'Standard methods for the examination of water and wastewater / prepared and

published jointly by American Public Health Association, American Water Works

Association, Water Pollution Control Federation ; joint editorial board, Arnold E.

Greenberg, Lenore S. Clesceri, Andrew D. Eaton.' (APHA-AWWA-WPCF:

Washington, D.C)

Arthington AH, Pusey BJ (2003) Flow restoration and protection in Australian rivers.

River Research and Applications 19, 377-395.

Asaeda T, Trung VK, Manatunge J, Van Bon T (2001) Modelling macrophyte-

nutrient-phytoplankton interactions in shallow eutrophic lakes and the evaluation of

environmental impacts. Ecological Engineering 16, 341-357.

Asaeda T, Van Bon T (1997) Modelling the effects of macrophytes on algal blooming

in eutrophic shallow lakes. Ecological Modelling 104, 261-287.

Baker PD, Brookes JD, Burch MD, Maier HR, Ganf GG (2000) Advection, growth

and nutrient status of phytoplankton populations in the Lower River Murray, South

Australia. Regulated Rivers: Research and Management 16, 327-344.

Baldry I (2000) Effect of Common Carp (Cyprinus carpio) on Aquatic Restorations.

http://www.hort.agri.umn.edu/h5015/00papers/baldry.htm

Bart J (1995) Acceptance criteria for using individual-based models to make

management decisions. Ecological Applications 5, 411-420.

Bartsch D (1997) Impact of Irrigation Drainage on Sunnyside Wetland: A

Comparative Limnological Study. Honours thesis, The University of Adelaide.

http://www.hort.agri.umn.edu/h5015/00papers/baldry.htm


217

Bavor HJ, Davies CM, Sakadevan K (2001) Stormwater treatment: do constructed

wetlands yield improved pollutant management performance over a detention pond

system? Water Science and Technology 44, 565-570.

Beardall J, Young E, Roberts S (2001) Approaches for determining phytoplankton

nutrient limitation. Aquatic Sciences 63, 44-69.

Beck MB (1997) Applying systems analysis in managing the water environment:

towards a new agenda. Water Science and Technology 36, 1-17.

Bedford BL, Preston EM (1988) Developing the Scientific Basis for Assessing

Cumulative Effects of Wetland Loss and Degradation on Landscape Functions:

Status, Perspectives, and Prospects. Environmental Management 12, 751-771.

Benndorf J, Recknagel F (1982) Problems of application of the ecological model

salmo to lakes and reservoirs having various trophic states. Ecological Modelling 17,

129-145.

Bjornsson KT, Ostendorf B, Recknagel F (2003) Using wetland nutrient modelling to

estimate River Murray and floodplain wetland water exchange. In 'International

Congress on Modelling and Simulation Proceedings'. Townsville, Australia. (Ed. DA

Post) pp. 404-409. (Modelling and Simulation Society of Australia and New Zealand

Inc.)

Blindow I, Anderson G, Hargeby A, Johansson S (1993) Long-term pattern of

alternate stable states in two shallow eutrophic lakes. Freshwater Biology 30, 159-

167.

Boon PI, Bailey PCE (1997) Implications of nutrient enrichment for management of

primary productivity in wetlands. In 'Wetlands in a dry Land: Understanding for

Management'. (Ed. WD Williams). (Environment Australia and Land and Water

Resources Research and Development Corporation)

Boumans R (2001) Personal Communication,

Boumans RM, Villa F, Costanza R, Voinov A, Voinov H, Maxwell T (2001) Non

spatial calibrations of a General Unit Model for Ecosystem simulations. Ecological

Modelling 146, 17-32.


218

Bowden WB (1987) The biogeochemistry of nitrogen in freshwater wetlands.

Biogeochemistry 4, 313-313.

Bowles BA, Powling IJIJ, Burns FL (1979) 'Effects on water quality of artificial

aeration and destratification of Tarago Reservoir / by Barbara A. Bowles, I. Joan

Powling and Frank L. Burns.' (Australian Government Publishing Service: Canberra)

Braskerud BC (2002) Factor affecting phosphorus retention in small constucted

wetlands treating agricultural non-point source pollution. Ecological Engineering 19,

41-61.

Bunn SE, Arthington AH (2002) Basic Principles and Ecological Consequences of

Altered Flow Regimes of Aquatic Biodiversity. Environmental Management 30, 492-

507.

Burbridge PR (1994) Integrated planning and management of freshwater habitats,

including wetlands. Hydrobiologia 285, 311-322.

Burgoon PS (2001) Denitrification in free water surface wetlands receiving carbon

supplements. Water Science and Technology 44, 163-169.

Burton JR (1974) Hydrology of the Basin. In 'The Muray Waters: Man, Nature and a

River System'. (Eds HJ Frith and G Sawer) pp. 45-61. (Angus and Robertson PTY

LTD: Sydney)

Caffrey JM, Kemp WM (1992) Influence of the submerged plant, Potamogeton

perfoliatus, on nitrogen cycling in estuarine sediments. Limnology and Oceanography

37, 1483-1495.

Carpenter SR, Cottingham KL (1997a) Resilience and restoration of lakes.

Conservation Ecology [online] 1, 2.

Carpenter SR, Pace ML (1997b) Dystrophy and eutrophy in lake ecosystems:

implications of fluctuating inputs. Oikos 78, 3-14.

Caswell H (1988) Theory and models in ecology: A different perspective. Ecological

Modelling 43, 33-44.

Cetin L (2001) Contribution to a Generic Wetland Ecosystem Model (WETMOD) of

the Lower River Murray Floodplains. Honours thesis, The University of Adelaide.


219

Cetin L, Recknagel F, Boumans RM (2001) WETMOD: A Generic Wetland

Ecosystem Model for the Simulation of Floodplain Wetlands at the Lower River

Murray (South Australia). In 'International Congress on Modelling and Simulation

Proceedings'. Canberra, Australia. (Eds F Ghassemi, P Whetton, R Little and M

Littleboy) pp. 837-842. (Modelling and Simulation Society of Australia and New

Zealand Inc.)

Chambers PA, Prepas EE (1994) Nutrient dynamics in riverbeds: the impact of

sewage effluent and aquatic macrophytes. Water Research 28, 453-464.

Chen RL, Barko JW (1988) Effects of Freshwater Macrophytes on Sediment

Chemistry. Journal of Freshwater Ecology 4, 279-289.

Close AF (1986) Computer Modelling of the River Murray. In 'Hydrology and Water

Resources Symposium'. Griffith University, Brisbane. (National Conference

Publication No. 86/13)

Close AF (1996) A New Model of Flow and Solute Transport in the River Murray. In

'23rd Hydrology and Water Resources Symposium'. Hobart Australia

Costanza R, d‟Arge R, et al. (1997) The value of the world‟s ecosystem services and

natural capital. Nature 387, 253-260.

Costanza R, Ruth M (1998) Using Dynamic Modeling to Scope Environmental

Problems and Build Consensus. Environmental Management 22, 183-195.

Croucher D (2005) Personal Communication,

Crumpton WG (2001) Using wetlands for water quality improvement in agricultural

watersheds; the importance of a watershed scale approach. Water Science and

Technology 44, 559-564.

Dambacher JM, Li HW, Rossignol PA (2003) Qualitative predictions in model

ecosystems. Ecological Modelling 161, 79-93.

de Wit MJM (2001) Nutrient Fluxes at the river basin scale. I: PolFlow model.

Hydrological Processes 15, 743-759.


220

de Wit MJM, Pebesma EJ (2001) Nutrient fluxes at the river basin scale. II: the

balance between data availability and model complexity. Hydrological Processes 15,

761-775.

DeAngelis DL, Gross LJ, Huston MA, Wolff WF, Fleming DM, Comiskey EJ,

Sylvester SM (1998) Landscape Modeling for Everglades Ecosystem Restoration.

Ecosystems 1, 64-78.

DWLBC (2004) Department of Water Land and Biodiversity Conservation,

Elliott JA, Irish AE, Reynolds CS, Tett P (2000) Modelling freshwater phytoplankton

communities: an exercise in validation. Ecological Modelling 128, 19-26.

Fitz HC, DeBellevue EB, Costanza R, Boumans R, Maxwell T, Wainger L, Sklar FH

(1996) Development of a general ecosystem model for a range of scales and

ecosystems. Ecological Modelling 88, 263-295.

Forgan BW (2001) Australian Solar Radiation Data (CD).

Frears A (2006) Wetland Project Officer, Lower Murray

South Australian Murray-Darling Basin Natural Resources Management Board,

Gibbs MM, White E (1994) Lake Horowhenua: a computer model of its limnology

and restoration prospects. Hydrobiologia 275/276, 467-477.

Goodall DW (1972) Building and testing ecosystem models. In 'Mathematical Models

in Ecology'. (Ed. JNJ Jeffers) pp. 173-194. (Blackwell: Oxford)

Goonan PM, Beer JA, Thompson TB, Suter PJ (1992) Wetlands of the River Murray

flood plain, South Australia. 1. Prelimenary survey of the biota and physio-chemistry

of the ten wetlands from Chowilla to Mannum. Transactions of the Royal Society of

South Australia 116, 81-94.

Graneli W, Solander D (1988) Influence of aquatic macrophytes on phosphosus

cycling in lakes. Hydrobiologia 170, 245-266.

Griffin SL, Herzfeld M, Hamilton DP (2001) Modelling the impact of zooplankton

grazing on phytoplankton biomass during a dionoflagellate bloom in the Swan River

Estuary, Western Australia. Ecological Engineering 16, 373-394.


221

Guerrin F, Dumas J (2001) Knowledge representation and qualitative simulation of

salmon redd functioning. Part I: qualitative modeling and simulation. Biosystems 59,

75-84.

Hamilton DP, Mitchell SF (1996) An empirical model for sediment resuspension in

shallow lakes. Hydrobiologia 317, 209-220.

Harrison J (1994) 'Water Resources Series No 11: Review of Nutrients in Irrigation

Drainage in the Murray-Darling Basin.' CSIRO Australia - Division of Water

Resources.

Hecky RE, Kilham P (1988) Nutrient limitation of phytoplankton in freshwater and

marine environments: a review of recent evidence on the effect of enrichment.

Limnology and Oceanography 33, 796-822.

Hills SE (1974) The Physiographic setting. In 'The Muray Waters: Man, Nature and a

River System'. (Eds HJ Frith and G Sawer) pp. 1-12. (Angus and Robertson PTY

LTD: Sydney)

Holland K, Jolly I, Overton I, Miles M, Vears L, Walker G (2005) 'The Floodplain

Risk Methodology (FRM): A suite of tools to rapidly assess at the regional scale the

impacts of groundwater inflows and benefits of improved inundation on the

floodplains of the lower River Murray.' CSIRO Land and Water, South Australian

Department for Environment and Heritage, CSIRO Land and Water Technical Report

27/05.

Holtan H, Kamp-Nielsen L, Stuanes AO (1988) Phosphorus in soil, water and

sediment: an overview. Hydrobiologia 170, 19-34.

Hulbert L, Wallis J, Peterson S, Richmond B (2000) STELLA Research Software. In.

(HPS High Performance Systems Inc.)

Hunter RG, Combs DL, George DB (2001) Nitrogen, Phosphorus and Organic

Carbon Removal in Simulated Wetland Treatment Systems. Archives of

Environmental Contamination and Toxicology 41, 274-281.

Jacobs T (1990) River Regulation. In 'The Murray'. (Eds N Mackay and D Eastburn)

pp. 39-60. (Murray Darling Basin Commission: Canberra, Australia.)

Jensen A, Paton P, Mowbray T, Simpson D, Kinnear S, Nicols S (1996) 'Wetlands

Atlas of the South Australian Murray Valley: A summary of current knowledge of


222

Murray Valley wetlands as a basis for integrated catchment management.' (South

Australian River Murray Wetlands Management Committee, Department of

Environment and Natural Resources)

Johnson LB, Gage SH (1997) Landscape approaches to the analysis of aquatic

ecosystems. Freshwater Biology 37, 113-132.

Johnston CA (1991) Sediment and Nutrient Retention by Freshwater Wetlands:

Effects on Surface Water Quality. Critical Reviews in Environmental Control 21, 491-

565.

Johnston CA, Detenbeck NE, Niemi GJ (1990) The cumulative effect of wetlands on

stream water quality and quantity. Biogeochemistry 10, 105-105.

Kadlec RH (1997) An autobiotic wetland phosphorus model. Ecological Engineering

8, 145-172.

Kadlec RH, Reddy KR (2001) Temperature effects in treatment wetlands. Water

Environment Research: a Research Publication Of The Water Environment

Federation 73, 543-557.

Keenan LW, Lowe EF (2001) Determining ecologically acceptable nutrient loads to

natural wetlands for water quality improvement. Water Science and Technology 44,

289-294.

