Patterns of water use by the riparian tree Melaleuca argentea in …€¦ · Melaleuca New Phytologist 192: 664–675. I was the major contributor to this paper, and completed the

Patterns of water use by the riparian tree Melaleuca

argentea in semi-arid northwest Australia

Elizabeth Helen McLean

BSc (Hons)

This thesis is presented for the degree of Doctor of Philosophy

The University of Western Australia

School of Plant Biology

2014

i

ABSTRACT

This thesis examines the water use physiology of the riparian tree Melaleuca argentea, and

the ways in which this species may respond to anthropogenic disturbances to hydrologic

processes. The research investigated patterns of water use by M. argentea in relation to key

aspects of the riparian environment, including in response to flooding and drought as well

as to groundwater drawdown.

Study sites focused on the remote Pilbara region of northwest Australia

(encompassing ~500,000 km2, ~1700 km north of Perth, the nearest major city). The

Pilbara is currently undergoing rapid economic development and population growth due to

expansion in the mining sector. There are concerns that intensification of mining efforts,

which often involves dewatering for extraction of ore below the water table, and abstraction

of water to supply regional populations and industry could produce undesirable

environmental outcomes of a long-term or even permanent nature. However,

comparatively little is known of the functioning and especially the water requirements of

many of the key species of the Pilbara and the ecosystems in which they occur. M. argentea

is considered an obligate phreatophyte and is confined to riparian zones where there is

permanent surface and near-surface water. M. argentea is thus very likely to be vulnerable

to changes in groundwater levels, recharge patterns and altered water dynamics that may

arise through abstraction and dewatering. Clearly, an understanding of tree water use

patterns of M. argentea will provide much-needed information for ecosystem management,

while also contributing to our fundamental understanding of riparian tree water use

physiology along ephemeral waterways and their resilience to both natural and imposed

dynamics in water availability.

Extreme vapour pressure deficits (VPD) and highly variable rainfall are typical of the

Pilbara region, but M. argentea typically has access to an unlimited supply of water.

Accordingly, water use patterns of M. argentea trees as determined by sap flow

measurements appeared to be largely driven by the atmospheric variables of VPD, wind

speed and light intensity. However, multiple regression analysis revealed that 20 to 30% of

seasonal variation in water use was not directly attributable to seasonal climatic

ii

fluctuations. Results suggest that M. argentea utilises internal water storage over both

diurnal and seasonal cycles, perhaps due to rates of water transport reaching their maximum

limit under high evaporative demand.

While M. argentea mostly occurs where there is access to permanent water, the

distribution of water in riparian zones in arid and semi-arid environments nevertheless

exhibits considerable spatial and temporal variability. M. argentea displayed plastic root-

level strategies to maintain high rates of resource capture under heterogeneous and dynamic

distributions of water around the root system. Sap flow measurements in the field suggested

that roots redistribute water from the saturated zones into drier regions, maintaining fine

root function in the dry zones for periods of at least six months. When the spatial

distribution of water was manipulated in glasshouse experiments, the roots rapidly

compensated for drying of one-third of the root system by increasing fine root hydraulic

conductance, aquaporin content and growth in the hydrated zone. I suggest that these

compensatory responses are likely to occur in a wide range of other species, particularly

those adapted to heterogeneous environments. During partial root zone drying of M.

argentea there was no evidence of root-derived signals inducing stomatal closure, as

stomatal conductance, water use (measured gravimetrically) and tissue abscisic acid (ABA)

content remained unchanged. Plastic root-level strategies and internal water storage appear

to enable M. argentea to maintain high rates of water and nutrient uptake and

transpiration, and therefore high rates of photosynthesis and growth, in an environment

characterised by high VPD above ground and heterogeneity below ground.

Since M. argentea is confined to the wettest positions in the landscape, it was

expected this species would be highly sensitive to disturbances that reduce water

availability, such as groundwater drawdown, but tolerant of disturbances increasing water

availability, such as prolonged flooding during discharge of water into creeklines. In a

glasshouse experiment exposing M. argentea seedlings to a variety of hydrologic conditions,

the seedlings were highly tolerant of flooding; rates of photosynthesis and most aspects of

shoot growth were undiminished after eight weeks of waterlogging, although total leaf area

was slightly lower than in drained plants. Roots responded plastically to waterlogging, with

reduced total root biomass but proliferation of shallow roots. In contrast, under even a mild

water stress, shoot growth of M. argentea seedlings rapidly ceased and greater biomass was

iii

allocated to roots, and leaf shedding soon occurred under more severe water stress. M.

argentea seedlings are therefore unlikely to establish in locations without abundant water

supply, or where near-surface water availability is reduced, while rates of recruitment could

increase in flooded locations under groundwater discharge.

The responses of larger trees to groundwater abstraction were explored at abstraction

and control sites at the Yule River Borefield on the coastal plain of a large ephemeral river,

50 km southwest of Port Hedland, in northwest Australia. The water table at the

abstraction site fell by 4.3 m over a 13-month period, compared with a 1.5 m decline over

the same period at an upstream control site. Large, mature trees were generally tolerant of

the abstraction, reflected by only small increases in leaf δ13C and slightly reduced sap flow

rates relative to the control trees, primarily when the depth to the water table exceeded

5.5 m. However, smaller, younger trees appeared more susceptible to lowering of the water

table, most likely due to reduced access to water as a result of their shallower root systems.

This finding highlights the potential vulnerability of M. argentea recruitment to long-term

reductions in water availability in the riparian zone while also demonstrating some

resilience of more mature trees to the range of conditions experienced throughout the study

period.

Overall, M. argentea displays highly plastic root-level responses to heterogeneous

water availability and to waterlogging, facilitating high rates of water use and growth in the

riparian wetland habitats of the Pilbara. Mature M. argentea trees also appear to tolerate

groundwater drawdown of at least several metres, most likely by employing the same plastic

root strategies to access deeper water. A particularly useful finding is that M. argentea can

also withstand short periods of severe drought, by adopting a ‘waiting’ strategy of ceasing

growth and shedding leaves to avoid moisture loss, a state from which they can then

recover. Nevertheless, M. argentea populations are unlikely to thrive under large and

prolonged reductions in water availability. This study clearly demonstrates the need for

careful consideration of the complex responses of riparian tree species to changes in water

availability, when planning water management in the riparian environments of the Pilbara.

iv

ACKNOWLEDGEMENTS

This PhD would not have been possible without the assistance and advice I received from

numerous people, and I am immensely grateful to them all.

Thank you most of all to my supervisors, Pauline Grierson, Martha Ludwig and

Mark Adams, for their guidance throughout this PhD from project conception to the final

thesis editing. To my primary supervisor Pauline, thank you for keeping me on track while

also allowing me the independence to pursue the research topics and directions of my

choosing. Through the breadth of your knowledge I always found fresh perspectives and

new ideas during our discussions. To Martha, thank you of course for your expert advice,

but thank you also for your general encouragement, support and friendly chats, even in

topics outside your field.

I thank Rio Tinto Pty Ltd and BHP Billiton for collaboration and support of much

of the work contained in this thesis, funded through an Australian Research Council

Industry Linkage grant (LP0776626). Thanks particularly to Tim Eckersley, Sally Madden,

Sam Luccitti and Nicole Gregory at Rio Tinto, and to Stephen White and Esme Winks at

BHP, for their advice and assistance, including in the organising of field sites, sampling

trips at Marillana Creek, and sharing datasets. I also thank the Water Corporation of

Western Australia and the Department of Water for collaboration and support of the work

at the Yule River. Thanks go especially to Jacquie Bellhouse and Andrew Jones at the Water

Corporation, and to Mike Baimbridge at the Department of Water, for their assistance

with data collection in the field and many useful discussions. I am also grateful to have

received a Prescott postgraduate scholarship from the University of Western Australia.

Thank you to Sebastian Pfautsch, Mike Kemp, Tim Bleby and Steve Burgess for

extensive advice and practical assistance with sap flow equipment, installation and

calculations. I also acknowledge Robert Willows and Tony Jerkovic at Macquarie

University for conducting the abscisic acid assays presented in Appendix C. To the many

people who helped me in the field, lab and glasshouse – chiefly Tom Wright, Chloe

Flaherty, Steve Naughton, Bruno Rosado, Gerald Page, Rachel Argus, Doug Ford, Kate

Bowler and Mark Waters – thank you for your diligence and companionship during long

v

hours of repetitive work, and in the heat of the Pilbara summer. I am grateful to the

workshop and glasshouse staff, especially Rob Creasy and Ray for their help in solving the

problems surrounding intricate split-root experiments on trees in huge pots. Thanks to all

the members of the ecosystems research group for your friendship, especially Gerald, Doug,

Ali and Alex with whom I have shared the PhD journey. I will remember these years most

of all for the coffees and lunches by the river. Thanks to Gerald Page for the many helpful

lunchtime discussions of plant physiology, among other topics.

And finally, thank you to my family for your patience, understanding and interest in

my work. Most of all to Mark – thank you for your seemingly endless patience and love,

and undying IT technical support. This has of course been only one small part of our life

together, but your willingness to move to the other side of the world and to support me

throughout this PhD means more to me than you know.

vi

STATEMENT OF CANDIDATE CONTRIBUTION

This thesis is my own work, except where otherwise acknowledged, and has been

substantially completed during the course of my enrollment for the degree of Doctor of

Philosophy at the University of Western Australia. It has not previously been accepted for

any other degree, at this or any other institution.

This thesis contains published work; Chapter 5 was published as: McLean E.H.,

Ludwig M. & Grierson P.F. (2011) Root hydraulic conductance and aquaporin abundance

respond rapidly to partial root-zone drying events in a riparian Melaleuca species. New

Phytologist 192: 664–675. I was the major contributor to this paper, and completed the

data collection, analysis and manuscript preparation. My co-authors (also my PhD

supervisors) provided methodological, technical and editorial advice. I have permission

from all co-authors to include this work in my thesis.

vii

CONTENTS

CHAPTER ONE: General introduction ..................................................................... 1

ECOHYDROLOGY OF the DRYLANDS of NORTHWEST AUSTRALIA .............................. 3

Patterns of rainfall and recharge across the Pilbara .......................................................................... 4

Hydrologic disturbance and management of groundwater resources in the Pilbara .................. 6

ECOPHYSIOLOGY OF RIPARIAN TREES IN DRYLAND REGIONS .................................... 9

OBJECTIVES OF THIS THESIS ....................................................................................................... 13

CHAPTER TWO: Seasonal patterns of water use by the riparian tree Melaleuca

argentea in semi-arid NW Australia ............................................................................ 15

INTRODUCTION ................................................................................................................................. 15

MATERIALS & METHODS ............................................................................................................... 18

Study sites ............................................................................................................................................. 18

Meteorological data ............................................................................................................................ 20

Tree water use ...................................................................................................................................... 20

Canopy physiological measurements ............................................................................................... 22

Statistical analysis ................................................................................................................................ 22

RESULTS .................................................................................................................................................. 23

Seasonal variation in tree water use and meteorological conditions ........................................... 23

Diurnal patterns of tree water use physiology across seasons ....................................................... 24

Relationships of tree daily water use with meteorological conditions ........................................ 26

The effects of site, rainfall and season on tree water use .............................................................. 30

Tree hydraulic conductance .............................................................................................................. 33

DISCUSSION .......................................................................................................................................... 34

Factors determining water use of M. argentea ............................................................................... 34

Seasonality of water use in M. argentea ........................................................................................... 36

CHAPTER THREE: Quantifying water fluxes at tree and stand scales from sap flow

measurements in the riparian tree Melaleuca argentea ................................................ 43

INTRODUCTION ................................................................................................................................. 43

MATERIALS & METHODS ............................................................................................................... 47

Study sites ............................................................................................................................................. 47

Sap flow measurement ....................................................................................................................... 47

viii

Tree sapwood area ............................................................................................................................... 48

Tree hydraulic conductance ............................................................................................................... 48

Analysis of variation in sap flow ........................................................................................................ 48

RESULTS ................................................................................................................................................... 49

Sap flow variation within the sapwood ............................................................................................ 49

Temporal variation in sap flow ......................................................................................................... 52

Variation between trees and sites ...................................................................................................... 53

DISCUSSION ........................................................................................................................................... 57

CHAPTER FOUR: Redistribution of water through root systems of a semi-arid

riparian tree during a flooding-drying cycle in an intermittent waterway.................. 64

INTRODUCTION ................................................................................................................................. 64

METHODS ............................................................................................................................................... 66

Study site ............................................................................................................................................... 66

Meteorological measurements ........................................................................................................... 66

Sap flow measurements ...................................................................................................................... 66

Statistical analysis ................................................................................................................................. 69

RESULTS ................................................................................................................................................... 69

Patterns of root sap flow with drying conditions ........................................................................... 69

Root responses to rainfall ................................................................................................................... 72

DISCUSSION ........................................................................................................................................... 76

CHAPTER FIVE: Root hydraulic conductance and aquaporin abundance respond

rapidly to partial root-zone drying events in a riparian Melaleuca species ................. 80

INTRODUCTION ................................................................................................................................. 80

MATERIALS & METHODS ................................................................................................................ 83

Plant material and growth conditions .............................................................................................. 83

Experimental design ............................................................................................................................ 83

Stomatal conductance and tissue water status ................................................................................ 85

Plant water use ..................................................................................................................................... 85

Root hydraulic conductance .............................................................................................................. 86

Xylem sap osmolarity .......................................................................................................................... 86

Root growth.......................................................................................................................................... 87

Protein extraction and immunoblotting .......................................................................................... 87

Statistical analysis ................................................................................................................................. 88

RESULTS ................................................................................................................................................... 89

ix

Stomatal conductance and leaf water status were unaffected by PRD ....................................... 89

Survival of the dried root portion .................................................................................................... 90

Water uptake and Lp of the hydrated roots increased during PRD ............................................ 91

Growth of fine roots in the hydrated root portion ........................................................................ 94

Abundance of PIP1 and 2 aquaporins in the hydrated roots was altered by PRD .................. 94

DISCUSSION .......................................................................................................................................... 94

Adaptive responses to heterogeneous water supply ....................................................................... 94

Adjustments in root hydraulic conductance during PRD ............................................................ 97

Conclusions ........................................................................................................................................ 100

CHAPTER SIX: Morphological and physiological adjustments across a hydrologic

gradient by seedlings of a semi-arid riparian Melaleuca ............................................ 101

INTRODUCTION ............................................................................................................................... 101

MATERIALS & METHODS ............................................................................................................. 104

Plant material and growth conditions ........................................................................................... 104

Experimental design ......................................................................................................................... 104

Plant growth and biomass ............................................................................................................... 105

Morphology and anatomy of leaves and roots ............................................................................. 105

Plant physiological status ................................................................................................................. 107

Statistical analysis .............................................................................................................................. 108

RESULTS ................................................................................................................................................ 108

Seedling growth and biomass allocation ....................................................................................... 108

Seedling leaf shedding and survival ................................................................................................ 111

Morphological responses.................................................................................................................. 112

Physiological responses ..................................................................................................................... 114

DISCUSSION ........................................................................................................................................ 118

Growth and responses under saturated conditions ..................................................................... 118

Drought responses ............................................................................................................................ 121

Implications for response thresholds under field conditions ..................................................... 124

CHAPTER SEVEN: Response of Melaleuca argentea trees to ground water

abstraction from the coastal plain aquifer of a large ephemeral river ........................ 126

INTRODUCTION ............................................................................................................................... 126

MATERIALS & METHODS ............................................................................................................. 129

Study site ............................................................................................................................................ 129

Groundwater levels and abstraction ............................................................................................... 131

x

Soil moisture ....................................................................................................................................... 131

Isotope analysis of xylem sap, ground and soil water .................................................................. 133

Tree water use .................................................................................................................................... 134

Leaf physiological measurements .................................................................................................... 134

Statistical analysis ............................................................................................................................... 135

RESULTS ................................................................................................................................................. 135

Water table recession with groundwater abstraction ................................................................... 135

Tree water uptake from soil, surface and ground water sources ................................................ 135

Tree water use .................................................................................................................................... 136

Leaf physiology of M. argentea in relation to lowering of groundwater .................................. 138

DISCUSSION ......................................................................................................................................... 139

Tree water use patterns in response to abstraction ....................................................................... 143

Effects of abstraction on small trees ................................................................................................ 144

Implications for vegetation management under increased abstraction ..................................... 145

CHAPTER EIGHT: General discussion: temporal and spatial patterns of water use

by Melaleuca argentea ................................................................................................ 146

ADAPTATIONS OF M. ARGENTEA ............................................................................................... 146

Extreme VPD and temperature conditions ................................................................................... 146

Heterogeneous belowground environment ................................................................................... 150

ECOSYSTEM SCALE IMPLICATIONS OF WATER USE PATTERNS ............................... 155

POTENTIAL IMPACTS OF ALTERED WATER REGIMES ON M. ARGENTEA ............. 157

CONCLUSION ..................................................................................................................................... 160

REFERENCES .......................................................................................................... 161

APPENDIX A: Verification of methods used to measure stomatal conductance in the

mid-canopy of mature trees ....................................................................................... 193

APPENDIX B: Supplementary information for Chapter Five .................................. 196

APPENDIX C: Tissue ABA content during partial root zone drying (Chapter Five)

................................................................................................................................. 200

APPENDIX D: Examples of micrographs used for anatomical measurements in

Chapter Six ............................................................................................................... 202

1

CHAPTER ONE: General introduction

The aim of this thesis is to increase understanding of the patterns and constraints in water

use by the tree Melaleuca argentea W. Fitzg. (silver leaved paperbark or cadjeput), in the

semi-arid Pilbara region of northwest Australia. In particular, I sought to characterise the

water use physiology of M. argentea and to assess the potential impacts of hydrological

disturbances upon this species. The thesis thus examines aspects of the ecohydrological

functioning of riparian ecosystems of the Pilbara.

Rivers and streams in arid and semi-arid regions (drylands) provide oases of surface

and groundwater within an otherwise water-limited environment. The relative abundance

and accessibility of water in these riparian zones support ecosystems of high biodiversity

and productivity, including large trees (Sabo et al., 2005; Sheldon et al., 2010; Figure 1.1).

However, relatively little is known about the ecological water requirements and water use

characteristics of riparian vegetation across northern and inland Australia.

The Pilbara is remote and mostly sparsely populated, but is currently experiencing

rapid population growth and development of resource industries. The region contains

internationally significant iron ore deposits, as well as smaller deposits of other metals

including gold, copper and manganese, and offshore oil and gas reserves (Department of

Regional Development and Lands & Pilbara Development Commission, 2011). The

ongoing development is increasing the demand for water supplies, as well as impacting on

hydrologic regimes across the region (Abbott et al., 2010; Department of Water, 2010).

Consequently, there is an urgent need for a clearer understanding of the Pilbara’s riparian

vegetation, and its links with the local hydrology, to guide the management of water

resources. Among the currently unknown aspects are the resilience of key species to water

abstraction, and their physiological functioning against the natural background of episodic

floods and droughts.

Melaleuca argentea occurs widely throughout northern Australia, and is one of the key

riparian tree species in the Pilbara region (Figure 1.1c–e). M. argentea is found exclusively

in locations with permanent water either at or close to the surface (Masini, 1988). The

Chapter One

2

Fig

ure

1

.1:

Wa

terw

ay

s in

th

e

Pil

ba

ra

reg

ion

, su

pp

ort

ing

ta

ll,

de

nse

, w

oo

dy

ve

ge

tati

on

, in

clu

din

g

sta

nd

s o

f M

ela-

leuc

a ar

gent

ea

tre

es.

D

ense

rip

aria

n ve

geta

tion

alon

g (a

) M

arill

ana

Cree

k (v

isib

le a

long

the

base

of t

he h

ills)

, and

(b)

Wee

li W

olli

Cree

k (p

hoto

grap

h Ri

o Ti

nto

Pty

Ltd)

. (c

& d

) A

sta

nd o

f M

. arg

ente

a tr

ees

frin

ging

the

wid

e, s

andy

rive

r bed

of

the

coas

tal p

lain

sec

tion

of th

e Yu

le R

iver

, a

larg

e in

term

itten

t riv

er. P

hoto

grap

hed

in

June

200

9, a

few

mon

ths

afte

r a

flood

ing

even

t, w

ith t

empo

rary

sur

face

poo

ls s

till

pres

ent

in t

he r

iver

bed

. (e

) Yo

ung

M.

arge

ntea

tre

es a

t a

perm

anen

t po

ol i

n D

ales

Gor

ge, K

araj

ini N

atio

nal P

ark.

(a)

(b)

(c)

(d)

(e)

General introduction

3

presence of M. argentea is also linked with places of cultural significance, as sites of

permanent water in particular are vital to indigenous identities, beliefs and livelihoods

across the Pilbara (Barber & Jackson, 2011). Owing to its dependence on permanent water,

M. argentea might be expected to be highly vulnerable to changes in hydrologic regimes

associated with regional development. An understanding of the water requirements and

functioning of this species may therefore be of particular value in ecological monitoring,

and in identifying critical thresholds of hydrological disturbance.

This chapter outlines the processes that determine patterns of temporal and spatial

water availability in the Pilbara, and the ways in which the hydrology may be affected by

the disturbances associated with regional development, such as groundwater abstraction,

and mining of ore bodies below the water table. I then provide an overview of the current

understanding of the water use physiology of riparian trees in dryland zones, with particular

reference to M. argentea. This background provides context for the experimental rationale

that is developed in chapters 2–7 of this thesis, including the potential applications of this

research to improve water management strategies amidst rapid development of semi-arid

and arid regions of Australia.

ECOHYDROLOGY OF THE DRYLANDS OF NORTHWEST AUSTRALIA

Ecohydrology refers to the understanding of the hydrologic mechanisms that underlie the

distribution, structure and functioning of ecosystems, including the effects of biotic

processes, such as tree water use, on the water cycle (Rodriguez-Iturbe, 2000; Nuttle,

2002). The availability of water to plant roots, as soil moisture and accessible groundwater,

is one of the key links between climate, hydrology and vegetation. Water availability is a

major determinant of terrestrial vegetation structure and productivity on a global scale (Box

& Fujiwara, 2009), but the coupling between water and vegetation is particularly close in

dryland environments (Caylor et al., 2006; Jackson et al., 2009b). Drylands are drought

prone regions, defined by an annual potential evaporation rate that exceeds annual

precipitation (Newman et al., 2006). Hot dryland regions, including northwest Australia,

are also typically subject to extreme variability in rainfall patterns both within and among

years (Schwinning & Sala, 2004; Bunn et al., 2006).

Chapter One

4

Patterns of rainfall and recharge across the Pilbara

Droughts, floods and violent storms are not exceptional events across northwest Australia,

rather they are integral features of the region with major roles in shaping the vegetation. In

the Pilbara, temperatures and vapour pressure deficits (VPD) can be extreme, and rainfall is

highly erratic (Figure 1.2). Rainfall patterns are dominated by tropical low pressure systems

and cyclones between December and April, with smaller amounts contributed by

thunderstorms (Leighton, 2004). Rainfall averages 200–350 mm per year, but inter-annual

variation is considerable, for instance, ranging from 44 mm to 630 mm total annual rainfall

at Port Hedland during 1943–2012 (Bureau of Meteorology, 2013; Figure 1.2). Long

periods of little rainfall are punctuated by large, episodic rainfall events, during which over

100 mm commonly falls within a single day. High evaporation rates also contribute to the

aridity of the region, with annual potential evaporation up to 3000 mm, exceeding mean

annual rainfall by as much as ten-fold (Dogramaci et al., 2012).

As is common in warm dryland regions, most waterways in the Pilbara are ephemeral,

typically with short-lived, high velocity flows, occurring only following large rainfall events

(Johnson, 2004; Young & Kingsford, 2006). The Pilbara contains seven major river

catchments, with numerous smaller creeklines (Department of Water, 2010; Figure 1.3).

High volume and velocity floods erode material from the rugged inland ranges, much of

which is deposited on the coastal plains, producing extensive floodplains with deep, duplex

soils. On the coastal plain, riverbeds can be very wide (up to 10 km), with braided

Figure 1.2: Average climatic conditions and inter-annual rainfall variability recorded at Port

Hedland, in the Pilbara region. (a) Mean monthly rainfall (bars) and temperature minima and maxima (lines), over the period 1948–2011. (b) Actual monthly rainfall during 2005–2010. Data from the Australian Bureau of Meteorology.

0510152025303540

0

20

40

60

80

100

120

Temperature (°C)Ra

infa

ll (m

m)

J F M A M J J A S O N D0

100

200

300

400

2005 2006 2007 2008 2009 2010

Rain

fall

(mm

)

(a) (b)


5

channels. Waterways in the ranges tend to be more confined (typically less than 1 km

wide), often within deep rocky gorges, although shallower sections with floodplains are also

present (Haig, 2009; McKenzie et al., 2009).

Groundwater is present throughout the Pilbara region. Beneath the coastal

floodplains groundwater typically occurs within the alluvium, often with underlying

calcrete or limonite, while aquifers in the ranges may be of porous calcrete or pisolite,

unconsolidated sedimentary rock (valley fill of alluvium and colluvium), or fractured

sedimentary and igneous rock (Johnson & Wright, 2001; Johnson, 2004). Groundwater

recharge occurs primarily through floods caused by large rainfall events, with little or no

contribution from rainfall events of less than 20 mm (Dogramaci et al., 2012). Recharge is

typically rapid; flows through the waterways saturate the shallow alluvium of valleys, gorges

and coastal plains within days, followed by seepage into fractured aquifers within weeks

(Dogramaci et al., 2012). The extent of aquifer recharge depends on both the size and

duration of the flows. Groundwater levels decline naturally between recharge events,

through discharge and evapotranspiration, with recession most rapid when the aquifer is at

Robe River

Cane River

We e li Wol

li C

kMarrillana CkMarsh

Jigalong C kTur e

e Ck

Turee Ck East

Maitland R

Fortesc

ue S

th

Yule River

Geo

rge

R

Sh erlock RiverPeaw

ah R

Turner RiverDe Grey River

Oak

ove

r Ri

ver

Talga R

Davis Rive

r

Nullagi

ne R

Yiga

long

Ck

Tongo

lo

l o C

k

MARBLE

ROEBOURNEKARRATHADAMPIER

ONSLOW

NEWMAN

PORT HEDLAND

TOM PRICE

Fort

escu

e Riv er

Ashburton River

Henry Rive

r

Har

d ing

R

Yannar ie River

Rouse Ck

Duck Ck

W Str

ell e

y R

E St

relle

y R

Shaw

River

Coo

ngan

R

West Yule

R

Figure 1.3: Major waterways of the Pilbara region. Modified from Department of Water (2010) and Van Vreeswyk et al. (2004). Locations of the field sites used for this study are indicated in red.

F ortescue River

Har dey River

BAR

Waterway Catchment boundary 100 kmTown Study site

Chapter One

6

its highest level, and slowing as the depth to groundwater increases (Antao & Braimbridge,

2009).

Along sections of many waterways, a saturated zone is commonly present within

metres of the riverbed surface, and surface pools may also feature. Most pools are superficial

and disconnected from the groundwater; they are maintained only by ongoing rainfall, and

dry out during longer drought periods (Fellman et al., 2011; Dogramaci et al., 2012).

Other pools are fed by groundwater springs and thus are essentially permanent, such as the

Millstream pools on the Fortescue River (Antao & Braimbridge, 2009). A single creekline

may contain both superficial pools largely isolated from the water table, and pools with

consistent input from shallow alluvium water (Fellman et al., 2011).

The shallow groundwater and ‘waterholes’ of the Pilbara’s riparian zones often

support unique ecosystems, including large trees and dense vegetation, in stark contrast

with the surrounding landscape. Most of the vegetation of the Pilbara is xerophytic,

displaying a range of adaptations to the predominantly arid conditions; spinifex grassland

(hummock grasslands dominated by Triodia) is the most common vegetation type, with

large areas also occupied by open Acacia shrublands and low woodlands (Beard, 1975;

McKenzie et al., 2009). Floodplains are mostly colonised by grasses, and by mangroves on

the coastal flats (McKenzie et al., 2009). Most waterways of the Pilbara support woodlands

of Eucalyptus camaldulensis and E. victrix, while M. argentea is present only in locations

with permanent surface or shallow ground water, where the trees may form denser forests

and closed woodlands (Masini, 1988; McKenzie et al., 2009). Waterways also support

unique endemics, such as the Millstream fan palm (Livistonia alfredii), a relict species

restricted to a few waterhole sites within the Pilbara (Antao & Bambridge 2009). Many of

the wetland and pool ecosystems of the Pilbara are of high conservation value and are

classified as nationally significant, such as the Millstream pools, the gorges of Karijini

National Park, and the Ramsar nominated Fortescue Marsh (Environment Australia, 2001;

Jackson et al., 2009a).

Hydrologic disturbance and management of groundwater resources in the Pilbara

Many fast-growing populations around the world are located in dryland regions, including

in southwestern North America, northern Africa, northern China and southern India,

creating particular challenges in sourcing and managing water with scarce and


7

unpredictable supplies (Batchelor et al., 2003; Newman et al., 2006; Scanlon et al., 2006;

Mays, 2009). Securing water supplies frequently involves hydrological disturbance, in the

form of groundwater abstraction, and the modification of river flow patterns. Most of the

species that inhabit the wetland and riparian zones of drylands, including large riparian

trees, are highly dependent on the ground and surface water and could not otherwise

survive in an arid environment, and thus can be highly sensitive to any disturbance

affecting native patterns of water availability (Ridolfi et al., 2006).

The potential effects of hydrological disturbance on dryland riparian ecology are

clearly demonstrated in rivers with long histories of extensive modification. Reducing flow

variability, to provide more consistent, secure water supplies and to limit destructive floods,

induces common sets of ecological changes (Poff et al., 2007; Poff & Zimmerman, 2010).

For example, in semi-arid southwestern USA, dams and diversions are present on most

waterways (Stromberg, 2001), leading to reduced riparian biodiversity, declines in native

Populus tree species due to recruitment failure, and facilitating invasion by exotic Tamarix

tree species (e.g. Busch & Smith, 1995; Tiegs et al., 2005; Williams & Cooper, 2005;

Merritt & Poff, 2010). On the Murray River in southeastern Australia, the construction of

dams and numerous weirs has reduced flow magnitudes to the lower reaches, and has

permanently flooded some previously ephemeral wetlands, while preventing flooding to

others (Kingsford, 2000; Walker, 2006). Among the documented ecological effects of these

altered hydrologic conditions are lack of recruitment by native trees such as Eucalyptus

largiflorens, invasion by exotic weeds, migration of plants normally confined to floodplains

into the channel regions, and declining numbers of water birds, native fish and

invertebrates (e.g. George et al., 2005; Walker, 2006; Catford et al., 2011). It is clear that

the native flow patterns, including the sizes, velocities and frequencies of floods, are integral

aspects of riparian environments, and are vital in maintaining riparian ecosystem

functioning and biodiversity.

Given the scarcity of permanent surface water across northwest Australia,

groundwater is the primary water source used to supply residential and industry needs.

Ongoing industry and population growth across the Pilbara will require sourcing of

additional water supplies, expected to be achieved primarily through further groundwater

developments (Haig, 2009; Department of Water, 2010). Greatest demand for secure and

Chapter One

8

reliable water supply is concentrated in towns along the Pilbara coast, including Karratha,

Dampier and Port Hedland (Figure 1.2). A large proportion of the water is thus sourced

from borefields on the coastal river floodplains, where groundwater typically occurs within

the alluvium (Johnson, 2004). In the western Pilbara, water is also obtained from the

Millstream dolomite aquifer, located south of Karratha on the Fortescue River, and from

Harding Dam, south of Roebourne on the Harding River (Antao & Braimbridge, 2009).

In the Hamersley Basin of the inland Pilbara, water abstraction occurs mainly in

association with iron ore mining. Mining is increasingly focused on ore bodies located

below the water table that occur as paleoalluvial deposits beneath modern day drainage

lines. Localised hydrological disturbances are created by the abstraction of large volumes of

groundwater to access the ore, and the subsequent disposal of the water elsewhere in the

landscape (Abbott et al., 2010). Since iron ore mining occurs inland, the long distances

make it impractical to use ore dewatering abstraction for town water supplies on the coast.

While some of the water is utilised by mining operations, the volumes abstracted are

usually far in excess of onsite requirements (Abbott et al., 2010). In 2008 approximately 60

GL of water in total was consumed in the Pilbara, while a further 66 GL of groundwater

was released to the environment during ore dewatering; both these figures are expected to

double by 2030 (Department of Water, 2010).

The riparian and wetland ecosystems of the Pilbara, most of which are considered

groundwater dependent, are likely to be highly vulnerable to groundwater alterations.

Declines in the water table during abstraction, and the long term flooding of riparian zones

during discharge of excess water, are both disturbances that can adversely impact associated

riparian vegetation (Hatton & Evans, 1998). However, the nature of responses to

disturbance, and the degree of tolerance and resilience to different disturbance types, are

still largely unknown for most of the Pilbara’s riparian species.

Although their frequency and magnitude is increasing, the impacts of groundwater

alterations in the Pilbara remain fairly localised, with many of the Pilbara’s waterways still

largely unmodified, in contrast with most other semi-arid regions of the world. The

floodwaters of some dryland rivers such as the Euphrates, Nile and Huang He (Yellow)

rivers, have been harvested and managed since ancient times; however the past 150 years

have seen the management of rivers on an unprecedented scale (Petts, 2005). The flow


9

patterns have now been substantially altered in the majority of dryland rivers, through

dams, diversions, abstraction, and complete flow regulation, to sustain irrigated agriculture,

domestic and industry water supplies, and hydroelectric plants (Kingsford et al., 2006). The

riparian zones of the Pilbara have been impacted by widespread grazing and trampling by

livestock, and invasion of exotic weeds (Van Vreeswyk et al., 2004; Grice, 2006), but the

flow patterns and ecosystems of many waterways are still relatively close to their ‘wild’ state

(Halse et al., 2007; Department of the Environment, 2008). The Pilbara region therefore

provides increasingly rare opportunities to study the natural dynamics of semi-arid riparian

systems.

ECOPHYSIOLOGY OF RIPARIAN TREES IN DRYLAND REGIONS

Understanding water fluxes through trees is a crucial component in developing an

understanding of the ecohydrological functioning of riparian systems (Asbjornsen et al.,

2011). Trees can transpire large volumes of water, with estimates of up to 200 litres per day

in a range of species, for trees of approximately 20 m height (e.g. Wullschleger et al., 1998;

Schaeffer et al., 2000; Gazal et al., 2006). The extensive root systems of trees can also re-

distribute water throughout soil profiles, linking deeper water with surface soil layers and

the atmosphere (Domec et al., 2010). Consequently, trees substantially influence whole

ecosystem productivity, patterns of evapotranspiration (ET), runoff, groundwater levels,

and climates (Lee et al., 2005; Amenu & Kumar, 2008; Bonan, 2008; Scott et al., 2008;

Domec et al., 2010).

