Carl Fitz Naiming Wang Everglades Landscape Fred Sklar

Everglades Division, SFWMD

EvergladesEvergladesLandscapeLandscape

ModelModel

Status of theEverglades Landscape Model

Carl FitzNaiming WangJason Godin

Fred Sklar & othersin the

Everglades Division

Aug 7, 2002

http://www.sfwmd.gov/org/wrp/elm

Everglades Division, SFWMD http://www.sfwmd.gov/org/wrp/elm

August 7 Goals

• Part I: Overview (this slide show)

§ Model objectives

§ Model structure

§ Calibration performance

§ Application performance measures

• Part II: Details (using web site)

§ Model tools

§ Code

§ Data

§ Performance (calibration)

Background for detailed review of ELM


§ Model objectives

§ Model structure




§ Model tools

§ Code

§ Data



Everglades Landscape Model(ELM) Objectives

• Understand ecosystem dynamics at regional scale

• Develop predictions of landscape response to altered water &

nutrient management

• One tool to aid in Everglades restoration

Integrate hydrology, biology and nutrient cycling in spatially

explicit simulation


August 7 Goals


§ Model objectives

§ Model structure




§ Model tools

§ Code

§ Data




Everglades"response"

Hydrology

Ecological interactions

SFWMM (&ELM) hydrology

Everglades"response"

HydrologyNutrients

+ ELM water quality

Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

+ ELM ecology


Landscape Pattern

EcologicalInteractions

Flows acrosspattern

Cypress

Wet Prairie

Sawgrass

Cattail

Slough

Canal

Legend

EcologicalInteractions

Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

Spatial interactions


WestPalmBeach

Miami

SFWMM10.4 km2

ELM1.0 km2

ELM0.25 km2

Florida Bay

LakeOkeechobee

WCA2A

Model domains

Outdated (v.1.0)map ofL-31N &C-111 boundary


Model Structure

• Flows:

§ Spatial: similar in general algorithmic structure to SFWMM

§ Incorporate explicit feedbacks and altered canal network/scales

• Ecology

§ 13 physical, chemical, biological modules; selectable at runtime

• Utilities

§ Water, phosphorus budgets

§ Summary stats

Integrated spatial and non-spatial modules of hydrology,

biology and nutrient cycling


Hydrologic data

• Data shared with SFWMM§ Elevation: filtered from 10 to 1 km2 resolution§ Hydraulic conductivity: filtered to 1 km,§ Rainfall: direct application of 2x2 mi daily data§ Other meteorological: dynamic interpolation of cloud, dew pt,

etc

• Canal/levee vectors managed in GIS using precisecoordinates

• Water control structure attributes managed in relationaldatabase


Hydrologic modules

• Vertical solutions (1 day time step)§ ET from evap model, solar rad model, daily data on temp,

dewPt, cloud, wind; variable LAI§ Rainfall directly from daily spatial time series from SFWMM§ 3 layers: infiltration, percolation, upflow


Hydrologic modules (con’t)

• Horizontal solutions (explicit, 2 hr. step)§ Overland: finite difference, Manning’s equation, ADE§ Groundwater: finite difference, simple Darcy’s, ADE§ Surface-groundwater interaction: every time step, solve for

available storages


Hydrologic modules (more)

• Water management network ( 2 hr step)§ Canal reaches: mass balance, iterative solution as in SFWMM§ Water control structs: daily flows for all structures, from

observations or SFWMM


Hydrologic modules (con’t)

• Mass balance and budget§ Basin & domain-wide budgets & error-checks

• Post-processing§ Consistency checks with SFWMM (maps and budgets, stage

hydrographs)


Phosphorus modules

• Overland, groundwater, and canal transport flows• Choice 1: Strict net settling rate module (poor performance,

unused)• Choice 2: Ecological dynamics

§ P uptake, mineralization, particulate settling§ Periphyton, macrophytes, soil dynamics§ Surface water - soil interactions