Kirk J (1998) 'GIS and Irrigation: An Inventory of Projects in the Murray-Darling

Basin.' (Murray-Darling Basin Commission)

Kobayashi T, Gibbs P, Dixon PI, Shiel RJ (1996) Grazing by a River Zooplankton

Community: Importance of Microzooplankton. Marine & Freshwater Research 47,

1025-1036.

Lemly AD (1997) Risk Assessment as an Environmental Management Tool:

Considerations for Freshwater Wetlands. Environmental Management 21, 343-358.

Lennox LJ (1984) Lough Ennell: laboratory studies on sediment phosphorus release

under varying mixing, aerobic and anaerobic conditions. Freshwater Biology 14, 183-

187.


223

Li X, Jongman RH, Xiao D, Bert HW, Bregt AK (2002) The effect of spatial pattern

on nutrient removal of a wetland landscape. Landscape and Urban Planning 60, 27-

41.

Li X, Xiao D, Jongman RH, Bert HW, Bregt AK (2003) Spatial modeling on the

nutrient retention of an esutary wetland. Ecological Modelling 167, 33-46.

Lijklema L (1994) Nutrient dynamics in shallow lakes: effects of changes in loading

and role of sediment-water interactions. Hydrobiologia 275/276, 335-348.

Lüderitz V, Gerlach F (2002) Phosphorus Removal in Different Constructed

Wetlands. Acta Biotechnology 22, 91-99.

Ludwig DB, Walker B, Holling CS (1997) Sustainability, stability, and resilience.

Conservation Ecology [online] 1, 7.

Mackay N, Eastburn D (1990) 'The Murray.' (Murray Darling Basin Commission:

Canberra, Australia)

Marsden MW (1989) Lake restoration by reducing external phosphorus loading: the

influence of sediemnt phosphorus release. Freshwater Biology 21, 139-162.

Marsh F (1997) A comparative study of the impacts of carp on phytoplankton and

water quality in two Lower River Murray wetlands. Honours thesis, The University of

Adelaide.

Maxwell T, Costanza R (1997) An Open Geographic Modeling Environment.

Simulation Journal 68, 175-185.

Mayer DG, Butler DG (1993) Statistical validation. Ecological Modelling 68, 21-32.

McComb A, Qiu S (1997) The effects of drying and reflooding on nutrient release

from wetland sediments. In 'Wetlands in a Dry Land: Understanding for

Management'. (Ed. WD Williams). (Environment Australia, Biodiversity Group:

Canberra)

McIntosh BS (2003) Qualitative modelling with imprecise ecological knowledge: a

framework for simulation. Environmental Modelling & Software 18, 295-307.


224

McIntosh BS, Muetzelfeldt RI, Legg CJ, Mazzoleni S, Csontos P (2003) Reasoning

with direction and rate of change in vegetation state transition modelling.

Environmental Modelling & Software 18, 915-927.

McIntyre NR, Wagener T, Wheater HS, Yu ZS (2003) Uncertainty and risk in water

quality modelling and management. Journal of Hydroinformatics 5, 259-274.

McIntyre NR, Wheater HS A tool for risk-based management of surface water

quality. Environmental Modelling & Software In Press, Corrected Proof.

Meijer ML, Jeppesen E, et al. (1994) Long-term responses to fish-stock reduction in

small shallow lakes: interpretation of five-year results of four biomanipulation cases

in The Netherlands and Denmark. Hydrobiologia 275/276, 457-466.

Mitasova H, Mitas L (1998) Process Modeling and Simulations, NCGIA Core

Curriculum in GIScience.

Mitchell SF (1989) Primary production in a shallow eutrophic lake dominated

alternately by phytoplankton and by submerged macrophytes. Aquatic Botany 33,

101-110.

Mitsch WJ, Gosselink JG (2000) 'Wetlands.' (John Wiley & Sons, Inc.: New York,

USA)

Morris JT (1991) Effects of Nitrogen Loading on Wetland Ecosystems with Particular

Reference to Atmospheric Deposition. Annual Reviews of Ecology and Systematics

22, 257-279.

Moss B (1990) Engineering and biological approaches to the restoration from

eutriphication of shallow lakes in which aquatic plant communities are important

components. Hydrobiologia 200/201, 367-377.

Muhammetoglu AB, Soyupak S (1997) A tree-dimensional water quality-macrophyte

interaction model for shallow lakes. Ecological Modelling 133, 161-180.

Murdoch B, Greenwood A, Cresswell D (2004) 'SIWM Model Evaluation Progress

Report to July 2004.' South Australian Department of Water, Land and Biodiversity

Conservation Report DWLBC 2005/08, Report DWLBC 2005/08.


225

Murray AG (2001) The use of simple models in the design and calibration of a

dynamic 2D model of a semi-enclosed Australian bay. Ecological Modelling 136, 15-

30.

Naiman RJ, Bunn SE, Nilsson C, Petts GE, Pinay G, Thompson LC (2002)

Legitimising Fluvial Ecosystems as Users of Water: An Overview. Environmental

Management 30, 455-467.

Nichols S (1998) 'Wetlands Management Study report Wetland Care Australia.'

Wetland Care Australia.

Nielsen DL, Chick AJ (1997) Flood-mediated changes in aquatic macrophyte

community structure. Marine & Freshwater Research 48, 153-157.

Nürnberg GK (1984) The prediction of internal phosphorus load in lakes with anoxic

hypolimnia. Limnology and Oceanography 29, 111-124.

Nürnberg GK (1998) Prediction of Phosphorus Release Rates from Total and

Reductant- Soluble Phosphorus in Anoxic Lake Sediments. Canadian Journal of

Fisheries and Aquatic Sciences 45, 453-461.

Olila OG, Reddy KR, Harris WG (1995) Forms and distribution of inorganic

phosphorus in sediments of two shallow eutrophic lakes in Florida. Hydrobiologia

302, 147-161.

Oliver RL (1993) 'Phosphorus Dynamics and Bioavailability in Turbid Waters.' Land

and Water Resources Research and Development Corporation, Canberra.

Oreskes N, Shrader-Frechette K, Berlitz K (1994) Verification, validation, and

confirmation of numerical models in the earth sciences. Science 263, 641-646.

Overton I (2000) Personal Communication,

Overton I (2005) Modelling floodplain inundation on a regulated river: integrating

GIS, remote sensing and hydrological models. River Research and Applications 21,

991 - 1001.

Parsons S, Dohnal M (1995) The qualitative and semiqualitative analysis of

environmental problems. Environmental Software 10, 75-85.


226

Peijl MJvd, Oorschot MMPv, Verhoeven JTA (2000a) Simulation of the effects of

nutrient enrichment on nutrient and carbon dynamics in a river marginal wetland.

Ecological Modelling 134, 169-184.

Peijl MJvd, Verhoeven JTA (1999) A model of carbon, nitrogen and phosphorus

dynamics and their interactions in river marginal wetlands. Ecological Modelling 118,

95-130.

Peijl MJvd, Verhoeven JTA (2000b) Carbon, nitrogen and phosphorus cycling in river

marginal wetlands; a model examination of landscape geochemical flows.

Biogeochemistry 50, 45-71.

Philcox MB (1997) 'Overview if the Lower River Murray Reclaimed Swamps

Irrigation Areas.' Primary Industries, South Australia.

Pimm SL (1997) The value of everything. Nature 387, 231-232.

Poff NL, Allan JD, Bain MB, Karr JR, Prestegaard KL, Richter BD, Sparks RE,

Stromberg JC (1997) The natural flow regime: a paradigm for river conservation and

restoration. BioScience 47, 769-784.

Power M (1993) The predictive validation of ecological and environmental models.


Pressey B (1990) Wetlands. In 'The Murray'. (Eds N Mackay and D Eastburn) pp.

167-182. (Murray Darling Basin Commission: Canberra, Australia.)

Pressey RL (1987) 'Murray Valley Management Review: Background paper number

5. The Murray Wetlands in SA: Management considerations and Research needs.'

South Australia.

Preston EM, Bedford BL (1988) Evaluating Cumulative Effects on Wetland

Functions: A Conceptual Overview and Generic Framework. Environmental

Management 12, 565-583.

Reckhow KH (1994) Water quality simulation modeling and uncertainty analysis for

risk assessment and decision making. Ecological Modelling 72, 1-20.

Recknagel F (nd) Personal Communication,


227

Recknagel F, Bartsch D, Marsh F, Schiller N (nd) Degradation of Lower Murray

Floodplain Wetlands by Common Carp and Irrigation Drainage: Impacts on the River

Murray and Implications for Restoration.

Recknagel F, Benndorf J (1982) Validation of the Ecological Simulation Model

"SALMO". Hydrobiologia 67, 113-125.

Recknagel F, Hosomi M, Fukushima T, Kong D-S (1995) Short- and Long-Term

control of external and internal phosphorus loads in Lakes - A scenario analysis.

Water Research 29, 1767-1779.

Recknagel F, Marsh F, Matthews S, Schiller N (1997) Common carp in natural

wetlands: impacts and management. In 'Wetlands in a Dry Land: Understanding for

Management'. (Environment Australia and Land and Water Resources Research and

Development Corporation)

Recknagel F, van der Wielen M (2000) Ecological Restoration of Shallow Eutrophic

Lakes by Drawdown: Concept and Implications for Lake Dianchi (China). In

'Proceedings of the International Conference on Eutrophication in China. Kunming'

Reddy KR, Diaz OA, Scinto LJ, Agami M (1995) Phosphorus dynamics in selected

wetlands and streams of the lake Okeechobee Basin. Ecological Engineering 5, 183-

207.

Reddy KR, Patrick JWK, Lindau CW (1989) Nitrification-denitrification at the plant

root-sediment interface in wetlands. Limnology & Oceanography 34, 1004-1013.

Rengasany P, Olsson KA (1991) Sodicity and soil structure. Australian Journal of

Soil Research 29, 935 - 952.

RMCWMB (2002) 'Water Allocation Plan for the River Murray Prescribed

Watercourse (as amended 12th January, 2004).' River Murray Catchment Water

Management Board, Government of South Australia, Berri, South Australia.

Roberts J, Ganf GG (1986) Annual production of Typha orientalis Presl. in inland

Australia. Australian Journal of Marine and Freshwater Research 37, 659-668.

Roberts J, Ludwig JA (1991) Riparian vegetation along current-exposure gradients in

flooplain wetlands of the River Murray, Australia. Journal of Ecology 79, 117-127.


228

Rousseau DPL, Vanrollenghem P, A., De Pauw N (2004) Constructed wetlands in

Flanders: a performance analysis. Ecological Engineering 23, 151-163.

Rykiel J, Edward J. (1996) Testing ecological models: the meaning of validation.


Sand-Jensen K (1998) Influence of submerged macrophytes on sediment composition

and near-bed flow in lowland streams. Freshwater Biology 39, 663-679.

Sand-Jensen K, Madsen TV (1992) Patch dynamics of the stream macrophyte,

Callitriche cophocarpa. Freshwater Biology 27, 277-282.

Sand-Jensen K, Mebus TJR (1996) Fine -scale patterns of water velocity within

macrophyte patches in streams. Oikos 76, 169-180.

Sand-Jensen K, Moller J, Olsen BH (1988) Biomass regulation of microbenthic algae

in a Danish lowland stream. Oikos 53, 332-340.

Scheffer M (1998) 'Ecology of Shallow Lakes.' (Chapman & Hall: London)

Scheffer M, Hosper SH, Meijer M-L, Moss B, Jeppesen E (1993) Alternative

equilibria in shallow lakes. Trends in Ecology & Evolution 8, 275-279.

Scheffer M, Rinaldi S (2000) Minimal models of top-down control of phytoplankton.

Freshwater Biology 45, 265-283.

Schindler DW (1977) Evolution of Phosphorus Limitation in Lakes. Science 195, 260-

262.

Scholz O, Gawne B, Ebner B, Ellis I (2002) The effect of drying and re-flooding on

nutrient availability in ephemeral deflation basin lakes in western New South Wales,

Australia. River Research and Applications 18, 185-196.

Schulz C, Gelbrecht J, Rennert B (2004) Constructed wetlands with free water surface

for treatment of aquaculture effluents. Journal of Applied Ichthyology 20, 64-70.

Schwoerbel J (1993) 'Einfuehrung in die Limnologie.' (UTB Fuer Wissenschaft;

Gustav Fischer Verlag Jena und Stuttgart)


229

Shafron M, Croome R, Rolls J (1990) Water Quality. In 'The Murray'. (Eds N Mackay

and D Eastburn) pp. 147-166. (Murray Darling Basin Commission: Canberra,

Australia.)