Plant water use is determined by water availability to the roots, the evaporative

demand of the atmosphere, and the hydraulic resistance and capacitance of the flow

pathway through the soil, plant tissues, and leaf boundary layer (e.g. Sperry et al., 2003).

The leaves, fine roots and rhizosphere are frequently the most vulnerable points in the flow

pathway, where cavitation is most likely to occur under high flux rates (Sperry et al., 2002;

Sperry et al., 2003; Brodribb et al., 2007). The factors that limit transpiration rates

ultimately place limits on the rates of carbon assimilation and growth. Such limitations are

most commonly discussed with respect to the availability of water to the roots, particularly

in dryland environments (e.g. Gazal et al., 2006; O'Grady et al., 2009; Du et al., 2011).

Drought resistant plants are generally able to withstand more negative xylem water

Chapter One

10

potentials (Ψ), and thus to draw water out of drier (more negative Ψ) soils. However, water

is relatively abundant in most riparian zones, and most riparian tree species have relatively

low cavitation thresholds and thus are sensitive to dry conditions. For instance, all riparian

Populus species of the northern hemisphere appear to be reliant on shallow groundwater,

and with reduction in water availability display rapidly increasing water stress responses,

initially stomatal closure and reduced photosynthesis, progressing to canopy loss and

branch abscission (e.g. Cooper et al., 2003; Rood et al., 2003; Hultine et al., 2010). Due to

the similarity in habitat type, M. argentea is likely to exhibit similar water stress response

patterns to riparian trees in other dryland regions.

In undisturbed habitat, trees such as M. argentea typically have constant access to

abundant water, while the canopy may be exposed to extreme VPD conditions. Under

these conditions, very high rates of transpiration would be expected, with water use patterns

driven largely by VPD. However, the water must be transported efficiently from the soil

reservoir to the leaf surfaces, and under some conditions, the total transport capacity may

become the limiting factor in water use (Tyree, 2003). Transpiration rates may therefore be

restricted at extreme levels of VPD via stomatal control, to prevent water potentials

reaching a critical threshold, even when soil water content is high. For example,

transpiration appeared restricted when VPD exceeded 1.2 kPa in montane tropical

rainforest trees in Ecuador (Motzer et al., 2005), and stomatal conductance and

photosynthesis began to decrease at VPD greater than 2 kPa in irrigated orange trees in

Spain (Martin-Gorriz et al., 2011). However, limitations on water use imposed by

transport capacity can be difficult to distinguish from those of soil water availability, since

high VPD conditions often co-occur with soil drought. The study of M. argentea in

undisturbed conditions, with confirmed presence of shallow groundwater, provides an

opportunity to examine the physiology of tree water use under extreme evaporative

demand, but with unlimited water availability.

In addition, any changes over time in fine root and canopy surface areas also

influence water use patterns, by affecting the quantities of water that can be taken up from

the soil and lost from the leaf surfaces. Many of the well-studied northern hemisphere

riparian species are winter-deciduous, while most Australian species are evergreen, including

M. argentea. As a result, the temporal patterns of leaf and fine root flushing and senescence


11

can differ markedly, creating very different seasonal water use dynamics (e.g. Eamus et al.,

2000; Goodrich et al., 2000; Hultine et al., 2004; Scott et al., 2004; Zeppel & Eamus,

2008). Therefore, while the water use physiology of M. argentea is expected to be similar in

many respects to other semi-arid riparian trees worldwide, some differences in water use

patterns are also expected reflecting differing phenology, with growth and transpiration

continuing year-round.

Some estimates of water use have been made for M. argentea in the wet-dry tropics of

northern Australia, near Darwin (O'Grady et al., 2006a). This study found considerable

variation in tree water use among three different sites along the Daly River, but no

significant seasonal variation in water use. However, water use patterns and physiological

functioning of M. argentea have not been evaluated in the semi-arid Pilbara region, where

the environment can be quite different, with an overall drier climate, more variable rainfall

and flooding patterns, and greater seasonal fluctuations in VPD. Thus the patterns of water

use and physiological functioning of M. argentea in the Pilbara may show greater seasonal

variability. With increasing pressures on water resources in the Pilbara, the ability to

calculate ecosystem water requirements and hydrological budgets will be important to

manage and allocate water appropriately. There is therefore a need to quantify water use

patterns for key species like M. argentea, including defining the relationships with climatic

variables, and factors such as tree size.

Hydraulic redistribution (HR) of water through tree roots is another important water

flux mediated by trees, occurring when the root system spans a gradient of soil water

potential, leading to passive transfer of water from wetter to drier soil portions (Burgess et

al., 1998). These fluxes can be particularly significant in semi-arid environments, where

groundwater may be present at depth while surface soil layers are dry, and many tree species

have both deep and shallow roots which link these soil layers (Ryel et al., 2008; Scott et al.,

2008; Bleby et al., 2010; Hao et al., 2010). HR can allow root survival in dry soil, thereby

increasing the total root surface area, enabling greater total tree water use and greater

nutrient availability (Domec et al., 2004; Bauerle et al., 2008; Armas et al., 2012). Water

transferred to shallow layers can also be used by shallow rooted plants, so can influence

productivity and functioning of the whole ecosystem (Brooks et al., 2006; Hawkins et al.,

2009). For example, HR by Pinus taeda in North Carolina, USA, maintained moisture in

Chapter One

12

the surface soil for longer periods of the year, increasing the total transpiration by both the

trees and understorey species by 30-50%, and leading to substantially higher ecosystem

productivity (Domec et al., 2010). With their permanent access to groundwater, mature M.

argentea trees may redistribute water to surface layers during the long rainless periods, when

surface soils are dry. This redistribution is likely an important process both in the patterns

of water flux through the trees themselves, and more broadly in the ecohydrology of the

Pilbara’s riparian ecosystems.

Although water may be relatively abundant in riparian zones compared with the

surrounding landscape, riparian trees must nevertheless contend with the high spatial and

temporal heterogeneity in water and nutrient availability that characterises dryland

waterways. Plant responses to, and impacts upon, environmental heterogeneity are

currently topics of much research interest, owing to the increasing realisation that

heterogeneous resource distributions can have consequences beyond those that might be

predicted based solely on studies conducted under homogeneous conditions (e.g. Hodge,

2010; García-Palacios et al., 2012). Plant root density in riparian zones is usually highly

variable in both time and space, reflecting responses of plants to heterogeneity in resource

distribution (Kiley & Schneider, 2005). Trees lining dryland streams can also be subject to

high energy floods that move large amounts of material and alter channel topography,

scouring some zones and depositing sediment and debris in others (Naiman & Decamps,

1997; Friedman & Lee, 2002). M. argentea is frequently found along central channels

where groundwater is close to or above the surface of the river bed, and where resource

heterogeneity and flood velocities are typically greatest. Portions of M. argentea root

systems are frequently exposed, so that prostrate trees destabilised by flood erosion are

common along channel banks. In other locations M. argentea trunks are partially buried in

sediment, and the trees often produce roots along the submerged stems. M. argentea also

commonly forms extensive root mats within surface pools left by flood waters, which then

die back as the pools dry. Due to the high levels of resource heterogeneity and disturbance,

M. argentea might be expected to display plastic and opportunistic responses to the

availability of water around the root system.

The nature of groundwater and surface flows are integral to the ecology of riparian

and wetland environments. With changes to the water regime, such ecosystems appear to


13

be vulnerable to abrupt shifts in vegetation state (Ridolfi et al., 2006). Knowledge of critical

thresholds to identify when abrupt and potentially irreversible ecological change may occur

is highly desirable for optimal water management, although true physiological thresholds

for say, tree survival, are rarely defined (Dodds et al., 2010; Oliveira, 2013). Nonetheless,

there is increasing interest in identification and use of thresholds for the purposes of early

intervention and preventative management actions (Groffman et al., 2006; Dodds et al.,

2010). As a key overstorey species in the wetland environments of the Pilbara, the responses

of M. argentea to disturbances may have flow-on effects for understorey species and for

whole ecosystem functioning, and this species may be considered an early indicator of any

local change in water status. As outlined above, many riparian ecosystems of the Pilbara are

subject to abstraction of underlying groundwater, while others are subject to increased

flows and prolonged flooding, where excess water is discharged. Given its wet habitat, I

expected that M. argentea would be sensitive to prolonged drought and to declines in

groundwater levels during abstraction, but would be tolerant of prolonged flooding.

Identification of physiological tolerance thresholds in M. argentea would be valuable in

water management strategies in the Pilbara’s riparian ecosystems.

OBJECTIVES OF THIS THESIS

The general objective of this thesis was to characterise the temporal and spatial patterns of

water use by M. argentea trees in the Pilbara, and to assess the ways these patterns may be

affected by disturbances to the native water regime.

I examined water use in M. argentea trees through sap flow measurements at two

contrasting sites in the Pilbara. I sought to assess the major factors controlling the observed

patterns of water use, such as atmospheric conditions and aspects of tree hydraulic

physiology (Chapter 2). I also aimed to characterise the patterns of water use over diurnal

and seasonal cycles, within different parts of the sapwood, and across tree sizes, information

that will be required to calculate ecosystem scale water budgets, and ecological water

requirements (Chapters 2 & 3).

Responses of M. argentea to spatiotemporal heterogeneity in water availability were

also investigated, with the aim of assessing characteristics of the root system that enable the

species to thrive in highly changeable riparian habitats. Sap flux was measured in deep and

Chapter One

14

shallow roots of M. argentea trees during a 12-month period to assess temporal variation in

water fluxes, and assess whether redistribution of water through tree roots may be an

important process in the Pilbara’s riparian ecosystems (Chapter 4). Controlled glasshouse

experiments were also conducted, using a split-root set up to mimic the heterogeneous

water distribution of riparian zones. The experiments aimed to determine whether

compensatory responses are displayed by M. argentea during drying of parts of the root

system, a situation that can occur during the flooding and drying cycles of ephemeral

waterways (Chapter 5).

Finally, I examined aspects of the tolerance and response of M. argentea to flooding,

drought and groundwater abstraction, including potential threshold responses with

decreasing water availability. Seedling responses to water availability were investigated in

controlled experiments, with the aim of understanding the possible effects of altered water

regimes on recruitment patterns (Chapter 6). Responses of trees to groundwater abstraction

were assessed through a case study at the Yule River borefield, on the coastal plain region of

the Pilbara. This study aimed to evaluate the changes in tree water use physiology that

occur with increasing groundwater drawdown (Chapter 7).

Each experimental chapter of this thesis is presented as a self-contained paper, and as

such, a small amount of repetition is inevitable across the body of work presented. Chapter

8 provides a general discussion of the outcomes of this research, their relevance to the

ecological understanding and water management of semi-arid northwest Australia, and

suggests avenues for further research.

Knowledge of water use patterns in M. argentea will contribute to the fundamental

understanding of the water use physiology of riparian trees. Insights into riparian

ecohydrology of key indicator species such as M. argentea gained in the Pilbara, particularly

from a more mechanistic perspective, are also likely to provide much needed baseline or

reference information that will inform the development of ‘best practice’ water

management strategies not only in the Pilbara, but in dryland regions worldwide.

15

CHAPTER TWO: Seasonal patterns of water use by the riparian tree

Melaleuca argentea in semi-arid NW Australia

INTRODUCTION

Tree water use can vary over annual cycles in response to seasonal climatic fluctuations.

Studies of seasonality in tree water use physiology have mostly focused on temperate or dry-

tropical regions with clearly defined and regular hydroclimatic seasons (e.g. Greco &

Baldocchi, 1996; Ford et al., 2004; Kelley et al., 2007; Zeppel & Eamus, 2008; Kunert et

al., 2010). In such environments, trees require physiological strategies to cope with the

seasonal extremes in temperature or water availability. For example, dry-tropical and

Mediterranean deciduous species shed their leaves to avoid transpiration during the dry

season, while co-occurring evergreen species maintain water transport to their canopy year-

round through more conservative water use and hydraulic properties such as greater

cavitation resistance (e.g. Eamus, 1999; Baldocchi et al., 2010; Fu et al., 2012). There has

been less study of seasonal responses of trees in more arid regions, partly because rainfall

and water availability are less predictable both in space and time. However, trees within the

riparian zones of dryland regions usually co-occur with groundwater and thus can

experience relatively predictable and constant water availability (e.g. O'Grady et al.,

2006b). Melaleuca argentea is one such tree species in the semi-arid Pilbara region of

northwest Australia. Temperatures and vapour pressure deficits (VPD) in arid climates

often display predictable seasonal fluctuations; the water use and physiology of riparian

trees in these regions might therefore be expected to display corresponding seasonal

patterns.

Tree water use is determined primarily by water availability to the roots, the

evaporative demand of the atmosphere, and the hydraulic resistance and capacitance of the

flow pathway through the soil, plant and leaf boundary layer (e.g. Sperry et al., 2003).

Theoretically, transpiration rates should be high under conditions of abundant water

supply and exposure to high VPD, with transpiration largely defined by atmospheric

conditions until the limit of safe hydraulic function is approached (Brodribb, 2009). The

soil and roots are often weak points in the flow pathway (Sperry et al., 1998), but where

Chapter Two

16

trees have roots in the saturated zone, such as many riparian trees, xylem hydraulic

conductivity (K) may be the primary limiting factor for water use (Brodribb, 2009). In tall

trees, transpiration can be limited by the rate at which water can be transported from the

soil reservoir to the leaves, rather than by water availability or demand (Goldstein et al.,

1998; Williams et al., 2001). In riparian zones, where trees are not limited by water

availability, transpiration rates might also be limited by transport rates under extreme VPD.

Although K can vary with xylem anatomy, the rates of sap flux in riparian trees in high

VPD environments, such as Populus and Salix species in semi-arid Arizona, USA, typically

fall within the same range of flux rates observed in non-riparian species (Wullschleger et al.,

1998; Schaeffer et al., 2000). Over diurnal cycles, transpiration is maintained within safe

limits by the adjustment of stomatal aperture, leading to a ‘saturation’ of transpiration

above a certain VPD (Goldstein 1998, Brodribb 2003). For example, in well-watered

Populus fremontii, rates of sap flow saturated at VPD between 4 and 6 kPa (Gazal et al.,

2006). In addition, trees can utilise internal water storage as a buffer over diurnal cycles to

partly overcome the limitation of xylem K (Scholz et al., 2011). Eucalyptus victrix, a riparian

tree that often co-occurs with M. argentea in the Pilbara region, displayed withdrawal and

refilling of stem water stores over diurnal cycles, even with access to shallow ground water

(Pfautsch et al., 2011). Thus, while M. argentea occurs only with access to a shallow water

table, water use might nonetheless be restricted by tree K and/or capacitance under the

extreme VPD conditions that can occur during the middle of the day in the summer

months (frequently greater than 5 kPa).

Over annual cycles, seasonal fluctuations in the major driving variables of evaporative

demand and soil water availability often largely explain water use patterns in evergreen

trees, such as coniferous species and eucalypts (e.g. Oren et al., 1999; Ford et al., 2004;

Zeppel et al., 2004; Small & McConnell, 2008). However, these seasonal water use

patterns are also modified by tree adjustments in canopy cover, as well as by variation in

hydraulic resistance and capacitance. Leaf area is clearly a major determinant of water use.

For example, many of the well-studied forests of the northern hemisphere, including semi-

arid riparian systems in the south-west of the USA, consist largely of deciduous trees in

which seasonal water use patterns tend mostly to reflect the patterns of leaf flushing and

senescence (e.g. Goodrich et al., 2000; Scott et al., 2004). The canopy cover of evergreen

Seasonal patterns of water use by Melaleuca argentea

17

trees can also vary seasonally as part of a species’ water use strategy. For instance, in the

savannas of northern Australia, Eucalyptus species display patterns of canopy thinning and

flushing, coinciding with the pronounced wet and dry seasons, where canopy cover may

vary by as much as 50% in any given year (Duff et al., 1997; O'Grady et al., 1999). Xylem

hydraulic conductivity can also vary seasonally, typically due to xylem embolism induced

by water stress during dry periods or during freeze-thaw cycles in climates with cold winters

(e.g. Magnani & Borghetti, 1995; Tognetti et al., 2004). While the techniques used for

measurement of xylem embolism and recovery have recently come into question,

researchers nevertheless agree that embolism does occur at least under conditions of severe

water stress and freezing (Cochard & Delzon, 2013; Wheeler et al., 2013). Stem water

storage can further significantly influence seasonal water use patterns in many species, by

enabling the tree to draw upon internal reserves during dry periods (e.g. Zweifel & Häsler,

2001; Baker et al., 2002; Betsch et al., 2011). However, riparian trees such as M. argentea

in the Pilbara occur in conditions that are favourable throughout the year i.e., with warm

temperatures and a perennial water supply. It might therefore be expected that M. argentea

would display little seasonal variation in leaf area, K or capacitance as it would be well

buffered against seasonal variation in water availability. Without seasonal variation in these

physiological factors, or in water availability to the roots, then the seasonal variation in

atmospheric demand would remain as the sole factor driving patterns of tree water use

across the year.

The objective of this chapter was to characterise temporal patterns of water use by M.

argentea, in the semi-arid Pilbara region of NW Australia, by determining the relationships

between tree water use and meteorological conditions. I measured tree sap flow, canopy

physiology (leaf water potential and stomatal conductance) and meteorological conditions

over periods of at least 12 months, at two contrasting and spatially separated sites within

the Pilbara, enabling a detailed analysis of tree water use patterns over time. Aspects of

canopy physiology were also examined at intervals, including leaf water potential, stomatal

conductance and canopy cover. I expected that due to its warm habitat and constant access

to groundwater, M. argentea trees in undisturbed habitat would show little seasonal

adjustment in canopy cover, K or capacitance, and would be relatively insensitive to

additional pulses of surface soil moisture. I therefore hypothesised that seasonal patterns of

Chapter Two

18

Melaleuca water use would strongly reflect variations in atmospheric conditions (largely

VPD), perhaps with some limitation of water use at extreme VPD, as has been observed in

other well-watered trees over diurnal cycles. Without the complicating factors of seasonal

variations in water supply, or in tree physiological or morphological adjustments, I

anticipated that the relationships between water use and atmospheric variables would

remain constant over annual cycles.

MATERIALS & METHODS

Study sites

Sites were located on two intermittent waterways in the Pilbara region of north-west

Australia. The first site was at the Yule River (20.656° S 118.295° E), at a near-coastal

location and the second site at Marillana Creek (22.721° S, 118.958° E), in the inland

Hamersley Basin (Figure 2.1). The two sites are geographically separated (approximately

250 km apart), located within two distinct sub-regions of the Pilbara, with differing

underlying hydrogeology. By conducting the study at these contrasting sites, more

confidence can be placed in the broader applicability of the findings. At both sites,

Melaleuca argentea co-occurred with Eucalyptus camaldulensis and E. victrix, two other

common riparian tree species. The Yule River site consisted of a narrow, dense band of

trees bordering a wide, sandy riverbed, with vegetation beyond the river bank composed

largely of grassland. At the Marillana Creek site, the riparian vegetation was confined

within the (usually dry) creek bed, between steep, rocky banks, and the trees clustered along

deeper sub-channels. Spinifex grassland and open Eucalyptus woodland surrounded the

creekline.

The climate of the Pilbara region is classed as semi-arid to arid and is characterised by

highly variable rainfall, with extreme temperatures (up to 49 °C) and VPD (up to 10 kPa)

during November to March (Australian Bureau of Meteorology data). Cooler temperatures

occur between June and August (daily minima typically 5–18 °C and maxima 18–32 °C).

Cyclones and other low pressure systems can pass through the region during December to

April, bringing rainfall of up to 400 mm over a few days that results in flooding; such large

events typically occur every 4–6 years. In contrast, small, localised rainfall events (25 mm or

less) can occur throughout the year (Dogramaci et al., 2012). River flows are intermittent


19

and highly unpredictable, reflecting the variable rainfall patterns. The coastal subregion of

the Pilbara (containing the Yule River site) is slightly warmer and more humid compared

with the inland subregion (containing the Marillana Creek site). Winter (June–August)

temperatures inland are typically 3–6oC cooler than the coast, while summer (December–

February) temperatures are similar between subregions (minima typically 20–29oC and

maxima 32–42oC), but relative humidity is greater on the coast (45–60%) than inland (20–

40%). Seasonal periods are referred to in this chapter as: summer (December–February),

autumn (March–May), winter (June–August) and spring (September–November).

Just prior to this study, cyclone activity resulted in substantial flows in most Pilbara

waterways during the December 2008–Feb 2009 summer. Large surface pools were present

in the riverbeds within 10 m of the study trees at both sites when this study commenced (in

March 2009 at Yule River, and in June 2009 at Marillana Creek), but had dried down by

November 2009. Five M. argentea trees were selected across a size range of 100–500 mm

Marsh

DAMPIER

NEWMAN

Fort

escu

e Riv er

Waterway

West Yule

R

100 km

Figure 2.1: Study site locations. Modified from Department of Water (2010) and Van Vreeswyk et al. (2004).

F ortescue River

Town

Study site

Major road

KARRATHA ROEBOURNE

TOM PRICE Marillana Ck

For te s

cue Sth

We e li Wol

li C

k

Yule River

PORT HEDLAND

Chapter Two

20

diameter at breast height (1.2 m), within a 900 m2 area at each study site. Measurements

were conducted on the same ten trees throughout the study.

Meteorological data

At the Yule River, weather stations were installed approximately 18 km NW of the study

site (HOBO; Onset Computer Corporation, Bourne, MA, USA during March–September

2009; and ET107; Campbell Scientific Inc., Logan, Utah, USA during September 2009–

May 2010). Stations were positioned in grassland on a flat area near the top of the main

river channel bank, more than 100 m from any trees. Temperature, humidity, light

intensity, wind speed, wind direction and rainfall data were logged half hourly by the

HOBO station, and hourly by the ET107 station. Due to problems with the light sensors,

daily solar radiation data modelled from satellite images were also obtained from the

Australian Bureau of Meteorology (http://www.bom.gov.au/climate/data-services/) for the

entire period.

At Marillana Creek a weather station (ET107; Campbell Scientific Inc., Logan, UT,

USA) was installed in the open at the top of the creek bank, approximately 300 m from the

study trees, between September 2009 and September 2010. Data were logged as for the

Yule River site. For the period June–September 2009, daily data were provided by Rio

Tinto Pty Ltd., from a weather station approximately 16 km ESE of the study site at the

Yandicoogina Mine.

Vapour pressure deficit was calculated from the temperature and humidity readings

according to Howell et al. (1995), as:

VPD = (1 – RH/100) 0.6108 e17.27 T / (T + 237.3)

where VPD has units of kPa, RH is the relative humidity (%) and T is the temperature

(°C).

Tree water use

Tree water use patterns were determined with sap flow probes (HRM-30; ICT

International, Armidale, NSW, Australia), via the heat ratio method (Burgess et al., 2001).

Probes were installed into the main stem of each tree at 1.2 m height. Bark was first peeled

back from the insertion sites, to leave a 5 mm layer over the cambium, then holes were

drilled 0.6 mm apart using a drill guide (ICT International, Armidale, NSW, Australia)


21

following the grain of the wood, and the probes inserted. Sensor units and probes were

covered with bubble wrap and insulation foil to minimise effects of ambient temperature

fluctuations. Data were obtained from a single measurement point positioned at 7.5 mm

depth into the sapwood, which corresponded to the approximate centre of the sapwood

band. Heat pulses were set at 30 min intervals, and the data processed to heat pulse velocity

(Vh) values by the “Smart Logger” (SL5; ICT International, Armidale, NSW, Australia).

Sap flow was logged continuously at the Yule River site from 29th March 2009 to 12th

May 2010, and at Marillana Creek from 18th July 2009 to 13th September 2010, apart

from three periods at Marillana Creek, when equipment was damaged by livestock despite

being fenced (1st–22nd September 2009; 20th June–22nd July 2010; and 26th July–1st

Sept 2010).

All calculations of Vh, and conversion to sap flow velocity (Vs) were performed

according to methods of Burgess et al. (2001) and Bleby et al. (2004). For each probe, the

minimum value of Vh recorded on nights with low VPD and wind speed was set to zero

(after filtering of the datasets), to correct baseline errors caused by slight probe

misalignment. Wood properties (density and volumetric water content) were measured on

blocks of sapwood 2–5 cm3 in size, cut from five trees at each site. Volume was measured

by the displacement weight of water by fresh samples. Fresh and dry mass were determined

by weighing before and after drying at 60 °C. Mean values for sapwood properties were

then determined for each site to calculate Vs. Wounding was examined in additional blocks

of sapwood cut from three probe insertion sites from trees at Marillana Creek at the end of

the study period; sapwood was shaved back and the diameter of darkened wood around the

insertion site measured under a dissecting microscope. The mean diameter of 0.22 cm was

used in all Vs wounding corrections. Wood cores were also taken from each study tree with

an increment borer (Haglöf, Långsele, Sweden), to assess wood appearance and sapwood

depth. Surfaces of the fresh cores were shaved back and stained with methyl orange (1%

w/v in 4% v/v methanol); the sapwood stained deep red and the heartwood yellow.

The Vs data presented here were collected from the same positions within the same

trees over time as my objective for this chapter was to assess temporal patterns of water use

in detail. Spatial variation in Vs and volumetric quantification of water use in M. argentea

are examined separately in Chapter 3.

Chapter Two

22

Canopy physiological measurements

Leaf measurements were conducted on sunlit, outer canopy leaves cut from 5–7 m height

with extendable clippers. At each time point, measurements were made on 2–3 leaves per

tree, each from branchlets cut separately from the canopy, then averaged to give a single

value for each parameter per tree. Leaf water potential (Ψl) was measured with a

Scholander-type pressure chamber (PMS instruments, Albany, OR, USA) immediately

after cutting from the trees. Stomatal conductance (gs) was measured with a leaf porometer

(Decagon Devices, Pullman, WA, USA) immediately after severing from the tree. Testing

showed that if the measurement was complete within 2 min of excision, the mean gs did

not differ between excised and attached leaves, although the variance of gs was greater for

excised leaves (Appendix A).

Statistical analysis

Statistical analysis was performed in SAS version 9.2 (SAS Institute Inc.; Cary, NC, USA).

Mixed-effect ANOVA was used to test for site and seasonal differences in water use, canopy

cover and nocturnal sap flow, with the individual trees as a random effect. ANCOVA was

used to test for site and seasonal differences in the relationships between tree water use and

meteorological parameters.

For multiple regression analysis, fixed effects (meteorological variables) were selected

using the least absolute shrinkage and selection operator (LASSO) method (PROC

GLMSELECT), using minimum cAIC as the model selection criterion (Flom & Cassell,

2007). The selected fixed effects were then fitted in a mixed model (PROC MIXED)

including a random effect for the individual tree subjects over time (Fernandez, 2007). R2

statistics were computed from models containing the fixed effects only, both with and

without individual tree subjects also included as fixed effects, to assess the amount of

variation explained by the model when among-individual variation was included, and when

among-individual variation was eliminated.

The selected meteorological models were then used to test the significance of the

effects of recent rainfall (considered here as a proxy for moisture of upper soil layers), site

and season, after controlling for the meteorological differences. These additional factors

were included individually in the models as fixed effects, and their significance assessed by

the resulting change in cAIC values, log-likelihood tests, and type-3 F-tests.


23

RESULTS

Seasonal variation in tree water use and meteorological conditions

Patterns of most meteorological variables were similar between the Yule River and

Marillana Creek study sites during the study period (Figure 2.2). Seasonal variation in

VPD and temperature was pronounced with lowest values in June–July (daily maxima 0.8–

3.5 kPa) and highest November–February (daily maxima 2.5–8.2 kPa). Marillana Creek

received five times more rainfall than the Yule River site during the study period (160 mm

over 517 total measurement days, c.f. 26 mm over 423 days). No cyclones occurred during

the study period and all rainfall events were relatively small with the largest 37 mm

occurring in November 2009 at Marillana Creek.

0

1

2

3

4

5

V s (m

day

-1)

Figure 2.2: Seasonal

patterns of water use

by Melaleuca argentea

trees, and the major

driving meteorological

variables during the

study period. (a) Daily sap flow velocity (Vs) at each study site, aver-aged by month. Error bars are standard error, calculated to include variance both between days and between the five measured trees. (b)

Daily solar irradiance and maximum VPD at each study site, aver-aged by month. Closed symbols; Yule River site, open symbols; Maril-lana Creek site.

Marillana CreekYule River

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC0

2

4

6

8

Irrad

ianc

e (k

Wh

m-2

day-

1)

VPDm

ax (kPa)

0

2

4

6

8

(a)

(b)

Month

Chapter Two

24

Rates of tree water use varied seasonally, with daily Vs lowest in June and highest in

December at both sites (p<0.0001; Figure 2.2a). Water use patterns reflected the seasonal

variation in the major driving variables of solar irradiance and VPD (Figure 2.2). Although

seasonal patterns were similar between the two sites, Vs was higher at the Yule River site

than at Marillana Creek throughout the year. The seasonal fluctuation in water use was

greater at Marillana Creek; mean daily Vs in June was only 50% of Vs in December at

Marillana Creek, but 70% of the December Vs at the Yule River (Figure 2.2a).

Diurnal patterns of tree water use physiology across seasons

Diurnal patterns of Vs, Ψl and gs were assessed separately on days in mid-winter, late spring

and late autumn (Figure 2.3). Vs increased rapidly in the mornings, coinciding with sunrise

and stomatal opening at both sites and on all measurement dates. Stomatal conductance

(gs) peaked mid-morning, declining toward midday. In November (late spring) at both

sites, a second peak in gs then occurred late afternoon, while in cooler months only the

mid-morning peak occurred. VPD was greatest mid-afternoon, and so was higher during

the afternoon than in the mornings. Stomatal conductance therefore correlated negatively

with VPD during much of the day, which meant a fairly constant Vs was maintained

throughout the middle of the day (Figure 2.3). This plateauing of sap flow generally

occurred once VPD exceeded ~3 kPa. Minimum Ψ l occurred between midday and 3 pm,

but remained above -2 MPa on all dates measured. Minimum VPD usually occurred a few

hours before dawn, and sap flow frequently continued at low rates throughout the night,

reaching a minimum sometime between midnight and dawn. Nocturnal sap flow at

Marillana Creek ranged between 16 and 38% of the total flow over the 24 hr period

compared to between 9 and 41% at Yule River. Predawn Ψ l was close to zero on all

measured dates.

The relationships between Vs and the driving variables of irradiance and VPD

displayed marked hysteresis over diurnal cycles during much of the year (Figure 2.4). On

most days in winter and spring, Vs was lower during the morning than at the same light

intensity in the afternoon (anti-clockwise hysteresis), most likely due to the higher VPD

conditions in the afternoon driving higher water use (Figure 2.4a, d). However, the degree

of hysteresis between Vs and irradiance decreased during the summer and autumn (Figure

2.4b, c, e, f), i.e., rates of afternoon water use were not much higher than in the mornings,


25

0

10

20

0

2

4

6

0

100

200

300

-2

-1

0

400

800

1200

June 2009 November 2009 May 2010

12:006:00 18:000:00 0:00 12:006:00 18:00 0:00 12:006:00 18:00 0:00

VPD (kPa)

Irrad

ianc

e (W

m-2)

V s (c

m h

-1)Ψ

leaf

(MPa

)

gs (m

mol m

-2 s-1)

Time of day (h)

November 2009 July 2010

-2

-1

0

400

800

1200

0

2

4

6

0

10

20

0

100

200

300

12:006:00 18:000:00 0:00 12:006:00 18:00 0:00

Time of day (h)

Irrad

ianc

e (W

m-2)

V s (c

m h

-1)Ψ

leaf

(MPa

)

VPD (kPa)

gs (m

mol m

-2 s-1)

Figure 2.3: Diurnal patterns of irradiance, vapour pressure deficit (VPD), tree sap flow

velocity (Vs), stomatal conductance (gs) and leaf water potential (Ψ) at (a) Yule River and (b)

Marillana Creek sites. Each vertical column of panels shows values measured over a single 24 h period during the month indicated at the top of the column. Values of Vs, Ψ and gs are means ± standard error of five trees.

(a)

(b)

Chapter Two

26

despite the higher VPD in the afternoons. By contrast, Vs displayed clockwise hysteresis

with VPD on most days year round (Figure 2.4g–l), meaning that water use was higher

during the morning than at the same VPD in the afternoon on any given day.

Maximum rates of sap flow reached each day were largely consistent across sites and

seasons. However, the number of hours that maximum rate was maintained during the

middle of the day differed between summer (on average 10.5 h per day) compared to

winter (on average 7 h per day). Thus, daily total sap flow was linearly correlated with both

daily irradiance and VPD (Figure 2.5).

Relationships of tree daily water use with meteorological conditions

Daily Vs was significantly correlated with daily values of nearly all meteorological variables

measured (Figure 2.5a–d, Table 2.1). Many of the meteorological variables were highly

correlated with one another (not shown), but the variables giving the strongest correlations

with Vs were the average daily VPD, its interaction with average wind speed (VPDavg x uavg),

solar irradiance, and maximum temperature (Table 2.1). The correlation between Vs and

recent rainfall (falling within the previous 10 days) was significant though weak at

Marillana Creek. Nocturnal sap flow also correlated with meteorological conditions at both

sites. The best correlates for nocturnal Vs were the average nocturnal VPD (giving positive

correlations, with R2 of 0.32 at the Yule River and 0.24 at Marillana Creek), and nocturnal

VPDavg x uavg (giving positive correlations, with R2 of 0.34 at the Yule River and 0.25 at

Marillana Creek). The maximum and total daily Vs differed between the two sites, with

trees at the Yule River having higher rates of water use for a given value of a meteorological

parameter (such as VPD or light intensity), with site differences more pronounced at lower

Vs (Figure 2.5). For example, daily Vs was on average 1.8 times greater at the Yule River

than Marillana Ck on days with maximum VPD of 2 kPa, and 1.3 times greater at the Yule

River on days with maximum VPD of 8 kPa.

The relationship between daily Vs and VPD differed among seasons (Figure 2.5a, b).