Slo

ug

hs

Rid

ges

Pra

irie

s

Up

lan

ds

Pla

ins

Below-ground:

water flowsP transformations

macrophyte gain/losspeat gain/loss

Above-ground:

water flowsP transformations

macrophyte gain/lossperiphyton gain/loss

Major vertical mass flows,w/o horizontal flows orinformation linkages

Atmosphere

SaturatedWater

OrganicSoil P&OM

PorewaterP

Floc P&C

UnsaturatedWater

SurfaceWater

Mac

ro-

ph

ytes

NonPhoto-Bio P&C

Photo-BioP&C Standing

Detr. P&C

PeriphytonP&C

SurfaceWater P


Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

• integrated ecosystem approach

• process-orientation avoids statistical “lock-in”of model application

• significant efforts in balancing mechanisticcomplexity vs. parameter availability

• when stocks and rates of multiple variablesperform well, indicative of effective modelstructure and parameterization

• mass balance among fully integrated complexsystem provides checks on system function

• can “calibrate” surface water quality,

• but does not arrive at that calibration atexpense of excessive soil accretion, plantgrowth, etc.

Refinement &Calibration Process

Simple modules,complex spatio-temporal

interactions


Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

• 2 critical Landscape Drivers are 1) hydrology and 2)surface water quality

• PROVIDED that other Landscape Attributes are inLevel 3 to 2 calibration:

• low-intermediate model sensitivity tofeedbacks between surface water quality andsoils & biota

• calibrate hydrology to Level 1 to 2

• calibrate surface water quality to Level 2 to 1

• AND verify performance of other variables toLevel 3 to 2

Refinement &Calibration Process

Levels of calibration

• Level 1: “Nirvana”, w/ hi quality &quantity of data across space & time

• Level 2: “Common”, with varying dataquality/quantity across space & time

• Level 3: “Ballpark”, with sparse or poordata across space and time

• Level 4: “None”, with little/no effort tocompare model to observations


Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

• ecological modules OFF

• model-observed stages: CALM*

• model-observed stages: ELM

• SFWMM-ELM basin flow budgets

• SFWMM-ELM spatial depth & hydroperiod

• if “OK”, turn on all ecological modules

• repeat above steps for ELM

• (generally relatively little regional diffs)

• aim for Level 1-2

• if “OK”, evaluate surface water quality & ecology

CalibrationEvaluation Process

Hydrology

1: Hydrology

*CALM (Conservation Area Landscape Model)= ELM implemented at finer scale in WCA-2A






Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

• surface nutrients

• model-observed TP concentrations: CALM*

• model-observed TP concentrations: ELM

• aim for Level 2-1

• other ecological dynamics

• soil, periphyton, macrophytes

• at minimum, ensure Level 2-3 in CALM*

• at minimum, ensure Level 3-2 in ELM

• if “OK”, fully evaluate/refine system dynamics

CalibrationEvaluation Process 2: “Water Quality”






Nutrients


August 7 Goals


§ Model objectives

§ Model structure




§ Model tools

§ Code

§ Data




Start out “small”...

ELM is scalable in spatial grain and extent


WCA1

WCA2

WCA3A

EvergladesAgricultural

Area

EvergladesNational

Park

Hi

25 km

LakeOkeechobee

N

S-10D

S-10CS-10A

S-7

S-10E

F0

U3

F1

F4F5

2-17

F3

F2

Water Conservation Area 2A

S-146S-145

S-144

S-38S-11C

S-11B

S-11A S-143

LowSoil Phosphorus

Concentration


Model refinement...

1.00 92.00 183.00 274.00 365.004.94

5.72

6.501

1

1

12

2

2

2

3

3

3 3

Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

AnInfiniteLoop?

• Modified:§ surface-ground water

integration§ plant/soil nutrient kinetics§ others....