Shiel RJ, Walker KF, Williams WD (1982) Plankton of the Lower River Murray,

South Australia. Australian Journal of Marine and Freshwater Research 33, 301-327.

Siebentritt MA (2003) The Influence Of Water Regime On The Floristic Composition

Of Lower River Murray Wetlands. Ph.D. thesis, University of Adelaide.

Siebentritt MA, Ganf GG, Walker KF (2004) Effects of an enhanced flood on riparian

plants of the River Murray, South Australia. River Research and Applications 20,

765-774.

Simpson D (2003) Water Loss Calculator.

Sklar FH, Costanza R, Maxwell T (1993) State-of-the-Art Spatial Modeling for

Wetland Management. In ' Barataria-Terrebonne National Estuary Program,

Scientific-Technical Committee Data Inventory Workshop Proceedings'. Thibodaux,

LA 70310 pp. 354-366. (BTNEP)

SKM (2004) 'River Murray Wetlands Baseline Survey.' South Australian Murray

Darling Basin Natural Resources Management Board.

SKM (2006) 'River Murray Wetlands Baseline Survey - 2005.' South Australian

Murray Darling Basin Natural Resources Management Board.

Soendergaard M, Kristiansen P, Jeppesen E (1992) Phosphorus release from

resuspended sediment in the shallow and wind exposed Lake Arresoe, Denmark.

Hydrobiologia 228, 91-99.

Stephen D, Moss B, Phillips G (1998) The relative importance of top-down and

bottom-up control of phytoplankton in a shallow macrophyte-dominated lake.

Freshwater Biology 39, 699-713.

Strager JM, Yuill CB, Wood PB (2000) Landscape-based Riparian Habitat Modeling

for Amphibians and Reptiles using ARC/INFO GRID and ArcView GIS. In 'ESRI

International User Conference 2000'

Sullivan TJ (1997) Ecosystem Manipulation Experimentation as a Means of Testing a

Biogeochemical Model. Environmental Management 21, 15-21.


230

Thomann RV (1998) The Future "Golden Age" of Predictive Models for Surface

Water Quality and Ecosystem Management. Journal of Environmental Engineering

124, 94-103.

Timms RM, Moss B (1984) Prevention of growth of potentially dense phytoplankton

populations by zooplankton grazing, in the presence of zooplanktonivotus fish, in a

shallow wetland ecosystem. Limnology and Oceanography 29, 472-486.

Tsang C (1991) The Modelling Process and Model Validation. Ground Water 29,

825-831.

van der Molen DT, Boers PCM (1994) Influence of internal loading on phosphorus

concentration in shallow lakes before and after reduction of the external loading.

Hydrobiologia 275-276, 379-389.

van der Wielen M (2001) Drying cycles as a switch between alternative stable states

in wetlands. In 'SIL 2001 conference'. Melbourne

van der Wielen M (nd) Personal Communication,

Voinov A, Costanza R (1999a) Watershed management and the Web. Journal of

Environmental Management 56, 231-245.

Voinov A, Costanza R, Wainger L, Boumans R, Villa F, Maxwell T, Voinov H

(1999b) Pautuxent landscape model: integrated ecological economic modeling of a

watershed. Environmental Modelling and Software 14, 473-491.

Walker KF (1979) Regulated Streams in Australia: The Murray-Darling River

System. In 'The Ecology of Regulated Streams'. (Eds JV Ward and JA Stanford) pp.

143-162. (Plenum Press: New York)

Walker KF (1985) A review of the ecological effects of river regulation in Australia.

Hydrobiologia 125, 111-129.

Walker KF, Boulton AJ, Thoms MC, Sheldon F (1994) Effects of Water-level

Changes Induced by Weirs on the Distribution of Littoral Plants along the River

Murray, South Australia. Australian Journal of Marine and Freshwater Research 45,

1421-1438.


231

Walker KF, Hillman TJ (1982) Phosphorus and Nitrogen Loads in Waters Associated

with the River Murray near Albury-Wodonga, and their Effects on Phytoplankton

Populations. Australian Journal of Marine and Freshwater Research 33, 223-243.

Walker KF, Thoms MC (1993) Environmental Effects of Flow Regulation on the

Lower River Murray, Australia. Regulated Rivers: Research and Management 8, 103-

119.

Wallach D, Genard M (1998) Effect of uncertainty in input and parameter values on

model prediction error. Ecological Modelling 105, 337-345.

Webster IT, Maier H, Baker P, Burch M (1997) Influence of wind on water levels and

lagoon-river exchange in the River Murray, Australia. Marine & Freshwater

Research 48, 541-550.

Wen L (2002a) Mechanisms for Phosphorus Elimination in Constructed Wetlands: A

Pilot Study for the Treatment of Agricultural Drainage Water from Dairy Farms at the

Lower River Murray, South Australia. PhD thesis, The University of Adelaide.

Wen L (2002b) Personal Communication,

Wen L, Recknagel F (2002) In situ removal of dissolved phosphorus in irrigation

drainage water by planted floats: preliminary results from growth chamber

experiment. Agriculture, Ecosystems and Environment 90, 9-15.

Weymouth G, Le Marshall J (1994) An operational system to estimate insolation over

the Australian region. In 'Proc. Pacific Ocean Remote Sensing Conference'.

Melbourne, Australia 1-4 Mar 1994 pp. 443-449. (Bureau of Meteorology Research

Centre,)

Wood SN (2001) Partially Specified Ecological Models. Ecological Monographs 71,

1-25.

Young WJ, Lam DCL, V. R, Wong IW (2000) Development of an environmental

flows decision support system. Environmental Modelling & Software 15, 257-265.


232

Glossary

Terminology

Wetland categories The division of wetlands into very simplified hydrological

connectivity classification, i.e. wetlands of similar type

“Exemplar” The monitored data of a wetland of a given category

Category wetlands wetland with no driving variable data within a give wetland

category for which “exemplar” data will be used as driving

variables, i.e. wetlands of a particular category

GIS Geographical Information System

DEM Digital Elevation Model

SME Spatial Modelling Environment – A GIS based modelling

environment

Simulation Running the model based on a management scenario

Scenario Hypothetical management situation which is modelled by

WETMOD 2 at a simulation run. One run of the model

Development Construction of the model including adapting WETMOD 1,

spatial data, wetland monitored data and river data followed by

calibration and validation of the model.

Calibration Fitting the model output to monitored data and adjusting

parameters such as thresholds

Validation Testing the model with data not used during the model

development to determine the degree of agreement between a

model and the real system.

State variables Model output (Phosphorus as PO4-P, nitrogen as NO3-N,

macrophytes, phytoplankton and zooplankton)

Driving variables Model time-series input (water temperature, turbidity, Secchi

depth and solar radiation)

Calibration Set parameters adjusted within the model to fit the model to

monitored data (e.g. turbidity sedimentation threshold,

zooplankton mortality rate, maximum phytoplankton growth

rate)

Retention Nutrient retain within a wetland

Uptake The reduction of nutrient load in a wetland through

phytoplankton and macrophyte growth ≈ Retention

Load Amount of suspended nutrient in the wetland, irrigation

drainage or river (resulting in inflow load to the wetland).

Directly related to the concentration simulated.

NTU Nephelometric Turbidity Units


233

Organisations

MDBC Murray Darling Basin Commission

BOM Bureau of Meteorology

DWLBC South Australian Department of Water, Land and Biodiversity

Conservation

DEH South Australian Department for Environment and Heritage

Equations

t

N R[mg/day] Nutrient Retention

τ [1/day] Turnover rate

D Average linear deviation from the measured values as a fraction

of the average observed values

ID Total Irrigation Drainage load

IC Concentration of irrigation drainage nutrient

I Irrigation Drainage flow in litres/day

∆ID Change in total Irrigation Drainage load after management

RF Total River Inflow load

%RO Percentage Reduction in Outflow

OF Total Outflow load

∆OF Change in total Outflow load post management.

CR and CW Concentrations of nutrients in the river

CR and CW Concentrations of nutrients in the wetland

R River flow rate

ƒ Represents a fraction of the river flow rate R.

%RO: Change in outflow due to management when compared to the

status quo (no management).

%RI: Effective change in wetland nutrient inflow due to nutrient

reduction scenario as compared with the status quo.

RL Initial river nutrient load

RN Change in wetland retention due to management

%RL Percentage River Load removed due to the wetland

management

Appendix

234

Appendix A: WETMOD differential equations

The initial concentrations for each wetland category are fixed as in Table 17.

Table 17: Initial values

Category Macrophyte

(MAC_BIOMASS),

Phytoplankton, Zooplankton, PO4-P, NO3-N

1 5 0.0001 1.2 0.00011 0.0003

2 15 0.0001 0.001, 0.00275 0.0004

3 5 0.0001 1.2 0.000133 0.00011

4 0.1 7.04 1.2 0.00026 0

5 0.1 2.51 1.2 0.000109

5

0.00026

6 15 Look at Data 1.2 Look at

Data

Look at

Data

The descriptions of the Macrophyte, Phytoplankton and Nutrient sectors were adapted

from (Cetin 2001).

Appendix

235

$Macrophytes

Equations Source

MAC_BIOMASS(t) = MAC_BIOMASS(t - dt) +

(Mac_Gross_PP - Mac_mortality - Mac_respiration) * dt

INFLOWS:

Mac_Gross_PP = if Turbidity<TurbGrowthLimiting then

Mac_GPP*mac_prod_cf*MAC_BIOMASS else 0.001

(Boumans 2001)

OUTFLOWS:

Mac_mortality = Mac_mort_rate*MAC_BIOMASS (Asaeda et al.

1997)

Mac_respiration = Mac_resp_rate*MAC_BIOMASS (Asaeda et al.

1997)

mac_net_prod = Mac_Gross_PP-Mac_respiration

mac_nut_cf =

(NO3N/(NO3N+mac_Ks_N))*(PO4P/(PO4P+mac_Ks_P))

Jorgensen 1986

mac_prod_cf = underwater_light_cf*mac_temp_cf*mac_nut_cf (Boumans 2001)

mac_temp_cf = EXP(0.2*(water_temp-mac_temp_opt))*((40-

water_temp)/(40-mac_temp_opt))^(0.2*(40-mac_temp_opt))

(Boumans 2001)

reflection = 0.9*(SolarRadiationInCalculation*100) (Recknagel et al.

1982)

surface_light = 0.5*reflection (Recknagel et al.

1982)

Turbidity2Secchi = IF (2.4355*(Turbidity)^-0.5675) =0 Then

0.000001 Else (2.4355*(Turbidity)^-0.5675)

underwater_light_cf = surface_light*EXP(-

(4.6/Zeu_Calculated)*1)

(Recknagel et al.

1982)

Zeu_Calculated = IF(Manual_Secchi_Overide=0)

THEN(1.7*(Manual_Secchi_Overide+0.001))

ELSE(1.7*Manual_Secchi_Overide)

(Recknagel et al.

1982)

Parameters Units Source

Mac_GPP = 0.005 kg/m3/d (Boumans 2001)

mac_Ks_N = 0.0001 kg/m3 Calibrated

mac_Ks_P = 0.00005 kg/m3 Calibrated

Mac_mort_rate = 0.01 kg/m3/d (Asaeda et al. 1997)

Mac_resp_rate = 0.018 cm3/m

3/d (Asaeda et al. 1997)

TurbGrowthLimiting = 70 NTU Calibrated

Appendix

236

Model terms Definition

MAC BIOMASS The biomass of the photosynthetic portion of the

macrophytes.

Mac GPP The gross primary production rate for the total plant

biomass.

Mac Gross PP The gross primary productivity of the photosynthetic

biomass.

Mac Ks N The half-saturation constant for the uptake of nitrates by

macrophytes.

Mac Ks P The half-saturation constant for the uptake of phosphate by

macrophytes.

Mac mort rate Mortality rate for the photosynthetic biomass

Mac mortality The mortality of the photosynthetic biomass.

Mac net prod The net primary productivity for total macrophyte biomass.

Mac nut cf The macrophyte nutrient coefficient.

Mac prod cf The macrophyte production coefficient.

Mac resp rate Respiration rate of photosynthetic biomass

Mac respiration The respiration of photosynthetic biomass.

Mac temp cf Macrophyte temperature coefficient

Mac temp opt The optimum temperature for macrophyte growth

Manual Secchi

override

Switch between sources of Secchi depth.

Manual vs Monitored

Secchi

Switch between sources of Secchi depth.

Reflection Determines the proportion incoming solar radiation reflected

from the water surface.

Secchi Selection of calculated or measured Secchi depth

Site assumed Secchi

Manual

Manual input of Secchi depth (fixed)

Surface Light Defines the proportion of light entering the surface water.

Turbidity2Secchi The calculation of the Secchi depth based on turbidity (see

Methodology)

Underwater light cf The underwater light coefficient.