The range of VPD was similar between spring and autumn, but the regressions of VPD

with daily Vs differed between these two periods; at the Yule River the intercept of the

relationships differed significantly (p<0.0001), and at Marillana Creek the slope of the

relationships differed (p=0.0005). Tree water use was greater during the spring (season of

increasing VPD), than during autumn (season of decreasing VPD), i.e., clockwise hysteresis


27

SPRING SUMMER

10

20

30

0

AUTUMN

0 200 400 600 800 1000

(a) YR

10

20

0

5

25

V s (c

m h

-1)

(d) MC

am

pm

amam

pm

am

pm

Irradiance (W m-2)

0 200 400 600 800 1000

pm

0 200 400 600 800 1000

am

pm

(b) YR (c) YR

(e) MC (f) MC

10

20

30

0

V s (c

m h

-1)

(g) YR (h) YR (i) YR

pm

pm

am

pmam

pm

amam

pmpm

0 2 4 6 0 2 4 6 0 2 4 6

10

20

0

5

25

VPD (kPa)

SPRING SUMMER AUTUMN

(j) MC (k) MC (l) MC

Figure 2.4: Patterns of hysteresis over the diurnal cycle in the relationship between sap flux

velocity (Vs) and the major driving variables of (a–f) light intensity and (g–l) vapour pressure

deficit (VPD). Data are from representative days in mid spring (left hand panels), summer (centre panels) and autumn (right hand panels), for trees at the Yule River (YR; upper panels in each block), and Marillana Creek (MC; lower panels in each block). The arrows indicate the direc-tion of hysteresis. Values are means +/- standard error of five trees.

Chapter Two

28

Figure 2.5: Comparisons of sap flux velocity (Vs) between seasons. Relationships of (a & b) total daily Vs with vapour pressure deficit (VPD), (c & d) total daily Vs with solar irradiance, and (e & f) nocturnal Vs averaged over the dark period with nocturnal VPD, at (a, c & e) the Yule River and (b, d & f) Marillana Creek. Each datapoint is a single day of the study period, and Vs values are means of five trees. Where relationships differed significantly between seasons, regression lines and their R2 values are shown for spring and autumn.

VPD (kPa)

Marillana Creek

R2= 0.82

R2= 0.84

Autumn

Spring

Summer

Winter

Yule River

R2= 0.38

R2 = 0.32

0

1

2

3

4

5

0

4

8

12

16

0 2 4 6 8 0 2 4 6 8

V s (m

day

-1)

0

1

2

3

4

5

V s (m

day

-1)

0 2 4 6 8 0 2 4 6 8

Noc

turn

al V

s (cm

h-1)

0 2 3 41 0 2 3 41 5

Irradiance (kWh m-2 day-1)

(a) (b)

(c) (d)

(e) (f) R2= 0.73 R2= 0.64

R2= 0.91

R2= 0.84

Nocturnal VPD (kPa)

10

10


29

Table 2.1: Correlations between tree daily sap flow and individual meteorological variables. Pearson coefficients (r) and probability values (p) are provided. VPD; vapour pressure deficit, T; temperature, RH; relative humidity, u; wind speed, R; solar irradiance, max; daily maximum, min; daily minimum, avg; daily mean of hourly measurements, tot; daily total, 1 day; value for the same day, 10 days; total during the previous 10 days.

Meteorological parameter

Yule River Marillana Creek Sites pooled

r p r p r p

VPDmax 0.373 <.0001 0.397 <.0001 0.301 <.0001 VPDmin 0.322 <.0001 0.355 <.0001 -0.023 0.1905 VPDavg 0.422 <.0001 0.398 <.0001 0.157 <.0001 Tmax 0.347 <.0001 0.390 <.0001 0.364 <.0001 Tmin 0.179 <.0001 0.378 <.0001 0.300 <.0001 Tavg 0.293 <.0001 0.399 <.0001 0.326 <.0001 RHmax -0.240 <.0001 -0.175 <.0001 0.206 <.0001 RHmin -0.299 <.0001 -0.251 <.0001 -0.132 <.0001 RHavg -0.328 <.0001 -0.220 <.0001 0.122 <.0001 umax 0.213 <.0001 0.193 <.0001 0.142 <.0001 umin 0.251 <.0001 0.182 <.0001 0.086 <.0001 uavg 0.383 <.0001 0.268 <.0001 0.175 <.0001 umax x VPDmax 0.337 <.0001 0.365 <.0001 0.277 <.0001 uavg x VPDavg 0.462 <.0001 0.407 <.0001 0.193 <.0001 Rtot 0.390 <.0001 0.427 <.0001 0.370 <.0001 Rain1day -0.040 0.0945 -0.084 0.0014 -0.112 <.0001 Rain10days -0.004 0.8641 0.074 0.0048 -0.069 <.0001

between VPD and Vs exists at both diurnal and seasonal scales. In contrast, the regression

between irradiance and daily Vs did not differ among seasons (p>0.05; Figure 2.5c, d). The

seasonal difference in the relationship between Vs and VPD was evident during both the

light and dark periods, with greater Vs in spring than autumn throughout the diurnal cycle.

Nocturnal Vs was also greater in spring than autumn when averaged to a per-hour-of-

darkness basis, to account for the seasonal difference in day length (Figure 2.5e, f).

However, the proportional contribution of nocturnal flow to the total daily flow over a

24 h period remained similar across the year, ranging from 6 to 48% of daily sap flow at

both sites. The proportional contribution of nocturnal flow increased with nocturnal VPD,

but this relationship did not vary seasonally (not shown).

Chapter Two

30

Multiple regression analysis was used to assess the extent to which meteorological

conditions could explain temporal patterns and site differences in tree water use. To assess

whether the observed seasonal differences in the relationships of Vs with VPD were due

simply to other meteorological factors varying between sites and seasons, or whether tree

physiological or growth responses are involved, I first constructed meteorological models

based on atmospheric variables, then added other variables individually (site and season).

These multiple regression models also allowed testing of whether the apparent effect of

rainfall (or lack thereof) on tree water use was related to changes in surface soil moisture, or

due to the influence of other meteorological effects that tend to accompany periods of

rainfall, such as reduced light intensity and cooler temperatures. Given daily Vs was linearly

related to the daily values of the meteorological variables, I employed linear regression to

construct the models.

All variables shown in Table 2.1, except for rainfall, were considered as fixed effects

in the regression analysis, as well as interactions between the VPD, wind speed and light

parameters (29 variables in total). Reduced models were selected with the LASSO

procedure (Flom & Cassell, 2007), and the resulting models are shown in Table 2.2.

Greater relative importance cannot necessarily be attributed to the variables listed in Table

2.2, owing to the problems inherent in any model selection process and the high degree of

multicollinearity amongst the predictor variables (Graham, 2003; Bolker et al., 2009).

Rather, the selected models are considered here to be adequate meteorological descriptions

for the tree water use patterns, for the purposes of testing the effects of further factors.

Overall, I found that meteorological conditions explained 67–80% of the temporal

variation in tree water use when among-tree variation was accounted for by including the

tree individuals as additional factors in the models (Table 2.3). However, the variability in

water use among individual trees was large and when not accounted for, the amount of

variation explained by the meteorological conditions was only 23–42%.

The effects of site, rainfall and season on tree water use

The two sites still differed significantly in rates of water use after controlling for

meteorological differences; the addition of site as a fixed-effect variable significantly

improved the pooled site model (Table 2.4a). The Vs of trees at the Yule River site was on


31

Tabl

e 2.

2: M

ulti

ple

regr

essi

on m

odel

s fo

r tr

ee d

aily

sap

flow

(cm

day

-1)

wit

h m

eteo

rolo

gica

l va

riab

les.

Est

imat

es f

or t

he i

nter

cept

and

reg

ress

ion

coeffi

cien

t pa

ram

eter

s ±

stan

dard

err

ors,

and

type

-3 F

-tes

t pr

obab

ilitie

s ar

e pr

ovid

ed. V

PD; v

apou

r pr

essu

re d

efici

t (k

Pa),

T; t

empe

ratu

re (

°C),

RH; r

elat

ive

hum

idity

(%),

u; w

inds

peed

(m s-1

), R;

sol

ar ir

radi

ance

(kW

h m

-2),

max

; dai

ly m

axim

um, m

in; d

aily

min

imum

, avg

; dai

ly m

ean

of h

ourly

mea

sure

men

ts, t

ot; d

aily

to

tal. (a

)

Y

ule

Rive

r

Mar

illan

a Cr

eek

Si

tes

pool

ed

Effec

t Co

-effi

cien

t

p >

F

Effec

t Co

-effi

cien

t

p >

F

Effec

t Co

-effi

cien

t

p >

F

Inte

rcep

t 24

1 ±

57

0.

0243

Inte

rcep

t -1

61 ±

172

0.41

73

In

terc

ept

31 ±

80

0.

7084

T avg

-0

.34

± 0.

72

0.

6351

T min

4.

39 ±

0.2

7

<.00

01

VP

Dm

ax

-11.

4 ±

9.4

0.

2257

RH

max

1.

05 ±

0.2

2

<.00

01

RH

min

0.

94 ±

0.2

1

<.00

01

VP

Dm

in

48 ±

26

0.

0687

RH

avg

-2.2

7 ±

0.25

<.00

01

RH

avg

-1.1

5 ±

0.13

<.00

01

VP

Dav

g -6

8 ±

24

0.

0049

u m

ax

5.9

± 1.

2

<.00

01

u m

ax

2.30

± 0

.59

<.

0001

T max

6.

1 ±

1.1

<.

0001

u m

in

5.5

± 8.

7

0.53

05

R t

ot

22.7

± 1

.4

<.

0001

T min

1.

04 ±

0.5

6

0.06

10

R tot

13

.4 ±

2.0

<.00

01

RHm

ax

1.74

± 0

.18

<.

0001

VP

Dm

ax x

um

ax

-0.2

8 ±

0.44

0.53

05

RHm

in

-0.8

3 ±

0.23

0.00

03

VPD

min

x u

avg

6.1

± 2.

8

0.03

32

RHav

g -2

.90

± 0.

22

<.

0001

VP

Dav

g x u

max

-2

.9 ±

1.2

0.01

12

u max

12

.3 ±

1.6

<.00

01

VPD

avg x

uav

g 4.

4 ±

1.1

<.

0001

u a

vg

-15.

8 ±

2.5

<.

0001

VP

Dm

ax x

um

in

-1.7

± 2

.9

0.

5661

R t

ot

2.4

± 2.

8

0.38

71

VPD

min

x u

min

-9

.6 ±

8.6

0.26

22

VPD

max

x u

max

-2

.69

± 0.

95

0.

0046

VP

Dav

g x u

min

13

.8 ±

7.6

0.06

85

VPD

max

x u

avg

3.2

± 1.

5

0.04

09

R tot

x u

min

-2

.7 ±

1.2

0.02

26

VPD

min

x u

max

3.

6 ±

2.2

0.

1006

VP

Dm

in x

uav

g -3

.0 ±

3.1

0.33

37

VPD

avg x

um

ax

-1.3

± 2

.3

0.

5902

VP

Dav

g x u

avg

4.5

± 3.

5

0.20

01

VPD

min

x R

tot

-5.5

± 2

.5

0.

0302

VP

Dav

g x

R tot

7.

8 ±

1.7

<.

0001

Chapter Two

32

Table 2.3: Proportion of variance explained by the multiple regression models of tree daily sap flow shown in Table 2.2. R2 values were computed including among tree variation (accounting for meteorological effects only), and also with among tree variation accounted for (by including tree individuals as factors in the model; meteorological + tree effects).

Meteorological effects

Meteorological + tree effects

Marillana Creek 0.23 0.79Yule River 0.27 0.67Sites pooled 0.42 0.80

Table 2.4: Tests for differences in tree sap flow (a) between sites and (b) after rainfall, after adjusting for meteorological factors. Site or recent rainfall (mm falling within the previous 10 days) was added as a fixed-effect variable to the meteorological models shown in Table 2.2. Where significant, the regression coefficient estimate and standard error are shown. ΔcAIC is the reduction in score when the additional factor is added to the model. The log likelihood ratio (LLR) and type-3 F-test (p>F) probabilities are also provided.

(a) Effect of site Coefficient Δ cAIC LLR p > χ2 p > F Yule River relative to Marillana Creek 134 ± 51 15.1 0.0005 0.0343

(b) Effect of rainfall

Yule River - 3.6 0.0544 0.1423 Marillana Creek - 1.1 0.2942 0.1388

average 1.34 ± 0.51 m day-1 faster than trees at Marillana Creek, if all meterological

variables were identical between the sites. However, recent rainfall had no significant effect

on sap flow at either Marillana Creek or Yule River once the other meteorological variables

were controlled for, indicating no effect of surface soil moisture on rates of tree water use

(Table 2.4b).

Seasonality in tree water use was still highly significant (p<0.0001) after accounting

for meterological and site effects (Table 2.5). The residuals from the pooled site multiple

regression model, averaged over each month, are shown in Figure 2.6. The actual water use

was less than predicted by the model during February to May, and higher than predicted

during July to December. Water use adjusted for site and meteorological differences was

also significantly lower in all autumnal months compared to spring months.


33

Table 2.5: Tests for seasonal differences in tree water use after adjusting for meteorological variables. Month was added as a fixed-effect variable to the meteorological models shown in Table 2.2, with site and rainfall variables also included where significant (Table 2.4). ΔcAIC is the reduction in score when month is added to the model. The log likelihood ratio (LLR) and type-3 F-tests (p>F) for the effect of month are also provided.

Δ cAIC LLR p > χ2 p > F

Yule R 219.2 <0.0001 <0.0001 Marillana Creek 450.7 <0.0001 <0.0001 Sites pooled 492.3 <0.0001 <0.0001

Tree hydraulic conductance

Since seasonal variation in tree sap flow was not fully explained by meteorological

conditions, whole tree hydraulic conductance (K) was examined for evidence of seasonal

variation. Changes in factors such as canopy cover, quantity of conducting sapwood or root

absorptive surface area can alter tree K and affect rates of tree water use. There was no

detectable variation in tree K seasonally, as the relationship between Vs and Ψl across the

diurnal cycle remained constant, with data points from different seasons falling within the

same line (examples in Figure 2.7).

-30

-20

-10

0

10

20

30

Resi

dual

Vs (

cm d

ay-1)

Figure 2.6: Difference

between the measured tree

sap flux (Vs), and the sap flux

predicted by a multiple

regression model of mete-

orological and site effects.

Values are residual daily Vs averaged by month (mean ± standard error). Negative values indicate lower rates of sap flux than predicted by the model, and positive values higher rates than predicted. Different letters indicate significant differences between months.JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

ac

ce

bb

b

b

ab

cd cd

dde de

Chapter Two

34

DISCUSSION

Factors determining water use of M. argentea

My results confirm that water use by M. argentea in the Pilbara is primarily determined by

atmospheric evaporative demand, with up to 79% of day to day variation explained by

atmospheric conditions. Pulses of surface soil moisture resulting from sporadic rainfall

events had no discernible effect on rates of sap flow. Pre-dawn Ψl was also close to zero

throughout the year and these findings collectively confirm that water use was not limited

by availability of water to the roots. These findings are consistent with observations of other

tree species in similar environments. For example, transpiration was strongly related to

meteorological factors and unaffected by rainfall in Populus deltoides forests above a shallow

water table, on the semi-arid Middle Rio Grande river system, New Mexico (Cleverly et al.,

2006). In contrast, where trees do not have access to permanent water, particularly in semi-

arid and arid regions, rates of tree water use typically increase following precipitation events

(e.g. Hubbard et al., 1999; Zeppel et al., 2008; Du et al., 2011).

As water supply was constant, this study provided an opportunity to more closely

examine the effects of atmospheric variables on tree water use over time. Evaporative

demand, wind speed and direction, temperature and solar irradiance can all influence

0

5

10

15

20

0

10

20

30

0 0-0.5-1.0-1.5-2.0-2.5 -0.5-1.0-1.5-2.0

V s (c

m h

-1)

Ψleaf (MPa)

Vs (cm

h-1)

June

NovemberMay

July

November

(a) (b)

Figure 2.7: Examples of relationships between sap velocity (Vs) and leaf water potential

(Ψleaf), used to assess whole tree hydraulic conductance. Data are shown for (a) one tree at the Yule River in winter (June 2009), spring (November 2009) and autumn (May 2010), and (b) one tree at Marillana Creek in winter (July 2010) and spring (November 2009).


35

transpiration rates, but since these are all inter-correlated, it is generally difficult to

determine which factors might be the most important drivers of water use (Leng et al.,

2006). However, VPD was the variable that best correlated with sap flow in M. argentea,

suggesting that this may be the primary driver of tree water use. Wind speed may play an

additional role, since the interaction between VPD and wind speed explained sap flow

patterns significantly better than VPD alone. Tree water use is frequently influenced by

wind speed, as under conditions of low air movement, transpiration can be inhibited by the

buildup of a humid boundary layer around the leaves (e.g. Chu et al., 2009; Zheng &

Wang, 2012). Leaf boundary layer conductance may therefore be a significant factor in the

water use of M. argentea, as is commonly the case for broad-leafed species (Meinzer et al.,

1997a; Cleverly et al., 2006). My results also indicate that atmospheric controls on tree

water use may have differed between the two study sites. Water use by trees at Marillana

Creek appeared more closely linked to meteorological drivers compared to the Yule River

site (79 % compared to 67%) despite the more complex multiple regression equation used

to describe water use in the Yule River trees. Relationships between climate and water use

by other riparian tree species have been shown previously to differ between sites (Cleverly et

al., 2006). Reasons for such variation among stands may relate to differing site

characteristics, but remain largely unknown.

While seasonal patterns of sap flow were largely explicable by atmospheric variables

in all the M. argentea trees examined, the absolute rates of sap flow were highly variable

among individuals. Daily sap flow varied by two- to three-fold among trees that were

growing under the same environmental conditions. Daily rates of sap flow also differed

substantially between the two study sites. Inter-tree and site variability in water use by M.

argentea trees has been previously noted by O’Grady et al. (2006a), on the Daly River in

the wet tropics of Northern Australia. Likely reasons for inter-tree variability both within

and among sites include tree size (such as differences in sapwood area and foliage area), and

variation in sap flow rates within different regions of the sapwood (e.g. Gebauer et al.,

2008; Zeppel & Eamus, 2008).

In the Pilbara region, M. argentea grows in extreme VPD conditions, while having

constant access to a reliable water supply, and the rates of sap flow were therefore relatively

high. The maximum observed site averages of Vs in the present study were 23 cm h-1 and

Chapter Two

36

3.9 m day-1 at Marillana Creek, and 31 cm h-1 and 5.2 m day-1 at the Yule River. These

rates are similar to those measured on M. argentea trees in the wet tropics of the Northern

Territory (O'Grady et al., 2006a). The sap flow rates of M. argentea are also within the

range of values reported for riparian trees in other semi-arid regions. For example, studies

in semi-arid Arizona, USA, recorded peak flow rates for Populus species with access to

permanent water of 15–35 cm h-1, and for Salix species up to 60 cm h-1 (Schaeffer et al.,

2000; Gazal et al., 2006). Nocturnal sap flow in M. argentea was also similar to rates

reported for semi-arid poplars (10–25% of total daily water use; Gazal et al. 2006). Water

use and water use physiology of semi-arid riparian trees therefore appears to be similar

across the continents.

Seasonality of water use in M. argentea

I found a distinct seasonal pattern in the water use of M. argentea in the semi-arid Pilbara,

largely corresponding with seasonal fluctuation in atmospheric conditions. This result is in

contrast to O’Grady et al. (2006), who did not observe any significant seasonal variation in

M. argentea water use in the wet-dry tropics at the Daly River. I attribute this partly to

climate differences between regions; water use was lower during the cool winter months in

the Pilbara, while the VPD in the tropics can be highest during the ‘winter’ dry season

(June to August). The long period of continuous measurement over this study also allowed

me to analyse water use over time in much greater detail. O’Grady et al. (2006) collected

sap flow measurements during three two-week periods across the year, which may not have

been sufficient to detect any subtle seasonal patterns that may have been present. Seasonal

climatic fluctuations in the Pilbara clearly drive seasonal variation in the water use of M.

argentea.

Although water use by M argentea was largely explained by meteorological variation,

the ‘shape’ of the relationships of water use with the major driving variables changed across

the year, contrary to expectations. Hysteresis was evident in the relationships over both

diurnal and seasonal cycles. M. argentea water use exhibited counter-clockwise hysteresis

with irradiance on most days in winter and spring, with the extent of hysteresis decreasing

during summer and autumn, and moving to clockwise hysteresis on a few days in summer.

These changes in diurnal hysteresis suggest that Vs was increasingly limited by water

availability to the canopy during summer and autumn. Diurnal hysteresis of water use with


37

respect to VPD and solar irradiance is commonly observed, and can be explained by the

time lag that exists between the irradiance and VPD conditions, stomatal responses to light,

as well as changes in the availability of water to the leaf surfaces over the diurnal cycle (e.g.

Zeppel et al., 2004; Zheng & Wang, 2012). With abundant water availability, if stomatal

conductance is primarily a function of light intensity, then the fact that maximum light

levels occur at midday, while VPD peaks mid-afternoon, will generate clockwise hysteresis

of water use with VPD and counter-clockwise hysteresis with light (Zeppel et al., 2004). At

a given VPD, the light intensity and therefore gs is greater during the morning than at the

same VPD in the afternoon, driving greater Vs in the increasing phase of VPD than the

decreasing phase generating clockwise hysteresis. Similarly, at a given light intensity, the

VPD is lower in the morning than in the afternoon, driving higher Vs during the decreasing

phase of light intensity than during the increasing phase, resulting in counter-clockwise

hysteresis. However, if water use is also influenced by a reduction of water supply over the

daylight period, such as that caused by depletion of water in the rhizosphere or in internal

water stores, or through increasing xylem embolism over the course of the day, then gs will

tend to be higher in the morning, than at the same VPD or light intensity in the afternoon.

Thus, diurnal regulation of transpiration by a combination of both light intensity and

water limitation ought to lead to a reduced afternoon water use, manifesting as an increase

in the extent of the clockwise hysteresis with VPD, and a reduction in the extent of the

counter-clockwise hysteresis with light, eventually moving into clockwise hysteresis with

light, as the severity of water deficit increases.

A seasonal shift in the diurnal pattern of hysteresis between water use and irradiance

was seen in M. argentea despite continuous availability of water to the roots. Strong

counter-clockwise hysteresis with irradiance has been shown in a well irrigated Populus

hybrid in Washington, USA (Meinzer et al., 1997b), and also in numerous tree species in

wet, high humidity tropical forests in Costa Rica (O'Brien et al., 2004) and Indonesia

(Horna et al., 2011). Patterns of hysteresis with light also show clear changes during

drought. For example, water use by Eucalyptus crebra in northwest New South Wales,

Australia, displayed counter-clockwise hysteresis with irradiance during the wetter summer,

but little or no hysteresis during a prolonged drought (Zeppel et al., 2004). Water use by

Haloxylon ammodendron trees in arid China also displayed counter-clockwise hysteresis with

Chapter Two

38

irradiance after rainfall, shifting to clockwise hysteresis under dry conditions (Zheng &

Wang, 2012). The similar pattern seen in M. argentea, of counter-clockwise hysteresis of

water use with irradiance in the winter and spring, moving to no hysteresis or clockwise

hysteresis in summer and autumn indicates seasonal changes in the supply of water to the

leaf surfaces, such as through a limitation in the transport capacity of the trees.

Water use by the M. argentea trees exhibited clockwise diurnal hysteresis of Vs with

VPD year round, with no clear patterns evident in the extent of hysteresis. Similarly,

clockwise hysteresis in the relationship between water use and VPD appears almost

ubiquitous across species, and the extent of the hysteresis is usually not clearly attributable

to water availability. For example, hysteresis with respect to VPD did not alter in response

to changing soil water content in Eucalyptus globulus in Tasmania, Australia (O'Grady et

al., 2008). Hysteresis with VPD also did not vary with water availability in any of four tree

species planted in an urban park in Liaoning Province, China (Chen et al., 2011).

However, hysteresis did increase during the dry season in E. tetrodonta and E. miniata in

northern Australia (O'Grady et al., 1999). The extent of diurnal hysteresis with VPD

appears to depend most strongly on the level of VPD itself, perhaps reflecting more

complex regulation of stomatal and leaf conductance by combined factors of VPD, light

and water availability (O'Grady et al., 2008; Guyot et al., 2012). Hysteresis in the

relationship between water use and irradiance therefore appears to be a more useful

indicator of water limitation than hysteresis with respect to VPD.

In the present study, hysteresis was also observed over the annual cycle in the water

use of M. argentea trees with respect to VPD, but not irradiance. Water use patterns over

annual cycles might display patterns of hysteresis for similar reasons as over diurnal cycles;

the annual peaks in irradiance and VPD are offset in a similar manner as seen diurnally,

and water availability can also vary seasonally. Therefore, by the same reasoning as applied

to diurnal cycles, the lack of hysteresis in the relationship between Vs and irradiance

suggests restricted water use during the summer and autumn (the period of decreasing

irradiance). The annual patterns of hysteresis were still present after accounting for other

meteorological factors; compared with that predicted from meteorological conditions, tree

water use was higher in spring and early summer, declined from mid-summer and through

autumn, then began to increase in the winter. Seasonal differences in the relationships of


39

tree water use with VPD were also observed in E. crebra which were attributed to the

marked seasonal differences in soil water content and plant water status (Zeppel et al.,

2004). Similarly, in M. argentea, both seasonal patterns, and the seasonal variation in

diurnal patterns, suggest that water use by M. argentea was restricted during summer and

autumn.

The apparent water deficit during parts of the year observed in this study may be

related to limitations in the rate of water transport to the canopy. The seasonal variation in

tree water use patterns are unlikely to be due to changes in water availability to the roots

due to soil drying, since the M. argentea trees appeared to be accessing a perennial shallow

water table at both study sites, and there was no discernible tree response to small rainfall

events. Sap flow readings can decline over time due to the wounding response of xylem

tissue (Hatton et al., 1995; Hogg & Hurdle, 1997). However, a wounding artefact is

unlikely to have produced the observed patterns of Vs, since the pattern was consistent

across study sites, regardless of whether the sequence of measured seasons ran from autumn

2009 to autumn 2010 (Yule River), or winter 2009 to spring 2010 (Marillana Creek). In

addition, progressive wounding over the study period should not affect diurnal patterns,

and so would not explain the seasonal variation in diurnal hysteresis.

Alternative explanations for the seasonal response here are changes to tree hydraulic

conductance, which may occur with phenology of growth (canopy cover and conducting

xylem tissue in particular), or changes in tree capacitance. While not measured directly at

my sites during the time of the study, long-term monitoring at Weeli Wolli Creek,

upstream of Marillana Creek, showed no seasonal changes of canopy cover of M. argentea

(Figure 2.8; Rio Tinto Pty Ltd, unpublished data). This is in contrast with data from

locations experiencing groundwater abstraction, where clear reductions in canopy cover

were detectable in M. argentea trees over the same period (Rio Tinto Pty Ltd, unpublished

data). Canopy cover of M. argentea therefore appears not to alter as a regular seasonal

adjustment at undisturbed locations, but partial canopy loss may occur as a safety

mechanism under water stress. There were also no visible growth rings in the wood of the

studied M. argentea trees, indicating relatively continuous growth, similar to many wet

tropical forest trees (Martı́nez-Ramos & Alvarez-Buylla, 1998). Seasonal variation in

production and loss of leaves and conducting xylem are therefore unlikely at the sites used

Chapter Two

40

in the present study. Likewise, no difference was detectable in whole tree resistance of M.

argentea among seasons, either in the present study, or by O’Grady et al. (2006a) in the

wet-dry tropics. Seasonal changes in tree K that were not detectable from the measurements

of Vs and Ψ conducted in the present study or by O’Grady et al. (2006a) cannot be entirely

ruled out. However, changes in the use of internal water stores by M. argentea across the

year appear to be the most likely explanation for the seasonal variation in water use patterns

observed here.

My results suggest that M. argentea may utilise water storage over annual cycles. M.

argentea is likely to have a reasonably high wood capacitance, due to its moderate to low

wood density (Borchert & Pockman, 2005). Eucalyptus victrix, a co-occurring riparian

species in the Pilbara region, has been shown to utilise internal water storage over diurnal

cycles, despite constant access to groundwater, to meet the water transport demands of the

high VPD environment (Pfautsch et al., 2011), and the same process might occur in M.

argentea. Seasonal fluctuation in stem water storage is commonly observed in tree species

subjected to seasonal drought, where water withdrawn from internal reserves during dry

periods and replenished during wetter months enables a greater total water use (e.g.

Loustau et al., 1996; Baker et al., 2002; Domec et al., 2005; Betsch et al., 2011). Although

the M. argentea trees had a constant water supply, under the extreme VPD conditions

during parts of the diurnal and annual cycles the rate of water transport appeared to

Figure 2.8: Leaf area index (LAI) of

Melaleuca argentea trees at undis-

turbed sites on Weeli Wolli Ck, across

seasons (Rio Tinto Pty Ltd, unpub-

lished data). Values are relative changes over time in canopy cover, determined at long term monitoring points by digital cover photography (Macfarlane et al., 2007). Values are means ± standard error of eight to ten trees, and did not differ significantly between seasons.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

JUN2010

OCT2010

DEC2010

MAR2011

LAI

(cov

er p

ropo

rtio

n)

a aa

a


41

become the limiting factor in tree water use; high capacitance might partly overcome this

limitation.

Internal water stores are depleted during daylight hours and refilled during the dark

period, but full replenishment does not always occur by dawn (Bucci et al., 2004; Scholz et

al., 2007). Nocturnal sap flow comprises refilling as well as nocturnal transpiration (Phillips

et al., 2010). The correlation between nocturnal Vs and VPD in M. argentea indicates night

time transpiration losses, however, flows still occurred on nights when VPD was close to

zero, suggesting that parts of the nocturnal flow were due to refilling. Nocturnal Vs was also

greater during spring than autumn for any given VPD, suggesting the capacity to replenish

stores may be lower in summer and autumn. High VPD and rates of water use during

summer may progressively deplete internal stores in M. argentea, despite constant access to

groundwater. The lower VPD during winter may provide a respite that allows the trees to

refill under conditions of reduced leaf level water loss. However, further data would be

needed to verify seasonal changes in water storage and capacitance in M. argentea, such as

sap flow measurements from sensors distributed along the trunk and upper branches

(Meinzer et al., 2004b; Phillips et al., 2009), measurements of changes in wood water

content over time and/or diurnal changes in stem circumference (Čermák et al., 2007;

Betsch et al., 2011).

Overall, the relationships between temporal patterns in water use and meteorological

factors (light, VPD and wind speed) were more complex than expected. I observed distinct

seasonal variation in water use patterns, which was not accounted for by seasonal

meteorological variation. With access to a perennial water supply, but in a high VPD

environment, rates of water transport appear to limit the water use of M. argentea during

high demand periods, over both diurnal and annual cycles. Stored water likely enables the

higher rates of transpiration by the trees during spring and early summer. However, I

hypothesise that the long periods of high VPD progressively deplete water stores, leading to

reduced rates of water use during late summer and autumn. The short period of low VPD

conditions in the winter may allow recharging of stored water. Nonetheless, water use by

M. argentea in the Pilbara region can be described to a large extent by the meterological

variables of light, VPD and wind speed, as would be expected for a tree with constant

access to groundwater. Water use patterns over time are thus relatively predictable, which

Chapter Two

42

might allow relatively simple methods of calculating total water fluxes through trees over

annual and multi-annual scales. However, there was considerable unexplained variation in

sap flow rates among trees and between the study sites. Further study of this variability

would clarify spatial patterns in tree water use, and aid in quantifying the hydrological

fluxes through trees.

43

CHAPTER THREE: Quantifying water fluxes at tree and stand scales

from sap flow measurements in the riparian tree Melaleuca argentea

INTRODUCTION

Volumetric estimates of tree water use at the scale of whole trees and stands are important

in both assessing vegetation water requirements for resource management, and for

understanding tree and ecosystem functioning. For instance, whole tree water use

measurements are important in understanding physiological controls on transpiration, such

as the roles of stomata and hydraulic architecture (Meinzer et al., 2013). Furthermore, trees

usually play a dominant role in transpiration and redistribution of water within ecosystems,

and so tree water fluxes are an important component of most ecosystem water budgets

(Baird & Maddock, 2005; Eamus et al., 2006b; Domec et al., 2010). However, few

estimates of water requirements or fluxes have been made for any of the tree species of

northern or inland Australia.

Sap flow measurements are often used to obtain accurate point measurements of rates

of water flow through trees, which may then be scaled up to the level of whole trees and

stands (e.g. Schaeffer et al., 2000; Doody & Benyon, 2010; Tateishi et al., 2010). However,

for such scaled estimates to be accurate, the sampling must adequately encompass the

spatial and temporal variation in sap velocity (Fiora & Cescatti, 2006; Ford et al., 2007;

Kume et al., 2010; Tsuruta et al., 2010). Failure to account for the variability can lead to

significant margins of error in water use estimates. For example, estimates of whole-tree

water use by the mangrove Avicennia marina varied by up to 240% depending on where

sap flow sensors were located within the sapwood (Van de Wal et al., 2013).

Spatially, sap flow varies within trees throughout different regions of the sapwood,

between trees within a stand, and between stands. Within the sapwood, flow velocities are

generally higher in the younger, outer wood and decrease moving inward toward the

heartwood, as vessels become increasingly blocked by tyloses with age. The shape of this

radial flow distribution appears to be species specific, with both linear and curved patterns

observed depending on properties such as the pattern of vessel distribution across the

sapwood (Cohen et al., 2008; Gebauer et al., 2008). The relationship between sap flow

Chapter Three

44

velocity and radial depth generally needs to be verified for each species, if water use

estimates are to be accurate, by placing sensors at multiple depths. Sap flow also varies

azimuthally, around the tree stem circumference, in a manner that is usually less consistent

among trees than radial variation (Cohen et al., 2012). Azimuthal variation may relate to

variation in vessel anatomy and the positioning of major roots and branches (e.g. Tateishi et

al., 2008; López-Bernal et al., 2010). Accounting for sapwood variation is crucial in

obtaining volumetric estimates of water use at the whole tree scale.