• Added:§ soil organic P storage§ variable C:P stoichiometry§ soil flocculent layer

• Calibrated v2.1:§ WCA-2A ecosystem

dynamics§ Everglades-wide

hydrology and surfacewater quality


Better Performance

ELM porewater TP: WCA-2A F-transectCALM end-of-simulation '96

0

200

400

600

800

1000

1200

0 2 4 6 8 10 12

Distance from S-10C, Hillsboro levee (km)

Model v2.1 Model v1.0 Unpubl. SFWMD obs.

single point from Reddy et al. 1991 soil core porewater


New Dynamics

ELM floc TP: WCA-2A F-transectCALM end-of-simulation '96

0

200

400

600

800

1000

1200

1400

1600

0 2 4 6 8 10 12

Distance from S-10C, Hillsboro levee (km)

Model v2.1


ELM water management

40

41

42

43 45

4772

63

4666

29 28

70

2221

1915

10

25

16

1768

1112

67

2724

50

71

51

52

5369

37

36

35

34

33

646532 38

39

30

5954

60

Reach ID # on canal-side of levee, or oncanal if dual/no levee

ELM boundary

57

5855

61

62

8079

two separateELM reaches

External seepageborrow*, stagemaintained bycalc'd structure

16

Water ControlStructure(SFWMM flows)

Water ControlStructure(ELM calc'd)

56


Stage monitoring regions/points

50 Kilometres

NESRS3NESRS2

NESRS1

G1502

RUTZKE

EPSWNP207

NP67

THSO

NP201

G620NP205

NP34 NP33

NP206NP36

NP35NP44

NP38

3A-28

3A-4

3B-SE

3-71

3-763A-SW3A-S

3A-NE

3A-11

3A-2

3A-12 3A-3

3A-NW

1-7

1-9

1-8T

2A-159

2A-17

2A-300

2B-Y

ROTTSHOLG

HOL1

HOL2

SHARK


Stage calibration examples (v.2.1)

NP-206 (East ENP)

0.00

0.50

1.00

1.50

2.00

2.50

Jan'80 Jan'81 Jan'82 Jan'83 Jan'84 Jan'85 Jan'86 Jan'87 Jan'88 Jan'89 Jan'90 Jan'91 Jan'92 Jan'93 Jan'94 Jan'95

Observed Land elev. ELM SFWMM

NP-36 (Shark Slough)

0.00

0.50

1.00

1.50

2.00

2.50


Jan'96

Jan'96

Observed Land elev. ELM SFWMM


NESRS1

EPP37

TSB

NP34 NP33

NP36

NP35

S31

3

S12-A-DS333

CA315

S9

S151

CA311L28I

CA38

C123SR84

CA34

CA36

S140 S144-6

CA32

S11A-C

G204-6

G155CA35

CA33

S10A-D

S5

L7L40

5

9 87 6

14

111312

15 16

S6

4

10

S332

S18C

S8 S150

LOX3-16

COOPER.

50 Kilometres

Surface water TP monitoring regions/points

S7


Surface water TP calibration examples (v.2.1)S-11 A-C Mean

0

50

100

150

200

250

300


S-11s observed S-11s simulated

Jan'96

S12 A-D Mean

Jan'80 Jan'81 Jan'82 Jan'83 Jan'84 Jan'85 Jan'86 Jan'87 Jan'88 Jan'89 Jan'90 Jan'91 Jan'92 Jan'93 Jan'94 Jan'95 Jan'96

S-12s observed S-12s simulated

0

50

100

150

200

250

300


Periphyton

Macrophytes

Nutrients Hydrology

Soil

StandingDetritus

• integrated assessment

• data dependent, varying Levels dependingon region and on simulation time period

• aim for Level 2 in general, Level 1-2 insubregions, Level 3

• communicating uncertainty

• show users at least qualitative level ofconfidence for different PerformanceMeasures across space

• taking advice of “Model UncertaintyWorkshop” (Jan 02) panelists on analysis

• when “OK”, release ELM v3.0

CalibrationEvaluation Process 3: Complex system






Nutrients Hydrology

Periphyton

Macrophytes

Standing Detritus

Soil


Hierarchical Sensitivity Analysis

• Large number of parameters

• 1) Evaluate many- parameter response at local (non-spatial) scale

of “unit” model

• 2) Evaluate subset of parameters in simple (WCA-2A) spatial basin

• 3) Evaluate subset of parameters in entire Everglades domain


Summary: Model Performance

• Evaluate ecological process/mechanisms in subregions with highdata (spatial & temporal) quality