Zeu Calculated Defines euphotic zone at 1 metre depth.

Appendix

237

$Phytoplankton

Equations Source

PHYTOPLANKTON(t) = PHYTOPLANKTON(t - dt) +

(pht_Gross_PP + Phytoplankton_In - Pht_grazing -

pht_respiration - pht_mortality - pht_sedimentation -

Phytoplankton_Out) * dt

INFLOWS:

pht_Gross_PP = if PHYTOPLANKTON>pht_max or

Turbidity>TurbGrowthLimiting then pht_max else

pht_prod_cf*pht_GPP*PHYTOPLANKTON

(Boumans 2001)

Phytoplankton_In = PhytoplanktonInflow_cm3m3

OUTFLOWS:

Pht_grazing = PHYTOPLANKTON*(zoo_growth_rate-

Zoo_resp_rate)

(Recknagel et al.

1982)

pht_respiration =

pht_resp_rate*pht_temp_cf*PHYTOPLANKTON

pht_mortality = pht_mort_rate*PHYTOPLANKTON (Asaeda et al.

1997)

pht_sedimentation = pht_sed*PHYTOPLANKTON (Recknagel et al.

1982)

Phytoplankton_Out = PhytoplanktonOutflow_cm3m3

pht_max = IF Cat_Cal_Used=6 THEN (pht_max_6) ELSE IF

Cat_Cal_Used = 5 THEN (pht_max_5) ELSE

((IF(Cat_Cal_Used = 1) THEN(pht_max_1) ELSE

((IF(Cat_Cal_Used = 2 ) THEN (pht_max_2) ELSE

((IF(Cat_Cal_Used =3) THEN (pht_max_3) ELSE

((IF(Cat_Cal_Used = 4) THEN (pht_max_4) ELSE 2))))))))

(Recknagel et al.

1982)

pht_net_prod = pht_Gross_PP-pht_respiration

pht_nut_cf =

(NO3N/(NO3N+pht_Ks_N))*(PO4P/(PO4P+pht_Ks_P))

Jorgensen 1986

pht_prod_cf = underwater_light_cf*pht_temp_cf*pht_nut_cf (Boumans 2001)

pht_sed = IF Cat_Cal_Used=6 THEN (pht_sed_6) ELSE IF

Cat_Cal_Used = 5 THEN (pht_sed_5) ELSE

((IF(Cat_Cal_Used = 1) THEN(pht_sed_1) ELSE

((IF(Cat_Cal_Used = 2 ) THEN (pht_sed_2) ELSE

((IF(Cat_Cal_Used =3) THEN (pht_sed_3) ELSE

((IF(Cat_Cal_Used = 4) THEN (pht_sed_4) ELSE 0.2))))))))

(Recknagel et al.

1982)

pht_temp_cf = 1.08^(water_temp-20) Hamilton and

Schladow 1997

Appendix

238

Equations Source

ZOOPLANKTON(t) = ZOOPLANKTON(t - dt) + (Pht_grazing

- Zoo_mortality) * dt

INFLOWS:

Pht_grazing = PHYTOPLANKTON*(zoo_growth_rate-

Zoo_resp_rate)

(Recknagel et al.

1982)

OUTFLOWS:

Zoo_mortality =

ZOOPLANKTON*zoo_mort_rate*(1.05^(water_temp-20))

(Recknagel et al.

1982)

dark_grazing = grazing_temp_cf*zoo_grazing_cf (Recknagel et al.

1982)

day_length = 12-7*COS(Time_period) (Recknagel et al.

1982)

grazing_temp_cf = IF(water_temp=0) THEN(1.05*EXP(-

2*ABS(LOGN((water_temp+0.001)/20))+0.26))

ELSE(1.05*EXP(-2*ABS(LOGN(water_temp/20))+0.26))

(Recknagel et al.

1982)

pht_grazing_rate = dark_grazing*(24-

day_length)/24+0.8*dark_grazing*day_length/24

(Recknagel et al.

1982)

pht_Ks_grazing = If PHYTOPLANKTON>0 THEN

4*0.4*PHYTOPLANKTON^1.5 Else

4*0.4*(PHYTOPLANKTON+0.00001)^1.5

(Recknagel et al.

1982)

zoo_grazing_cf = if ZOOPLANKTON>0 then

PHYTOPLANKTON

*pht_pref/ZOOPLANKTON/(5/pht_Ks_grazing

+PHYTOPLANKTON*pht_pref/pht_Ks_grazing

+5/ZOOPLANKTON+PHYTOPLANKTON

*pht_pref/ZOOPLANKTON) else 0.001

(Recknagel et al.

1982)

zoo_growth_rate = if MAC_BIOMASS>10 then ((0.8-

0.4/1.3)*pht_grazing_rate) else 0.05

(Recknagel et al.

1982)

zoo_mort_rate = IF Cat_Cal_Used=6 THEN (ZooMortRate_6)

ELSE IF Cat_Cal_Used = 5 THEN (ZooMortRate_5) ELSE

((IF(Cat_Cal_Used = 1) THEN(ZooMortRate_1) ELSE

((IF(Cat_Cal_Used = 2 ) THEN (ZooMortRate_2) ELSE

((IF(Cat_Cal_Used =3) THEN (ZooMortRate_3) ELSE

((IF(Cat_Cal_Used = 4) THEN (ZooMortRate_4) ELSE

0.3))))))))

(Recknagel et al.

1982)

Zoo_resp_rate = (((0.22-0.08/1.3)*pht_grazing_rate)*0.36)

*(0.17*(water_temp/20)^2+0.05)

(Recknagel et al.

1982)

Appendix

239


pht_GPP = 1.8 cm3/m

3/d (Boumans 2001)

pht_Ks_N = 0.00001 kg/m3 Hamilton and Schladow

1997

pht_Ks_P = 0.00001 kg/m3 Hamilton and Schladow

1997

pht_max_1 = 0.1

pht_max_2 = 0.1

pht_max_3 = 1

pht_max_4 = 1

pht_max_5 = 0.5

pht_max_6 = 2

Calibrated

pht_mort_rate = 0.019 cm3/m


pht_pref = 2.5 dimless (Recknagel et al. 1982)

pht_resp_rate = 0.047 cm3/m


pht_sed_1 = if Turbidity >TurbSed_pht

then 0.1 else 0.01


then 0.05 else 0.01


then 0.05 else 0.01


then 0.5 else 0.2


then 0.5 else 0.2


then 0.5 else 0.2

Fraction of

biomass

(Where 1 is

100%)

Calibrated

TurbSed_pht = 95 NTU Calibrated

ZooMortRate_1 = 0.2

ZooMortRate_2 = 0.2

ZooMortRate_3 = 0.5

ZooMortRate_4 = 0.2

ZooMortRate_5 = 0.6

ZooMortRate_6 = 0.3

cm3/m

3/d Calibrated

Appendix

240


PHYTOPLANKTON The biomass of phytoplankton (defines Chl-a concentration

in terms of biomass).

Dark grazing Defies the grazing rate of zooplankton during night-time

feeding on phytoplankton.

Day length Defines the length of the day.

Grazing temp cf Temperature coefficient for grazing.

Pht GPP The phytoplankton gross primary production rate.

Pht grazing The grazing of phytoplankton by zooplankton.

Pht grazing rate Determines the grazing rate dependent on the time of day.

Pht Gross PP The phytoplankton gross primary productivity.

Pht Ks grazing The half-saturation constant for zooplankton grazing on

phytoplankton.

Pht Ks N The half-saturation constant for the uptake of nitrates by

phytoplankton.

Pht Ks P The half-saturation constant for the uptake of phosphate by

phytoplankton.

Pht max The maximum possible biomass of phytoplankton, i.e. the

carrying capacity.

Pht mort rate The phytoplankton mortality rate.

Pht mortality The phytoplankton mortality.

Pht net prod The phytoplankton net primary productivity.

Pht nut cf The phytoplankton nutrient coefficient.

Pht pref The zooplankton preference factor for phytoplankton

grazing.

Pht prod cf The phytoplankton production coefficient.

Pht resp rate The phytoplankton respiration rate.

Pht respiration The phytoplankton respiration.

Pht sed The sedimentation rate of phytoplankton, which is

dependent on turbidity.

Pht sedimentation The sedimentation of phytoplankton.

Pht temp cf The phytoplankton temperature coefficient.

Phytoplankton in The inflow of phytoplankton into the wetland.

Phytoplankton out The inflow of phytoplankton into the river.

PhytoplanktonInflow

cm3m3

The phytoplankton inflow concentration in cm3/m3

PhytoplanktonOutflow

cm3m3

The phytoplankton outflow concentration in cm3/m3

Appendix

241

ZOOPLANKTON The biomass of zooplankton.

Zoo grazing cf The grazing coefficient of zooplankton, which changes with

the phytoplankton biomass.

Zoo growth rate The growth rate of zooplankton.

Zoo mort rate The mortality rate for zooplankton.

Zoo mortality The zooplankton mortality.

Zoo resp rate The respiration rate of zooplankton.

Appendix

242

$Nutrients

Equations Source

PO4P(t) = PO4P(t - dt) + (P_loading + P_sed_release +

P_IN_gL - P_uptake - P_soil_coprecip - P_OUT) * dt

INFLOWS:

P_loading =

(P_from_land+P_loading_rate)/Wetlandvolume_Liters

Jorgensen 1986

P_sed_release = Turbidity/900*P_from_land (Recknagel et al.

1982)

P_IN_gL = PInflowAmount_mgL/1000

OUTFLOWS:

P_uptake =

PO4P*((pht_net_prod*pht_PC)+(mac_net_prod*Mac_PC))

(Boumans 2001)

P_soil_coprecip = P_sed*PO4P (Recknagel et al.

1982)

P_OUT = POutflow_Amount_gL

P_sed = IF Cat_Cal_Used=6 THEN (P_sed_6) ELSE IF

Cat_Cal_Used = 5 THEN (P_sed_5) ELSE ((IF(Cat_Cal_Used

= 1) THEN(P_sed_1) ELSE ((IF(Cat_Cal_Used = 2 ) THEN

(P_sed_2) ELSE ((IF(Cat_Cal_Used =3) THEN (P_sed_3)

ELSE ((IF(Cat_Cal_Used = 4) THEN (P_sed_4) ELSE

0.05))))))))

(Recknagel et al.

1982)

pht_PC = IF Cat_Cal_Used=6 THEN (pht_PC_6) ELSE IF

Cat_Cal_Used = 5 THEN (pht_PC_5) ELSE

((IF(Cat_Cal_Used = 1) THEN(pht_PC_1) ELSE

((IF(Cat_Cal_Used = 2 ) THEN (pht_PC_2) ELSE

((IF(Cat_Cal_Used =3) THEN (pht_PC_3) ELSE

((IF(Cat_Cal_Used = 4) THEN (pht_PC_4) ELSE 0.05))))))))

(Boumans 2001)

Equations Source

NO3N(t) = NO3N(t - dt) + (N_loading + N_sed_release +

N_IN_gL - N_uptake - N_soil_coprecip - N_OUT -

Denitrification) * dt

INFLOWS:

N_loading =

(N_from_land+N_loading_rate)/Wetlandvolume_Liters

Jorgensen 1986

N_sed_release = Turbidity/2500*N_from_land (Recknagel et al.

1982)

N_IN_gL = NInflowAmount_mgL/1000

Appendix

243

OUTFLOWS:

N_uptake =

NO3N*((pht_net_prod*pht_NC)+(mac_net_prod*Mac_NC))

(Boumans 2001)

N_soil_coprecip = N_sed*NO3N (Recknagel et al.

1982)

N_OUT = NOutflow_Amount_gL

N_sed = IF Cat_Cal_Used=6 THEN (N_sed_6) ELSE IF

Cat_Cal_Used = 5 THEN (N_sed_5) ELSE ((IF(Cat_Cal_Used

= 1) THEN(N_sed_1) ELSE ((IF(Cat_Cal_Used = 2 ) THEN

(N_sed_2) ELSE ((IF(Cat_Cal_Used =3) THEN (N_sed_3)

ELSE ((IF(Cat_Cal_Used = 4) THEN (N_sed_4) ELSE

0.1))))))))

(Recknagel et al.