When seeking to scale up water use to the stand and landscape scales, variability in

flux among trees and between stands can become more important than within tree

variation. Variation in tree size, quantity of conducting sapwood area, and canopy area can

all contribute to significant differences in water use between trees. For instance, in Japanese

cedar stands, between-tree variation was identified as the most important level to account

for in estimating stand water use (Kumagai et al., 2005; Shinohara et al., 2013). In twelve

trees within one stand, sap flux density varied more than 2.6-fold between trees for

unknown reasons, with the variation not explained by factors such as tree size (Kumagai et

al., 2005). In Pinus strobus plantations in North Carolina, USA, variation between different

stands in tree sapwood area and stand density was the source of greatest variability in

estimating catchment scale water fluxes (Ford et al., 2007). A continuing challenge in

estimating landscape scale water fluxes is to understand and quantify the variability in water

use among trees and locations.

For stand and landscape scale tree water use estimates, universal scaling relationships

have been proposed based on factors such as tree structure, stand density, and moisture

availability (e.g. Enquist, 2002; Meinzer, 2003; Meinzer et al., 2004a; Zeppel & Eamus,

2008; O'Grady et al., 2009), but these require significant validation. Inherent species

differences in properties such as stand structure and sensitivity to water deficit may mean

that these scaling relationships do not always hold, and species and/or site specific

relationships may be required (Aranda et al., 2012). In addition, the convergence of water

use among species tends to be related to the constraint of limited soil moisture availability

(e.g. Meinzer et al., 2004a; Kelley et al., 2007; O'Grady et al., 2009). Thus, it remains

unclear whether the proposed universal scaling relationships might apply to species in

wetland-type ecosystems, where water is constantly available. Direct measurements of

Quantifying water use by Melaleuca argentea

45

individual tree sap flow are therefore valuable to determine patterns and variability of tree

water use, particularly in non-water-limited environments.

Temporally, sap flow varies over diurnal and seasonal cycles as well as inter-annually,

with changes in environmental conditions. For instance, transpiration is regulated by

stomata in response to irradiance and water availability, and total tree water use also

depends on the canopy area, which can display phenological cycles in response to seasonal

fluctuation in temperature or water availability (e.g. Hutley et al., 2001; Gazal et al., 2006;

Zheng & Wang, 2012). At the stand and landscape scales, water flux estimates for time

periods of a year or longer are frequently required. Measurements taken during shorter time

windows distributed across the year are commonly scaled up to annual fluxes, for example

measurements may be taken for a few weeks during the wet and dry seasons (e.g. Irvine et

al., 2004; O'Grady et al., 2006a; Zeppel et al., 2008). However, this approach has rarely

been tested. It might also be possible to estimate tree water use over time from

environmental conditions such as atmospheric demand and soil water availability, since

these factors are the main drivers of tree water use (Granier et al., 2000). If species and/or

site specific relationships can be determined, monitoring of the appropriate environmental

variables might allow ongoing estimates of tree water flux. For species and sites where water

use is driven largely by atmospheric demand, then it might be possible to accurately

estimate tree water use from meteorological data alone, which is usually widely available

and easily accessed.

Water resources throughout the world are under increasing demand from human

populations, requiring greater degree of water management and accounting, in order to

balance competing human and ecological requirements (e.g. Arthington et al., 2010; Shen

& Chen, 2010). In the semi-arid Pilbara region of northwest Australia, rapid population

growth and groundwater perturbations related to mining activity are placing increasing

pressure on the riparian ecosystems (Department of Water, 2010). However, the region

remains relatively remote and there has been limited quantitative assessment of vegetation

water requirements against the background of extreme hydroclimatic variability (Leighton,

2004; Dogramaci et al., 2012). M. argentea is one of the dominant riparian tree species in

the Pilbara region, found only in locations with perennial surface or near-surface water.

Rates of water use by this species are relatively high (Chapter 2), leading to high vegetation

Chapter Three

46

water requirements. In some locations, water use by M. argentea is therefore likely to

contribute significantly to total ecosystem hydrological fluxes. Volumetric estimates of tree

water use would be valuable in assessing vegetation water requirements. Obtaining such

volumetric estimates from sap flow measurements requires characterisation of the spatial

and temporal patterns of sap flow variability, and validation of approaches to extrapolate

localised measurements obtained over short time periods to larger spatial and temporal

scales.

In Chapter 2, I demonstrated that sap flow of M. argentea was highly variable among

trees, both within and among sites, a finding consistent with studies undertaken on this

same species in the tropics (O'Grady et al., 2006). My objective here was to assess spatial

variation in sap flow of M. argentea, including within and between tree variation. I

examined patterns of sap velocity within the sapwood and across tree sizes, as well as tree

allometric ratios, to identify relationships that could facilitate water use estimates at whole

tree and stand scales, from fewer point measurements of sap flow. I also demonstrated in

Chapter 2 that sap flow of M. argentea is strongly related to the atmospheric evaporative

demand, with no detectable influence of variation in soil moisture content, due to constant

access to a shallow water table. I generated regression models from meteorological variables

at two contrasting sites within the Pilbara region, which explained up to 80% of the daily

variation in tree sap flow (Chapter 2). In the present study, I hypothesised that accurate

estimates of total annual tree sap flow could be derived from these models. In the present

study, I extracted annual estimates of sap flow from these existing models, and compared

these with the measured values. I also tested whether these models may be broadly

applicable, or whether different models would need to be determined for different stands of

trees, by applying the models to a different stand of trees from which they were derived. In

addition, the long period of continuous sap flow measurements (more than 12 months) at

each site allowed me to test the approach of scaling up sap flow measurements from short

time windows to total annual flows. Four-week windows were selected at evenly spaced

time points across the year, and the scaled estimates compared with the measured total

annual sap flow. Characterisation of spatial sap flow variability, and validation of temporal

scaling approaches, will assist in determining the extent of sap flow sampling required to


47

obtain acceptable estimates of water use by M. argentea at the scale of whole trees, stands

and landscapes.

MATERIALS & METHODS

Study sites

The study was conducted at undisturbed locations on the Yule River and Marillana Creek,

two intermittent waterways in the Pilbara region of NW Australia as described in Chapter

2. Measurements were conducted on the same ten Melaleuca argentea trees as the study

presented in Chapter 2.

Sap flow measurement

Sap flow measurements were performed with heat pulse velocity probes via the heat ratio

method (HRM30; ICT International, Armidale, NSW, Australia), as described in Chapter

2. The measurement of wood properties, and conversion of pulse velocity to sap velocity

(Vs) are also described in Chapter 2.

Variation in Vs within the sapwood was assessed in four trees at Marillana Creek. In

two trees, multiple sensors were installed on the eastern side of the stem at 1.2 m height,

with measurement points positioned at 5 mm depth increments into the sap wood from

2.5 mm to 27.5 mm, to assess radial variation in Vs. In another two trees at Marillana

Creek, three or four sensors were installed at evenly spaced positions around the stem

circumference, to assess azimuthal variation. Measurements were recorded every 30 min for

two weeks during July 2010. To estimate volumetric sap fluxes, the cross-sectional sap

wood area (As; see next section) was divided into sections (concentric radial bands or

azimuthal quadrants) delineated by the midpoints between adjacent sensor points.

Volumetric flow was then calculated as the product of Vs and the As of the section in which

the sensor point was located (Burgess et al., 2001). Total tree water use (Q) was estimated

as the sum of the volumetric fluxes of all sap wood sections in the tree.

The variation in Vs among trees, and over time, was quantified from the data

presented in Chapter 2. Briefly, sensors were installed in five trees at the Yule River and five

at Marillana Creek, and data collected from a measurement point positioned at 7.5 mm

Chapter Three

48

depth (the approximate centre of the sap wood band). Vs measurements were recorded

every 30 min for approximately 12 months.

Tree sapwood area

The cross-sectional sapwood area (As) at 1.2 m height was determined for each study tree as

a measure of tree size relevant to water use, and for estimation of volumetric fluxes (see

previous section). The sapwood depth (ws) was determined from 3–4 wood cores taken at

equally spaced positions around the tree circumference with an increment borer (Haglöf,

Långsele, Sweden). Surfaces of the fresh cores were shaved back and stained with methyl

orange (1% w/v in 4% v/v methanol); the sapwood stained deep red and the heartwood

yellow. The position of the colour change corresponded well with the boundary of

conducting sap wood as determined by adjusting the depths of the sap flow probes. Bark

depth (wb) was measured with a needle-type gauge at a minimum of four points around the

tree, and overbark diameter (d) measured with a diameter-tape. The cross-sectional

sapwood area was determined from the above measurements assuming circularity of the

stem cross section:

As = π(½d – wb)2 – π(½d – wb – ws)2

Tree hydraulic conductance

Conductance of the whole tree hydraulic pathway (K) was calculated as the slope of the

linear part of the relationship between Vs and leaf water potential (Ψl), obtained from the

data presented in Chapter 2 (examples in Figure 2.8).

Analysis of variation in sap flow

Statistical analysis was performed in SAS version 9.2 (SAS Institute Inc.; Cary, NC, USA).

The effects of within-sapwood variation in sap flow on the calculation of volumetric tree

water use was assessed, by comparing the total tree water use calculated using a single

measurement point, with that calculated using the weighted average of all measurement

points (positioned at intervals radially or azimuthally). Relationships of As with sap flow, K,

Ψl and tree diameter were assessed by linear regression, and differences in relationships

between seasons and sites tested by ANCOVA.

Extrapolation of water use over time from shorter time periods was tested, by

assuming the measurements from equally spaced selected months across the year were


49

representative of an equal portion of the year. Estimation of water use from meteorological

data was also tested. Models of daily tree sap flow based on meteorological variables for

each site were described in Chapter 2. When used to derive daily estimates of water use, the

models gave slight over- or under-estimates depending on the season (Chapter 2). In the

present chapter, I assessed these models for their utility in obtaining water use estimates

integrated across longer periods of time, by comparing the measured monthly and annual

total sap flow with the values predicted by the models. I also tested whether these

meteorological models may be applicable to different sets of trees from those from which

they were derived, by using the model produced for one study site to predict water use at

the other. The fixed (meteorological) effects from the models were used for these tests, and

tree As was also added as an additional fixed effect, to account for any effect of tree size on

rates of water use. The accuracy of the water use estimates derived from calculated and

extrapolated values are reported as percentage deviance from the actual measured values of

sap flow.

RESULTS

Sap flow variation within the sapwood

Across the radial sapwood profiles Vs was highest toward the outer sapwood, and showed

and almost linear decrease to zero toward the central heartwood. However in one tree Vs

peaked at 7.5 mm depth, while in the other tree highest Vs was recorded by the sensor at

2.5 mm depth (Figure 3.1). The linear portion of the decline across the radial profile was

significantly steeper in tree 1 than tree 2 (p=0.038). The proportional contribution of sap

flow at each depth remained constant over time, over diurnal cycles (Figure 3.1c, d) and

over the two week measurement period (Figure 3.1e, f).

Azimuthally, Vs varied both over time and between trees (Figure 3.2). In tree 3,

morning sap flow was greatest on the northern and eastern sides of the stem, midday flow

was greatest on the eastern side, and afternoon flow on the southern side (Figure 3.2a).

Over the course of a day the eastern side contributed the most to water use and the western

side the least. In contrast, in tree 4 morning flow was greatest on the eastern side, while

midday and afternoon flow were greatest on the southwest (Figure 3.2b). Over the course

of a day, the southwestern side contributed the most to sap flux and the northern the least.

Chapter Three

50

There

was an approximately 2-fold difference in Vs between the fastest and slowest azimuths

during the middle of the day, in both trees. The proportional contribution to flux of each

azimuth therefore varied over the course of a diurnal cycle (Figure 3.2c, d), although the

overall diurnal pattern was similar throughout the sapwood, and between days (Figure

3.2e).

Due to the variation in Vs throughout the sapwood of these study trees, calculating

total tree water use (Q) from measurements at a single sensor point would be inaccurate for

many of the points. Daily Q values calculated as the average of multiple sensor points

(weighted by the As of the azimuth or radial band in which each was located) are compared

0

2

4

6

8

10

12

14

16

18

12:0018:00

0:00

8:00

0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20 25 0 5 10 15 20 25 30

V s (c

m h

-1)

total daily

Cont

ribut

ion

to to

tal (

prop

ortio

n)

(a) (b)

(c) (d)

Tree 1 Tree 2

Radial depth (mm)

Figure 3.1: Variation in sap flow velocity (Vs) with radial depth (0 is the cambium) into the

sapwood. (a & b) Vs across the radial profile, and (c & d) the proportional contribution of each radial position to the total sapflux across the profile (weighted by the sapwood area of each radial band), in (a & c) tree 1 and (b & d) tree 2. Data shown are from July 24th 2010, as a repre-sentative example, for mid-morning, midday, mid-afternoon, midnight and for the cumulative sap flux over 24 hours (total daily).


51

with Q calculated from each sensor point alone in Table 3.1. The values presented in Table

3.1 are for a single day as an example, but similar results were obtained for the other days

during the two week measurement period. Failing to account for radial variation when

calculating Q in these study trees resulted in potentially larger errors (over- or under-

estimates of up to 70% of the weighted average) than failing to account for azimuthal

variation (which lead to estimates differing from the weighted average by up to 35%).

Sensors positioned in the radial band of fastest Vs yielded the least accurate single-point

estimates of Q.

Figure 3.2: Variation in sap flow velocity (Vs) with azimuthal position around the stem

circumference, at 7.5 mm depth into the sapwood. (a & b) Vs with azimuthal position, and (c

& d) the proportional contribution of each azimuth to the total sap flow around the stem, in (a

& c) tree 3 and (b & d) tree 4. Data shown are from July 24th 2010, as a representative example, for mid-morning, midday, mid-afternoon, midnight and for the cumulative sap flow over 24 hours (total daily).

0

10

20

30

40

0

0.1

0.2

0.3

0.4

0.5

0.6

12:0018:00

0:00

8:00

total daily

(a) (b)

(c) (d)

N E S W

V s (c

m h

-1)Co

ntrib

utio

n to

tota

l (pr

opor

tion)

N E S WAzimuthal position (compass direction)

Tree 3 Tree 4

Chapter Three

52

Temporal variation in sap flow

Although the absolute Vs and diurnal Vs patterns varied among different regions of the

sapwood, the relative patterns of daily Vs over time remained consistent throughout the

Table 3.1: Comparison of estimates of whole tree water use (litres) during a single representative day (24th July 2010), calculated as the weighted average of multiple sensor positions or from single sensor points. Sensors were placed at (a) 3 radial depths or (b) 3–4 azimuthal positions within the sapwood.

(a) radial Sensor position (mm depth)

weighted average

2.5 7.5 12.5

Tree 1 calculated water use 21.1 12.2 33.6 17.5 % difference from weighted average -42 59 -17

Tree 2 calculated water use 14.5 24.6 13.0 7.9 % difference from weighted average 70 -10 -46

(b) azimuthal Sensor position

weighted average N E S W

Tree 3 calculated water use 29.1 30.1 34.5 32.1 20.0 % difference from weighted average 3 19 10 -31

Tree 4 calculated water use 117.1 76.1 109.4 157.3 (SW) % difference from weighted average -35 -7 34

sapwood. The measurements taken at a single point within each of the ten study trees

therefore appear to be valid for assessment of temporal variation in Vs, over the scale of days

and months.

The accuracy of estimating total annual sap flow by extrapolating from shorter

measurement periods was tested by extrapolating from the sap flow data during April and

October only (beginning and end of the dry season), or during January and July only (mid-

summer and mid-winter), or from data during all four of these months. The estimates

deviated from the measured annual sap flow by at most 16% for Marillana Creek trees and

9.8% for the Yule River trees (Table 3.2). Using January and July measurements gave

slightly better estimates than using April and October measurements, particularly for the

Yule River trees. Using quarterly measurements did not yield substantially better estimates

than using bi-annual data.


53

The possibility of quantitatively estimating sap flux over time from meteorological

variables was assessed using linear regression models previously determined for these study

trees (Chapter 2). Estimates for individual months deviated from the measured values by

up to 5% at the Yule River, and by up to 16% at Marillana Creek. However, over the year

as a whole the estimates deviated from the measured by less than 3% at both sites (Table

3.3).

Table 3.2: Accuracy of estimates of annual sap flow (% difference from the measured value), derived by extrapolating from measurements taken during selected months. The total range of accuracies of five trees (range) and the mean ± standard deviation of accuracies of five trees (mean) are provided. Negative values indicate under-estimates, positive over-estimates.

Yule River Marillana Creek

Selected months Range Mean Range Mean

April & October -8.4–9.8 2.3 ± 7.4 5.8–15.4 10.0 ± 3.5

January & July -5.3–5.2 0.2 ± 4.6 -1.3–16.1 9.2 ± 6.6

January, April, July & October -1.5–4.1 1.2 ± 2.4 3.6–15.8 9.6 ± 4.3

Table 3.3: Accuracy of estimating monthly and annual sap flow (% difference from the measured value) from meteorological variables, through linear multiple regression models. Negative values indicate underestimates, positive values overestimates. YR; Yule River, MC; Marillana Creek. See text for details of the models used.

J F M A M J J A S O N D Annual

YR 1.6 -4.4 -2.4 -1.5 -5.1 -4.8 -2.0 -0.5 -1.4 -2.3 -0.2 -0.8 -2.42 MC -1.1 15 5.8 9.3 2.5 -1.8 -8.6 -5.1 14 -16 -9.2 -8.4 -2.00

Variation between trees and sites

Sap flow velocity varied considerably between trees, both within and between sites, as

measured at a single point near the centre of the sapwood band in each tree. Tree size

appeared to partially, but not completely, explain the difference among trees. Daily sap

flow varied approximately 2-fold among trees during cooler months, and 3-fold during

hotter months (Figure 3.3a). Daily flow was significantly correlated with tree size (As) on

most days during the hottest months (e.g. 10th Jan 2010; p=0.040, R2=0.427), but not

during cooler months (e.g. 24th July 2009; p=0.291, R2=0.138). On all days the

correlation was positive, i.e. flow was greater in larger trees. Cumulative sap flow over the

Chapter Three

54

course of a year did not correlate significantly with tree size (p=0.09), although there may

be a trend toward greater annual flow in larger trees (Figure 3.3b). The annual sap flow at

the Marillana Creek site (during July 2009–June 2010) varied nearly 3-fold among trees,

with values ranging from 454 to 1304 m year-1 (Figure 3.3b). Annual flow at the Yule River

(during April 2009–March 2010) varied 1.5-fold among trees, ranging from 1197 to 1772

m year-1. Since all these inter-tree comparisons were measured at a single point within each

tree, and within-sapwood variation can also be considerable (as described above), clearly

part of the among-tree variation in Vs will be due to the within-sapwood component of the

variability. However, the correlation between daily sap flow and tree size during warmer

months indicates a likely additional among-tree component of variation, related to the size

of the tree.

Tree Vs was correlated more strongly with As than with stem diameter. Nonetheless,

stem diameter may still provide an acceptable estimate of As, since the two were strongly

correlated in the studied M. argentea trees (Figure 3.4a). However, the relationship between

diameter and As differed between the study sites, with As increasing more steeply with

diameter in the Yule River trees than at Marillana Creek (p=0.002). The width of the

sapwood band (ws) tended to be greater at the Yule River than Marillana Creek (p=0.02),

Figure 3.3: Correlation between sap flow velocity (Vs) and tree size (cross sectional

sapwood area; As). (a) Daily sap flow of trees at the Yule River (closed symbols) and Marillana Creek (open symbols), on a representative day in winter (July 2009; blue) and in summer (January 2010; red). (b) Annual sap flow measured over April 2009−March 2010 at the Yule River, and August 2009−July 2010 at Marillana Creek. Regression lines are shown for significant relationships (p<0.05).

0

1

2

3

4

5

6

0

400

800

1200

1600

2000

Marillana CkYule River

0 100 200 300 0 100 200 300

R2 = 0.41

V s (m

day

-1)

V s (m

yea

r-1)

As (cm2)

(a) (b)

(sites combined)


55

explaining the greater As of the Yule River trees (Figure 3.4b). However ws did not correlate

with tree diameter (p=0.63), while bark depth did increase significantly with diameter

(p=0.01).

Whole tree hydraulic conductance (K) varied widely among individuals, from 6 to 31

cm h-1 MPa-1, but did not relate clearly to tree size (Figure 3.5a), and so appears not explain

the higher Vs in larger trees. Likewise, leaf water potential (Ψl), which could drive higher Vs,

appeared not to vary with tree size at either site (Figure 3.5b). However, Vs displayed a

plateau of high flow during the middle of the day (Chapter 2, Figure 2.3), and the length

of the plateau phase was correlated with tree size (Figure 3.5c). Thus, larger trees sustained

low Ψl and high flow rates for longer periods. In addition, nocturnal flows were greater in

larger trees (Figure 3.5d), although the contribution of nocturnal flow was much smaller

than that of the daytime plateau phases.

Inter-tree variation in Vs was considerable within each site, but Vs also differed

between the two study sites, with Yule River trees generally having higher flow rates than

trees at Marillana Creek, even after accounting for meteorological and tree size differences

between the sites. While Vs was largely predictable from meteorological conditions in M.

argentea, the regression models that were created for Vs against meteorological parameters

Figure 3.4: Relationships of tree stem dimensions at 1.2 m height, for trees at the Yule

River (closed symbols) and at Marillana Creek (open symbols). (a) Correlation between stem diameter and sapwood cross sectional area (As), (b) correlations of stem diameter with bark width (triangles), and sapwood width (circles). Regression lines are shown for the significant relationships (p<0.05) of diameter with As at each site, and with bark depth for sites pooled.

0

100

200

300

A s (c

m2)

0

1

2

3

4

5

6

Width (cm

)

0 20 40 010 30 50 20 4010 30 50 60

bark sapwood

y = 7.51x - 90.81R2 = 0.96

y = 4.20x - 40.35R2 = 0.88

(a) (b)

Diameter (cm)

R² = 0.56(sites combined)

Chapter Three

56

differed between the two study sites (Chapter 2). When the meteorological models

generated for each site were used to estimate sap flow at the other site, the predictions were

much less accurate (Table 3.4). The model produced from the Yule River data generated a

predicted annual sap flow that deviated only 7.6% from the measured value when it was

applied the same data set. However, when the Yule River model was applied to the

Marillana Creek meteorological data, the predicted annual sap flow for the Marillana Creek

trees deviated 36% from the measured value. Similarly, the Marillana Creek model

produced an annual sap flow estimate that deviated by 13% when re-applied to the

Marillana Creek data, but deviated 27% when applied to the Yule River data. Many of the

individual month estimates were even less accurate, deviating from the measured values by

up to 128%.

Figure 3.5: Relationships between water use physiology and tree size (sapwood area; As).

(a) Whole tree hydraulic conductance (Ktree), (b) midday leaf water potential (Ψleaf ), (c) length of the daytime plateau phase of maximum sap flux velocity (Vs-max), (d) total nocturnal sap flow. Data are from Yule River trees (closed symbols) and Marillana Creek trees (open symbols), in November (red) and June/July (blue symbols). Regression lines are shown for significant relationships (p<0.05).

-2.5

-2.0

-1.5

-1.0

-0.5

0

0

20

40

60

80

100

120

6

7

8

9

10

11

12

0

10

20

30

40

0 100 200 300

K tre

e (cm

h-1

MPa

-1)

Midday Ψ

leaf (MPa)

0 100 200 300

V s-m

ax p

late

au le

ngth

(h) N

octurnal Vs (cm

night-1)

As (cm2)

(a) (b)

(c) (d)

R2=0.50

R2=0.93

R2=0.76

R2=0.62R2=0.88

R2=0.74

R2=0.32


57

Table 3.4: Accuracy of estimates of monthly and annual sap flow (% difference from the measured value), derived from regression models of meteorological variables that were produced for trees at a different study site. Negative values indicate underestimates, positive values overestimates. See text for details of the models used.

J F M A M J J A S O N D Annual

Yule River (YR) model

YR* -0.8 16 13 27 32 40 9.7 7.9 36 -11 -13 -12 7.60

MC+ -20 -28 -27 -20 -25 -16 -19 -24 -29 -28 -23 -23 -36.5

Marillana Creek (MC) model MC* -19 -18 -15 0.0 -7.3 -10 -7.0 -11 -14 -15 -17 -18 -12.8 YR+ 17 40 34 54 78 83 128 19 81 6.0 12 5.0 27.0

*Predictions generated for the same data the model was derived from, as controls. +Predictions generated for data from the other study site, as a test of potential applicability of the model to a new site

DISCUSSION

Rates of sap flow in M. argentea appear to be highly spatially variable at multiple scales,

from within-tree variation throughout the sapwood, to variation between trees and sites.

Sap flow also varied over time, but the temporal variation was more predictable, and

estimates derived from meteorological data or extrapolation from shorter time periods are

likely to be adequate, especially for estimates of total sap flux over longer periods (e.g.

annual fluxes). Further examination of spatial variation in water use in this species would be

required in order to fully characterise and understand the basis for spatial patterns, and to

obtain sufficiently accurate estimates of stand and ecosystem water fluxes.

Flow rate varied across the radial sapwood profile in M. argentea, showing a linear

decrease from the cambium to the heartwood in one tree, while in another tree Vs peaked at

a depth of approximately 7 mm, then appeared to decrease linearly to the heartwood. Both

these types of radial pattern are commonly observed in other tree species (e.g. Cohen et al.,

2008; Gebauer et al., 2008). The formation of tyloses increasingly blocks vessels with age,

reducing the conductance toward the heartwood. The temporary increase in conductance

with age that occurs in the peaked radial pattern may be related to morphological changes

Chapter Three

58

of the pit membranes (Gartner & Meinzer, 2005). Given that the slope of the radial flow

profile differed between the two M. argentea trees in our study, further data would be

required to establish whether there may be common patterns across trees in the radial

profile that would allow total flux estimates from fewer sensor points. For example, the

radial decline in Vs can vary with tree size in some species (Delzon et al., 2004). However,

the width of the sapwood band did not appear to alter with tree size in M. argentea, and

was only 15–30 mm, which is quite narrow compared with most other species. Some

Eucalyptus species also have sapwood widths in the range 15–40 mm, which also remain

constant with age (e.g. Miranda et al., 2009; Pfautsch et al., 2010). In contrast, sapwood

widths of species including Populus fremontii, Populus canescens, Fagus sylvaticus, Acer

pseudoplatanus and Pinus taeda usually increase with tree size and are typically 70–200 mm

(Nadezhdina et al., 2002; Ford et al., 2004; Gebauer et al., 2008; Hultine et al., 2010).

Although the patterns of radial flow require further analysis, the relatively uniform sapwood

width among trees should simplify future sampling strategies and flux calculations.

As well as the usual decrease across the radial profile, I also observed an

approximately two-fold variation in flow rates around the circumference of each measured

tree. While some species show negligible azimuthal variation in sap flow, such as Pinus

taeda (Domec et al., 2010), two-fold azimuthal variation may be broadly typical of a wide

range of other species, including Quercus glauca (Tateishi et al., 2008) and Chamaecyparis

obtusa (Tsuruta et al., 2010). This azimuthal variation might be related to the positions of

large roots and branches, and coupling of certain stem sapwood portions with sections of

the root system and canopy (Čermák et al., 2008). In addition, M. argentea stems are

frequently asymmetrical and non-vertical as a result of the high levels of physical

disturbance in the riparian zone, which can lead to variation in vessel anatomy and K

around the stem (e.g. Vertessy et al., 1997). Azimuthal variation has been previously

highlighted as a factor that can cause large errors in estimates of tree water use, and remains

an ongoing methodological challenge (e.g. Saveyn et al., 2008; Tsuruta et al., 2010; Cohen

et al., 2012). It is not clear whether there may be simple ways of predicting azimuthal

variation, and thus for tree scale water use measurements might need to be assessed

individually on every tree. When quantifying stand scale water use, varying the azimuthal


59

positioning of sensors among measured trees might help to overcome any systematic over-

or under-estimation due to aspect (Shinohara et al., 2013).

In M. argentea, the contribution of each radial depth remained fixed over diurnal

cycles, and between days over the measured period of two weeks. However, the

contribution of azimuthal portions varied over the course of the day, most likely related to

changing irradiance levels on different parts of the canopy, although the azimuthal

contributions to total daily water use remained constant over the two week measurement

period. In some species, patterns of sap flow within the sapwood can vary diurnally and

seasonally (Ford et al., 2004; Fiora & Cescatti, 2006). For instance, in Abies alba, Picea

abies and Fagus sylvatica the inner sapwood increases its relative contribution during high

demand periods, both diurnally and seasonally (Fiora & Cescatti, 2006; Saveyn et al.,

2008). However, other studies indicate that in most species the radial sap flow pattern is

likely to remain fairly constant over time, showing little variation in response to evaporative

demand or soil water availability, except under severe stress (Caylor & Dragoni, 2009;

Dragoni et al., 2009). The present study of M. argentea only observed within-sapwood

patterns over a few weeks, but contributions of each measured sapwood portion were

constant during this time. Nonetheless it is possible that azimuthal patterns could vary

seasonally, with sun positions changing patterns of irradiance and shading across the

canopy. It is also not known whether within-sapwood patterns might alter with severe

stresses such as major hydrological disturbance. However, it appears that variation in

sapwood contributions over time is unlikely to be a significant source of error in

quantifying whole tree water use in this species, at least in its normal habitat.

A wide variation in Vs was observed among the individual M. argentea trees in this

study. This was also noted by O’Grady et al. (2006a), who found inter-tree variation to be

much greater in M. argentea than in the co-occurring Corymbia bella, along the Daly River

in the wet tropics of Northern Australia. Nonetheless, similar variation in Vs (of two to

three fold) among trees has been observed in other species, such as Eucalyptus regnans in

southeastern Australia (Vertessy et al., 1997). Part of the observed variation in M. argentea

will be due to the within-sapwood variation, since most trees were sampled only at a single

point within the sapwood. However differences in Vs among M. argentea trees also appears

Chapter Three

60

to be related to tree size, and variation in the allometric ratios between tree diameter, As and

canopy area might also contribute to the variability (e.g. Cohen et al., 2012).

Sap flow velocity tended to increase with tree size (trunk diameter and As) in M.

argentea, a finding that is unusual given the relatively small size of the studied trees (less

than 50 cm diameter, and As 0.03 m2). Flow velocity frequently varies with tree size, as

xylem K, and leaf area to sapwood area ratio (Al:As) change with tree height and age, due to

altered flow path lengths and tortuosity, and effects of gravity (McDowell et al., 2002).

Total tree water use shows sigmoidal relationships with tree diameter and sapwood area, in

a variety of angiosperm and gymnosperm species (Meinzer et al., 2005). Across an As of

approximately 0.1–0.4 m2, water use generally increases disproportionately with As, (i.e. Vs

increased with tree size), while above and below this size range, changes in water use slows

and saturates (i.e. Vs remains constant or decreases with size) (Meinzer et al., 2005).

Observations in other studies also appear to correspond with these ranges of As. Sap velocity

increased with tree size in Populus fremontii of As up to 0.21 m2, while in Salix goodingii

with As less than 0.06 m2, Vs decreased with tree size (Schaeffer et al., 2000). In small

Eucalyptus crebra and Callitris glaucophylla trees (As of 0.001–0.03 m2), Vs and Al:As did not

vary with diameter or height (Zeppel & Eamus, 2008). The M. argentea trees in the present

study also had As less than 0.03 m2, yet daily sap flow increased with As during hotter parts

of the year.

However, the increase in Vs with tree size did not appear to be related to differences

in tree K in M. argentea, but rather was due to larger trees maintaining maximal rates of

flow for longer periods during the middle of the day. Larger trees often have greater

internal water storage capacity (e.g. Phillips et al., 2003; Scholz et al., 2011), which may

explain the greater daily water use of larger trees in M. argentea. Greater internal storage

capacity is associated with higher total rates of water use, due to the maintenance of high

rates of transpiration for longer portions of the diurnal and annual cycles (e.g. Goldstein et

al., 1998; Meinzer et al., 2003), as observed in M. argentea. The higher rates of nocturnal

flow in larger M. argentea trees are also consistent with depletion and replenishment cycles

of greater magnitude in these trees. Seasonal changes in capacitance, or a lower requirement

to utilise water stores during low VPD periods (Chapter 2), could also explain the seasonal


61

difference in the relationship between daily Vs and tree size, with a clear relationship present

during summer, but much less pronounced in winter.

The inter-tree variation in M. argentea observed in the present study and in O’Grady

et al. (2006) will also partly reflect within sapwood variation, since most of the trees were

sampled at only a single position. Accounting for variation within the sapwood, through

more intensive sampling within each tree, might therefore resolve some of the observed

noise in the relationship between tree size and sap flow rates. If the relationship between Vs

and tree size can be further clarified through increased sampling, the relationship could be

used to scale estimates of tree water use. In addition, Vs depended upon tree size primarily

during hotter months, and accounting for tree size appears to be less important for

cumulative estimates over longer time periods (e.g. annual flux). A robust linear

relationship between As and tree diameter was observed in this study, indicating that simple

tree diameter measurements may be a useful way of adjusting stand scale water use

estimates.

The trees at the Yule River displayed significantly higher rates of sap flow than trees

at Marillana Creek, even after accounting for meteorological conditions and tree sizes.

Variation in water use by M. argentea was also high among the study sites of O’Grady et al.

(2006), where sites were within 50 km of one another. The observed among site variation

could simply be a part of the inter-tree and within sapwood variation discussed above, and

thus inter-site variation might be resolved by better accounting for tree and sapwood

variation through increased sampling. For example, in Eucalyptus regnans in southeastern

Australia, sap velocity varied two to three fold among individuals, but total stand water use

varied by less than 20% among sites (Vertessy et al., 1995; Vertessy et al., 1997). A sample

size larger than the four to five M. argentea trees used in the present study and by O’Grady

et al. (2006) is likely to be needed. Ten to fifteen trees were needed to accurately estimate

stand scale water use of a Chamaecyparis obtusa plantation in Japan (Kume et al., 2010) and

at least 20 trees were required in a Pinus strobus plantation (Ford et al., 2007). In natural

ecosystems containing multiple species and variable tree sizes and stand densities, the

required numbers of trees may be even higher. Nonetheless, the different relationships of Vs

with meteorological conditions at each study site suggest some difference in tree

functioning (Chapter 2). The different relationships of tree diameter with As at each site

Chapter Three

62

also suggests a site difference in tree allometry. Differences may relate to specific site

conditions, or to genetic differences between trees on different waterways. Different

responses to atmospheric conditions among locations has been observed in other species,

for example in Populus deltoides in New Mexico (Cleverly et al., 2006). If variation in tree

structure and physiology exists among stands of M. argentea, it may be necessary to sub-

sample each location in order to estimate tree water use, unless underlying explanations for

the variation can be uncovered.