• Calibrate both targets (e.g., stage, TP conc.) and rates (e.g., ET, TPuptake)

• To extent possible, extend subregional understanding to otherregions with comparable antecedent conditions & dynamics


August 7 Goals


§ Model objectives

§ Model structure




§ Model tools

§ Code

§ Data




• Modified water deliveries...

• changes nutrient distributions…and influences soils and plants.

• How will the landscape pattern ofperiphyton and macrophytesrespond?

Model Application


Project Alternative Evaluation:Hydrologic Considerations

• For each Alternative, SFWMM provides (via scripted procedures)

§ daily input data on managed flows through water control structures

§ daily stages at boundary cells (ELM code not final for this)

• ELM hydrology (overland, groundwater, canal flows, etc)

otherwise independent of SFWMM

• ELM uses SFWMM data on rainfall, topography, others

• Verify that ELM hydrology is consistent with SFWMM (stage,

hydroperiod, budgets)


Performance measures(web-enabled)

• Regional maps§ Multi- year summaries: TP,

periphyton, vegetation, ...§ Animations (annual or monthly)§ Difference maps


• Track inflows from S-8

• Monthly mean concentrations

• Distribution via overland andcanal flows

Example: tracer flow in surface water

INFLOW


§ Water quality

• TP concentration & load• LOK water tracer

§ Soils

• peat accretion• TP concentration

§ Periphyton

• biomass & community type• tissue TP concentration

§ Macrophytes

• biomass & community type• tissue TP concentration

Performance measures(web-enabled)

50 Kilometres

Many Gradie

nts/Indic

ator Regio

ns

Subregional summaries


• Stormwater Treatment Areas

• Decreased P loading• Altered flows

• Landscape response

Model Application

Historical/currentmanaged flowsinto Everglades

Futuremanaged flowsinto Everglades

Principalmanaged flows

within Everglades 50 Kilometres

N

STA 5

STA 6

STA3/4

STA2

STA1W

STA1E


31- yr Phosphorus Accumulation

0

20

40

60

80

100

120

140

160

180

200

F1 F2 F3-F4 U3 S-12 up S-12 dnWCA-2A Gradient Location ENP Gradient Location

Mea

n a

ccu

mu

lati

on

rat

e(m

g/m

^2/y

r)

Atmosphericload

1995base

2050 base,50 ug/L

2050 base,10 ug/L


Conclusions

• Effectively simulated spatial and temporal interactions in complexecological system

• Calibrated ecological processes in WCA-2A, calibrated hydrologyand water quality throughout domain

• Confidence in current ELM dynamics allows evaluation of regionalsurface water quality

1.00 92.00 183.00 274.00 365.004.94

5.72

6.501

1

1

12

2

2

2

3

3

3 3


Status

• Available NOW for evaluations of surface water quality throughoutthe Everglades

TP in surface water:

monthly, 1979-1995

• Finalizing calibration/refinement to evaluate other ecological(soils, periphyton, macrophytes) responses

• Extending Period-of-Record for calibration/verification thru 2000


Documentation & Review

• Web site§ Data used in model development and model application

• Descriptions of data on web site• All data on ftp link at web site

§ Model structure• Documented in manuscripts and conceptual diagrams at web site• Detailed code descriptions on web site• All source code on ftp link at web site

§ Model results• Post-processed performance measures• Raw output data (to be) on ftp link at web site

• Peer Review§ May 9: USGS Model Informational Workshops, including ELM§ Aug 7: Inter-agency review, organized by CERP-RECOVER Model

Development and Refinement Team§ Soon?: SCT/RECOVER sponsoring independent peer review