1982)

pht_NC = IF Cat_Cal_Used=6 THEN (pht_NC_6) ELSE IF

Cat_Cal_Used = 5 THEN (pht_NC_5) ELSE

((IF(Cat_Cal_Used = 1) THEN(pht_NC_1) ELSE

((IF(Cat_Cal_Used = 2 ) THEN (pht_NC_2) ELSE

((IF(Cat_Cal_Used =3) THEN (pht_NC_3) ELSE

((IF(Cat_Cal_Used = 4) THEN (pht_NC_4) ELSE 0.05))))))))

(Boumans 2001)


P_loading_rate = 0.0005 g/L Walker and Hillman

N_loading_rate = 0.005 g/L Walker and Hillman

Mac_NC = 0.5 Ratio (Boumans 2001)

Mac_PC = 0.1 Ratio (Boumans 2001)

N_from_land = 0.0005 g/m2 Young et al 1996

P_from_land = 0.00003 g/m2 Young et al 1996

N_sed_1 = if Turbidity>TurbSedN

then 0.32 else 0.22


then 0.15 else 0.12


then 0.5 else 0.1


then 0.5 else 0.2


then 0.2 else 0.1

N_sed_6 = if Turbidity>70 then 0.2

else 0.1

Ratio Calibrated

P_sed_1 = if Turbidity>TurbSedP then

0.32 else 0.22

Ratio Calibrated

Appendix

244


0.15 else 0.12


0.5 else 0.1


0.5 else 0.2


0.2 else 0.1

P_sed_6 = if Turbidity>70 then 0.2

else 0.1

pht_NC_1 = 0.05

pht_NC_2 = 0.05

pht_NC_3 = 0.05

pht_NC_4 = 0.05

pht_NC_5 = 0.05

pht_NC_6 = 0.05

Ratio (Boumans 2001)

pht_PC_1 = 0.05

pht_PC_2 = 0.05

pht_PC_3 = 0.1

pht_PC_4 = 0.5

pht_PC_5 = 0.05

pht_PC_6 = 0.05

Ratio (Boumans 2001)

TurbSedN = 70 NTU Calibrated

TurbSedP = 70 NTU Calibrated


NO3N Nitrate as NO3-N

PO4P Orthophosphate as PO4-P

N sed Coprecipitation rate for NO3-N dependent on turbidity

P sed Coprecipitation rate for PO4-P dependent on turbidity

Mac NC N:C ratio required by macrophytes

Pht NC N:C ratio required by phytoplankton

N loading Non point source of NO3-N

N loading rate Non point source of NO3-N (minimal)

N from land Non point source of NO3-N (minimal)

Appendix

245

P loading Non point source of PO4-P

P loading rate Non point source of PO4-P (minimal)

P from land Non point source of PO4-P (minimal)

Mac PC P:C ratio required by macrophytes

Pht PC P:C ratio required by phytoplankton

N soil coprecip The coprecipitation rate for NO3-N O

P soil coprecip The coprecipitation rate for PO4-P

N in gL The inflow of NO3-N into the wetland.

P in gL The inflow of PO4-P into the wetland.

Ninflow Amount gL The NO3-N inflow concentration in g/L

Noutflow Amount gL The NO3-N outflow concentration in g/L

N sed release The NO3-N released from sediments.

N out The outflow of NO3-N to the river

P out The outflow of PO4-P to the river

Pinflow Amount gL The PO4-P inflow concentration in g/L

Poutflow Amount gL The PO4-P outflow concentration in g/L

P sed release The PO4-P released from sediments.

N uptake The uptake of NO3-N associated with macrophyte and algae

production.

P uptake The uptake of PO4-P associated with macrophyte and algae

production.

Appendix

246

$NutrientExchange

Equations/Rules Description/definition

DrainFlow_SunnyORPaiw =

IF(Category_Time_Series_Used=1)THEN(PDrainFlo

w_Paiwalla)

ELSE(IF(Category_Time_Series_Used=2)

THEN(PDrainFlow_Sunnyside) ELSE(0))

Gives the modeller the

option to simulate

irrigation inflow into

Paiwalla wetland.

Intended to test whether

the hypothesis that no

irrigation drainage was

affecting Paiwalla

wetland.

DrainFlow_PreMultiplication_Factor =

IF(IrrigationDrainage=1)

THEN(DrainFlow_SunnyORPaiw)

ELSE(IF(IrrigationDrainage=2)

THEN(PDrainFlow_REEDY) ELSE(0))

Selects the appropriate

drainage flow depending

to the wetland being

simulated.

DrainFlow_L = IF

(Drainage_Channel_multiplication_Factor=0) THEN

(DrainFlow_PreMultiplication_Factor) ELSE

((DrainFlow_PreMultiplication_Factor

*(Drainage_Channel_multiplication_Factor

*Seasonal_Flow_Pattern_SunnyORReedy)))

Calculates the drain flow

volume given the average

flow volume per day and

the seasonal flow pattern.

Therefore the average

flow can be increased

and the seasonal flow

pattern maintained.

Seasonal_Flow_Pattern_SunnyORReedy =

IF(Category_Time_Series_Used = 2)

THEN(Seasonal_Flow_Pattern_Sunnyside)ELSE(IF(

Category_Time_Series_Used = 4)

THEN(Seasonal_Flow_Pattern_Reedy) ELSE(1))

PDrainFlow_Paiwalla =

IF((Paiwalla_P_Drain_mg_perL+Paiwalla_N_Drain_

mg_perL)>0)

THEN(DrainFlowVolume_Liters_perDay_Sunnyside)

ELSE(0)

PDrainFlow_REEDY =

IF((Reedy_DrainPConc_mg_perL+REEDY_DrainNC

onc_mg_perL)>0)

THEN(DrainFlowVolume_Liters_perDay_REEDY)

ELSE(0)

PDrainFlow_Sunnyside =

IF((Sunnyside_P_Drain_mg_perL+Sunnyside_N_Drai

n_mg_perL)>0)

THEN(DrainFlowVolume_Liters_perDay_Sunnyside)

ELSE(0)

Selects the drain flow

volume from the

appropriate wetland data.

Appendix

247


Chla%_Removed_from_Drainage_Load = 0 Manual control to reduce

the Chl-a inflow.

Chla_DrainLoad_REEDY =

IF(REEDY_Chla_Drain_ugL>0)

THEN(REEDY_Chla_Drain_ugL

*DrainFlowVolume_Liters_perDay_REEDY)

ELSE(0)

Chla_DrainLoad_Sunnyside =

IF(Sunnyside_Chla_ugL>0)

THEN(Sunnyside_Chla_ugL*DrainFlowVolume_Lite

rs_perDay_Sunnyside) ELSE(0)

Calculates inflow load

from the concentration

and flow volume.

Chla_Drain_Load_Reedy2 =

(IF(IrrigationDrainage=1) THEN(0)


THEN(Chla_DrainLoad_REEDY)/100 ELSE(0)))*(IF

(Chla%_Removed_from_Drainage_Load >0) THEN

(100-Chla%_Removed_from_Drainage_Load) Else

100)

Chla_DrainLoad_Sunnyside2 =

(IF(IrrigationDrainage=1)

THEN(Chla_DrainLoad_Sunnyside)/100

ELSE(IF(IrrigationDrainage=2) THEN(0)

ELSE(0)))*(IF

(Chla%_Removed_from_Drainage_Load >0) THEN

(100-Chla%_Removed_from_Drainage_Load) Else

100)

Calculates the actual load

used in the simulation.

This is where the load is

reduced as per potential


REEDY_Chla_Drainage_divided_into_wetland =

IF(Drainage_Channel_multiplication_Factor=0)

THEN(Chla_Drain_Load_Reedy2/Wetlandvolume_Li

ters)

ELSE((Chla_Drain_Load_Reedy2/Wetlandvolume_Li

ters)*(Drainage_Channel_multiplication_Factor*Seas

onal_Flow_Pattern_SunnyORReedy))

Sunnyside_Chla_divided_into_wetland =


THEN(Chla_DrainLoad_Sunnyside2/Wetlandvolume

_Liters)

ELSE((Chla_DrainLoad_Sunnyside2/Wetlandvolume

_Liters)*(Drainage_Channel_multiplication_Factor*S

easonal_Flow_Pattern_SunnyORReedy))

Calculates the dispersal

of inflow load into the

wetland, i.e. to obtain

concentration.

Fits the concentration to


Chla_Accross_Wetland = Selects wether Reedy

Appendix

248

IF(Category_Time_Series_Used=2)

THEN(Sunnyside_Chla_divided_into_wetland)

ELSE(REEDY_Chla_Drainage_divided_into_wetland

)

Creek or Sunnyside

wetland data is to be used

depending on wetland

being simulated.

PhytoplanktonInflow_cm3m3 =

(((ChlaRiver_ugL/2.5)*Hypothetical_Inflow_m3)/(We

tlandvolume_Liters/1000))+(Chla_Accross_Wetland*

2.5)

Calculates the total

Phytoplankton inflow

into the wetland.

Merges Irrigation

drainage Chla-a inflow

and River Chl-a inflow.

Converts Chl-a into

phytoplankton.

PhytoplanktonOutflow_cm3m3 =

Hypothetical_Outflow_m3*PHYTOPLANKTON/(We

tlandvolume_Liters/1000)

Calculates the

concentration of outflow

depending on the outflow

volume and the

concentration within the

wetland.

Appendix

249


N%_Removed_from_Drain_Load = 0 Manual control to reduce

the NO3-N inflow

NDrainLoad_REEDY =

IF(REEDY_DrainNConc_mg_perL>0)

THEN(REEDY_DrainNConc_mg_perL*DrainFlowV

olume_Liters_perDay_REEDY) ELSE(0)

NDrainLoad_Sunnyside =

IF(Sunnyside_N_Drain_mg_perL>0)

THEN(Sunnyside_N_Drain_mg_perL*DrainFlowVol

ume_Liters_perDay_Sunnyside) ELSE(0)

NDrainLoad_Paiwalla =

IF(Paiwalla_N_Drain_mg_perL>0)

THEN(Paiwalla_N_Drain_mg_perL*DrainFlowVolu

me_Liters_perDay_Sunnyside) ELSE(0)



and flow volume

NDrainLoad_SunnyORPaiw =

IF(Category_Time_Series_Used=1)THEN(NDrainLoa

d_Paiwalla)


THEN(NDrainLoad_Sunnyside) ELSE(0))

Select the appropriate

drain load for either

Sunnyside or Paiwalla

wetlands.

NDrainLoad = (IF(IrrigationDrainage=1)

THEN(NDrainLoad_SunnyORPaiw)/100


THEN(NDrainLoad_REEDY)/100 ELSE(0)))*(IF

(N%_Removed_from_Drain_Load >0) THEN (100-

N%_Removed_from_Drain_Load) Else 100)

Calculate the actual load





N_Drain_Water_Inflow =


THEN(NDrainLoad/Wetlandvolume_Liters)

ELSE((NDrainLoad/Wetlandvolume_Liters)*(Drainag

e_Channel_multiplication_Factor*Seasonal_Flow_Pat

tern_SunnyORReedy))




concentration.



NInflowAmount_mgL =

((Hypothetical_Inflow_Liters*NRiver_mgL)/Wetland

volume_Liters)+N_Drain_Water_Inflow

Calculates the inflow

concentration as a

function of the wetland

volume of NO3-N into

the wetland.

NOutflow_Amount_gL =

(NO3N*Hypothetical_Outflow_Liters)/(Wetlandvolu

me_Liters)

Calculates the outflow

concentration as a


volume of NO3-N from

the wetland.

Appendix

250


P%_Removed_from_Drain_Load = 0 Manual control to reduce

the PO4-P inflow

PDrainLoad_REEDY =

IF(Reedy_DrainPConc_mg_perL>0)

THEN(Reedy_DrainPConc_mg_perL*DrainFlowVolu

me_Liters_perDay_REEDY) ELSE(0)

PDrainLoad_Sunnyside =

IF(Sunnyside_P_Drain_mg_perL>0)

THEN(Sunnyside_P_Drain_mg_perL*DrainFlowVolu


PDrainLoad_Paiwalla =

IF(Paiwalla_P_Drain_mg_perL>0)

THEN(Paiwalla_P_Drain_mg_perL*DrainFlowVolu




and flow volume

PDrainLoad_SunnyORPaiw =

IF(Category_Time_Series_Used=1)THEN(PDrainLoa

d_Paiwalla)


THEN(PDrainLoad_Sunnyside) ELSE(0))

Select the appropriate

drain load for either

Sunnyside or Paiwalla

wetlands.

PDrainLoad = ((IF(IrrigationDrainage=1)

THEN(PDrainLoad_SunnyORPaiw)/100


THEN(PDrainLoad_REEDY)/100 ELSE(0)))*(IF

(P%_Removed_from_Drain_Load >0) THEN (100-

P%_Removed_from_Drain_Load) Else 100))

Calculate the actual load





P_Drain_Water_Inflow =


THEN(PDrainLoad/Wetlandvolume_Liters)

ELSE((PDrainLoad/Wetlandvolume_Liters)*(Drainag

e_Channel_multiplication_Factor*Seasonal_Flow_Pat

tern_SunnyORReedy))




concentration.



PInflowAmount_mgL =

((Hypothetical_Inflow_Liters*PRiver_mgL)/Wetlandv

olume_Liters)+P_Drain_Water_Inflow

Calculates the inflow

concentration as a


volume of PO4-P into the

wetland.