Temporal variation in Vs was well characterised in the M. argentea trees (Chapter 2).

Our results show that extrapolating from shorter periods of measurements distributed

across the year could produce reasonable estimates for annual water fluxes in M. argentea in

undisturbed conditions. On average, extrapolating from two or four sampling periods, each

one month long, overestimated annual sap flow by a few per cent at the Yule River, and by

approximately 10% at Marillana Creek. Two sampling windows spaced six months apart

are likely to be sufficient, as four sampling windows did not improve the estimates at either

site. Using daily meteorological parameters to estimate annual sap flow via regression

models was even more accurate, leading to underestimates of only 2–3%. However, the

high spatial variation in flow rates means that the meteorological models were not readily

applicable to other locations, and so may not be a practical method of estimating tree water

use. In general, cumulative annual flows appear to be predictable enough in M. argentea to

allow reasonable estimates of tree water requirements to be derived from shorter windows

of Vs measurements, or from meteorological measurements. However, water use estimates

for shorter periods than one year may need to account for the additional seasonal

component of variation in Vs. Tree water requirements may vary seasonally in a manner

that is not entirely related to meteorological conditions, due to factors such as the potential

influence of internally stored water (Chapter 2).

While the temporal component of sap flow variation has been well quantified, the

present study has only touched upon the spatial aspects of sap flow variability in M.

argentea. The sample sizes were small and thus I have not thoroughly assessed the extent of

spatial variation, or the underlying explanations. Nonetheless, the findings demonstrate

that M. argentea is among the more heterogeneous species in sap flow rates, and more

spatially intensive sampling would be required to properly establish a basis for accurately


63

quantifying water use in this species. In particular, further investigation of azimuthal

variation, and of the relationship between tree size and flow rates, would contribute to a

more complete understanding of tree physiology, and improve our ability to estimate

hydrological fluxes at stand and landscape scales.

64

CHAPTER FOUR: Redistribution of water through root systems of a

semi-arid riparian tree during a flooding-drying cycle in an intermittent

waterway

INTRODUCTION

In dryland environments, soil moisture frequently varies throughout the soil profile. Along

intermittent, dryland waterways in particular, water may be present at depth, while surface

soil layers are dry. Plant root systems frequently transfer water from wet to dry soil regions,

a passive process driven by water potential (Ψ) gradients, known as hydraulic redistribution

(HR) (Burgess et al., 1998). HR appears to be an almost ubiquitous phenomenon,

occurring whenever root systems span gradients of soil water potential, during periods that

are not dominated by upward fluxes to above ground parts. In dryland areas, trees often

have strongly dimorphic root systems comprising both deep and shallow roots; these deep

roots can transport groundwater or soil water at depth into dry surface soils, a process

referred to as hydraulic lift (HL) (Ryel et al., 2008; Lubczynski, 2009; Bleby et al., 2010).

There is growing evidence that HR can have substantial hydrological and ecological

impacts in some ecosystems, particularly those that are seasonally dry (e.g. Lee et al., 2005;

Domec et al., 2010; Prieto et al., 2012).

Hydraulic lift can increase total ecosystem evapotranspiration and productivity via a

number of processes. First, nutrient release and microbial activity are usually concentrated

in the surface soil layers (~ top 20 cm or so) and can be strongly limited by moisture when

soils are dry (e.g. Grierson & Adams, 2000; McIntyre et al., 2009). HL can ‘irrigate’ surface

horizons in dry periods, thereby enhancing nutrient availability to the trees themselves as

well as to other plants (Aanderud & Richards, 2009; Armas et al., 2012). HL also increases

root longevity in the dry surface soils by maintaining root hydration (Bauerle et al., 2008),

which in turn allows the tree to rapidly take advantage of resource pulses, such as those

following rainfall events. Second, HL can increase the total root system water uptake

capacity, increasing transpiration by as much as 30–50% in some cases (e.g. Caldwell &

Richards, 1989; Scott et al., 2008; Domec et al., 2010). Third, HL in some cases supplies

water to understorey plants with shallower root systems that are unable to access deeper

Redistribution of water through root systems during a flooding and drying cycle

65

water directly (Hawkins et al., 2009; Pang et al., 2013). Consequently, HL by trees lining

ephemeral waterways may play an important role in resource capture by the trees

themselves, and in resource availability in the ecosystem as a whole.

Along the intermittent waterways that are common in dryland regions, the

distribution of water can fluctuate considerably over time. The Pilbara region of northwest

Australia is characterised by prolonged dry periods of up to several years, punctuated by

often extreme episodic flooding from cyclone activity (Chapter 1). Groundwater is

commonly present along waterways within ten metres of the ground surface, and is

recharged during flooding then usually progressively recedes during dry conditions

(Dogramaci et al., 2012). Given these conditions it might be expected that the root system

of the phreatophytic riparian tree, Melaeluca argentea, would redistribute deeper

groundwater to drier surface soil layers, and that this would maintain function of the

surface roots during dry periods.

Patterns of sap flow in different roots can be used to demonstrate redistribution of

water within the root system. HL manifests as acropetal flows (“reverse” flows away from

the stem base) in shallow lateral roots, combined with basipetal flows (water uptake) in the

descending roots with access to deeper water. In phreatophytes with constant access to

groundwater, HL sap flow patterns commonly occur whenever surface soils are dry (e.g.

Prieto et al., 2010). However, HL typically ceases if the surface soils are wet; for instance in

Quercus fusiformis, Bumelia lanuginosa and Prosopis glandulosa trees accessing deep water,

sap flow in surface lateral roots changed from acropetal to basipetal flows following rainfall,

and the rates of basipetal flows in deep roots concurrently decreased (Bleby et al., 2010). If

the deeper soils are not saturated, and surface soil moisture becomes greater than the

moisture present at depth, hydraulic descent can occur with basipetal flows through surface

roots and acropetal flows in deep roots (e.g. Hultine et al., 2004; Yu et al., 2013). Changes

in the soil moisture content of surface soils can clearly influence root water fluxes, even in

trees accessing deeper water supplies.

In this study, I used heat ratio method (HRM) probes to measure sap flows in lateral

roots, roots descending to the groundwater, and in trunks of M. argentea trees on an

intermittent creek, during 14 months following a major flooding event. I sought to

examine the dynamics of root spatial flows over the cycle of wetting and drying, and to

Chapter Four

66

assess the potential signficance of HL in the riparian ecosystems of the Pilbara. I

hypothesised that M. argentea would utilise floodwater immediately post-flooding via

shallow lateral roots. I also expected that as the surface soils subsequently dried out, water

from deeper in the soil profile would be redistributed into surface layers through these

shallow roots.

METHODS

Study site

The study took place at an undisturbed section of Marillana Creek (22.72° S, 118.96° E),

an intermittent waterway in the Pilbara region of northwest Australia (Chapter 2; Figure

2.1). Groundwater levels in the catchment fluctuate considerably between cyclonic recharge

events. Water level records from a bore ~1 km from the study site for the period 1991–

2004 show levels ranged from 5.4 to 17.1 m below the surface both within and among

years (Figure 4.1; data provided by Rio Tinto Pty Ltd). Following cyclonic rainfall and

flooding in 1996, 1999 and 2002, the groundwater level rose by 7–8 m. Unfortunately

there was no permanent monitoring bore adjacent to my site. However, cyclonic rainfall

resulted in substantial river flows and replenished groundwater in the region in February

2009, with surface pools still apparent at the site when the study commenced in July 2009

(Figure 4.2). There were no further major rainfall events during 2009 or 2010. Based on

personal observations at the site over the study period and general understanding of this

and similar creeks in the area (Rio Tinto, unpublished data), ground water levels at the

study site receded by between 4 and 7 m over the 14 month duration of the study.

Meteorological measurements

As described in Chapter 2, a weather station (ET107; Campbell Scientific Inc., Logan, UT,

USA) was installed at the top of the creek bank, approximately 300 m from the study trees.

Rainfall, temperature, relative humidity, light intensity, wind speed, and wind direction

were logged hourly.

Sap flow measurements

Three M. argentea trees along the bank of the main channel of the creek were selected for

sap flow measurements. The base of some large, downward angled (descending) roots were


67

already exposed on the channel side of the trees due to flood erosion (Figure 4.3) and were

excavated further where necessary. A small lateral root of each tree was also located by

excavating soil on the bank side of the trees. Sap flow was measured via the heat ratio

method with sensor units (HRM-30; ICT International, Armidale, NSW, Australia)

0

50

100

150

200

250

300

350

4000

2

4

6

8

10

12

14

16

18 1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

Rainfall (mm

)

Dep

th to

gro

undw

ater

leve

l (m

)

Figure 4.1: Groundwater level fluctuations and rainfall ~1 km from the study site from 1991

to 2004.

Figure 4.2: The main channel of Marillana Creek, beside the study trees. (a) The start of the study period (July 2009), five months after a flood, with surface pools still present. (b) The end of the study period (October 2010), twenty months after the flood event.

(a) (b)

Chapter Four

68

installed and operated as described in Chapter 2. Probes were inserted into a descending

root and a lateral root in each tree, within 30 cm of their joining with trunk. An additional

probe was also installed in the main trunk of each tree at 1.2 m height above the ground.

For each probe set, data were obtained from a single measurement point, positioned near

the centre of the sapwood band. The sizes of the instrumented roots and stems are shown

in Table 4.1. Sap flow data were logged from 19th July 2009–12th September 2010, apart

from periods when equipment was damaged by livestock (1st September–23rd October

2009, 20th June–23rd July 2010 and 26th July–1st September 2010).

Table 4.1: Sizes of the roots and stems fitted with sap flow sensors. Overbark diameter (d; cm) and sapwood cross-sectional area (As; cm2) at the sap flow measurement positions.

d As

Tree 1 stem 48.8 166.2 descending root 16.1 58.4 lateral root 5.3 8.55 Tree 2 stem 43.5 131.3 descending root 18.9 99.7 lateral root 3.9 5.11 Tree 3 stem 36.6 131.7 descending root 9.4 39.7 lateral root 3.1 3.30

Figure 4.3: Two of the study trees at Marillana Creek, showing the base of large decending

roots instrumented with sap flow probes. The stem and root bases on the creek channel side of the trees cwere already exposed due to previous flood scouring.


69

The calculation of heat pulse velocity (Vh) and its conversion to sap flow velocity (Vs)

were performed as described in Chapter 2 (Burgess et al., 2001; Bleby et al., 2004). To

correct for small misalignments of the probe needles, the zero flow baseline was determined

for each root at the end of the study period, by completely severing the roots from the tree,

while continuing to log sap flow for a further 12 hours. For the stem sap flow

measurements, the zero flow baselines were set to the minimum value of Vh recorded on

nights with low VPD and wind speed, after filtering of the datasets. Root acropetal flows

are displayed as negative values, and basipetal flows as positive values.

To determine sapwood width (ws) and cross sectional area (As), cross sectional discs

were cut from the roots through the probe insertion sites. The sapwood-heartwood

boundaries were determined by the distinct colour transition, and where necessary by

staining the wood with methyl orange (1% w/v in 4% v/v methanol), which stains the

sapwood deep red and the heartwood yellow. The root discs were imaged in a flatbed

scanner, and As measured from the images using Image J (Abramoff et al., 2004). Stem ws

and As were determined from wood cores, as described in Chapter 3.


Relationships between daily total sap flow and vapour pressure deficit (VPD) were assessed

via linear regression for each tree separately. The effects of recent rainfall on flow rates,

alone and in combination with VPD, were also assessed through linear models.

Relationships are reported as significant where p<0.05. All analyses were performed in R

version 2.13.0 (R Development Core Team, 2010).

RESULTS

Patterns of root sap flow with drying conditions

The surface lateral roots of trees 1 and 2 functioned as expected; they contributed to water

uptake at the beginning of the study while surface soils were wet, then as surface soils dried

with increasing time since the flood, displayed net acropetal (negative) flows on most days

(Figure 4.4a, b). The acropetal flows were often substantial, reaching a maximum of 1 m

per day in tree 1 and 3.5 m per day in tree 2, approaching some of the basipetal velocities

seen in the stems and descending roots. However, the due to the much smaller size of the

surface laterals compared with the descending roots (Table 4.1), any water effluxed from

Chapter Four

70

the lateral roots is likely to be only a small fraction of the total quantity of water taken up

by the tree. Net flow rates in the surface laterals then declined to zero in tree 1 by 12

months after the flood and in tree 2 by 15 months after the flood.

The lateral root of tree 3 behaved somewhat differently; this root showed mostly

basipetal flows, providing net contribution to water uptake (Figure 4.4c). The pattern and

rates of flow in this root were similar to the flows in the stem and descending root of the

same tree, during much of the study period. For instance, flows of up to 5 m day-1 were

recorded in December 2009. However, unlike stems and descending roots, the rates of flow

in the lateral of tree 3 declined over the study period and eventually reached zero flow by

May 2010, approximately 15 months after the flooding event (Figure 4.4c). While this root

arose at right angles to the base of the stem and appeared similar those instrumented on the

other trees, the flow pattern suggests this root likely descended close to the saturated zone.

The tree 3 lateral might therefore be better classified as an intermediate root, in between a

deep descending root and a surface lateral.

The flow patterns in the large descending roots corresponded closely with those in

the main stem of the same tree. Both diurnal and seasonal patterns of descending roots and

stems largely followed the patterns of vapour pressure deficit (Figure 4.4; see also Chapter

2). The high rates of basipetal flow in the descending roots throughout the study (up to 7

m day-1 in tree 1, 3 m day-1 in tree 2 and 6 m day-1 in tree 3 during the highest VPD

months) clearly demonstrate that these roots are the primary source for tree water uptake

and have constant access to an abundant water supply, most likely groundwater.

Flows in the laterals of all three trees had decreased to close to zero, with little or no

diurnal flow pattern evident, by 15 months after the flooding event (Figure 4.4). Zero flow

was recorded from the lateral of tree 1 from mid February 2010 onward. The absorbing

roots of the lateral of tree 1 may have died at that point, since flow in this root did not

respond to any rainfall events after mid-February (root responses to rainfall are described in

the next section below). However, the root wood appeared to be alive at the point of probe

insertion, based on the moisture content and pale, fresh appearance of the sapwood when it

was excised at the end of the study. Zero flow was recorded in the lateral roots of trees 2

and 3 from mid-May 2010, however both of these laterals responded to rainfall in


71

Figure 4.4: Daily sap flow velocity (Vs) in stems, descending roots and lateral roots

of three Melaleuca argentea trees. Measurements were conducted over July 2009−June 2010, as the site dried out following flooding in early 2009. Panel (d) shows daily rainfall (bars) and minimum and maximum vapour pressure deficit (VPD; lines) over the same period. Negative root Vs values indicate net acropetal flows (away from the stem base).

-10

2

4

6 descending rootlateral root

stem

-3

-2

0

2

-2

0

2

4

6

0

1

2

3

4

5

6

7

8

0

5

10

15

(a) Tree 1

(b) Tree 2

(c) Tree 3

AUG SEP OCT NOV DEC JAN FEB MAR APR MAY

VPD (kPa)

Rain

fall

(mm

)V s

(m d

ay-1)

V s (m

day

-1)V s

(m d

ay-1)

-1

1

3

5

-1

3

1

1

3

5

7

(d)

Chapter Four

72

September, four months later, suggesting maintenance of at least some of the absorptive

roots. The lateral of tree 3 was contributing to tree water uptake during most of the study,

until May 2010 (Figure 4.4c) when its initial water source may have no longer been

present, i.e. the water table may have receded beyond the range of this root.

The acropetal (negative) flows through the surface lateral roots of trees 1 and 2

showed clear diurnal patterns during the early and middle parts of the study period, but

were frequently negative throughout most or all of the diurnal cycle (Figure 4.5).

Maximum (or least negative) flows occurred between mid-morning and midday. The

diurnal cycling in these surface laterals shows that the acropetal flows were inhibited but

not entirely prevented by high transpiration rates in the canopy. However, on the scale of

total daily flows, the flows through the surface lateral roots appeared not to be affected by

VPD or transpiration (Figure 4.6). During the periods of active lateral root flows that were

not influenced by recent rainfall (omitting periods of zero flow and the twenty days

following rainfall events from the dataset), surface lateral root flows were in fact negatively

correlated with VPD and stem sap flow, with greater acropetal flow occurring on days with

higher VPD and rates of stem flow (Figure 4.6a). During the same time periods, sap flow

in the stems and in the roots involved in water uptake were positively correlated with VPD,

and with each other (Figure 4.6b–d). Similar results were obtained whether nocturnal,

daytime, or total daily flows were considered (not shown). The apparent negative

correlation of surface lateral root flow with VPD and stem flow is probably due to the fact

that both VPD and acropetal flows ‘co-incidentally’ decreased over the course of the study

period. The decline in lateral root flows may be related to the elapsed time since the last

large (cyclonic) rainfall event, and the associated decline in surface soil moisture and fine

root biomass over time. Thus the effects of soil moisture and root surface area may have

masked the effects of VPD on surface lateral root flow over the longer term.

Root responses to rainfall

While there were no major rainfall events during the study period, numerous rain showers

of <40 mm occurred (Figure 4.4d). Responses in lateral root Vs were apparent following the

three largest of these rainfall events (all >10 mm within a 7 day period; Figure 4.7). The

first of the larger rainfall events was 37 mm over nine days, in November 2009, the second

17 mm in a single day in February 2010, and the third 32 mm over two days, in September


73

2010. There were no responses in sap flow to any of the smaller showers that occurred

during the study.

Following the first (November) rainfall event, the lateral root of tree 1 switched from

net acropetal to basipetal flow, with positive daily net flows peaking at 1 m day-1, and

positive flows throughout most of the diurnal cycle for approximately ten days after the

event (Figure 4.4a, 4.7a). However, there was no detectable response in the lateral of tree 1

to either of the subsequent rainfall events; zero flows were recorded for this root for the

remainder of the study (Figure 4.7b, c). The lateral root of tree 2 responded to all three

events with basipetal flow during the days immediately following the rainfall of up to

4 cm h-1 and 3 m day-1, then flows gradually returned to net acropetal or zero (Figure 4.7a–

c). The lateral of tree 3 was already displaying net positive flows at the time of the first

rainfall event, and there was no detectable response to rainfall (Figure 4.7a). After the

Figure 4.5: Diurnal patterns of sap

flow velocity (Vs) in stems (black),

descending roots (blue) and lateral

roots (red) of Melaleuca argentea

trees. Data is shown from two differ-ent trees in panels (a) and (b) from the 3rd and 4th January 2010 as an example, with vapour pressure deficit (VPD) and solar irradiance for the same period shown in panel (c).

-10

0

10

20

30

40

-20

-10

0

10

0

2

4

6

0:00 12:00 0:00 12:00 0:00

200

400

600

800

1000

0

Irradiance (W m

-2)

VPD

(kPa

)V s

(cm

h-1)

V s (c

m h

-1)

(a)

(b)

(c)

Chapter Four

74

second (February) and third (September) rainfall events the lateral of tree 3 responded with

initial acropetal flow immediately after rainfall, followed by a period of net water uptake

(Figure 4.7b, c). The initial acropetal flows in the lateral of tree 3 were most rapid during

the middle of the day, when the strong negative water potentials generated in the canopy by

transpiration ought to prevent large acropetal flows. The explanation for these

measurements is unknown, but the values are assumed to be erroneous, possibly generated

by rainwater affecting the sensor unit.

In descending roots and stems, water uptake appeared to increase slightly following

the February (second) rainfall event, (Figure 4.4, 4.7). For instance, in the descending root

of tree 1 Vs initially increased from 1.4 to 3.6 m day-1, then over the next ten days declined

to approximately 2 m day-1. In the descending root of tree 2 Vs increased from 1 to 1.7 m

day-1 immediately after the rainfall, then declined to approximately 1.2 m day-1 over the

next ten days. The increased uptake occurred mainly through longer duration of the

-4

-3

-2

-1

0

1 (a) shallow lateral roots

Tree 1Tree 2

0

2

4

6

(c) descending roots

0

1

2

3

4

5

0 1 2 3 4 5 6VPD

(d) main stems

Tota

l dai

ly V

s (m

day

-1)

0

1

2

3

4

5(b) intermediate lateral root

5

3

1

7

0 1 2 3 4 5VPD

Total daily Vs (m

day-1)

Tree 3

Figure 4.6: Relationships between net daily sap flow velocity (Vs) and average daily vapour

pressure deficit (VPD). Flow rates in (a) Shallow lateral roots of trees 1 and 2, and (b) root of tree 3 intermediate between a shallow lateral and a deep descending root, (c) deep descend-ing roots, and (d) main stems. Data were excluded from periods of zero flow activity in the lateral roots, and from periods affected by rainfall pulses. Note differing y-axis scales.

R2 = 0.52

R2 = 0.61

R2 = 0.33

R2 = 0.38R2 = 0.05

R2 = 0.18

R2 = 0.60R2 = 0.08R2 = 0.68


75

Fig

ure

4.7

: C

ha

ng

es

in s

ap

flo

w v

elo

city

(V s

) in

re

spo

nse

to

sm

all

ra

infa

ll p

uls

es.

Ver

tical

col

umns

dis

play

resp

onse

s to

rain

fall

even

ts in

Nov

embe

r 200

9,

Febr

uary

201

0 an

d Se

ptem

ber 2

010,

in la

tera

l roo

ts (u

pper

pan

els)

, des

cend

ing

root

s (c

entr

e pa

nels

) and

ste

ms

(low

er p

anel

s). R

ainf

all i

s al

so d

ispl

ayed

in th

e up

per p

anel

s. D

ata

from

des

cend

ing

root

s in

Spe

tem

ber 2

010

is m

issi

ng d

ue to

equ

ipm

ent f

ailu

re.

-4-202468

tree

1tr

ee 2

tree

3

481216

0102030 0

1020

Vs (cm h-1)Rainfall (mm)

0

1312

1110

98

76

2021

2223

2425

2627

281

23

45

2524

2322

2120

1918

1716

1514

1312

1110

98

76

54

Dat

e (N

ovem

ber 2

009)

Dat

e (F

ebru

ary-

Mar

ch 2

010)

Dat

e (S

epte

mbe

r 201

0)

rain

fall

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

late

ral r

oots

late

ral r

oots

late

ral r

oots

desc

endi

ng ro

ots

desc

endi

ng ro

ots

stem

sst

ems

stem

s

Chapter Four

76

plateau phase of sap flow during the middle of the day, but also elevated rates of nocturnal

flow, particularly in tree 3 where nocturnal minima increased from 4.5 to 10.4 cm h-1 in

the descending root and from 7 to 15 cm h-1 in the stem (Figure 4.7e, g). There were no

discernible responses in stem or descending root flows to either of the other two rainfall

events (November 2009 and September 2010), apart from reduced flow rates while the rain

was actually falling.

DISCUSSION

Water uptake occurred through the lateral roots of the M. argentea trees at the beginning of

the study (July and August 2009) as hypothesised, when surface pools were present in the

creek bed around the base of the study trees. Also as expected, substantial acropetal flows

commenced in the surface lateral roots of two trees by November 2009, when surface pools

had dried down. However, over the following eight months the magnitude of flows in the

lateral roots diminished, as the site dried down further with increasing time since the last

cyclonic recharge event. The acropetal flows clearly indicate that surface fine roots were

maintained in dry surface soils, for a period of at least six months. However, the surface

roots were not maintained indefinitely, and the internal transfer of water to these roots

eventually ceased. During the period of acropetal flows and to some extent during the zero

flow period, the lateral roots showed dramatic responses to relatively minor rainfall (15–40

mm), evidenced by a transitioning to net water uptake for the ten to twenty days following

rainfall. Overall, this study clearly demonstrates the temporal variation in patterns of root

sap flow that can occur during the multi-year flooding and drying cycles, and suggest that

hydraulic lift could be an important process in the riparian ecosystems of the Pilbara.

The patterns of HL in M. argentea appear largely typical of those observed in other

phreatophytes. Redistribution of water from deeper reservoirs to surface roots is very

commonly observed when groundwater is accessible to deep roots while surface soils are

relatively dry. The rates of acropetal flow in the surface laterals of M. argentea were similar

to those seen in other studies, for instance, 1–3 cm hr-1 acropetal flows were reported in

Eucalyptus camaldulensis (Burgess et al., 1998), Fraxinus velutina and Juglans major (Hultine

et al., 2003b). HL generally ceases if groundwater becomes inaccessible (Brooksbank et al.,

2011) or if surface soils are re-wet by rainfall (Millikin Ishikawa & Bledsoe, 2000; Hultine


77

et al., 2004; Oliveira et al., 2005; Bleby et al., 2010), and HL also ceased in M. argentea for

short periods after rainfall pulses.

However, HL in M. argentea also declined and eventually ceased as the site became

drier over the course of the study period, despite constant access to groundwater by the

deeper roots and dry surface soils. As surface soil moisture declines from high to moderate

soil water content, HL usually increases in trees with access to groundwater, due to the

increasing water potential (Ψ) differential between the deep and shallow portions of the

root system (e.g. Meinzer et al., 2004a; Brooks et al., 2006). However, HL can be inhibited

if surface soils become very dry, due to declining soil conductivity, and increasing root

xylem cavitation as water potentials become more negative, leading eventually to fine root

mortality (Warren et al., 2007; Neumann & Cardon, 2012). In Retama sphaerocarpa shrubs

accessing groundwater, HL was greatest at intermediate Ψsoil, decreasing as Ψsoil fell below -4

MPa (Prieto et al., 2010). Similarly, in Acacia tortillis with access to groundwater during

the dry season in East Africa, HL ceased when surface soil Ψ fell below -5 MPa, and more

HL occurred during a wetter year than in a particularly dry year (Ludwig et al., 2003). In

the present study, saturation of surface soil layers by the cyclonic rainfall in early 2009 may

have triggered surface fine root proliferation and surface water uptake, then as the soil

moisture decreased, acropetal flows began to take place. The acropetal flows may have

diminished and eventually ceased, as the soil dried beyond a critical threshold.

Another somewhat unusual finding in my study was that acropetal flows in the lateral

roots were frequently observed throughout the diurnal cycle. This indicates that Ψ of the

surface roots were able to “compete” with the canopy for water, even under high rates of

upward stem flow. Acropetal root flows typically occur at night, facilitated by low

transpiration rates and leaf Ψ closer to zero, but can occur at any time the Ψ of the dry root

portions and/or surrounding soil becomes more negative than the canopy. A few studies

have reported HR concurrent with daytime transpiration, including HL in Blepharocalix

salicifolius in a Brazilian savanna (Scholz et al., 2002), and hydraulic descent in Juglans

major in the Chihuahuan desert, Arizona (Hultine et al., 2003a), but it seems to be a fairly

rare occurrence. Our data from M. argentea demonstrate that HL can continue throughout

the diurnal cycle for prolonged periods, even under extreme VPD (at times exceeding 7

kPa) and high rates of transpiration. Leaf Ψ in M. argentea usually do not fall below -2.5

Chapter Four

78

MPa despite high VPD (Chapter 2, 6), due to their constant access to groundwater and low

xylem cavitation resistance. These high Ψleaf, combined with high clay content in the

surface soils (which can require high pressures to release stored water), most likely enabled

the continuous and substantial hydraulic lift by the M. argentea trees in this study.

Even after long periods of zero flow in the surface lateral roots, immediate responses

to rainfall were still observed in two of the M. argentea trees. The immediate response

suggests that some of the fine roots were maintained, and therefore fine root cavitation and

mortality is unlikely to have been the sole cause of the cessation of lateral root sap flows.

Surface roots may have survived through internal transfer of water within the root system,

but low soil hydraulic conductivity or loss of contact between roots and soil may have

prevented measureable effluxes of water from the roots into the soil (Richards & Caldwell,

1987; Nobel & Cui, 1992). Alternatively, cell membrane water channels (aquaporins)

appear to play a role in modulating HR, although the extent of their role is unknown

(McElrone et al., 2007; Prieto et al., 2012). It is possible that efflux from the surface roots

of M. argentea toward the end of the study might have been prevented by reduced

quantities or gating of aquaporins, but this requires further investigation.

Hydraulic redistribution can sometimes benefit trees by increasing their total water

use capacity; transfer of water into dry soil portions can then enable the entire root system

to participate in uptake during high demand periods (Ryel et al., 2002; Domec et al.,

2010). However, in other cases HR contributes little or nothing to the total plant water use

(Hultine et al., 2003b). It cannot be discerned from sap flow data alone whether water was

effluxed to the soil, or only rehydrated the root tissues themselves. In the present study,

there was a total net loss of water from the two surface lateral roots over 8 months of

measurements, and net losses on most days, with daytime flows rarely positive in the

absence of rainfall inputs. If any water was effluxed to the surface soils in the present study,

it was not recaptured during high demand periods, and so HR appears not to play a role in

the water use strategy of M. argentea. Any water effluxed by the shallow roots of the M.

argentea trees may instead have evaporated, or been consumed by understorey plants.

The surface laterals of M. argentea in the present study switched from net acropetal

flows to net water uptake immediately following rainfall events, which is likely to facilitate

opportunistic nutrient capture. Even if water is not actually effluxed to the soil, internal


79

transfer of water can maintain surface roots (e.g. Bauerle et al., 2008), enabling a rapid

response to take advantage of any nutrient pulses. In dryland environments in general,

release of nutrients can be limited by water availability, and precipitation typically occurs in

short pulses which tend to trigger a pulse of nutrients at the surface (Ivans et al., 2003; Ryel

et al., 2008). Similar patterns of lateral root flow and responses to rainfall have been

observed in other semi-arid phreatophytes, including Prosopis velutina in Arizona (Hultine

et al., 2004) and Populus euphratica in northwest China (Yu et al., 2013). HL during dry

periods might therefore be important for nutrient capture in a wide range of semi-arid

riparian trees, including M. argentea. Nonetheless, the extent to which HL actually

improves plant nutrient status has not been well established, and warrants further study to

determine whether HL might play a major role in nutrient uptake (Hodge, 2010; Prieto et

al., 2012).

Hydraulic lift by M. argentea might prolong the periods for which surface roots and

surface soil moisture are maintained in the Pilbara’s intermittent waterways. However, HL

did not continue indefinitely in the present study, despite dry surface soils and access to

groundwater, as the soil at the study site appeared to become too dry for HL to continue.

The extent and duration of HL by M. argentea will most likely depend on the surface soil

type, which can vary considerably in the Pilbara’s riparian zones, ranging from coarse

gravels to finely textured clays (Johnson, 2004). Fine textured soils can typically reach more

negative Ψsoil, so are likely to support higher rates of HL initially, but are also more likely to

reach values of Ψsoil too low for HL to occur. Further analysis of fine root conductance and

mortality, and of soil water content and matric potential surrounding surface roots, will

clarify the extent to which HR by M. argentea contributes to whole ecosystem

evapotranspiration throughout the Pilbara.

80

CHAPTER FIVE: Root hydraulic conductance and aquaporin abundance

respond rapidly to partial root-zone drying events in a riparian Melaleuca

species*

INTRODUCTION

Partial root-zone drying (PRD) occurs when a part of the root system dries out, while water

remains available to other portions. Plant responses to PRD have been studied primarily in

agricultural settings, in attempts to reduce irrigation while maintaining crop yields and

quality (Davies et al., 2002; Kang & Zhang, 2004). However, heterogeneity of water

availability around root systems is a far more general phenomenon, occurring widely in

natural environments. River banks provide an extreme example of a heterogeneous

environment, although most if not all plants may experience some degree of heterogeneity

of water supply throughout their lifespan. Riparian trees can be subject to natural cycles of

drying and re-wetting of portions of their root systems, especially in arid and semi-arid

environments where stream flows are often intermittent or strongly seasonal (Dettinger &

Diaz, 2000). This heterogeneous water availability around root systems, both spatially and

temporally, can result in frequent PRD events.

Partial root-zone drying often results in partial stomatal closure, leading to increased

water use efficiency. In some cases, stomatal closure appears to be due to a chemical signal

from the drying roots, and not simply water deficit, since the stomata re-open when the

dried root portion is severed (Gowing et al., 1990; Liu et al., 2001). Increased abscisic acid

(ABA) content of the dried roots and shoot xylem sap and increased xylem sap pH have

been observed during stomatal responses to PRD, and are considered likely signalling

mechanisms (Davies et al., 2002; Dodd et al., 2008). However, the extent of the stomatal

response to PRD varies considerably. For example, the stomatal conductance (gs) of Citrus

aurantium subject to PRD was ~70% that of fully watered control plants (Zekri & Parsons,

1990), while in another study the gs of Acer pseudoplatanus saplings under PRD declined to

~30% of controls (Khalil & Grace, 1993). In other experiments using Lupinus cosentinii,

*Chapter published as: McLean EH, Ludwig M, Grierson PF (2011). Root hydraulic conductance and aquaporin abundance respond rapidly to partial root-zone drying events in a riparian Melaleuca species. New Phytologist 192: 664-675.

Partial rootzone drying in a riparian Melaleuca

81

Quercus robur, Betula pendula, Olea europaea, Phaseolus vulgaris and Persea americana, no

stomatal response to PRD was observed; instead gs and rates of water use consistent with

pre-PRD conditions were maintained (Gallardo et al., 1994; Fort et al., 1997; Fort et al.,

1998; Centritto et al., 2005; Wakrim et al., 2005; Neuhaus et al., 2007). While differences

in PRD response can be partly explained by the variable conditions used within each study,

some of the variation may also reflect species differences (e.g. Croker et al., 1998). I

hypothesise that differences among species in PRD reflect adaptations to environmental

heterogeneity or homogeneity. If this hypothesis is correct, I would expect gs and water use

to be unaffected by PRD in species commonly subjected to natural PRD events, including

some riparian trees, provided adequate water is available to other parts of the root system. If

a plant is able to maintain water uptake via the remaining hydrated zones, the best adaptive

strategy may be to continue water use, and therefore photosynthesis and growth, at a

constant rate.

The mechanisms by which plants maintain water uptake via only a portion of the

root system are unclear. Even when total water use is reduced in response to PRD, the

uptake rates of the wet roots can increase, thus compensating partly for reduced uptake

from the dried parts of the root system (Lawlor, 1973; Zekri & Parsons, 1990; Kang et al.,

2003; Mingo et al., 2004). The increased flow rates in the wet root portion during PRD

may result solely from changes in hydrostatic gradients, or the roots might also adjust their

hydraulic conductance (Lp) according to demand. Kang et al. (2003) found that Pyrus

communis tree roots may have increased Lp in response to PRD, although this result was

inferred from sap flux data and not measured directly. However, there are many examples

of roots rapidly adjusting Lp to match demand under other circumstances. Partial root

excision in wheat plants resulted in increased Lp of the remaining roots (Vysotskaya et al.,

2004). Lp is known to fluctuate diurnally, in synchrony with transpirational demand

(Henzler et al., 1999; Beaudette et al., 2007; Vandeleur et al., 2009). Shading can also

induce a reduction in fine root Lp (McElrone et al., 2007). If plants do increase Lp under

PRD, then the adjustment must be rapid, and localised to the watered root portion.