August 7 Goals


§ Model objectives

§ Model structure




§ Model tools

§ Code

§ Data




Scalefor alloutputmaps

High

Low

Water depth

Hydroperiod

0.25 km2 cells1.0 km2 cells

Peat depth/elevation

Output E

xamples

Soildecomposition

Porewaterphosphorus

Macrophytebiomass


a) NativePeri: 65 g C m-2

0c) MacBio: 1.1 kg C m-2

b) P SfWat: 100 ug L-1

a) b) c)

Periphyton, phosphorus,and macrophyte patterns


Observed

0.002

Simulated

0.8

Interstitial P (mg L-1)

(ELM version 1.0)

Soil porewater P calibration


●●

●

❑

❑

❑

❍

❍ ❍

0.00

0.25

0.50

0.75

1.00

0 2 4 6 8 10 12

Pea

t ac

cret

ion

(cm

yr-1)

Distance from canal (km)

● Reddy

❑ Richardson

❍ Model

❑

❑

❑

❑❑

❍ ❍❍

❍

❍0.00

0.05

0.10

0.150

- 0 2 4 6 8 10 12Distance from canal (km)

❑ Observed

❍ Model

Su

rfac

e w

ater

P (

mg

L-1)

❑

❑

❑ ❑

❍

❍

❍ ❍0.00

0.25

0.50

0.75

1.00

0 2 4 6 8 10 12

Po

rew

ater

P (

mg

L-1)


❑ Observed

❍ Model

●

●

●

❍

❍

❍

0.00

0.25

0.50

0.75

1.00

1.25

1.50

0 2 4 6 8 10 12

Mac

rop

hyt

e b

iom

ass

(kg

C m-2

)


● Observed

❍ Model

(ELM version 1.0)P, accretion, macrophytes calibration


1973 1976 1982

1987 1991

Vegetation classes

Cattail &sawgrass

Cattail

Sawgrass

Cattail & brush

Cattail invasion:Water Conservation Area 2A


Vegetation change calibration

Observed

Cattail 90+%Cattail 50+%Cattail 10+%Sawgrass etc.

} lumpedfor simulation

Simulated

1991

Observed

Simulated

1995(ELM version 1.0)


Land use change:South Florida

1900

LakeOkeechobee

1973

LakeOkeechobee

Agriculture

Cur

rent

Eve

rgla

des

Sys

tem

Urb

an

U

rban


N

50 Kilometres

1 556 Open Water

2 1705 Sawgrass plain3 802 Sawgrass ridge4 834 Sawgrass slough deep5 848 Sawgrass slough shallow6 982 Sawgrass marl prairie

7 518 Gramminoid mix 8 475 Wet prairie9 9 Slough w/ gramminoids10 26 Slough w/ non-gramminoids

11 157 Cattail (high density)12 43 Cattail (med density)13 35 Cattail (low density)

14 90 Muhly grass

15 58 Salt marsh

16 25 Hardwood-mixed17 861 Swamp forest18 477 Mangrove forest19 99 Buttonwood forest

20 290 Pineland savannah21 594 Cypress savanna

22 89 Brush23 95 Hardwood Scrub 24 59 Cypress scrub 25 489 Mangrove Scrub 26 121 ButtonWood Scrub

27 35 Brazilian Pepper 28 9 Melaleuca 29 13 Human Influence

HabitatCode

Area(km2) Description

ELM vegetation classes, ca. 1995


0

250

500

750

1000

50 Kilometres

mg P • kg-1 soil(0-10 cm depth)

Soil Total Phosphorus (ca.1995)


0-2 km2-4 km

TamiamiTrail

Mean P Accumulation

0

10

20

30

40

50

mg

/m^

2 p

er

ye

ar

1995 Base 2050 Base Modwaters Alt120-2 km 2-4 km 0-2 km 2-4 km 0-2 km 2-4 km

Rainfall Loading

50 Kilometres

Example: P Accumulation in upper NESS

Documents

Carl Fitz Naiming Wang Everglades Landscape Fred Sklar