POutflow_Amount_gL =

(PO4P*Hypothetical_Outflow_Liters)/(Wetlandvolum

e_Liters)

Calculates the outflow

concentration as a


volume of PO4-P from

the wetland

Appendix

251

$Wetland&RiverFlowExchange


Percentage_of_River_Flow_regarded_as_exchange =

1

Manual control of the

exchange volume as

percentage of the

wetland.

River_Exchange_Below_1% = 1 To reduce the exchange

volume below 1% of

river flow

FlowExchange%ofRiverFlow =

((FlowRiver_m3_per_Day/100)*Percentage_of_River

_Flow_regarded_as_exchange)/River_Exchange_Belo

w_1%

Calculates the volume

exchanged.

Hypothetical_Inflow_m3 =

IF(Flow_In_No1_ManualInput2_Wetland3_River4 =

2) THEN(ManualControlFlowIn_m3)

ELSE(IF(Flow_In_No1_ManualInput2_Wetland3_Ri

ver4 = 3) THEN(FlowExchangeInVolumeDependent)

ELSE(IF(Flow_In_No1_ManualInput2_Wetland3_Ri

ver4 = 4)THEN(FlowExchangeInRiverDependent)

ELSE(0)))

Selects the source of the

control for volume

exchange. Possible to

manually set exchange

volume.

Hypothetical_Outflow_m3 =

IF(Flow_Out_No1_ManualInput2_Wetland3_River4

= 2) THEN(ManualControlFlowOut_m3)

ELSE(IF(Flow_Out_No1_ManualInput2_Wetland3_R

iver4 = 3)

THEN(FlowExchangeOutVolumeDependent)

ELSE(IF(Flow_Out_No1_ManualInput2_Wetland3_R

iver4 =

4)THEN(FlowExchangeOutRiverDependent+(DrainFl

ow_L/1000)) ELSE(0)))

Selects the source of the

control for volume

exchange. Possible to

manually set exchange

volume.

Adds the irrigation drain

inflow volume to the

outflow volume.

Appendix

252

$SpatialRelevantTimeSeries

Solar Radiation see Methodology

$RiverNutrients

See Methodology

$WetlandsTimeseriesUpdateMeasuredValues

Extra wetland data and future wetland data.

$WetlandTimeseriesUpdate

Extra wetland data and future wetland data.

$RiverTimeseries4WetlandUpdateTimeseries

Same as $RiverNutrients but for extra wetland data and future wetland data.

$PotentialContributionToRiver

See Methodology

Appendix

253

Appendix B: Driving Variables

Appendix

254

Figure 74: Data - Model Driving Variables; From Figure 9 in section 2.3

T urbidity

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

1 4 0

1 6 0

1 8 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

NT

U


0

5

1 0

1 5

2 0

2 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

de

g C

Solar R adiation Paiwalla & Sunnyside We tlands

0

5

1 0

1 5

2 0

2 5

3 0

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

MJ

pe

r s

qu

are

me

ter

A

B

C

P a iwa lla W e tland 1997 S unnys ide W e tland 1997

Appendix

255


T urbidity

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

NT

U


0

5

1 0

1 5

2 0

2 5

3 0

3 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

de

g C

Solar R adiation P ilby C re e k & Lock 6 We tlands

0

5

1 0

1 5

2 0

2 5

3 0

3 5

Fe

b-9

8

Ma

r-9

8

Ap

r-9

8

Ma

y-9

8

Ju

n-9

8

Ju

l-9

8

Au

g-9

8

Se

p-9

8

MJ

pe

r s

qu

are

me

ter

D

E

F

L o ck 6 we tla nd 1 9 9 7 P ilb y C re e k W e tla nd 1 9 9 7

Appendix

256


T urbidity

-5 0

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

NT

U


0

5

1 0

1 5

2 0

2 5

3 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

de

g C

Solar R adiation R e e dy C re e k We tland

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

MJ

pe

r s

qu

are

me

ter

G

H

I

Reedy C reek W e tland 2000 -2001

Appendix

257

Figure 77: Time Series Irrigation Drainage ; From Figure 10 section 2.3.1

Dra in ag e Ph yto p lan kto n

0

0 .0 5

0 .1

0 .1 5

0 .2

0 .2 5

0 .3

0 .3 5

0 .4

0 .4 5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

cm

3/m

3

D rainage N O 3-N

0

0 .2

0 .4

0 .6

0 .8

1

1 .2

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

D rainage PO 4-P

0

0 .5

1

1 .5

2

2 .5

3

3 .5

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

A

B

C

S unnys ide W e tland

Appendix

258

Figure 78: Time Series Irrigation Drainage; From Figure 10 section 2.3.1

Se aso n al Drain ag e Patte rn Re e d y Cre e k Su b catch me n t

0 .0 0

0 .2 0

0 .4 0

0 .6 0

0 .8 0

1 .0 0

1 .2 0

1 .4 0

1 .6 0

1 .8 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

Re

lati

ve

Ra

te P

er

Mo

nth

-

-

D

Appendix

259

Figure 79: Time Series Irrigation Drainage ; From Figure 10 in section 2.3.1

D rainage Phytoplankton

-2 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

1 4 0

1 6 0

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

D rainage N O 3-N

-0 .4

-0 .2

0

0 .2

0 .4

0 .6

0 .8

1

1 .2

1 .4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

Drainage PO4-P

-1

0

1

2

3

4

5

6

7

8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

E

F

G

Re e d y C re e k W e tla nd

Appendix

260

Figure 80: River Data; From Figure 11 in section 2.3.2

PO4-P

-1

0

1

2

3

4

5

6

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

NO3-N

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

mg

/L

Phytoplankton

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Feb-9

7

Mar-

97

Apr-

97

May-9

7

Jun-9

7

Jul-97

Aug-9

7

cm

3/m

3

A

B

C

P a iwa lla W e tla nd S unnys id e W e tla nd

Appendix

261


PO4-P

-0.05

0

0.05

0.1

0.15

0.2

Feb-9

7

Mar-

97

Apr-

97

May-9

7

Jun-9

7

Jul-97

Aug-9

7

Sep-9

7

mg

/L

NO3-N

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Feb-9

7

Mar-

97

Apr-

97

May-9

7

Jun-9

7

Jul-97

Aug-9

7

Sep-9

7

mg

/L

Phytoplankton

0

2

4

6

8

10

12

14

Fe

b-9

7

Ma

r-9

7

Ap

r-9

7

Ma

y-9

7

Ju

n-9

7

Ju

l-9

7

Au

g-9

7

Se

p-9

7

cm

3/m

3

D

E

F

L o ck 6 we tla nd 1 9 9 7 P ilb y C re e k W e tla nd 1 9 9 7

Appendix

262


PO4-P

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

NO3-N

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oct-

00

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

mg

/L

Phytoplankton

0

2

4

6

8

1 0

1 2

1 4

Ju

n-0

0

Ju

l-0

0

Au

g-0

0

Se

p-0

0

Oc

t-0

0

No

v-0

0

De

c-0

0

Ja

n-0

1

Fe

b-0

1

Ma

r-0

1

Ap

r-0

1

Ma

y-0

1

cm

3/m

3

G

H

I

Reedy C reek W e tland

Appendix

263

Appendix C: Key to wetland numbers

Table 18: Wetlands simulated as category 3 wetlands

Wetlands

ID

Australian

Wetland

Number

Wetland

Name

Used

depth

Volume

m3

Category

managed

703 S0070 CAURNAMONT 1.5 1353858 3

690 S0075 WALKER FLAT SOUTH LAGOON 0.8 710419 3

1107 S0076 LAKE BYWATERS 0.8 310292 3

685 S0082 DEVON DOWNS SOUTH 0.92 493457 3

1102 S0093 YARRAMUNDI 2 195388 3

1101 S0093 YARRAMUNDI 2 617098 3

663 S0094 YARRAMUNDI NORTH 2 704688 3

651 S0103 ARLUNGA 0.9 1497057 3

646 S0104 ROONKA 0.9 147172 3

644 S0105 REEDY ISLAND FLAT 1.2 266973 3

645 S0106 McBEAN POUND SOUTH 0.65 42489 3

642 S0107 McBEAN POUND NORTH 0.65 121855 3

641 S0108 SINCLAIR FLAT 0.92 20053 3

640 S0108 SINCLAIR FLAT 0.92 513745 3

1044 S0109 DONALD FLAT LAGOON 1.25 1760260 3

391 S0110 IRWIN FLAT 2 881564 3

383 S0111 MURBPOOK LAGOON COMPLEX 0.92 32620 3



379 S0112 MURBKO SOUTH 0.9 1147222 3

375 S0113 MURBKO FLAT COMPLEX 0.7 75477 3



Appendix

264

Wetlands

ID

Australian

Wetland

Number

Wetland

Name

Used

depth

Volume

m3

Category

managed

367 S0115 WOMBAT REST BACKWATER 0.7 264111 3

294 S0142 BOGGY FLAT 1.5 89373 3

324 S0149 BIG TOOLUNKA FLAT 2.3 848443 3

262 S0160 YARRA COMPLEX 2 1717745 3

1036 S0174 LOCH LUNA and NOCKBURRA CREEK 2 127894 3

190 S0174 LOCH LUNA and NOCKBURRA CREEK 2 6146303 3

631 S0189 PYAP LAGOON 2 904144 3

492 S0201 AJAX ACHILLES LAKE 1.2 22764 3

486 S0201 AJAX ACHILLES LAKE 1.2 262527 3

471 S0203 SALT CREEK AND GURRA GURRA LAKES 1.5 78987 3

1048 S0207 LYRUP CAUSEWAY WEST 0.92 17437 3

1039 S0214 RUMPAGUNYAH CREEK 2 230689 3

1031 S0214 RUMPAGUNYAH CREEK 2 371340 3

1007 S0218 GOAT ISLAND AND PARINGA PADDOCK 0.92 227636 3

1006 S0218 GOAT ISLAND AND PARINGA PADDOCK 0.92 235651 3

997 S0219 PARINGA ISLAND 0.92 12402 3

996 S0219 PARINGA ISLAND 0.92 39096 3

995 S0219 PARINGA ISLAND 0.92 111272 3

93 S0219 PARINGA ISLAND 0.92 227075 3

92 S0219 PARINGA ISLAND 0.92 25097 3

91 S0219 PARINGA ISLAND 0.92 72376 3

90 S0219 PARINGA ISLAND 0.92 17157 3

89 S0219 PARINGA ISLAND 0.92 10223 3

956 S0220 RAL RAL CREEK AND RAL RAL WIDEWATERS 2 6785374 3

69 S0227 HORSESHOE SWAMP 1.2 327432 3

978 S0229 WOOLENOOK BEND COMPLEX 1.2 2111925 3

Appendix

265

Wetlands

ID

Australian

Wetland

Number

Wetland

Name

Used

depth

Volume

m3

Category

managed



67 S0230 MURTHO PARK COMPLEX 0.92 31733 3




32 S0242 SLANEY OXBOW 1.25 90869 3

1134 XR001 Lock 6 Wetland 0.92 164860 3

Table 19: Wetlands simulated as category 4 wetlands

Wetlands

ID

Australian

Wetland

Number

Wetland

Name

Used

depth

Volume

m3

Category

managed

766 S0035 TAILEM BEND 0.8 765545 4

1110 S0052 REEDY CREEK 0.8 591799 4

310 S0148 LITTLE TOOLUNKA FLAT 1.4 739622 4

329 S0151 RAMCO LAGOON 0.3 279446 4

209 S0179 KINGSTON COMMON 0.92 340410 4

1029 S0180 WACHTELS LAGOON 0.92 6259251 4

583 S0185 YATCO LAGOON 0.5 1729378 4

Appendix

266

Appendix D: Cumulative Management Scenarios

Appendix

267

Table 20: Change in PO4-P wetland loading and percentage outflow due to management; category 3 wetland scenarios

PO4-P Net Loading to wetland kg/annum % Reduction in Outflow

Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0070 CAURNAMONT 703 1.5 1353858 3 541 535 540 542 -13 -2 3