Root water uptake occurs via parallel apoplastic, symplastic and transcellular

pathways (Steudle, 1994). The relative conductance of these pathways depends heavily on

the extent of hydrophobic apoplastic barriers and the nature of the gradients involved (i.e.,

Chapter Five

82

hydrostatic or osmotic), but the apoplast is generally considered the path of least resistance

(Steudle & Peterson, 1998; Steudle, 2000). Rapid adjustment in apparent Lp can either

result directly from changes in the gradients driving water flow, by changing the

contribution of each flow pathway (Steudle, 2000), or through changes in the permeability

of the root tissues themselves. Although anatomical changes can affect apoplastic Lp over

longer timescales (Enstone et al., 2003), the only known means of rapid permeability

adjustment is via transcellular Lp, through changes in the quantity or activity of root cell

plasma membrane intrinsic protein (PIP) aquaporins. There is growing evidence that short-

term changes in Lp may be generated primarily through aquaporins (Maurel et al., 2010).

Transient increases in root Lp during early drought may be due to increased PIP

permeability (Hose et al., 2000), reduction of root Lp under anoxia can occur through

proton gating of PIPs by cytosolic acidification (Tournaire-Roux et al., 2003), and the

reduction and subsequent recovery of root Lp during chilling is associated with changes in

expression and phosphorylation states of PIPs (Aroca et al., 2005). If rapid root Lp

adjustments do occur in the wet root portion during PRD, aquaporins may be involved.

The riparian tree Melaleuca argentea is widespread across northern Australia, in

locations where permanent shallow ground water is accessible. Following flood events, M.

argentea commonly grows extensive aquatic root mats within temporary pools that form in

the creek beds. Sap flux studies have shown that these aquatic roots can provide the

majority of the water used by the tree, especially when adjacent bank soils are relatively dry

(Graham, 2001). However, as the pools dry out, the aquatic roots die back and water

uptake must occur via different parts of the root system. Given the natural frequency of

drying events, I hypothesised that M. argentea would not respond to PRD with stomatal

closure but would maintain water uptake by compensating with other parts of the root

system. I also hypothesised that such a compensatory effect would involve rapid increases in

root hydraulic conductance, which may be mediated by changes in aquaporin content or

permeability. I tested these hypotheses with seedlings of M. argentea in a split-root system,

mimicking the conditions frequently observed in the field with a portion of the roots in

soil, and a portion in an aquatic pool. This design allowed us to test the effects of draining

the aquatic pool on the shoot, and on the remaining hydrated part of the root system.


83

MATERIALS & METHODS

Plant material and growth conditions

Seedlings of Melaleuca argentea W.Fitzg. were collected from an intermittent creek in the

semi-arid Pilbara region of Northwestern Australia (22.9114° S, 119.2139° E). Seedlings

were planted into pots of river sand with slow release fertiliser (OsmocoteR Native Gardens,

Scotts, Baulkam Hills, Australia), and grown for several months before transplanting into

split-root pots for PRD experiments. Genetically identical plants were generated for

Experiment 3 (below) by taking stem segment cuttings from a single seedling. Cuttings

were grown for 9 months before transplanting into split-root pots. Seedlings of ~1 m height

and 1–2 cm basal stem diameter were used in all experiments. Experiments took place

between October 2008 and May 2010, in a glasshouse at the University of Western

Australia (31.9828° S, 115.7997° E).

Experimental design

The PRD responses of M. argentea were tested in three experiments. Experiment 1 assessed

field collected seedlings for stomatal responses, tissue water status and root growth

responses to PRD over a 14-day period. Water stress treatments were also conducted as

comparisons to verify the sensitivity of our measurements. Four treatments were included,

with five to six replicate plants in each; a PRD treatment, PRD with restricted water (PRD-

RW), complete root drying (CRD) and two fully watered control treatments. Experiment 2

examined survival and recovery of the dried root portion after 28 days of PRD-RW

treatment, to determine whether rapid root death may explain a lack of signalling from the

drying roots. The control treatment consisted of six replicates, and the PRD-RW eight

replicates, four of which were re-wet at 28 days for the recovery treatment. Finally,

Experiment 3 tested for rapid responses in root Lp and aquaporin expression. Genetically

identical plants derived from cuttings were used to reduce variability among replicates.

Plants were treated with PRD for either 24 or 48 h, or remained fully watered as controls.

Four replicate plants of each treatment were measured for root Lp, and a second set of four

replicates harvested for aquaporin analysis. Treatments were applied across a three-week

period to allow measurements and sample collection to be conducted on all plants between

11:00 and 14:00 hours. Conditions within the glasshouse for each experiment are shown in

Chapter Five

84

Table 5.1. Details of the treatments for all experiments are described at the end of this

section.

Table 5.1: Glasshouse conditions during the partial root-zone drying experiments. Values give the total range of daily observations during the experimental periods. See text for descriptions of experiments.

Experiment 1 Experiment 2 Experiment 3

Maximum temperature (°C) 28–42 27–36 24–29

Minimum temperature (°C) 12–21 18–24 8–13

Maximum relative humidity (%) 45–71 nd 57–67

Minimum relative humidity (%) 26–49 nd 28–44

Maximum PAR (μmol m-2 s-1) 690–1451 1000–1600 800–1332 nd; not determined

Split-root pots consisting of two 15 litre compartments were used for Experiments 1

and 2, for Experiment 3 two 7 litre compartments were used, as the plants were required to

be transportable. For all experiments, one compartment of each pot was fully sealed and

filled with a diluted modified Hoagland’s nutrient solution (Hoagland & Arnon, 1950) to

form the aquatic compartment, and the second compartment had drainage holes in the

base and was filled with river sand to form the soil compartment. The moisture release

properties of the river sand as determined in a pressure chamber are shown in Appendix B*.

Approximately one third of the root system was placed in the aquatic compartment, with

the remaining two thirds including the taproot in the soil compartment. Soil compartments

were watered with ¼ strength modified Hoagland’s solution. The dilution of the

Hoagland’s solution in the aquatic compartment was calculated from the field capacity of

the river sand (~200 ml l-1), so that the quantity of nutrients per volume was equal between

the two compartments. The aquatic compartments were replenished daily, and the solution

changed every 7–10 days. Soil surfaces were covered with a 1 cm layer of white plastic

beads, and aquatic compartments covered tightly with white plastic to minimise

evaporation and light penetration. Plants were allowed to establish in the split-root pots for

3–5 weeks before the start of treatment, by which time there was copious new root growth

in both compartments.

*Appendix B was included with the published version of this chapter as Online Supporting Information


85

Plants were allocated randomly across treatments, and treatments were begun at

midday, by draining the aquatic compartments. The soil compartments of PRD plants

were watered twice daily with nutrient solution. For the PRD-RW treatment, the soil

compartment was watered only once per day, which allowed the water content to fall below

80% (v/v) of field capacity. M. argentea is a riparian species that has highest rates of growth

and transpiration under semi-saturated conditions; 80% of field capacity of river sand

would thus restrict water for this species. For the CRD treatment, the aquatic compartment

was drained, and watering of the soil compartment ceased. Two groups of controls were

used, with soil compartments watered either once per day as controls for PRD-RW, or

twice per day as controls for PRD. The moisture content of the soil compartments during

Experiment 1 is shown in Appendix B. The aquatic compartments of control plants

remained filled and maintained as described above.

Stomatal conductance and tissue water status

Stomatal conductance (gs) was measured on the abaxial surface of young, mature, exposed

leaves with a leaf porometer (Decagon Devices, Pullman, WA, USA). At each sampling,

measurements were taken on three leaves per plant then averaged to give a single value for

each plant. Leaf water potential (Ψl) was measured with a Scholander-type pressure

chamber (PMS Instruments, Albany, OR, USA). A single mature leaf from the upper half

of the plant was cut and the balancing pressure determined immediately. Measurements of

gs and Ψl were conducted in all experiments to verify consistency of plant response to PRD.

Leaf and fine root (< 2 mm diameter) water content were determined from 0.2–1 g

samples collected mid-morning. Roots were washed clean of sand and surface dried

thoroughly on paper towel. Samples were weighed immediately upon harvesting, then

frozen in liquid nitrogen and freeze-dried. Samples were re-weighed and original moisture

contents calculated, before processing of the samples for other analyses.

Plant water use

Plant water uptake was measured as the change in pot weight over the time intervals

between scheduled watering. Details of the method used to determine uptake from each

compartment are provided in Appendix B. Water uptake is expressed relative to pre-

Chapter Five

86

treatment values, to normalise for variation in plant size. Water uptake measurements were

conducted in all experiments to verify consistency of plant response.

Root hydraulic conductance

Hydraulic conductance of roots (denoted L prior to normalisation against root surface area)

in both split-root pot compartments was measured with a High Pressure Flow Meter

(HPFM; Dynamax Inc., Houston, TX, USA), using the “transient” method (Tyree et al.,

1995). All measurements of L were conducted between 11:00 and 14:00. Portions of roots

of diameter up to 10 mm were excavated in each soil compartment, the selected root cut

under water, and the cut end attached to an HPFM coupling with a watertight seal.

Transient pressure ramps were applied immediately. Water flow into the root and applied

pressure were logged every 3 s while applied pressure was increased at a rate of 3–7 kPa s-1,

from 0 to 550 kPa. At least three successive ramps were applied to each root. L was

calculated as the slope of the linear region (between 200 and 550 kPa for most

measurements) of the relationship between flow and applied pressure. Successive ramps

applied to the same root gave similar values of L, and no consistent differences among

samples in the direction of change between successive ramps were observed. The

temperature of the laboratory during measurement periods remained between 15.5 and

16.5 °C.

The roots measured for L were separated from the rest of the root system and their

fine root portions (<2 mm diameter) dried at 60 °C. The values of root L were normalised

against fine root dry mass as a proxy for root surface area, to yield values of Lp. Correlation

between fine root dry mass and surface area was established by scanning samples of fine

roots and analysing with WinRhizo software (Regent Instruments Inc, Quebec City,

Canada), prior to drying (Appendix B).

Xylem sap osmolarity

Xylem sap was collected between 12:00 and 13:00 h on sampling days to assess the osmotic

gradients from soil solution to xylem sap. Branchlets approximately 20 cm in length were

cut from the seedlings and inserted into a Scholander-type pressure chamber. Bark was

removed from the end 5 mm of the stump, to avoid contamination with phloem contents.

The stem was then pressurised to no more than 0.4 MPa above the balancing pressure


87

using a single pressurisation, to avoid dilution of the sap with cellular water (Berger et al.,

1994). The first droplet of xylem fluid was discarded and the cut surface blotted clean to

remove cellular contents of wounded cells. A few drops of fluid were then allowed to flow

into a vial, which was sealed and frozen at -80 °C until analysis. Osmolarity of 20 μl

samples was determined by freezing point depression with a Fiske 210 micro-osmometer

(Advanced Instruments Inc, Norwood, MA, USA).

Root growth

Photographs of roots in the soil compartments of the split-root pots were taken through

windows cut into the pot walls, to assess growth rates of the hydrated root portion.

Windows were kept sealed between measurement timepoints to prevent roots from being

exposed to light. Each photograph was taken of the same region, at the mid-point of each

pot (50 cm depth), and cropped to a final image for analysis of 10–20 cm2. All fine roots

visible in the images were traced in Adobe Illustrator (Adobe Systems Inc, San Jose, CA,

USA), and the tracings analysed in WinRhizo (Regent Instruments Inc, Quebec City,

Canada) to obtain root length measurements.

Protein extraction and immunoblotting

Treatments were staggered so that all samples for protein analysis could be harvested on the

same day, between 12:00 and 13:00 h. Frozen fine root tissue was ground under liquid

nitrogen and mixed with protein extraction buffer (50 mM HEPES-KOH pH 7.1, 1 mM

EDTA, 1 mM dithiothreitol, 5 mM MgCl2, 1 mM PMSF, 1% polyvinylpolypyrrolidone

(w/v)) using 200 μl of buffer per 100 mg of tissue powder. Samples were further

homogenised in the buffer, then sodium dodecyl sulphate (SDS) added to 2% (w/v). The

samples were heated to 95 °C for 5 min and clarified by centrifugation at 12 000 x g. Total

protein concentration of the samples was determined using a bicinchoninic acid

quantification kit (Pierce, Rockford, IL, USA), with bovine serum albumin as a standard,

according to the manufacturer’s instructions.

Specificity of primary antibody binding was tested by immunoblot analysis

(Appendix B). Protein samples were separated on 12% polyacrylamide gels (SDS-PAGE)

with molecular mass standards (GE Healthcare, http://www.gelifesciences.com) and then

transferred to Hybond C nitrocellulose membrane (GE Healthcare,

Chapter Five

88

http://www.gelifesciences.com) by semi-dry electroblotting (Bio-Rad, http://www.bio-

rad.com) for 1.5 h at 0.54 mA cm-2. Membranes were blocked for 1 h in 5% (w/v) skim

milk powder in TBS-T (Tris-buffered saline pH 7.9 containing 0.1% (v/v) Tween 20),

then incubated overnight at 4 °C with either anti-PIP1 (AS09487; Agrisera, Vännäs,

Sweden) or anti-PIP2 (AS09491) at 1:1000 dilution in blocking buffer. Membranes were

then washed in TBS-T, incubated with anti-rabbit IgG conjugated to horseradish

peroxidase (GE Healthcare, http://www.gelifesciences.com) at 1:3000 dilution in TBS-T

for 1 h, and washed again in TBS-T with a final wash in TBS. The signal was detected with

chemiluminescence substrate (ECL; Pierce, Rockford, IL, USA) according to the

manufacturer’s instructions, and imaged with a LAS3000 LumiImager (Fujifilm Corp.,

Tokyo, Japan).

Immunoassays for quantification were performed as dot-blots (Heinicke et al., 1992).

Samples were diluted to a uniform total protein concentration, and 2 μl of each applied to

a grid on Hybond C nitrocellulose membrane (GE Healthcare,

http://www.gelifesciences.com). One sample was used to create a standard curve, against

which the relative signal intensity of other samples was measured. Antibody binding and

detection were performed as described for the immunoblots. The dot-blots were imaged

with dark and flat frame corrections (Gassmann et al., 2009), with several images taken

across a range of exposure times. Densitometry was performed with ImageJ software

(Abramoff et al., 2004). Three replicate dot-blots were quantified with each primary

antibody, and the relative signal intensities averaged for each sample.


Differences between control and treatment groups were analysed by repeated measures

ANOVA for time series data (gs, Ψl, tissue water content, water uptake and root growth),

and by single-factor ANOVA at selected time points. Differences among treatments in Lp

and PIP expression were analysed by single-factor ANOVA with Tukey’s HSD. Analysis

was performed in Microsoft Excel 2003 and SAS version 9.1 (SAS Institute Inc., Cary, NC,

USA). Differences are reported as significant where p<0.05.


89

RESULTS

Stomatal conductance and leaf water status were unaffected by PRD

The gs of the PRD plants did not differ significantly from the controls at any measurement

point during 14 days of treatment (Figure 5.1a, c). However, gs began to decrease after

about 24 h in the restricted water supply treatments, in which either less frequent watering

allowed soil moisture content to fall below 80% of field capacity each day (PRD-RW), or

watering of the soil ceased completely (CRD) (Figure 5.1b). After the initial decrease, gs of

the PRD-RW plants recovered to control levels at about day 7, while gs of the CRD plants

continued to decrease (Figure 1d).

200

400

600

800

0 12 24 36 48

200

400

600

800 Control PRD

CRD

0

0

PRD-RW

g s (m

mol

m-2

s-1)

Time after treatment (h) Time after treatment (days)

(a)

(c)

(b)

(d)

0 2 4 6 8 10 12 14

Figure 5.1: Stomatal conductance (gs) of Melaleuca argentea seedlings under partial root-

zone drying (PRD) and water stress treatments. (a) gs of control and PRD plants during the first 48 h of treatment. The PRD treatment was applied at midday (time zero, indicated by the arrow), by draining the aquatic portion of the root system. Black bars on the x-axis indicate dark periods. Values are means ± SE of five plants. (b) Midday gs of the same plants shown in (a), over 14 days of treatment. (c) gs of control plants, PRD plants with restricted water supply to the wet roots (PRD-RW) and completely droughted plants (CRD) during the first 48 h of treatment. (d) Midday gs of the same plants shown in (b), over 14 days of treatment. Values are means ± SE of six plants.

Chapter Five

90

Even with restricted water supply to the wet root portion (PRD-RW), the Ψl of PRD

plants did not differ from the controls at any measurement point, with midday Ψl

remaining between -0.7 and -1.5 MPa, and predawn Ψl between -0.05 and -0.5 MPa

(Figure 5.2a, b; only PRD-RW and control treatments shown). Predawn Ψl of the

droughted (CRD) plants was significantly lower than controls by day 7 (Figure 5.2a), at

which point soil water content in the CRD treatment had fallen to 38 ± 3% (v/v) of field

capacity. However, midday Ψl of the CRD plants was lower than controls from 24 h of

treatment Figure 5.2b). Leaf water content was also unaffected by PRD-RW, but decreased

slightly in the CRD plants and was lower than controls by day 14 (Figure 5.2c).

Survival of the dried root portion

Water content of fine roots in the hydrated zone remained the same as the control in PRD-

RW plants, but was lower than the control in CRD plants at day 14 (Figure 5.3a). The

drying aquatic roots rapidly lost water content relative to the control in both PRD-RW and

-2.0

-1.5

-1.0

-0.5

-1.5

-1.0

-0.5

0.0

ControlPRD-RWCRD

40

60

80

0 2 4 6 8 10 12 14 Time after treatment (days)

Pred

awn

ψle

af (M

Pa)

Mid

day

ψle

af (M

Pa)

LWC

(%)

(a)

(b)

(c)

0

Figure 5.2: Leaf water status of plants

under partial root-zone drying (PRD)

with restricted water supply, and

drought treatment. (a) Predawn leaf water potential (Ψl) of control, PRD with restricted water supply to the wet roots (PRD-RW) and completely droughted (CRD) plants. (b) Midday Ψl of control, PRD-RW and CRD plants. (c) Leaf water content (LWC) of control, PRD-RW and CRD plants. Values are means ± standard error, n=6 for Ψl, n=5 plants for leaf water content. Where error bars are not visible they are smaller than the symbols.


91

CRD treatments (Figure 5.3b). However, at least some of the dry roots were able to survive

for as long as four weeks because when re-wet after 28 days of PRD-RW, aquatic root water

uptake resumed within 24 h, and returned to control levels within days (Appendix B).

Water uptake by roots in the soil compartment concurrently began to decline, also

returning to control levels.

Water uptake and Lp of the hydrated roots increased during PRD

In untreated (control) plants eight weeks after transplanting, the dry mass of fine roots in

the aquatic compartment was 32%, and in the soil compartment 67% (SE 2%, n=4) of the

total fine root biomass. However, just prior to the start of treatments, water uptake via the

aquatic roots was 58% (SE 5%, n=17) of the total plant water use. Therefore, draining the

aquatic compartments to induce PRD dried one third of the root system, but amounted to

removing nearly two thirds of the plants’ water supply.

Total water use by PRD plants did not differ significantly from controls at any

measurement point during 14 days of treatment (Figure 5.4a). During the first 5 h after the

aquatic compartments were drained, there was a significant decline in the total water

uptake by PRD-RW plants relative to the control, but by 24 h total water uptake by PRD-

RW plants had increased and did not differ significantly from controls for the remainder of

60

70

80

90

0

20

40

60

80

100

50

0 2 4 6 8 10 12 14

(a)

(b)

Time after treatment (days)

Root

wat

er c

onte

nt (%

)

0

ControlPRD-RWCRD

Figure 5.3: Fine root water content in

Melaleuca argentea seedlings during

partial root-zone drying (PRD). Seed-lings had roots divided between an aquatic compartment and a soil com-partment. The PRD treatment was applied by draining the aquatic compart-ment, while the soil compartment remained well watered. (a) Moisture content of the soil roots of control, PRD and completely droughted (CRD) plants. (b) Moisture content of roots in the aquatic compartment of control, PRD and CRD plants. All values are means ± standard errors of four plants. Where error bars are not visible they are smaller than the symbols.

Chapter Five

92

the experiment (Figure 5.4b). Water uptake by the CRD plants also declined during the

first 5 hours of treatment, and continued to fall thereafter (Figure 5.4b).

Water uptake by roots in the hydrated soil compartment of the PRD plants did not

differ from those of control plants during the first 5 hours of treatment. However, by 24 h,

uptake by the hydrated root portion of the PRD plants had increased. By day 2, uptake was

on average more than three-fold greater than the control plants’ uptake from that

compartment (Figure 5.4c). Since the roots in the soil compartment of control plants were

on average responsible for 42% of the total plant water uptake (see above), the trebling of

their uptake rate under PRD enabled the total water uptake of the PRD plants to remain at

control levels.

0

20

40

60

80

100

120

0

20

40

60

80

100

120

ControlPRD

0

100

200

300

400

500

PRD-RWCRD

Time after treatment (days)0 2 4 6 8 10 12 14

ControlPRD

(a)

(b)

(c)

Tota

l wat

er u

se (%

of c

ontr

ol)

Soil

wat

er u

se (%

of c

ontr

ol)

Time after treatment (h)0 12 24 36 48

Figure 5.4: Water uptake during

partial root-zone drying (PRD). (a)

Change in the total water uptake by control and PRD plants, relative to pre-treatment water uptake. Values are means ± standard error of five plants. (b) Change in total plant water uptake relative to pre-treatment water uptake, by control plants, PRD plants with restricted water supply to the remaining wet part of the root system (PRD-RW), and completely droughted plants (CRD). Values are means ± stan-dard error of six plants. (c) Water uptake from the hydrated compart-ment by control and PRD plants, relative to pre-treatment uptake, during the first 48 h of treatment. The PRD treatment was applied at midday (at time zero), and water uptake mea-sured as the change in pot mass during 4–5 h time intervals. Values are plotted at the mid-points of the mea-surement intervals, and the start and end points of the intervals are indicated by vertical dashed lines. The black bars on the x-axis indicate dark periods. Values are means ± standard error of four plants.


93

Midday Lp did not differ significantly between roots growing in the aquatic and soil

compartments in plants that did not experience drying (control treatment; Figure 5.5).

After 24 h of partial drying, the Lp of the wet soil roots was on average more than three-fold

that of the control, the same magnitude of change as seen in the rates of water uptake by

those roots. However, after 48 h of partial drying, Lp of the wet soil roots had returned to

control levels. The Lp of the dried aquatic roots of PRD plants was not significantly affected

by the treatment (Figure 5.5). Xylem sap osmolarity was also unchanged by the PRD

treatment (Table 5.2), so the radial osmotic gradient across the roots (between the external

root medium and the xylem) is unlikely to have differed between treatments.

Table 5.2: Osmolarity (mOsmol kg-1) of xylem fluid collected at midday from stems of partial root-zone dried (PRD) and control plants. Values are means ± SE of five plants.

Time after treatment

Treatment 24 h 48 h 96 h PRD 4.8 ± 0.5 4.6 ± 0.4 3.4 ± 0.3 Control 4.6 ± 0.8 4.4 ± 0.5 5.0 ± 0.9

0

10

20

30

40

50

Control 24 h 48 h

Aquatic roots Soil roots

L p (m

l s-1

MPa

-1 g-

1 )

Treatment

a

a

a a

a

bFigure 5.5: Root hydraulic conduc-

tance (Lp) during partial root-zone

drying (PRD) treatment. Seedlings had roots divided between two pot com-partments, an aquatic and a soil com-partment. The PRD treatment was applied by draining the aquatic com-partment, while the soil compartment remained well watered. Values are means ± standard error of four plants. Different letters indicate a significant difference (p<0.05).

Chapter Five

94

Growth of fine roots in the hydrated root portion

The rate of root length increase in the wet soil compartments as measured at transparent

pot windows is shown in Figure 5.6. The difference between control and PRD root growth

was not significant; however there is an overall trend in the data. Root growth rate in the

wet compartments of PRD plants was slightly higher during the first two days of treatment,

then returned to the growth rate seen in control plants.

Abundance of PIP1 and 2 aquaporins in the hydrated roots was altered by PRD

The relative abundance of aquaporin proteins in fine root samples was determined by

immunoassays using antibodies generated against conserved regions of Arabidopsis thaliana

PIP1 or PIP2 sequences. Binding of the antibodies to M. argentea total protein samples was

highly specific (Appendix B). After 24 h of PRD, there was a small but significant increase

in the total PIP1 pool in the hydrated root portion, compared with the control treatment

(Figure 5.7a). By 48 h of partial drying, PIP1 content in the hydrated roots was more

variable, and was not significantly different from either the control or the 24 h level. Total

PIP2 content did not differ from the control at 24 h of PRD treatment, but at 48 h of

PRD PIP2 levels had declined relative to 24 h (Figure 5.7b).

DISCUSSION

Adaptive responses to heterogeneous water supply

M. argentea responded to PRD with a rapid increase in water uptake by the remaining

hydrated roots, which allowed leaf function and water status to remain unchanged. This

-2

0

2

4

0–1

1–2

2–4

4–7

7–14

Control

PRD

Time interval (days after treatment)

Root

gro

wth

(mm

cm

-2 da

y-1)

Figure 5.6: Fine root growth rate in

the watered root portion during

partial root-zone drying (PRD). Root growth rates measured in control and PRD plants, at windows in the pot walls. Values are means ± standard error of four plants.


95

manner of responding to PRD events most likely represents an adaptation to the highly

variable environments in which M. argentea is found. I used a well-watered, sandy substrate

in this study to minimise the influence of changes in soil conductance that occurs with soil

drying, particularly in clays (Nobel & Cui, 1992). Our experimental soil conditions are also

consistent with those in which M. argentea naturally occurs. However, the lower gs observed

when water supply was restricted to the wet root portion of M. argentea clearly shows how

the wet roots were unable to compensate below a certain threshold of soil water availability.

Many agricultural PRD experiments use deficit rates of irrigation (below the

evapotranspiration rate of untreated plants) and have in fact measured ‘drought’ response,

making it difficult to determine to what extent the plants are responding to hydraulic or

non-hydraulic signals. It is likely that more species than currently recognised would in fact

fully compensate for PRD events if soil conductance and water content around the wet root

portion were not limiting.

In cases where leaf responses to PRD are not observed, the plant is clearly displaying

root-level compensatory adjustments. Drying a portion of the roots of M. argentea seedlings

induced an increased Lp in the remaining hydrated roots within 24 h, which may have

mediated the rapid increase in the rate of water uptake by those roots. An increase in fine

root growth in the wet soil compartment may have provided longer-term compensation.

Observations of root growth in other species during PRD have yielded variable findings;

however, increased wet-zone root growth under PRD was observed in oak saplings and

peach trees, along with little or no gs response (Fort et al., 1997; Goldhammer et al., 2002;

Abrisqueta et al., 2008). Partial root-zone drying also weakly stimulated root growth in

(a) PIP1

0

0.5

1

1.5

2

Control

(b) PIP2

24 h 48 h Control 24 h 48 h

Treatment

Rela

tive

abun

danc

e

a

b ab

aba

b

Figure 5.7: Abundance of PIP aquapo-

rins in the watered root portion

during partial root-zone drying (PRD).

Relative abundance of (a) PIP1 and (b)

PIP2 aquaporin subfamily proteins in untreated (control) plants, and plants subjected to 24 and 48 h of PRD. Values are means ± standard error of four plants. Different letters indicate a significant difference (p<0.05) within each panel.

Chapter Five

96

Citrus aurantium, corresponding with a mild stomatal response (Zekri & Parsons, 1990;

Mingo et al., 2004). Increased root Lp and fine root growth in wet zones may be important

means by which species adapted to soil moisture heterogeneity compensate for partial

drying events.

Production of a signal for stomatal closure by the dried roots (most likely ABA, or

changes in xylem sap pH that then induce changes in ABA concentrations or

compartmental distribution within leaf cells) appears to be the primary PRD response in

some plants. Observations of stomatal re-opening after severing of dried roots demonstrate

that the wet roots were capable of compensating, but were prevented from doing so

through signalling of stomatal closure (Gowing et al., 1990; Croker et al., 1998; Liu et al.,

2001). An increase in ABA is frequently, although not always, detectable in the drying

roots, xylem sap and leaves during stomatal responses to PRD (e.g., Neales et al., 1989; Liu

et al., 2001). In contrast, in PRD studies where no stomatal response was detected, ABA

content did not change, suggesting that in some cases the signal is simply not produced by

the drying roots (Gallardo et al., 1994; Fort et al., 1997; Fort et al., 1998; Wakrim et al.,

2005). Consequently, while I did not measure ABA*, I argue it is likely that in M. argentea

seedlings under PRD, the dry roots did not generate any changes in ABA. Species in which

the absence of stomatal PRD responses has been observed still produce an ABA efflux from

roots under drought, so do not lack the ability to produce or respond to ABA (Gallardo et

al., 1994; Fort et al., 1997; Fort et al., 1998). Rather, it seems that in some species, the

extent of root ABA signalling is dependent upon the integrated soil moisture across the

whole root system, or on interactions with shoot responses to water stress. The response of

a species to soil moisture heterogeneity most likely has implications for the way in which it

responds to water stress.

The variations in PRD response that have been observed among plants may in part

be due to adaptations to soil moisture heterogeneity. The numerous crop species with

strong leaf-level PRD responses may have adapted to relatively homogeneous soil

conditions during their domestication. In homogeneous soil environments, localised soil

drying might indicate imminent drought, making root signalling to conserve soil water a

Measurements of ABA were completed subsequent to the publication of this chapter, which confirmed our expectations. The ABA results are presented in Appendix C.


97

useful adaptation. Stomatal conductance and growth of cultivated lettuce (Lactuca sativa)

declined when upper soil layers were dried, while a wild lettuce (L. serriola) native to

spatially heterogeneous soil habitats did not show any leaf level responses to the treatment

(Gallardo et al., 1996). It might also be expected that large plants would encounter greater

heterogeneity simply due to the scale of their root systems, and indeed many of the species

reported to respond weakly or not at all to PRD are trees. Croker et al. (1998) observed a

weak, negative correlation between PRD stomatal response and drought sensitivity in six

temperate tree species. Drier environments tend to be more heterogeneous in a range of soil

properties (Jackson & Caldwell, 1993; North & Nobel, 2000), so adaptation to soil

moisture heterogeneity might account for the weaker PRD stomatal responses in some

drought tolerant species. Contrary to our hypothesis, Salix dasyclados, the only other

riparian tree examined under PRD to date, showed a 30–40% reduction in gs during PRD

(Liu et al., 2001). However, S. dasyclados is native to regions where waterways are

considerably more consistent than in northern Australia, the region where M. argentea

occurs. More data are needed on the PRD responses of species with well-described habitats

before any firm conclusions can be made as to the adaptive nature of the responses;

however, our study with M. argentea provides a clear example.

Adjustments in root hydraulic conductance during PRD

Our results show that increased water uptake by the hydrated roots during PRD, rather

than simply being a passive process, may involve an active change in the Lp of the hydrated

roots. Adjustments of root Lp could occur in any species that compensates fully or partly for

partial drying events via the hydrated roots. In our study, the hydrated root portion was

required to triple its previous water uptake in order to fully compensate for the partial

drying. Notably the measured Lp of these roots was also approximately three times that of

the control, suggesting a causal relationship between the increase in water uptake and the

increase in Lp. However the Lp of the hydrated roots then returned to control values after

48 h, while the water uptake rate remained high. Increased fine root growth during the first

few days of PRD may have contributed enough to water uptake capacity to allow the

decline in Lp. Root growth in the soil compartment had certainly occurred by 48 h, and

although not significantly greater in the PRD plants than the controls, our measurements

were made via transparent windows which do not always accurately reflect total root

Chapter Five

98

growth (Glinski et al., 1993). Water use from the hydrated compartment increased within

hours of partial drying, but total water uptake was not identical to control levels until day

4, suggesting some gradual continuing adjustment, which would be consistent with the

growth of new roots.

Changes in aquaporin content were detected in the hydrated root portion, which

corresponded with the measured root Lp. PIP1 abundance increased after 24 h of PRD,

followed by a decline in PIP2s and possibly also PIP1s by 48 h. The magnitude of change

in PIP levels was small, but could still be sufficient to account for the observed changes in

Lp. Individual PIP isoforms vary in their location, conductance to water and response to a

given treatment, suggesting that each may have a slightly different biological function (Jang

et al., 2004; Hachez et al., 2006; Galmés et al., 2007). Since the PIP antibodies used in this

study will have detected many isoforms, the signal of change from one or a few isoforms

will be diluted within the total pool. The identity, location within the root, and the

proportion of proteins in un-gated states in our study are all unknown, and await further

examination. Nonetheless, the small change in total PIP protein levels observed in this

study could be physiologically significant, and the correspondence with Lp suggests that

changes in PIP levels could be involved in root responses to PRD.

PIP2s generally have high water permeability, while PIP1s often have little or no

measurable permeability (Chaumont et al., 2000; Kaldenhoff et al., 2008). However,

interaction between PIP1 and PIP2 isoforms in the formation of hetero-tetramers can

dramatically increase total permeability (Fetter et al., 2004; Mahdieh et al., 2008;

Vandeleur et al., 2009). A correlation between root Lp and PIP1 levels with no change in

PIP2 was observed in bean and grapevine. ABA application to the roots of bean (Phaseolus

vulgaris) plants increased both Lp and the PIP1 protein pool (Aroca et al., 2006). In

grapevine (Vitis vinifera), VvPIP1;1 is proposed to regulate water transport to match

demand diurnally and under water stress via its interaction with VvPIP2;1, which did not

change in transcript levels (Vandeleur et al., 2009). A similar mechanism may operate in M.

argentea during response to PRD, with PIP1 and 2 isoforms interacting to produce the

observed effects on Lp. The difference between the PIP1 and PIP2 subfamilies in the

direction and timing of their response hints at the complexity that may exist at the

molecular level in root responses to PRD.


99

In our study, stomata of M. argentea did not close during PRD and there was

increased water uptake from the wet roots. Lp and aquaporin adjustments in the wet roots

under PRD raises the question of how these responses are signalled. Although ABA can

upregulate Lp and PIP abundance (Mariaux et al., 1998; Hose et al., 2000; Aroca et al.,

2006), it is unlikely to be the signalling mechanism in this case since ABA concentrations

appear not to change in wet roots under PRD, whether or not stomatal responses occur

(e.g. Zhang et al., 1987; Fort et al., 1997; Fort et al., 1998). PIP abundance and gating can

be regulated by several other stimuli, including Ca+, pH and pressure pulses (Wan et al.,

2004; Luu & Maurel, 2005). Direct hydraulic ‘signalling’ may be the dominant mechanism

of controlling water use, particularly in woody plants (Brodribb, 2009). Consequently, the

hydraulic pressure effects of shoot demand may act directly to modulate aquaporin levels

and Lp in the hydrated roots of M. argentea.

The results presented here indicate a possible role for aquaporins in root water uptake

under a sudden increase in transpirational demand. The widely accepted composite

transport model (CTM) predicts that under the strong hydrostatic gradients produced by

transpiration, water flow should be largely apoplastic, with aquaporin-mediated flow

playing a lesser role (Steudle, 2000). Under the CTM, observations of rapid adjustment of

root Lp to match transpirational demand are explained by changes in the apoplastic

contribution as a direct result of gradient changes (Steudle & Frensch, 1996; Steudle &

Peterson, 1998). However, as Vandeleur et al. (2009) argue, gradient induced changes in

flow pathways cannot account for differences in Lp as measured under consistent gradients

and conditions. Transient pressure ramps by HPFM apply the same hydrostatic pressures

in each measurement, and the osmotic gradient appears consistent since xylem sap

osmolarity and applied nutrient solutions did not differ between PRD and control

treatments. HPFM measurements can themselves affect the osmotic gradients within the

root, since water is injected into the root xylem, potentially polarising solutes (Knipfer et

al., 2007). However, the consistent Lp observed between successive pressure ramps in our

experiment indicates that the degree of polarisation caused by measurements was likely to

be small, and similar among treatments. If the driving forces are the same, the remaining

possibility to explain a difference in Lp is a change in the permeability of the root tissues

themselves. The rapid but not immediate response in the rate of root water uptake in our

Chapter Five

100

study also argues against an entirely passive mode of compensation, but rather is consistent

with a process requiring several hours for its induction. Adjustments in aquaporin quantity

or permeability are the only currently known mechanisms by which rapid and reversible

changes in tissue permeability can occur (Maurel et al., 2010), and may explain our

observations; however, I cannot entirely rule out other explanations.

There are several other lines of evidence suggesting that aquaporins may at times play

a larger role in root water uptake than is usually supposed. In some plants, purely apoplastic

root radial flow appears not to be possible, requiring water to enter cell-to-cell pathways

(Knipfer & Fricke, 2010). Numerous studies using aquaporin inhibitors point to a

substantial contribution of aquaporins to total root Lp in a wide range of species and

conditions (Javot & Maurel, 2002). Transgenic approaches have also revealed significant

aquaporin contributions, with one Arabidopsis isoform even appearing to increase

hydrostatic root Lp, but not osmotic Lp (Martre et al., 2002; Siefritz et al., 2002; Javot et al.,

2003; Postaire et al., 2010). The levels of aquaporin transcripts in roots also peak at

midday, corresponding with the time of highest transpiration (Henzler et al., 1999;

Beaudette et al., 2007; Vandeleur et al., 2009). Whether aquaporins are in general

important during high rates of water uptake is not yet clear, but evidence is growing that

they play a significant part.

Conclusions

Plant responses to PRD events vary enormously as recorded in the literature; some of the

variation may be due to species adaptations to soil moisture heterogeneity or homogeneity.

Species appear to fall onto a continuum in PRD response, from those responding primarily

at leaf level, to those responding mainly via root-level compensation. Some degree of

compensatory increase in water uptake from the wet-zone during PRD occurs in many

species, and our results with M. argentea show that this may involve short-term increases in

root Lp, possibly via changes in aquaporin levels. Our findings form part a growing body of

evidence that aquaporin-mediated water flow may be significant when root water uptake

rates are high.

101

CHAPTER SIX: Morphological and physiological adjustments across a

hydrologic gradient by seedlings of a semi-arid riparian Melaleuca

INTRODUCTION

The availability of water and the depth to the underlying water table in and around dryland

streams can vary considerably, both spatially and temporally (Lite et al., 2005; Larned et al.,

2010). Recruitment of riparian trees in such environments generally occurs only after flood

events, since the flood waters disperse propagules and deposit sediments providing the most

optimal conditions for germination, establishment and survival (Scott et al., 1997; Pettit &

Froend, 2001; Friedman & Lee, 2002; Stromberg et al., 2010a). The few years after

germination is a critical phase, as the availability of water during this establishment period

can be the major determinant of survival to maturity (Horton & Clark, 2001; Dixon,

2003). In intermittent waterways, water availability typically declines post-recruitment due

to the recession of surface and ground water (Donovan & Ehleringer, 1991; Rood et al.,

2003). Riparian tree seedlings must therefore grow rapidly during the initial saturated

conditions, to quickly establish a large and deep enough root system to follow the water

table downward as the system dries (Mahoney & Rood, 1998).

The vulnerability of riparian tree recruitment is made most apparent where water

regimes have been altered by human activities. For instance, dams on the Kootenai river in

western North America have reduced the recruitment rate of Populus species, by both

altering channel hydraulics and bed sediment distribution, as well as increasing the rate of

water table recession during seedling establishment periods (Burke et al., 2009). Riparian

and wetland ecosystems appear particularly susceptible to abrupt changes in community

composition, occurring at particular thresholds of change to the water regime (Ridolfi et al.,

2006). Such threshold shifts were observed in the semi-arid riparian systems of Arizona,

USA; when the depth to groundwater and surface flow frequency fell below particular

values, recruitment of native Salix and Populus seedlings was reduced, and replaced by

exotic Tamarix ramosissima seedlings (Lite & Stromberg, 2005; Merritt & Poff, 2010).

These changes in recruitment and survival patterns are likely in turn to have further

ecological effects. Understanding the tolerance of trees, particularly at the seedling stage,

Chapter Six

102

across the range of water availability conditions is important in understanding a species

ability to establish in the riparian environment, and the likely responses to anthropogenic

changes to the water regime.

Plant responses to water availability can also occur in a threshold pattern with

decreasing water availability, related to thresholds of physiological stress (Hacke & Sauter,

1996; Nardini et al., 2001; Froend & Drake, 2006). Soil waterlogging and drought create

different types of stresses to the root system, and thus require different root responses to

tolerate the stress. However, leaf-level responses to drought and waterlogging can be similar

since the shoot can effectively experience water stress under both conditions due to a lack of

water in one case, and impaired root functioning in the other. Under waterlogged

conditions the major stress is usually lack of oxygen in the root zone, and associated

changes in soil chemistry (Kozlowski, 1997; Pezeshki, 2001). The major adaptive traits

associated with waterlogging tolerance are production of shallow and adventitious roots,

which have better access to oxygen at the surface of the flood water, as well as aerenchyma

formation, which allows oxygen transport within the root system (Justin & Armstrong,

1987; Kozlowski, 1997; Pimenta et al., 1998). In contrast, in dry soils, the production of a

deeper root system and allocation of greater biomass to root growth are frequently adaptive

(e.g. Leiva & Fernandez-Ales, 1998; Bell & Sultan, 1999). Suberisation of the root

endodermis is often observed in dry conditions, and can provide a barrier to water loss from

the root into the soil (e.g. North & Nobel, 1998). These adaptations to soil drought and

waterlogging would be expected to increase the survival and growth rates of tree seedlings

that display them in response to the conditions in which they are growing.

There appears to be a tradeoff between waterlogging and drought tolerance, at least

for angiosperm trees and shrubs, which may be due to the different adaptive responses

required (Niinemets & Valladares, 2006). For example, Muehlenbeckia florulenta, a shrub

common to semi-arid floodplains of Australia, is more tolerant of drying than flooding, and

seedlings correspondingly showed plastic adjustments in many traits during drought, while

under flooded conditions seedling growth ceased and morphological changes were not

observed (Capon et al., 2009). Most tree species are vulnerable to waterlogging, and some

gain a tolerant status only because they survive flooded periods in a dormant state (Parolin

& Wittmann, 2010). However, species that must not only persist but also grow rapidly

Seedling responses across a hydrologic gradient in a riparian Melaleuca

103

during flooded periods must possess particularly strong adaptations to waterlogging. An

ability to achieve high rates of growth and transpiration in wet conditions is often

associated with poorer drought tolerance. For example, seedlings of Salix viminalis and

Salix triandra had higher mortality rates during a three week drought, but higher rates of

growth and transpiration when well watered, compared with more drought tolerant Salix

alba and Populus nigra (Van Splunder et al., 1996). Riparian species can be subject to both

drought and flooding, particularly at the seedling stage. However, adaptive traits enabling

rapid growth in waterlogged or moist conditions following floods, necessary for

establishment, may preclude a strong adaptation to drought stresses.

Melaleuca argentea is a tree species common to riparian habitats across Northern

Australia, including the Pilbara region. M. argentea, like many other Melaleuca tree species,

occurs in the wettest parts of the riparian zone (Masini, 1988) and might therefore be

expected to be tolerant of flooding, but sensitive to drought. In Pilbara waterways

undergoing groundwater drawdown (Chapter 1), M. argentea is likely to be one of the

earliest and most severely affected species, and consequently may be a useful ‘indicator’

species for broader ecological monitoring of these areas and for setting thresholds for

ecological water requirements. However, the adaptive characteristics that enable M.

argentea seedlings to both establish and survive remain poorly described. In the previous

chapter (Chapter 5) I showed that the roots of M. argentea seedlings can display rapid

responses to changes in water availability around the root system to maintain high rates of

water uptake. To further elucidate the tolerance and responses of seedlings across the full

range of soil water conditions, I grew seedlings under a range of water availability

treatments, from complete waterlogging to severe drought, in controlled glasshouse

experiments. Growth rates were evaluated as a measure of seedling performance, and the

likelihood of successful establishment under those conditions. A range of leaf and root

physiological and morphological characteristics, considered to be adaptive responses to

waterlogging and/or drought, were also examined across the gradient of treatments. I then

compared the functional responses between waterlogging and drought conditions, and

assessed thresholds of water availability for various responses, and sequence in which they

occurred across the gradient. I hypothesised that (i) M. argentea seedlings would show

greater tolerance and morphological and physiological response during waterlogging than

Chapter Six

104

under drought stress, and (ii) that seedling physiological performance and growth rate

would decline in threshold-type patterns with decreasing water availability.

MATERIALS & METHODS

Plant material and growth conditions

Melaleuca argentea W. Fitzg. seedlings were collected from the streambed of Weeli Wolli

Creek in the semi-arid Pilbara region of Western Australia (22.91° S, 119.21° E), and

transferred to a glasshouse at the University of Western Australia (31.985° S, 115.820° E).

Seedlings were planted into 5 litre pots of river sand with slow release fertiliser mixed

through (3 ml l-1; OsmocoteR Native Gardens, Scotts, Baulkam Hills, Australia), and

allowed to recover from transplanting for at least six weeks under well watered, drained

conditions. Experiment 1 was conducted in April–May 2008, and Experiment 2 in March–

May 2010. Glasshouse conditions during the experimental periods are provided in Table

6.1.

Table 6.1: Glasshouse conditions during the experimental periods. Values give the total range of daily observations; see text for descriptions of experiments.

Experiment 1 Experiment 2

Maximum temperature (°C) 19 –33 28–42

Minimum temperature (°C) 8 –18 12–21

Maximum relative humidity (%) 64 –85 45–71

Minimum relative humidity (%) 24 –69 26–49

Maximum PAR (μmol m-2 s-1) 264–1206 552–1161

Experimental design

Experiment 1 (hydrologic gradient) characterised morphological and physiological

responses across a range of soil water content. Four treatments were established, 100%

saturation (fully waterlogged), 80% of sand saturation by volume (partially waterlogged),

50% of saturation (close to field capacity) and 20% of saturation (low water), with six

replicates (the moisture release properties of the river sand used in this study are provided

in Appendix B, Table B-1). Seedlings were 40–64 cm in height, and were divided across

the four treatments so that each treatment contained a similar range of seedling sizes. To

impose the treatments, the pots were enclosed in plastic bags, which were fastened around


105

the base of the stem to prevent water loss by evaporation, and each pot placed inside a

larger container. Plants in the 20%, 50% and 80% saturation treatments were watered to

weight every 48 h. For the 100% saturation treatment, the outer container was filled with

water, maintained within 2 cm of the sand surface. Plants were positioned randomly on the

glasshouse benches, and rotated weekly. Treatments were maintained for eight weeks, after

which final measurements were taken and plants were harvested for biomass measurements.

Experiment 2 subjected plants to progressive drought; watering of seedlings ceased,

and sand was allowed to dry out, while control plants remained watered to field capacity.

For the first seven days of treatment, the soil water content of the droughted plants was

adjusted by weight so that soil water content remained approximately equal among

individuals. After the first week, soil was allowed to dry out naturally. The droughted plants

were re-watered after eight weeks of treatment to ascertain survival.

Plant growth and biomass

Plant height was measured weekly by running a measuring tape along the longest stem

from base to apical meristem. Leaf elongation was determined from measurements of the

distance between tip and the junction of the petiole with the stem taken at 24 h intervals

with digital calipers of three immature leaves per plant on the apical stem. Leaf elongation

rates were then averaged to give a single value per plant for each time point. Stem basal

diameter was determined with digital calipers, as the average of two measurements taken

per stem at 90° angles. After eight weeks, all plants from Experiment 1 were harvested and

the leaves, stems and washed roots of each plant dried at 60 °C and weighed for total

biomass and biomass allocation ratios. Leaf area of each plant was measured with an optical

leaf area meter (LI-3100C; LI-COR, Lincoln, NE, USA) before drying. In Experiment 2

(progressive drought), cardboard sheets were taped around the base of each pot in a cone

shape to catch fallen leaves. Leaves that had fallen from the plant or were visibly dead

(entire leaf dull coloured, shrivelled and brittle) were removed every 1–2 days, dried at

60 °C and weighed to measure the rate of leaf shedding.

Morphology and anatomy of leaves and roots

Leaf morphological characteristics were measured in Experiment 1 at the end of the eight

weeks treatment, on young, fully expanded leaves. Terminal leaves were tagged at the start

Chapter Six

106

of treatment to ensure that leaves collected for analysis had developed during the treatment

period. Specific leaf area (SLA) was calculated from the area of fully hydrated leaves

measured in an optical leaf area meter (LI-COR, Lincoln, NE, USA), and their mass after

drying at 60 °C. Thickness of leaves and of individual tissue layers were measured by taking

transverse sections by hand across the centre of fresh leaves. Leaf sections were imaged

through a light microscope fitted for epifluorescence (Zeiss, Oberkochen, Germany), under

both brightfield and ultraviolet (UV) illumination (example images are shown in Appendix

D). Measurements were taken from the resulting images using ImageJ software (Abramoff

et al., 2004), at ~2 mm from the midvein on each side; the cuticle was distinguishable by its

blue fluorescence under UV light, and other tissues were clearly visible under both types of

illumination. The cuticle, epidermis and pallisade mesophyll layers were not consistently

thicker on the adaxial or abaxial side, so for each tissue type the values were averaged to give

a single value per leaf. Leaf density was calculated by dividing leaf specific dry mass (mass

per leaf surface area) by the leaf thickness, to yield values of dry mass per volume of fresh

tissue.

Stomatal and leaf trichome density were determined by photographing leaf surface

epifluorescence under UV light. Images were taken of both adaxial and abaxial surfaces near

the centre of each leaf, for 3 fields of view per leaf surface. The guard cells and trichomes

fluoresced blue against the red chlorophyll background (Appendix D), enabling the

numbers of stomata and trichomes in each image to be counted with ImageJ (Abramoff et

al., 2004). Stomatal and trichome densities were determined in the same manner for an

additional series of leaves collected along the main stems of four separate seedlings, to assess

effects of developmental age.

At the end of the treatment period, pots were cut open and the soil brushed away

where necessary, to examine the macro-morphological characteristics of the root systems.

Root endodermal suberisation and aerenchyma were assessed in roots of diameter 0.6–1.6

mm. Transverse sections were taken by hand at 4–6 cm from the root tip. Sections were

photographed under UV illumination through a light microscope fitted for epifluorescence

(Zeiss, Oberkochen, Germany) to view the endodermis, which fluoresced blue. As

verification, several sections were also stained with Sudan red III or Sudan black B

(lipophilic dyes); the stained portions of the endodermis corresponded to the portions


107

fluorescing blue under ultraviolet (not shown). The extent of suberisation was measured as

the proportion of the endodermal ring showing fluorescence; up to a value of 200% (two

complete suberised rings, the maximum observed). Example images are shown in Appendix

D. Aerenchyma was measured as the area occupied by air spaces as a proportion of the total

cortical area (Sojka, 1988), in bright field images of the same sections.

Plant physiological status

Shoot water potential (Ψshoot) was determined with a Scholander-type pressure chamber

(PMS instruments, Albany, OR, USA). In Experiment 1, Ψshoot was measured at midday at

the end of the eight week treatment period; the top of the main stem with 4–6 leaves

attached was cut from each seedling, and the balance pressure determined immediately. In

Experiment 2, Ψshoot was determined both predawn and midday at time points throughout

the eight week drought period; measurements were conducted on individual leaves (Ψleaf)

until the majority of leaves had been shed by the droughted plants, after which point Ψstem

was measured for terminal branchlets in both the drought and control treatments. Leaf

relative water content (RWC) was determined in Experiment 1 at the end of the treatment.

Young, fully expanded leaves were cut between 11:00 and 12:00, weighed immediately for

fresh weight, then immersed in water for 24 h, surface dried and weighed for re-hydrated

weight, and then dried at 60 °C for dry weight.

Gas exchange was measured with an open flow infra-red gas analyser (LI-6400; LI-

COR, Lincoln, NE, USA) with a broad leaf chamber. Chamber conditions were controlled

at 380 ppm CO2, 1500 umol m-2 s-1 PAR, 25 °C and 40–60% relative humidity, and leaf

gas exchange recorded once stabilised within the chamber (typically within 5 min). The leaf

portion contained within the chamber was then traced onto a sheet of paper, and the area

measured in an optical leaf area meter (LI-3100C; LI-COR, Lincoln, NE, USA).

Leaf and root samples were analysed for carbon isotope ratio, and total C and N

contents (%). Leaf 13C is used as an indicator of intrinsic water use efficiency integrated

over the life span of the leaf; RuBisCO preferentially utilises 12CO2, but as stomatal

conductance decreases, CO2 within the sub-stomatal cavities and leaf tissues becomes

limiting, forcing increased fixation of 13CO2 (Dawson et al., 2002). Young, fully expanded

leaves and fine root samples collected at the end of treatments were dried at 60 °C and

ground to a fine powder in a ball mill. Samples were analysed at the West Australian

Chapter Six

108

Biogeochemistry Centre, using an Automated Nitrogen-Carbon Analyser Mass

Spectrometer (Europa Scientific Ltd., Crewe, UK). 13C contents are reported as delta values

in parts per thousand (‰), relative to the Vienna PeeDee Belemnite international standard.

External error (standard deviation) of the analysis was <0.15‰. The C and N contents are

given as percentages by weight, with external error of analyses <1.5% for C and <0.5% for

N.


For each parameter measured, the treatments were compared by ANOVA with Tukey’s

HSD, performed with SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). For each

analysis, homogeneity of variance among groups was examined with Levene’s test, and

residuals were tested for departure from a normal distribution using a Shapiro-Wilk test.

Differences are reported as significant where p<0.05.

RESULTS

Seedling growth and biomass allocation

Shoot growth (height, biomass accumulation, leaf area, stem basal diameter, and stem and

leaf elongation rates) in the hydrologic gradient experiment (Experiment 1) was least in the

low water (20% of saturation) treatment (Figure 6.1). Shoot growth in the 20% treatment

also showed a different pattern over time, where height increase was relatively rapid until

week three and then slowed; by week six, shoot growth had virtually ceased. In contrast,

shoot height in the 50, 80 and 100% soil saturation treatments all increased at similar rates

of ~32 cm over 8 weeks, compared to only 13.4 cm in the 20% treatment (Figure 6.1a).

The greatest increase in total biomass and leaf area occurred in the 80% saturation

treatment; this treatment had 8.5 g greater dry mass and 0.04 m2 greater leaf area than the

50 or 100% treatments by week eight (Figure 6.1b, c). Leaf elongation rate after eight

weeks of treatment were similar between the 80 and 100% treatments (3.3 mm day-1) but

lower under the 50% treatment (2.2 mm day-1) and lower still under the 20% treatment

(0.3 mm day-1) (Figure 6.1d). The basal stem diameter showed a linear relationship with

soil water content, with the greatest diameter at 100% saturation (Figure 6.1e).


109

0

0.05

0.10

0.15

0.20

0.25

Leaf area ( m2)

0

2

4

6

8

10

Basal diameter increase (m

m)

0

10

20

30

Leaf

/ roo

t/ s

tem

dry

mas

s (g

)

0

20

40

60root

stemleaf

whole plant

40020%50%80%100%

300

200

100

00 10 20 30 40 50 60

Hei

ght i

ncre

ase

(mm

) (a)

Time (days)W

hole plant dry mass (g)

(b)

Stem elongation (m

m day-1)0

1

2

3

0

2

4

6

8

10

Leaf

elo

ngat

ion

(mm

day

-1)

20 40 60 80 100

stemleaf

(d)

0

(c)

(e)

20 40 60 80 1000

Figure 6.1: Growth of Melaleuca argentea seedlings over eight weeks across a range of soil

water content. (a) Cumulative increase in shoot height from the start of treatment (day zero). (b) Biomass after eight weeks treatment of whole seedlings, and of each component organ. (c) Total leaf area of seedlings after eight weeks. (d) Rates of leaf and stem elongation meas-ured at the end of eight weeks. (e) Increase in stem basal diameter from the start of treatment. Note dual scales in panels (b) and (d). Values are means ± standard error of six plants. Different letters indicate significant differences (p<0.05) between treatments within a measured parameter. Within each panel, upper case letters refer to the parameter shown in open symbols, lower case letters the parameter shown in closed symbols.

Soil water content (% of saturation) Soil water content (% of saturation)

ab bcc

a

A

BCB

C

a

bb

b

a

ab abb

AA

BCB

C

A

BB

B

a

ab

b b

A

B

C D

Chapter Six

110

The relative biomass allocation of the seedlings after eight weeks differed significantly

among treatments (Figure 6.2). The 80% treatment had the greatest total biomass of both

roots and shoots (Figure 6.1b). However, the 20% treatment had the highest proportional

allocation to roots, with these organs comprising 43% of total dry mass (Figure 6.2a). The

lowest root:total biomass ratio occurred at 100% saturation where roots made up only 28%

of plant dry mass. Specific leaf area increased with soil water content until the 80%

treatment, and did not differ between the 80 and 100% treatments, indicating greater

investment in the leaf tissue produced under lower water availability (Figure 6.2b). Leaf

nitrogen content was lowest in the 100% saturation treatment at 1.7% (w/w), compared

with 2.7% in the three lower water treatments (Figure 6.2c). In contrast, the nitrogen

content of roots was slightly higher in both the 100% and 20% treatments (1.3 and 1.4%)

than in the 50 and 80% treatments (1.1 and 1.0%).

110

120

130

140

150

160

SLA (cm

2 g-1)

0

1

2

3

Soil water content (% saturation)

Leaf

N c

onte

nt (%

)

0

1.0

1.5

Root N content (%

)

0

10

20

30

40

50

Root

dry

mas

s (%

of t

otal

) (a)

0.5

20 40 60 80 1000 20 40 60 80 1000

Figure 6.2: Resource allocation by seedlings after eight weeks treatment across a range of

soil water content. (a) Root biomass as a proportion of total plant biomass. (b) Specific leaf area of young mature leaves that had developed during the treatment. (c) Nitrogen content of leaves (% w/w). (d) Nitrogen content of roots (% w/w). Values are means ± standard error of six plants, values labelled with the same letter are not significantly different (p>0.05).

(b)

(c) (d)

A AA

B

A

B

A

AB

A

B

AB

B

A

B B

C


111

Seedling leaf shedding and survival

In the progressive drought treatment of Experiment 2, shoot growth not only ceased, but

seedlings also shed leaves. Seedlings first began to shed leaves after 13 days of soil drying,

when soil water content was ~12% of saturation. By 35 days, when soil water content was

less than 3% of saturation, the seedlings had lost all leaves (Figure 6.3). No leaf shedding

was observed in the control (field capacity) treatment during the experimental period, or in

any of the treatments in Experiment 1 (data not shown).

All plants survived the eight week treatment periods, including the progressive

drought treatment (Experiment 2). Although all droughted seedlings in Experiment 2 were

leafless at five weeks of treatment, they all produced new leaves when re-watered after eight

weeks of drought. However, re-sprouting took up to one month to occur; the first shoot

buds appeared 22 days after re-watering, and buds were apparent on all plants by 32 days

after re-watering (data not shown). The new shoots were produced solely from the thickest

parts of the stem near the base of the seedlings, suggesting that thinner branches were killed

by the drought treatment.

0

10

20

30

40

50

Time (days)

20

40

60

80

100

00 10 20 30 40 50

Figure 6.3: Leaf shedding by Melaleuca argentea seedlings during severe drought treat-

ment, as soil water content declined from field capacity (50% of saturation) to undetect-

able levels. Leaf mortality values are the cumulative mass of dead leaves as a percentage of the total plant leaf mass, mean ± standard error of six plants.

Soil

wat

er c

onte

nt (%

sat

urat

ion)

Cumulative leaf m

ortality (% of total)

Chapter Six

112

Morphological responses

The density, thickness and water content of leaves differed between treatments in

Experiment 1. In both the 80 and 100% soil saturation treatments, leaf water content was

80% (w/w) at the end of the eight weeks treatment, declining to 78% leaf water content in

the 50% soil saturation treatment, and 74% at 20% soil saturation (Figure 6.4a). Leaf

tissue was denser under the two lower soil water content treatments (0.23 g cm-3),

compared with the two higher soil water treatments (0.19 g cm-3). The thickest leaves

occurred in the 100% treatment at 0.40 mm, and the thinnest of 0.35 mm in the 50%

treatment, with the 20% and 80% treatments intermediate (Figure 6.4b). The variation in

leaf thickness appeared to be mainly due to differences in the spongy mesophyll thickness;

the palisade mesophyll thickness did not differ between treatments (Figure 6.4c). Leaf

epidermal thickness also did not differ significantly across the soil water gradient, while

there was a small but significant increase in cuticle thickness with decreasing soil water

content from 4.4 mm to 6.2 mm (Figure 6.4d).

Leaf trichome and stomatal densities also differed between treatments (Figure 6.4e).

Trichome density was lowest under the 20% soil water treatment, while stomatal density

peaked in the 50% treatment, with fewer stomata under both 20% and 100% saturation.

Leaf size (of the same leaves measured for trichome and stomatal density) tended toward

smaller in both the 20% and 100% treatments, but the difference was not statistically

significant (Figure 6.4a), suggesting that the differing trichome and stomatal densities were

not simply due to differences in the extent of leaf expansion between treatments.

A series of leaves from along the stems of a separate set of seedlings (growing at field

capacity (i.e. ~50% soil saturation)) was examined to determine changes in surface

morphology with leaf age (Figure 6.4f). Immature, actively expanding leaves had more

densely packed trichomes and stomata, and densities reduced with age. The leaves

designated as the ‘youngest mature leaves’ (leaf 1) in the age series were at a similar

developmental stage to those collected from Experiment 1, and indeed densities of stomata

and trichomes were similar to the 50% treatment of Experiment 1. However, older leaves

had slightly lower densities, suggesting that although the leaves sampled in Experiment 1

were not expanding in length, some lateral expansion may still have been occurring.

Nonetheless the ratio of stomata to trichomes remained similar throughout leaf


113

Figure 6.4: Leaf morphology and anatomy of seedlings under a range of soil water

contents. (a) Leaf water content (LWC) following hydration to full turgor, (b) leaf thickness and density, (c) thickness of leaf spongy and palisade mesophyll layers, (d) thickness of leaf epider-mis and cuticle. Leaf stomatal and trichome density of (e) youngest fully expanded leaves on seedlings across a range of soil water content, and (f) leaves across an age sequence on seed-lings at 50% soil water saturation (leaves are numbered according to their relative position along the stem, where leaf 1 is the youngest fully expanded leaf, similar to the leaves used for the measurements in the other panels). Values are means ± standard error of four to six plants. Different letters indicate significant differences (p<0.05) between treatments within a meas-ured parameter. Within each panel, upper case letters refer to the parameter shown in open symbols, lower case letters to the parameter shown in closed symbols.

74

76

78

80

82

0.15

0.20

0.25

0.30

0.40

0

2

4

6

8

10

12

50

100

150

200

Leaf

thic

knes

s (m

m)

Leaf density (g cm-3)

Laye

r thi

ckne

ss (μ

m)

Laye

r thi

ckne

ss (μ

m)

0 20 40 60 80 100


0 20 40 60 80 100

(a) (b)

palisadespongy mesophyll(c)

cuticleepidermis

(d)


300

400

500

4

6

8

120

10 Leaf size (cm2)

Hyd

rate

d LW

C (%

)

0

5

10

15

20

25

600

Stom

ata

(per

mm

2)

Trichomes (per m

m2)

0 20 40 60 80 100Soil water content (% saturation)

0

20

40

60

80

100

0

200

400

600

-2 -1 1 2 3 4 5

Stom

ata

(per

mm

2)

Trichomes (per m

m2)

Leaf number along stem

0.35

0.45

(e) (f)

aa a

aA

AB B B

A AA

A

a

aa a

AA

AA

aa

a a

AAB

BAB

A

B

AB

AB

a

a

b b

a

b bb



Chapter Six

114

development and aging (Figure 6.4f), while the ratios clearly differed between the soil water

content treatments (Figure 6.4c, d). While trichome numbers were similar between the

50%, 80% and 100% saturation treatments, the 80% and 100% plants had fewer stomata,

suggesting that partial and complete soil saturation triggered formation of leaves with fewer

stomata. In the 20% treatment, however, both trichome and stomatal densities were lower

than in the 50% treatment, which is explained most simply by the 20% treatment leaves

being older; since growth rates were greatly reduced in the low water plants, the ‘youngest

mature leaves’ were undoubtedly older in terms of actual time than in the higher water

treatments.

Adjustments in root morphology were also evident across the soil water gradient. The

100% soil saturation plants produced a dense root mat in the top of the pot, and numerous

upward growing roots with tips just protruding from the surface of the flooding water

(Figure 6.5a, c). All roots in the lower half of the pots had blackened and died by the end of

the eight weeks. The 80% soil saturation plants had numerous fine roots near the sand

surface, including a few upward growing root tips protruding from the surface, and roots

emerging from the stem base, but also maintained a large quantity of roots in the lower half

of the pot (Figure 6.5b, d). The quantity of surface roots decreased markedly with

decreasing sand water content, with roots in the 20% treatment concentrated in the lower

portion of the pot (Figure 6.5c–f).

The extent of endodermal suberisation in fine roots at 5 cm from the root tip was

lower in the 100% treatment, compared with the 20% and 50% soil water content

treatments (Figure 6.6a). Endodermal suberisation did not differ between the 20% and

50% treatments. The quantity of root aerenchyma was similar between the 50%, 80% and

100% treatments, and lower in the 20% treatment, although the only significant difference

in aerenchyma was between the 20% and 100% treatments (Figure 6.6b).

Physiological responses

The midday Ψshoot became more negative with decreasing soil water content. In Experiment

1, there were small but significant differences between all treatments, leading to a linear

relationship between soil water content and Ψshoot from -1.50 MPa in the 20% treatment,

to -0.64 MPa in the 100% treatment (Figure 6.7a). Similarly, leaf RWC increased slightly


115

with soil water content, from 85.0% RWC in the 20% treatment, to 89.3% RWC in the

100% treatment (Figure 6.7a). In Experiment 2, midday Ψleaf was more negative than the

control by day 4 (soil moisture 29% of saturation), and predawn Ψleaf after day 7 (19% of

saturation) (Figure 6.7b). At 14 days of progressive drying (soil 8% of saturation),

corresponding with the initiation of leaf shedding, midday and predawn Ψleaf were -2.2

MPa and -1.4 MPa respectively. The Ψleaf of control (field capacity) plants remained above

-1.2 MPa at all timepoints. Once leaf loss was complete in the droughted plants at day 35,

the water potential of stems was -6.5 MPa at predawn, and below -7 MPa (the detection

(f) 20%(e) 50%(d) 80%(c) 100%

1 cm

(b) 80%

1 cm

(a) 100%

1 cm

Figure 6.5: Root macro-morphology of seedlings grown under soil water conditions rang-

ing from 100% saturation (waterlogged) to 20% of saturation (low water). (a) Surface roots of a fully waterlogged seedling, with upward growing root tips protruding from the soil surface. (b) Above-ground surface roots emerging from the stem of a partially waterlogged seedling (80% soil saturation). (c−f) Seedling root systems exposed by cutting away pot walls, illustrating the variation in root quantity and depth across the range of soil water content treatments (100, 80, 50 and 20% of soil saturation).

Chapter Six

116

limit of the pressure chamber) at midday, and below -7 MPa at both predawn and midday

thereafter.

Leaf gas exchange parameters measured in Experiment 1 at the end of the eight week

treatment period are shown in Figure 6.8. There were no detectable differences in the rate

of carbon assimilation (A) between treatments. However, transpiration (E) and stomatal

conductance (gs) were significantly affected by soil water content. At midday, E and gs were

highest in the 80% treatment and lowest in the 20% treatment (Figure 6.8a), although

earlier in the morning (08:00) and in the later afternoon (16:00) there were no differences

among treatments (not shown). As a result, the lowest photosynthetic water use efficiency

(WUE) occurred at 80% soil saturation throughout most of the day, and highest WUE

under 20% saturation (Figure 6.8b). The leaf δ13C indicated a similar pattern of WUE,

although it was a less sensitive measure; δ13C was higher in the 20% treatment, but did not

differ significantly among the three higher water treatments (Figure 6.8b). Leaf gas

exchange was measured weekly during the treatment period, and the relative differences

among treatments remained similar throughout (not shown).