S0075 WALKER FLAT

SOUTH LAGOON

690 0.8 710419 3 533 527 527 526 -12 -12 -14

S0076 LAKE BYWATERS 1107 0.8 310292 3 512 504 498 489 -11 -21 -34

S0082 DEVON DOWNS

SOUTH

685 0.92 493457 3 525 519 517 513 -11 -16 -23

S0093 YARRAMUNDI 1102 2 195388 3 493 483 471 496 -11 -25 4

1101 2 617098 3 530 524 523 531 -12 -14 2

S0094 YARRAMUNDI

NORTH

663 2 704688 3 532 527 527 533 -12 -12 1

S0103 ARLUNGA 651 0.9 1497057 3 512 505 511 514 -19 -4 5

S0104 ROONKA 646 0.9 147172 3 452 438 424 419 -15 -29 -34

S0105 REEDY ISLAND FLAT 644 1.2 266973 3 480 469 464 465 -16 -23 -20

Appendix

268


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0106 McBEAN POUND

SOUTH

645 0.65 42489 3 350 328 297 269 -11 -27 -41

S0107 McBEAN POUND

NORTH

642 0.65 121855 3 441 425 409 392 -14 -29 -44

S0108 SINCLAIR FLAT 641 0.92 20053 3 264 241 210 234 -8 -19 -10

640 0.92 513745 3 498 490 489 488 -16 -18 -20

S0109 DONALD FLAT

LAGOON

1044 1.25 1760260 3 514 506 513 516 -20 -2 6

S0110 IRWIN FLAT 391 2 881564 3 507 500 502 509 -18 -12 4

S0111 MURBPOOK LAGOON

COMPLEX

383 0.92 32620 3 321 298 266 278 -10 -24 -19

381 0.92 946764 3 508 501 503 506 -18 -12 -5

380 0.92 65777 3 393 373 348 346 -13 -29 -30

S0112 MURBKO SOUTH 379 0.9 1147222 3 510 503 506 510 -19 -10 -1

Appendix

269


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0113 MURBKO FLAT

COMPLEX

375 0.7 75477 3 405 386 362 344 -13 -30 -42

374 0.7 1135665 3 510 503 506 508 -19 -10 -6

371 0.7 65887 3 393 373 348 329 -13 -29 -41

S0115 WOMBAT REST

BACKWATER

367 0.7 264111 3 479 468 463 455 -16 -23 -35

S0142 BOGGY FLAT 294 1.5 89373 3 438 423 396 443 -11 -31 3

S0149 BIG TOOLUNKA FLAT 324 2.3 848443 3 532 526 526 533 -13 -14 2

S0160 YARRA COMPLEX 262 2 1717745 3 539 534 538 542 -15 -2 6

S0174 LOCH LUNA and

NOCKBURRA CREEK

1036 2 127894 3 492 475 456 534 -12 -27 32

S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 3 589 581 593 597 -25 11 24

S0189 PYAP LAGOON 631 2 904144 3 574 567 568 577 -15 -12 5

Appendix

270


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0201 AJAX ACHILLES LAKE 492 1.2 22764 3 283 259 225 419 -8 -20 48

486 1.2 262527 3 496 484 477 478 -16 -26 -25

S0203 SALT CREEK AND

GURRA GURRA LAKES

471 1.5 78987 3 420 401 375 444 -13 -30 16

S0207 LYRUP CAUSEWAY

WEST

1048 0.92 17437 3 250 227 195 228 -7 -17 -7

S0214 RUMPAGUNYAH CREEK 1039 2 230689 3 490 478 469 493 -15 -27 3

1031 2 371340 3 508 498 493 507 -16 -24 -1

S0218 GOAT ISLAND AND

PARINGA PADDOCK

1007 0.92 227636 3 490 478 468 461 -15 -27 -36

1006 0.92 235651 3 491 479 471 463 -15 -27 -36

S0219 PARINGA ISLAND 997 0.92 12402 3 169 154 129 145 -7 -18 -11

996 0.92 39096 3 263 248 224 219 -12 -32 -37

995 0.92 111272 3 321 311 298 290 -16 -36 -49

Appendix

271


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0219 PARINGA ISLAND 93 0.92 227075 3 343 336 329 324 -18 -33 -44

92 0.92 25097 3 229 213 186 188 -10 -28 -27

91 0.92 72376 3 301 289 271 263 -14 -36 -46

90 0.92 17157 3 197 181 155 161 -9 -23 -19

89 0.92 10223 3 153 138 115 137 -6 -16 -7

S0220 RAL RAL CREEK AND

RAL RAL WIDEWATERS

956 2 6785374 3 367 360 367 369 -37 4 14

S0227 HORSESHOE SWAMP 69 1.2 327432 3 350 343 340 339 -19 -30 -32

S0229 WOOLENOOK

BEND COMPLEX

978 1.2 2111925 3 365 359 364 366 -29 -3 7

84 1.2 29590 3 242 226 200 244 -11 -30 1

82 1.2 41520 3 267 253 229 253 -13 -33 -12

Appendix

272


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0230 MURTHO PARK

COMPLEX

67 0.92 31733 3 248 232 206 204 -11 -31 -32

61 0.92 24337 3 226 210 183 185 -10 -27 -26

60 0.92 50151 3 280 266 244 237 -13 -35 -41

47 0.92 250315 3 345 338 332 328 -18 -32 -44

S0242 SLANEY OXBOW 32 1.25 90869 3 313 302 286 291 -15 -37 -30

XR001 Lock 6 Wetland 1134 0.92 164860 3 364 358 363 364 -28 -6 2

Min 153 138 115 137

Max 589 581 593 597

Average 406 394 382 392

Median 438 423 396 419

Total 23140 22451 21795 22338

Appendix

273

Table 21: Change in NO3-N wetland loading and percentage outflow due to management; category 3 wetland scenarios

NO3-N Net Loading to wetland kg/annum % Reduction in Outflow

Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0070 CAURNAMONT 703 1.5 1353858 3 682 684 703 738 1 12 30

S0075 WALKER FLAT

SOUTH LAGOON

690 0.8 710419 3 665 662 675 700 -2 5 17

S0076 LAKE BYWATERS 1107 0.8 310292 3 626 609 616 650 -7 -4 10

S0082 DEVON DOWNS

SOUTH

685 0.92 493457 3 651 643 654 687 -4 1 16

S0093 YARRAMUNDI 1102 2 195388 3 590 563 567 795 -10 -8 73

1101 2 617098 3 660 655 668 759 -3 3 47

S0094 YARRAMUNDI

NORTH

663 2 704688 3 665 661 675 759 -2 5 46

S0103 ARLUNGA 651 0.9 1497057 3 709 705 728 753 -2 10 25

S0104 ROONKA 646 0.9 147172 3 583 537 547 640 -15 -12 19

S0105 REEDY ISLAND FLAT 644 1.2 266973 3 638 608 619 726 -12 -8 35

Appendix

274


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0106 McBEAN POUND

SOUTH

645 0.65 42489 3 405 333 342 349 -15 -13 -12

S0107 McBEAN POUND

NORTH

642 0.65 121855 3 561 510 520 531 -16 -12 -9

S0108 SINCLAIR FLAT 641 0.92 20053 3 282 215 227 617 -11 -9 56

640 0.92 513745 3 678 661 676 713 -8 -1 17

S0109 DONALD FLAT

LAGOON

1044 1.25 1760260 3 712 709 732 762 -2 12 29

S0110 IRWIN FLAT 391 2 881564 3 697 688 706 779 -5 5 43


COMPLEX

383 0.92 32620 3 361 289 299 610 -14 -12 48

381 0.92 946764 3 699 691 709 740 -4 5 22

380 0.92 65777 3 475 409 423 619 -16 -13 35

S0112 MURBKO SOUTH 379 0.9 1147222 3 704 698 717 746 -3 7 23

Appendix

275


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0113 MURBKO FLAT

COMPLEX

375 0.7 75477 3 496 433 446 495 -16 -13 0

374 0.7 1135665 3 704 697 716 739 -4 7 19

371 0.7 65887 3 475 410 423 477 -16 -13 0

S0115 WOMBAT REST

BACKWATER

367 0.7 264111 3 637 607 618 646 -12 -8 3

S0142 BOGGY FLAT 294 1.5 89373 3 516 466 462 803 -14 -15 79

S0149 BIG TOOLUNKA FLAT 324 2.3 848443 3 686 683 691 780 -2 2 49

S0160 YARRA COMPLEX 262 2 1717745 3 701 703 718 766 1 10 36

S0174 LOCH LUNA and

NOCKBURRA CREEK

1036 2 127894 3 611 560 569 947 -13 -11 88

S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 3 811 815 842 881 2 17 39

S0189 PYAP LAGOON 631 2 904144 3 777 770 790 880 -3 6 48

Appendix

276


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0201 AJAX ACHILLES LAKE 492 1.2 22764 3 304 234 242 805 -12 -11 85

486 1.2 262527 3 645 614 619 734 -13 -10 36


GURRA GURRA LAKES

471 1.5 78987 3 508 445 451 831 -16 -15 84


WEST

1048 0.92 17437 3 263 196 205 650 -11 -9 62

S0214 RUMPAGUNYAH CREEK 1039 2 230689 3 634 600 604 820 -13 -12 72

1031 2 371340 3 669 646 654 804 -10 -7 61


PARINGA PADDOCK

1007 0.92 227636 3 633 599 603 668 -13 -12 14

1006 0.92 235651 3 636 602 607 671 -13 -11 14

S0219 PARINGA ISLAND 997 0.92 12402 3 183 136 140 402 -12 -11 57

996 0.92 39096 3 308 258 259 394 -19 -18 33

995 0.92 111272 3 399 366 364 419 -19 -20 12

Appendix

277


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0219 PARINGA ISLAND 93 0.92 227075 3 436 416 416 449 -15 -15 9

92 0.92 25097 3 259 208 207 390 -17 -17 42

91 0.92 72376 3 366 326 324 405 -20 -21 19

90 0.92 17157 3 217 167 169 385 -14 -14 48

89 0.92 10223 3 164 121 124 430 -11 -10 65


RAL RAL WIDEWATERS

956 2 6785374 3 480 477 489 514 -3 9 38

S0227 HORSESHOE SWAMP 69 1.2 327432 3 449 433 436 477 -13 -11 23

S0229 WOOLENOOK

BEND COMPLEX

978 1.2 2111925 3 476 472 483 504 -5 8 30

84 1.2 29590 3 278 226 225 519 -17 -18 82

82 1.2 41520 3 314 265 266 506 -19 -19 75

Appendix

278


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0230 MURTHO PARK

COMPLEX

67 0.92 31733 3 285 234 233 392 -18 -18 38

61 0.92 24337 3 256 204 204 389 -16 -16 42

60 0.92 50151 3 333 286 286 397 -19 -20 27

47 0.92 250315 3 440 421 422 453 -15 -14 10

S0242 SLANEY OXBOW 32 1.25 90869 3 384 348 345 481 -19 -21 52

XR001 Lock 6 Wetland 1134 0.92 164860 3 475 470 481 499 -5 7 25

Min 164 121 124 349

Max 811 815 842 947

Average 513 481 490 622

Median 516 477 489 650

Total 29254 27445 27935 35477

Appendix

279

Table 22: Change in Phytoplankton wetland loading and percentage outflow due to management; category 3 wetland scenarios

Phytoplankton Loading to wetland m3/annum % Reduction in Outflow

Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0070 CAURNAMONT 703 1.5 1353858 3 -8 -9 -12 -10 -2 -19 -10

S0075 WALKER FLAT

SOUTH LAGOON

690 0.8 710419 3 -8 -9 -10 -12 -3 -11 -20

S0076 LAKE BYWATERS 1107 0.8 310292 3 -7 -8 -9 -11 -3 -11 -21

S0082 DEVON DOWNS

SOUTH

685 0.92 493457 3 -8 -8 -10 -12 -3 -11 -22

S0093 YARRAMUNDI 1102 2 195388 3 -7 -7 -8 -8 -3 -11 -11

1101 2 617098 3 -8 -8 -10 -3 -3 -11 27

S0094 YARRAMUNDI

NORTH

663 2 704688 3 -8 -9 -10 -6 -3 -11 13

S0103 ARLUNGA 651 0.9 1497057 3 -8 -8 -11 -11 -1 -20 -20

S0104 ROONKA 646 0.9 147172 3 -6 -6 -8 -6 -3 -12 -3

S0105 REEDY ISLAND FLAT 644 1.2 266973 3 -7 -7 -8 -7 -3 -11 -1

Appendix

280


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0106 McBEAN POUND

SOUTH

645 0.65 42489 3 -4 -4 -4 -5 -2 -4 -8

S0107 McBEAN POUND

NORTH

642 0.65 121855 3 -5 -6 -7 -8 -3 -12 -18

S0108 SINCLAIR FLAT 641 0.92 20053 3 -2 -2 -3 0 -1 -3 17

640 0.92 513745 3 -7 -8 -9 -11 -3 -11 -23

S0109 DONALD FLAT

LAGOON

1044 1.25 1760260 3 -8 -8 -11 -12 -1 -19 -25

S0110 IRWIN FLAT 391 2 881564 3 -8 -8 -9 -6 -2 -11 7


COMPLEX

383 0.92 32620 3 -3 -3 -4 -1 -1 -4 17

381 0.92 946764 3 -8 -8 -9 -11 -2 -11 -21

380 0.92 65777 3 -4 -5 -6 -7 -2 -14 -18

S0112 MURBKO SOUTH 379 0.9 1147222 3 -8 -8 -10 -11 -2 -11 -21

Appendix

281


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0113 MURBKO FLAT

COMPLEX

375 0.7 75477 3 -5 -5 -7 -7 -2 -13 -16

374 0.7 1135665 3 -8 -8 -9 -11 -2 -11 -18

371 0.7 65887 3 -4 -5 -6 -6 -2 -14 -13

S0115 WOMBAT REST

BACKWATER

367 0.7 264111 3 -7 -7 -8 -10 -3 -11 -20

S0142 BOGGY FLAT 294 1.5 89373 3 -5 -5 -7 -10 -3 -12 -35

S0149 BIG TOOLUNKA FLAT 324 2.3 848443 3 -8 -8 -10 -8 -3 -11 -3

S0160 YARRA COMPLEX 262 2 1717745 3 -8 -9 -12 -12 -2 -18 -18

S0174 LOCH LUNA and

NOCKBURRA CREEK

1036 2 127894 3 -6 -7 -8 -10 -2 -10 -26

S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 3 -9 -9 -12 -12 3 -12 -13

S0189 PYAP LAGOON 631 2 904144 3 -9 -9 -11 -6 -2 -10 15

Appendix

282


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0201 AJAX ACHILLES LAKE 492 1.2 22764 3 -2 -2 -3 0 -1 -3 18

486 1.2 262527 3 -6 -7 -8 -6 -3 -11 3


GURRA GURRA LAKES

471 1.5 78987 3 -5 -5 -6 -11 -3 -12 -42


WEST

1048 0.92 17437 3 -2 -2 -2 0 -1 -2 16

S0214 RUMPAGUNYAH CREEK 1039 2 230689 3 -6 -7 -8 -7 -3 -11 -3

1031 2 371340 3 -7 -7 -9 -4 -3 -11 16


PARINGA PADDOCK

1007 0.92 227636 3 -6 -7 -8 -7 -3 -11 -3

1006 0.92 235651 3 -6 -7 -8 -7 -3 -11 -3

S0219 PARINGA ISLAND 997 0.92 12402 3 -1 -1 -1 0 -1 -3 17

996 0.92 39096 3 -2 -2 -3 -4 -2 -14 -19

995 0.92 111272 3 -3 -4 -4 -3 -3 -12 -1

Appendix

283


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0219 PARINGA ISLAND 93 0.92 227075 3 -4 -4 -5 -5 -3 -11 -15

92 0.92 25097 3 -2 -2 -2 -1 -2 -4 16

91 0.92 72376 3 -3 -3 -4 -3 -3 -12 -4

90 0.92 17157 3 -1 -2 -2 0 -1 -3 16

89 0.92 10223 3 -1 -1 -1 0 -1 -2 15


RAL RAL WIDEWATERS

956 2 6785374 3 -4 -4 -6 -8 6 -13 -33

S0227 HORSESHOE SWAMP 69 1.2 327432 3 -4 -4 -5 -4 -3 -11 -4

S0229 WOOLENOOK

BEND COMPLEX

978 1.2 2111925 3 -4 -4 -6 -7 0 -15 -22

84 1.2 29590 3 -2 -2 -2 0 -2 -4 26

82 1.2 41520 3 -2 -2 -3 -6 -2 -14 -39

Appendix

284


Aus

wetland

#

Wetland

name

Wetlands

id

Used

depth

m

Volume

m3

Category

managed

Status Quo

Turbidity

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

25%

Turbidity

Reduction

50%

Turbidity

Reduction

75%

Turbidity

Reduction

S0230 MURTHO PARK

COMPLEX

67 0.92 31733 3 -2 -2 -2 -1 -2 -5 17

61 0.92 24337 3 -2 -2 -2 -1 -2 -4 16

60 0.92 50151 3 -3 -3 -4 -3 -2 -13 -11

47 0.92 250315 3 -4 -4 -5 -6 -3 -11 -23

S0242 SLANEY OXBOW 32 1.25 90869 3 -3 -3 -4 -4 -3 -12 -8

XR001 Lock 6 Wetland 1134 0.92 164860 3 -4 -4 -6 -6 0 -17 -18

Min -9 -9 -12 -12

Max -1 -1 -1 0

Average -5 -5 -7 -6

Median -5 -5 -7 -6

Total -293 -309 -380 -354

Appendix

285

Table 23: PO4-P comparison between Full year wet versus Summer wet Winter dry for three selected wetlands; category 3 wetland scenarios

Net Loading to wetland kg/annum % Reduction in Outflow

Aus wetland

#

Wetland name

Wetlands id

Used depth

Volume m

3

Status Quo Turbidity







Full Year Wet

S0174 LOCH LUNA and NOCKBURRA CREEK

190 2 6146303 589 581 593 597 -25 11 24

S0219 PARINGA ISLAND 93 0.92 227075 343 336 329 324 -18 -33 -44

XR001 Lock 6 Wetland 1134 0.92 164860 364 358 363 364 -28 -6 2

Sum Full Year Wet 1296 1274 1285 1286

Summer Wet; Winter Dry

S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 149 143 151 154 -48 15 36

S0219 PARINGA ISLAND 93 0.92 227075 71 68 70 72 -26 -5 8

XR001 Lock 6 Wetland 1134 0.92 164860 77 73 77 79 -48 4 24

Summer Wet Only 297 284 298 304

Less Loading to

wetland if Summer Wet Only

999 991 986 982

The load to the wetland, for the full year wet scenario, is calculated from the average retention in the scenario time period multiplied by 365. The

load to the wetland, for the summer wet winter dry management scenario, is calculated as a sum from the 88 days simulated in the model to be

the peak macrophyte growth period.

Appendix

286

Table 24: NO3-N comparison between Full year wet versus Summer wet Winter dry for three selected wetlands; category 3 wetland scenarios


Aus wetland

#

Wetland name

Wetlands id

Used depth

Volume m

3








Full Year Wet


190 2 6146303 811 815 842 881 2 17 39

S0219 PARINGA ISLAND 93 0.92 227075 436 416 416 449 -15 -15 9

XR001 Lock 6 Wetland 1134 0.92 164860 475 470 481 499 -5 7 25

Sum Full Year Wet 1722 1700 1740 1830


S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 374 375 393 417 1 30 71

S0219 PARINGA ISLAND 93 0.92 227075 182 173 186 206 -16 8 49

XR001 Lock 6 Wetland 1134 0.92 164860 200 198 208 217 -7 26 54

Summer Wet Only 756 746 787 841

Less Loading to


966 954 953 989




Appendix

287

Table 25: Phytoplankton comparison between Full year wet versus Summer wet Winter dry for three selected wetlands; category 3 wetland scenarios


Aus wetland

#

Wetland name

Wetlands id

Used depth

Volume m

3








Full Year Wet


190 2 6146303 -9.23 -8.57 -11.63 -11.81 3 -12 -13

S0219 PARINGA ISLAND 93 0.92 227075 -3.79 -4.10 -4.87 -5.30 -3 -11 -15

XR001 Lock 6 Wetland 1134 0.92 164860 -4.40 -4.41 -6.13 -6.29 0 -17 -18

Sum Full Year Wet -17 -17 -23 -23


S0174 LOCH LUNA and

NOCKBURRA CREEK

190 2 6146303 -2.34 -1.48 -2.71 -1.99 17 -7 7

S0219 PARINGA ISLAND 93 0.92 227075 -0.87 -0.85 -1.04 -1.01 1 -7 -6

XR001 Lock 6 Wetland 1134 0.92 164860 -0.95 -0.69 -1.43 -1.16 10 -20 -9

Summer Wet Only -4 -3 -5 -4

Less Loading to


13 14 17 19




Appendix

288

Table 26: Change in PO4-P wetland loading and percentage in and outflow due to management; category 4 wetland scenarios

PO4-P Net Loading to Wetland kg/annum % Reduction in Inflow % Reduction in Outflow

Aus

Wetland #

Wetland

name

Used

depth m

Volume

m3

Wetland

Category

Status

Quo Irrigation Drainage

Nutrient

25%

Irrigation Drainage Nutrient

Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

S0035 TAILEM

BEND

0.8 765545 4 7171 7487 7798 8104 0.0070 0.0140 0.0210 6.00 11.90 17.69

S0052 REEDY CREEK

0.8 591799 4 21778 22101 22518 22829 0.0014 0.0027 0.0041 1.23 2.81 3.99

S0148 LITTLE TOOLUNKA FLAT

1.4 739622 4 5088 5454 5727 6218 0.0089 0.0177 0.0266 7.67 13.42 23.70

S0151 RAMCO LAGOON

0.3 279446 4 4974 5015 5557 5861 0.0089 0.0177 0.0266 0.86 11.95 18.18

S0179 KINGSTON

COMMON

0.92 340410 4 5126 5665 5723 6273 0.0082 0.0165 0.0247 9.88 10.96 21.03

S0180 WACHTELS LAGOON

0.92 6259251 4 9140 9200 9259 9317 0.0082 0.0165 0.0247 4.13 8.26 12.37

S0185 YATCO LAGOON

0.5 1729378 4 7585 7798 7973 8110 0.0082 0.0165 0.0247 7.12 12.97 17.54

Min 4974 5015 5557 5861

Max 21778 22101 22518 22829

Average 8694 8960 9222 9530

Median 7171 7487 7798 8104

Total 60861 62720 64555 66712

Appendix

289

Table 27: Change in NO3-N wetland loading and percentage in and outflow due to management; category 4 wetland scenarios

NO3-N Net Loading to Wetland kg/annum % Reduction in Inflow % Reduction in Outflow

Aus

Wetland #

Wetland

name

Used

depth m

Volume

m3

Wetland

Category

Status


Nutrient

25%


Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

S0035 TAILEM

BEND

0.8 765545 4 19038 19102 19166 19230 0.0003 0.0006 0.0009 0.42 0.84 1.26

S0052 REEDY CREEK

0.8 591799 4 3965 4019 4074 4128 0.0007 0.0014 0.0021 0.69 1.38 2.07

S0148 LITTLE TOOLUNKA FLAT

1.4 739622 4 1782 1837 1891 1945 0.0009 0.0017 0.0026 0.51 1.02 1.52

S0151 RAMCO LAGOON

0.3 279446 4 8078 8180 8178 8125 0.0009 0.0017 0.0026 2.31 2.28 1.06

S0179 KINGSTON

COMMON

0.92 340410 4 7484 8196 6421 7977 0.0008 0.0016 0.0024 12.28 -18.31

8.51


0.92 6259251 4 5493 5507 5521 5536 0.0008 0.0016 0.0024 0.19 0.37 0.56

S0185 YATCO LAGOON

0.5 1729378 4 3160 3195 3230 3265 0.0008 0.0016 0.0024 0.35 0.70 1.04

Min 1782 1837 1891 1945

Max 19038 19102 19166 19230

Average 7000 7148 6926 7172

Median 5493 5507 5521 5536

Total 49000 50036 48481 50204

Appendix

290

Table 28: Change in Phytoplankton wetland loading and percentage in and outflow due to management; category 4 wetland scenarios

Phytoplankton Net Loading to Wetland

m3/annum % Reduction in Inflow % Reduction in Outflow

Aus

Wetland #

Wetland

name

Used

depth m

Volume

m3

Wetland

Category

Status


Nutrient

25%


Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

25%


Reduction

50%


Reduction

75%


Reduction

S0035 TAILEM BEND

0.8 765545 4 31 47 63 80 0.0014 0.0029 0.0043 4.13 8.24 12.37

S0052 REEDY CREEK

0.8 591799 4 33 49 65 81 0.0011 0.0022 0.0032 4.08 8.16 12.24

S0148 LITTLE

TOOLUNKA FLAT

1.4 739622 4 33 50 67 83 0.0014 0.0028 0.0042 4.10 8.20 12.28

S0151 RAMCO

LAGOON

0.3 279446 4 87 115 139 159 0.0014 0.0028 0.0042 7.92 14.61 20.12

S0179 KINGSTON COMMON

0.92 340410 4 77 100 123 146 0.0014 0.0027 0.0041 6.04 12.10 18.14


0.92 6259251 4 136 140 143 146 0.0014 0.0027 0.0041 1.05 2.10 3.16

S0185 YATCO

LAGOON

0.5 1729378 4 91 101 110 120 0.0014 0.0027 0.0041 2.63 5.26 7.91

Min 31 47 63 80

Max 136 140 143 159

Average 70 86 101 116

Median 77 100 110 120

Total 487 601 710 815

Appendix

291

Appendix E: WETMOD 2 Code

See attached CD.

Documents

Tumi Bjornsson Ph.D