Figure 6.6: Root anatomy of seedlings under

a range of soil water contents. (a) Percentage of endodermal ring suberised (up to a maxi-mum of 200%, indicating two complete suber-ised rings). (b) Root air space (as a percentage of the root cortex cross sectional area). Values are means ± SE of four plants, values labelled with the same letter are not significantly differ-ent (p>0.05).

0

40

80

120

160

Endo

derm

al s

uber

isat

ion

(%)

0

5

10

15

20

Cort

ical

air

spac

e (%

)

20 40 60 80 100Soil water content (% saturation)

0

(a)

(b)

A

AA

A

A A

AB

B


117

(field capacity; ~50% saturation) treatment. The x axis of panel (b) displays the soil water content of the droughted plants, with the times since the start of treatment indicated above the graph. All values are means ± standard error of six plants, where error bars are not visible they are smaller than the symbols.

-1.6

-1.2

-0.8

-0.4

0

76

80

84

88

92


(a)

0 20 40 60 80 100

Sho

ot ψ

(MPa

)

Leaf RWC (%

)(b)

-7

-6

-5

-4

-3

-2

-1

0

01020304050

Drought predawnDrought midday

0 2 4 7 14 21 35Time (days)

Drought treatment soil water content (% saturation)

Sho

ot ψ

(MPa

)

Control predawnControl midday

Figure 6.7: Shoot water status of Mela-leuca argentea seedlings under a range

of soil water content. (a) Shoot water potential (ψ) and leaf water content relative to full hydration (RWC) of seed-lings after eight weeks of treatment at constant soil water content. Different letters indicate significant differences between treatments within a parameter. (b) Shoot ψ of seedlings during eight weeks of progressive drought (soil allowed to dry out over time) or control

A

B

CDa

ab

b b

Figure 6.8: Leaf gas exchange param-

eters of seedlings after eight weeks

under a range of soil water content

conditions. Instantaneous rates of (a) carbon assimilation (A), transpiration (E), and (b) intrinsic water use efficiency (WUE), at midday under light satura-tion, and (b) leaf 13C content. Values are mean ± standard error of six plants. Different letters indicate significant differences (p<0.05) between treat-ments within a measured parameter. Within each panel, upper case letters refer to the parameter shown in open symbols, lower case letters the param-eter shown in closed symbols.

0

1

2

3

4

5

0

4

8

12

16

0

2

4

6

-33

-32

-31

20(a)

(b)

0 20 40 60 80 100


8 E (mm

ol m-2 s-1)

A (μ

mol

CO

2 m-2

s-1)

δ13C (‰)

WU

E (μ

mol

CO

2. m

ol H

2O-1)

A A A A

a

bb b

a

b

cbc

A

B

CBC

Chapter Six

118

DISCUSSION

Growth and responses under saturated conditions

M. argentea seedlings are clearly highly tolerant of saturated soil conditions, with the

highest growth rates observed in partially waterlogged (80% saturation) soil. Under

complete waterlogging (100% saturation) the seedlings also performed well, although some

aspects of growth were reduced relative to partial waterlogging, most likely due to a

limitation in root growth or functioning. As hypothesised, seedlings in the 80% and 100%

soil saturation treatments displayed numerous adjustments, including substantial change to

root system morphology with production of shallow and negatively gravitropic roots. The

morphological and physiological responses observed most likely enable this species to grow

rapidly under waterlogged and semi-waterlogged conditions, which is vital for

establishment in intermittent waterways.

Unlike non-flood-tolerant species (e.g. Pezeshki & Anderson, 1996; Glenz et al.,

2006), M. argentea seedlings were able to continue growth, and showed no signs of water

stress during eight weeks of stagnant waterlogging. Leaf RWC, Ψ and rates of

photosynthesis in waterlogged seedlings were all higher or no different from those in the

‘optimal’ partial waterlogging conditions. Species of Salix and Populus inhabit equivalent

habitats to M. argentea in the northern hemisphere, and might be expected to show similar

levels of flooding tolerance. Seedlings of S. gracilistyla in southeast Asia, and S. nigra in

North America appear to be less flood tolerant than M. argentea, displaying severely

reduced growth and rates of photosynthesis under stagnant waterlogging relative to drained

conditions, and in some cases inhibited growth under partial waterlogging also (Pezeshki et

al., 1998; Li et al., 2004; Nakai et al., 2009). On the other hand, S. exigua may be more

tolerant than M. argentea, as seedling growth increased during complete waterlogging

(Amlin & Rood, 2001). Most poplars are less tolerant, for instance waterlogging greatly

reduced growth in seedlings of P. angustifolia, P. balsamifera and P. deltoides, and mature P.

balsamifera and P. deltoides were killed by long term inundation in Alberta, Canada (Amlin

& Rood, 2001). In seedlings of fourteen lines of North American poplar hybrids (hybrids

among P. trichocarpa, P. deltoides, P. nigra, and P. maximowiczii), photosynthesis decreased

during seven weeks of waterlogging, and plant height, leaf area and biomass were reduced

in most genotypes, although a few performed well (Gong et al., 2007). M. argentea


119

seedlings therefore appear to have flooding tolerance equivalent to some of the most

tolerant willows and poplars.

Many other Melaleuca species are also highly tolerant of waterlogging, and some may

be considered semi-aquatic trees or shrubs (e.g. Lockhart, 1996). Establishment of seedlings

of M. quinquenervia, native to tropical northern Australia, was optimal in very moist to

saturated soils (Myers, 1983), and transpiration rates of mature M. quinquenervia, were

unaffected by a seasonal inundation period of several months (McJannet, 2008). An ability

to continue such high rates of transpiration and growth during waterlogging is relatively

rare among tree species (Parolin & Wittmann, 2010). The success of M. quinquenervia as a

highly invasive weed in wetlands of Florida, USA, is attributed largely to its rapid and dense

root growth, and ability to rapidly produce deep roots to follow a receding water table

(Myers, 1983; Lopez-Zamora et al., 2004). Seedlings of M. cajuputi, native to far northern

Australia and tropical regions of Asia, also showed no difference in rates of photosynthesis

or shoot growth during 19 days of root hypoxia, compared with aerated controls, although

gs was reduced (Kogawara et al., 2006). M. argentea appears to share with other wetland

Melaleuca species a plastic and rapidly responding root system, which plays a major role in

flooding tolerance.

The M. argentea seedlings displayed a range of root-level responses to waterlogging.

Waterlogged seedlings had a shallow root system and wide stem base, responses common to

flood adapted species due to higher oxygen levels near the surface of waterlogged soil

(Pezeshki et al., 1998; Glenz et al., 2006; Myster, 2009). The M. argenta seedlings also

produced negatively gravitropic roots, similar those produced by the highly flood tolerant

M. quinquenervia (Sena Gomes & Kozlowski, 1980). The proportion of root cortical

airspace (aerenchyma) in M. argentea was low, remaining at only ~15% of the cross

sectional area in the field capacity, partially and fully waterlogged seedlings. Aerenchyma

frequently makes up 40–60% of the root cortical cross section in wetland plants, including

in Salix viminalis, and tropical wetland grasses (Baruch & Mérida, 1995; Jackson &

Attwood, 1996). However, production of extensive aerenchyma may be less common in

trees due to development of secondary thickening, for instance most central Amazonian

floodplain trees do not form appreciable root aerenchyma during waterlogging, even

evergreen species that actively grow during long flooded periods (De Simone et al., 2002b).

Chapter Six

120

Increased levels of enzymes involved in anaerobic metabolism may be an alternate strategy

to compensate for low levels of root porosity (Benz et al., 2007). In a comparison of two

flood tolerant Amazonian tree species, the highly aerenchymatous Salix martiana produced

low levels of alcohol dehydrogenase (ADH), while Tabernaemontana juruana produced

little aerenchyma but higher levels of ADH (De Simone et al., 2002a). Other highly flood

tolerant Melaleuca species such as M. cajiputi produce high levels of ADH under hypoxia

(Yamanoshita et al., 2005), and M. argentea may also adopt this strategy, reducing the

requirement for root aerenchyma.

Shoot morphological responses were also observed in M. argentea during

waterlogging which may be adaptations for shoot submergence. These included leaf and

stem elongation, and production of lower density leaves with a thicker spongy mesophyll

layer and lower stomatal density. While trees are highly unlikely to be fully submerged by

floods in semi-arid Australia, their seedlings may be, so these responses could be important

for young plants. Leaf responses can be critical during shoot submergence, for example

Melaleuca halmaturorum displays few leaf responses to flooding, and seedlings are highly

resistant to anoxic soil conditions, but susceptible to shoot submergence (Denton & Ganf,

1994). Ethylene production in the roots during flooding can signal shoot elongation, a

response thought to aid the plant in escaping flood waters (e.g. Voesenek et al., 2003). In

seedlings of the semi-aquatic M. quinquinervia, increased height growth under submerged

conditions was associated with improved survival rates (Lockhart et al., 1999). The thicker,

lower density spongy mesophyll suggests development of airspaces allowing internal oxygen

transport through the leaves; this response is observed during waterlogging in small plant

species that experience shoot submergence, such as the chrysanthemum Dendranthema

zawadskii (Yin et al., 2010), although leaf airspace formation appears to be less common in

flood tolerant tree species (e.g. Herrera et al., 2009). Lower stomatal density may be

advantageous during shoot submergence. At the Mapire River in Venezuela, partial

submergence reduced stomatal density in a few tree species, such as Eschweilera tenuifolia,

although not in other species (Herrera et al., 2009). In heterophyllous semi-aquatic species,

including M. quinquenervia and Ranunculus flabellaris, leaves produced underwater have

fewer stomata than those produced in air (Young & Horton, 1985; Lockhart, 1996). The


121

leaf and shoot responses seen in M. argentea might confer benefits under soil waterlogging,

and indicate that the seedlings may also be tolerant of shoot flooding.

Drought responses

The results of this study highlight the high water requirements of M. argentea in early stages

of establishment. As hypothesised, seedlings displayed fewer adaptive drought responses

compared with responses to waterlogging, and were clearly susceptible to drying of the

substrate. Below soil water content of 80% saturation, growth decreased and leaf Ψ and

RWC indicated declining water status. Under more severe water stress, the seedlings

appeared to rapidly enter a leafless, dormant state, in which they were able to survive at

least eight weeks without water. Rather than ‘tolerating’ drought, the plants appeared to

adopt a ‘waiting’ strategy, which may give the seedlings the best chance of survival in the

event that the drought is short lived, and abundant water returns.

Despite the cessation of growth and the significant decreases in shoot water status

(RWC and Ψ), the M. argentea seedlings maintained high rates of carbon assimilation

under low soil water content (20% of soil saturation). A small further decline in soil water

content, to 10-15% of soil saturation in the severe drought experiment, triggered a sudden,

rapid decline in Ψleaf, and onset of leaf shedding. Although the severe drought was a

progressive treatment, and thus may not be entirely comparable with the constant soil water

gradient experiment (e.g. Hukin et al., 2005), it appears that the M. argentea seedlings

maintained carbon assimilation rates with decreasing soil water, down to a point very close

to the threshold of ‘catastrophic’ xylem cavitation, and therefore displayed very limited

stomatal control of water loss. Some stomatal response was evident since gs decreased and

WUE increased with decreasing soil water content. However, species with sensitive

stomatal control or efficient osmo-regulation show little change in leaf RWC or Ψ with

moderate reductions in water availability (Gonzalez-Rodriguez et al., 2005). Drought

tolerant trees such as olive (Olea europa) typically also display a tight correlation between gs

and photosynthetic rate with decreasing water availability, indicating stomatal limitation of

both water loss and carbon gain (e.g. Guerfel et al., 2009). The M. argentea seedlings did

not show these types of physiological responses, instead displaying several morphological

responses to drought, including canopy shedding.

Chapter Six

122

Many riparian Salix and Populus species also readily shed leaves and branches under

drought. Canopy abscission is due to hydraulic failure of the leaf, petiole or stem xylem,

and early canopy loss during drought is therefore seen in species with low cavitation

resistance (Tyree et al., 1994; Nardini et al., 2001; Salleo et al., 2002). Leaf shedding was

observed in seedlings of a range of Salix species at predawn Ψleaf just below -1 MPa (Savage

& Cavender-Bares, 2011), similar to the point at which M. argentea seedlings shed leaves in

this study. Rapid leaf drop reduces water loss and can prevent the development of more

extensive xylem cavitation, leading to improved rates of re-sprouting upon rewetting

(Brodribb & Cochard, 2009; Savage & Cavender-Bares, 2011). For example, Populus

balsamifera in Alberta, Canada, displayed rapid leaf and branch senescence during moderate

groundwater drawdown over a period of 2 years, but tree growth rate returned to control

levels within a year of the drawdown ceasing (Amlin & Rood, 2003). Carbon starvation can

also lead to drought mortality, often acting in combination with hydraulic failure, since

avoiding hydraulic failure by closing stomata occurs at the cost of carbon assimilation. A

stronger reduction of growth than of photosynthesis during early water stress, as seen in the

M. argentea seedlings, is associated with accumulation of carbohydrates, which can then

sustain cellular respiration as drought severity increases (McDowell, 2011; Mitchell et al.,

2013). The M. argentea seedlings all survived three weeks without water in a leafless state,

and re-sprouted within four weeks of re-watering. The maintenance of gs and

photosynthesis through early drought, allowing ‘precocious’ shedding of leaves and

branches, may represent a drought survival strategy for many riparian tree species (Rood et

al., 2000), including M. argentea. By retaining moisture and carbohydrate reserves and

preventing further drought damage, the trees may increase their chance of survival and

recovery if favourable conditions return.

Several leaf-level responses to drought were observed in the M. argentea seedlings;

SLA and leaf hydrated water content decreased, and cuticle thickness increased slightly.

However, the strongest drought responses to water stress were observed in the roots.

Resource allocation to roots increased under drought; root:shoot ratio, root depth and root

nitrogen content all increased below 80% soil saturation. Melaleuca species are known to

accumulate nitrogenous compounds (proline and analogues) during drought as osmo-

protectants (Naidu et al., 1987); while osmolytes were not measured in this study, the


123

increased nitrogen levels could indicate osmo-protection of root tissues during water stress

(see further discussion below). The extent of root endodermal suberisation was also greater

in the lower water treatments, a common response which can prevent water loss from the

roots into dry soils (North & Nobel, 1998; Enstone et al., 2003). Root growth is crucial for

semi-arid riparian tree seedlings, in order to maintain contact with groundwater as it

recedes. Many trees species increase investment in roots and rooting depth during drought

(Osunubi & Davies, 1981; Bongarten & Teskey, 1987; Abrams, 1990), but by no means

all species do so (Osunubi & Davies, 1981; Pallardy & Rhoads, 1993; Joslin et al., 2000).

The protection of root tissues via osmotic and anatomical adjustments, and increased

allocation to roots by the M. argentea seedlings are clear adaptive responses to low water.

While M. argentea seedlings displayed some adaptive responses to drought, several

responses common in more drought tolerant species were lacking. Increased stomatal

density under water stress, which provides greater control of transpiration, is seen in

drought tolerant Quercus species and olive (Abrams, 1990; Bosabalidis & Kofidis, 2002),

but was not observed in M. argentea. Olive leaves also increased trichome density under

water stress to ~100 trichomes per mm2, which can reduce water loss and reflect heat and

radiation (Guerfel et al., 2009), while M. argentea leaves had 20 or fewer trichomes per

mm2. Increased leaf osmolyte concentration is another commonly observed response to

water stress, including in Eucalyptus astringens in semi-arid parts of south-west Australia

(Arndt et al., 2008). Some other Melaleuca species are adapted to dry conditions in arid

parts of Australia, such as M. filifolia, M. scabra and M. globifera, and these species

accumulate proline analogue compounds as osmo-protectants, but the most effective

osmolytes were absent from drought susceptible species including M. argentea (Naidu et al.,

2000). Leaf nitrogen content of the seedlings in the present study did not change with

decreasing SWC, indicating that changes in proline based osmolytes with water stress is

unlikely. Even in the absence of nitrogenous osmo-protectants, increases in leaf nitrogen

content are frequently associated with improved drought tolerance, including in many Salix

species and in Quercus suber (Aranda et al., 2005; Weih et al., 2011). This response in leaf

nitrogen was not observed in the M. argentea seedlings.

Chapter Six

124

Implications for response thresholds under field conditions

Changes in population structure or physiological functioning of dominant species can have

implications for whole ecosystem function (Xu & Zhou, 2011). In particular, identification

of response thresholds (abrupt shifts which occur across a small change in conditions) can

be important in ecosystem monitoring and management (e.g. Polasky et al., 2011), since

some changes may be irreversible or require prolonged recovery times, even after removal of

the environmental stressor (Ridolfi et al., 2006). A series of response thresholds was evident

in the M. argentea seedlings, moving from the wettest to driest ends of the soil water

gradient. Many parameters displayed a threshold between 80 and 50% saturation,

including reductions in leaf elongation rate and leaf water content at full turgor, and

increases in leaf tissue density and endodermal suberisation. Further reduction in soil water

from 50 to 20% produced an abrupt increase in root:shoot ratio, a reduction in the

quantity of root aerenchyma, and cessation of shoot height increase. With a further

decrease in soil moisture to ~15% of saturation, the threshold for a rapid decline in shoot

water potential and leaf shedding was reached. Abrupt physiological responses to decreasing

water availability are usually associated with xylem cavitation, occurring at distinct

threshold water potentials (Horton et al., 2001a), and the threshold responses in leaf

elongation, shoot height and leaf shedding are most likely directly linked with water

potential (Cosgrove, 1986; Salleo et al., 2002).

Partially saturated soils appear to be the ‘ideal’ condition for M. argentea seedlings,

likely to facilitate the maximum rates of tree recruitment. This is consistent with the

observation that M. argentea is frequently found in partially flooded locations in the field,

close to central river channels where pools of surface water form. The tolerance of

waterlogging is most likely attributable to the numerous adaptive responses that were

observed in both the roots and shoot. In contrast, the seedlings displayed fewer adaptive

responses to water stress, corresponding with poor drought tolerance and rapid decline in

performance with decreasing soil water content. However, the ‘non-tolerance’ of water

stress may in itself be considered an adaptive response to survive temporary drought.

Nonetheless, M. argentea clearly could not readily establish in persistently dry conditions,

while under conditions of increased water availability, such as during discharge of excess

water into a creekline (Chapter 1), recruitment rate may increase.


125

The point at which drought causes whole-plant death was not reached in this study,

although irreversible damage was caused to some branches. Since M. argentea appears to

enter dormancy during severe drought, the survival rate will be a function of the length of

the drought, and may also vary with factors such as temperature and tree size (McDowell et

al., 2008; Adams et al., 2009). Thinner stems and branches are often more vulnerable to

cavitation, and have greater surface area to volume ratios for evaporation to occur (Sperry

& Ikeda, 1997). Large M. argentea trees, with thick layers of bark, might therefore survive a

relatively long drought period in a dormant state. However, the relationship between tree

size and drought survival in a dormant state appears to vary among species and sites,

possibly depending on factors such as root depth, canopy exposure and water requirements,

and in some cases larger trees are more susceptible (Van Nieuwstadt & Sheil, 2005).

Survival can also depend upon the extent of carbohydrate reserves, the ability to mobilise

the reserves, and the rate of depletion during the dormancy period (McDowell et al., 2008;

Sala et al., 2010). High temperatures frequently increase cellular respiration rates, and may

also increase the desiccation rate due to evaporation, and so could reduce drought survival

times (Pinheiro & Chaves; Chaves et al., 2002; Adams et al., 2009). The ability of trees to

survive a period of drought is a complex process, however, the early leaf shedding response

of M. argentea might enable the species to survive relatively long periods of severe drought.

Some of the morphological and physiological responses that occurred at specific

points along the soil water gradient may be useful in understanding the extent of ‘stress’

experienced by M. argentea trees following hydrologic perturbation. For example, leaf Ψ

and canopy loss are frequently used to monitor tree ‘health’. Leaf Ψ appears to respond to

water availability in a linear fashion, which may belie the larger responses occurring in other

parameters. On the other hand, leaf loss may not occur until an advanced level of water

stress, and progresses rapidly below a threshold of water availability, so cannot provide

advance warning of severe stress in M. argentea. However, the trait response patterns

observed in this study would need to be verified under field conditions, and for mature

trees.

126

CHAPTER SEVEN: Response of Melaleuca argentea trees to ground

water abstraction from the coastal plain aquifer of a large ephemeral river

INTRODUCTION

In many arid and semi-arid regions of the world, groundwater is exploited to support

agriculture, industry and direct human consumption. As riparian trees are generally

dependent on shallow groundwater for survival and growth, abstraction can result in

declining tree health, including canopy loss and eventual mortality (Cooper et al., 2003;

Rood et al., 2003). Seedling recruitment along semi-arid ephemeral rivers is also reliant on

the episodic availability of surface and near-surface water (Mahoney & Rood, 1998). Under

groundwater abstraction these conditions occur less frequently, as complete groundwater

recharge takes longer and requires larger volumes of rainfall. Groundwater abstraction can

therefore lead to long term changes in tree population structures (Smith et al., 1991;

Williams & Cooper, 2005; Evans, 2007). As described in earlier Chapters, regional growth

in northwest Australia is placing greater demands on groundwater resources (Department

of Water, 2010). For example, the Yule River coastal plain aquifer, 50 km to the southwest

of the Port Hedland township, is currently under abstraction of up to 6.5 GL per year for

town and industry water supplies, and an increase of 2 GL per year in abstraction is

proposed. Current management of impacts on the groundwater dependent ecosystems is

based upon historical groundwater level records, but an increased abstraction regime would

require revision of the minimum groundwater levels that are set as the ‘trigger’ to reduce

abstraction pressure (Baimbridge, 2010). Relatively little is known about the ecological

water requirements of the groundwater dependent riparian vegetation of northwest

Australia, and thus there is a need to quantify vegetation responses to reduced groundwater

levels.

The sensitivity of riparian tree species to declining groundwater varies. For example,

Melaleuca argentea, a tree common to the waterways of northern Australia, appears to have

a low tolerance to drought (Masini, 1988; Chapter 6), and to show signs of decline earlier

than co-occurring Eucalyptus camaldulensis and E. victrix (unpublished data). Similarly,

along rivers in semi-arid Arizona, USA, Salix goodingii suffered more severe canopy dieback

Responses of Melaleuca argentea to groundwater abstraction

127

and greater rates of mortality than the co-occurring Populus fremontii when the depth to

groundwater fell below 3 m, while Tamarix chinensis in the same areas was more tolerant of

groundwater decline than either S. goodingii or P. fremontii (Horton et al., 2001b). The

differential responses between species may relate to differences in drought tolerance, and

differences in ability to maintain contact with a receding water table via root elongation

responses.

The rate of groundwater decline is a critical factor in the survival and response of

dependent vegetation (Eamus et al., 2006a). Provided the decline is gradual enough, trees

may be able to compensate for seasonal fluctuation in groundwater levels, and for

additional groundwater declines due to various human activities, via root elongation to

follow the water table. Numerous controlled rhizopod studies have been conducted with

tree seedlings and cuttings under simulated groundwater decline, in order to estimate the

maximum rates of root elongation, and therefore the maximum tolerable rate of

groundwater decline. Trees accustomed to large groundwater fluctuations may show greater

tolerance of drawdown, for example cuttings of Populus balsamifera, native to mountain

regions where water table depths often change abruptly, survived 10 cm day-1 simulated

drawdown, while cuttings of P. deltoides and P. angustifolia, native to floodplains with more

stable water tables, were mostly killed by the rapid drawdown (Kranjcec et al., 1998).

Recession rates of 1–2 cm day-1 appear tolerable for seedlings of a wide range of riparian

species (e.g. Alnus incana, Populus fremontii, Salix exigua, S. gooddingii), with rates of >3 cm

day-1 frequently leading to higher mortality rates, and reduced growth of any surviving

seedlings (Segelquist et al., 1993; Hughes et al., 1997; Amlin & Rood, 2003; Stella &

Battles, 2010). However sensitivity can clearly vary even among co-occuring species, e.g.

Tamarix chinensis showed no change in survival rate with drawdown of >4 cm day-1, while

S. gooddingii can show increased mortality at any rate of drawdown above zero (Horton &

Clark, 2001).

If groundwater recession exceeds the tolerable rate, or if groundwater levels fall below

maximum rooting depths, the impact on riparian and floodplain vegetation can be equated

to gradually increasing drought stress (Eamus et al., 2006a). Decline of groundwater levels

will dry an increasing portion of the root system; in the early stages the tree may be able to

compensate by taking up water via deeper roots (Chapters 4 & 5), but eventually a critical

Chapter Seven

128

root volume will be dried, beyond which the tree can no longer compensate. Decreasing

leaf water potential (Ψl), reduced stomatal conductance (gs) and reduced rates of

transpiration are early symptoms of water stress, progressing to canopy loss and mortality

with increasing stress (e.g. Horton et al., 2001c; Cooper et al., 2003). While in well watered

environments the vapour pressure deficit (VPD) of the air, and the soil and tree hydraulic

conductance (K) are the main drivers of gs and transpiration rate, in drying soil the soil

matric potential (Ψsoil), becomes the major determinant of gs (Brodribb, 2009; Chapter 2).

Stress responses during groundwater recession frequently occur in a threshold pattern at a

particular depth to groundwater (DGW), most likely corresponding to the root system

depth (Horton et al., 2001b; Chapter 6). Since juveniles frequently have shallower root

systems, they can be exposed to greater variation in water availability in dry environments

(Donovan & Ehleringer, 1991). It is therefore likely that larger trees will be able to access

water from deeper in the soil profile than smaller trees, and so the critical extent of root

drying under drawdown will be reached earlier for small trees. Thus, at a given DGW,

small trees may be exposed to lower Ψsoil, and so display greater water stress than larger

trees.

Along the coastal plain regions of the Yule River, DGW fluctuates considerably

within and between years (Department of Water, unpublished data), due to the highly

episodic rainfall patterns characteristic of the region, which in turn leads to episodic surface

flows and aquifer recharge. There are also likely to be high rates of evaporative water loss

from the aquifer through the wide, sandy river bed exposed to the high VPD conditions

(see Chapter 2). The potential impacts of groundwater abstraction therefore need to be

placed within the context of the natural range of water levels within the aquifer. M. argentea

is one of the dominant riparian tree species along the Yule River. Although this species

appears sensitive to small reductions in water availability (Chapter 6), trees on the Yule

River coastal plain may be relatively tolerant of groundwater abstraction, due to their

natural exposure to large fluctuations in water level. However abstraction causing water

levels to fall beyond a critical threshold of DGW would still be expected to induce drought

stress.

In this Chapter, I aimed to define the relationships between changing DGW and

physiological functioning of M. argentea trees growing along the Yule River, a large


129

ephemeral river in semi-arid NW Australia. I expected that (i) groundwater abstraction

would increase DGW; (ii) that lowering of the water table would impact negatively upon

M. argentea trees and (iii) the degree of impact would depend on tree size, with smaller

trees having reduced access to deeper portions of the soil profile, and thus showing signs of

water stress with smaller DGW than larger trees. I assessed tree physiological functioning

and the depth of water accessed by the trees, during one year of groundwater abstraction

from the Yule River aquifer. Trees in close proximity to the abstraction bores were

compared with those at an upstream location exposed only to the natural seasonal

groundwater fluctuations.

MATERIALS & METHODS

Study site

The Yule River coastal plain aquifer is located in the Pilbara region of northwest Australia,

where the river traverses the Port Hedland coastal plain ~20 km from the river mouth

estuary (Figure 7.1). Water extracted from the aquifer forms one of the water supply

sources for the town of Port Hedland, ~50 km to the NE. Due to increasing population

and demand for water, in 2009–2010 a trial was conducted to assess the ability of the

riparian vegetation to tolerate an increased rate of abstraction from the Yule River borefield.

Soils within the borefield are layered alluvial sands, silts and clays. The control site

surface soil is a pindan sand-clay to 1 m depth, then sand below 1m, with shingle and sand

layers at 1–4 m and 5–6 m depths. The impact site has a sandy surface soil to 4 m depth,

then shingle and sand at 4–7 m, and clay and sand below 7 m (data provided by the Water

Corporation, Western Australia). The climate of the region is semi-arid, characterised by

high temperature and VPD conditions for much of the year, and highly variable and

episodic rainfall patterns. Meteorological data was recorded during the study period by a

weather station located at the impact site, and additional daily solar radiation data were

obtained from the Australian Bureau of Meteorology, as described in Chapter 2.

The Yule River is ephemeral, with intermittent flows reflecting the highly variable

rainfall patterns of the Pilbara. The region can be subject to cyclone activity in the

December–April period, which can bring large rainfall events and flooding; river flows and

ground water recharge occur only after these events. Depth to the groundwater in the Yule

Chapter Seven

130

Port Hedland

KarrathaYULE RIVER

YULE BOREFIELD

5 km

NorthWest

Coastal H

wy

Impact

Control

Study site

Production boreWatercourse

Long term monitoring bores

Figure 7.1: Location of the Yule River borefield and study sites. Modified from maps provided by the Water Corporation of Western Australia.


131

River borefield fluctuates, as water levels decline naturally between episodes of recharge.

The section of the Yule River within the borefield consists of a wide riverbed of alluvial

sand, bordered by narrow bands of riparian vegetation, with broad floodplains beyond the

river bed vegetated by grassland and shrub-steppe. The production bores are located mostly

within the northern half of the borefield. A control site was established upstream, on the

river bank near the southern boundary of the borefield at 20.6563° S 118.2948° E, and an

impact site near to the northern-most production bores at 20.5372° S 118.1750° E, ~20

km from the control site (Figure 7.1). The control site is described in Chapter 2. The

impact site consisted of a wider band of riparian vegetation (~200 m), with sparser canopy

cover and trees less densely distributed, compared with the control site. The dominant

riparian tree species of Eucalyptus camaldulensis, E. victrix, and Melaleuca argentea were

present at both sites. Five M. argentea trees were selected at each site, encompassing a

diameter range (at 1.2 m height) of 140–500 mm. All measurements were conducted on

these same ten trees throughout the study.

Groundwater levels and abstraction

Cyclone activity during December 2008–February 2009 generated significant flows in the

Yule River, which recharged the aquifer (data provided by the Department of Water,

Western Australia). The pumping trial began in early April 2009, with abstraction from the

northern-most bores near the impact site given priority. However, pump failure meant that

several times the southern bores were also used by the Water Corporation in order to meet

the water supply requirements of Port Hedland, and the total rate of abstraction was lower

than planned. Hourly values of DGW were provided by the Department of Water,

measured at monitoring bores installed at the control and impact sites, fitted with sensors

and loggers. Historical water table depths were also obtained from the Department of

Water, for bores near to each site (Figure 7.2).

Soil moisture

Soil moisture data were provided by the Water Corporation, Western Australia. Soil

moisture through the profile to the water table was measured monthly with a neutron

moisture probe at each study site. Soil cores were also drilled at each site in July and

November 2009, and the volumetric water content measured on samples from 0.25 m

depth increments were used to calibrate the neutron probe readings.

Chapter Seven

132

Figure 7.2: Groundwater levels and rates of decline at the Yule Borefield study sites. Depth to groundwater (DGW) is shown relative to the ground surface at each study site. (a) Ground-water levels during the pumping trial study period. (b) Daily rate of groundwater decline during the study period (negative values indicate a water table rise). (c) Groundwater levels over ten years; the study period (shown in (a) and (b)) is indicated with vertical grey lines. Solid lines show data from the bores at the study sites installed in March 2005, dashed lines show data from long term monitoring bores, located within 2.5 km of the study sites, but further away from the river bed.

Days after start of pumping

(a)

0

1

2

3

4

5

6

7

ImpactControl

-6

-4

-2

0

2

4

6

Rate

of d

eclin

e (c

m d

ay-1)

0 100 200 300 400

NA MM J J A S O D J F M A

(b)

DG

W (m

)

0

1

2

3

4

5

6

7

8

1999 2000 2001

ControlImpactControl- long termImpact- long term

(c)

2002 2003 2004 2005 2006 2007 2008 2009 2010Year

DG

W (m

)


133

Isotope analysis of xylem sap, ground and soil water

Isotope analysis was used to identify the depth of the water within the soil profile accessed

by the trees. The isotopic composition of water usually varies through the soil profile, due

to evaporation at the surface enriching surface and near-surface soil water in the heavier

isotopes. Water uptake by plant roots is generally indiscriminate with respect to isotopic

composition (Dawson & Ehleringer, 1991). The isotopic signature of tree xylem water can

therefore be compared to the isotopic signatures of soil and groundwater, to assess the likely

source water depth used by the trees.

Samples of xylem sap were collected from the ten study trees (five at each site) in

March, June and November 2009, and May 2010. The sap was vacuum extracted in the

field; branchlets of ~1 cm diameter were cut from ~5 m height above the ground in the

early morning, all leaves removed, and the bark removed from the first 5 mm of the cut end

to prevent contamination with phloem sap. Xylem contents were collected in 2 ml vials,

which were sealed well and kept refrigerated or in an ice-box during transportation and

storage until analysis.

Soil cores were drilled at each site in July 2009 and early December 2009, and soil

samples collected at 0.6 m depth increments to a total depth of 3.6 m. Additional soil

samples were collected by hand auger in November 2009, at 0.1–0.2 m depth increments

to a total of 1.6 m. Soil was tightly packed into 30 ml vials immediately after drilling or

augering, and kept well sealed and refrigerated or in an ice-box until analysis. Drilling and

augering took place within the riparian tree bands, close to the study trees, near the base of

the bank to the main riverbed. Groundwater samples were taken from the monitoring bores

at each site, in November 2009 and May 2010. Water from surface pools was also sampled

in June 2009 (pools had dried down by November 2009).

Soil water was cryogenically distilled, and the 2H and 18O content of all water samples

determined by mass spectrometry (L1115-I Cavity Ring-down Spectrometer, Picarro, Santa

Clara, USA) at the West Australian Biogeochemistry Centre. Isotopic contents are

expressed as delta values in parts per thousand (‰) relative to the Vienna Standard Mean

Ocean Water international standard. The external error (standard deviation) for δ2H

analysis was ~2‰ for distilled samples, and <1‰ for other samples. The external error for

δ18O analysis was ~0.2‰ for distilled samples, and <0.1‰ for liquid samples.

Chapter Seven

134

Tree water use

Sap flow sensors (HRM 30; ICT International, Armidale, NSW, Australia) were installed

on 28th March 2009, just before the start of the pumping trial in early April. A single

sensor was installed in the stem of each of the ten study trees (five at each site), at 1.2 m

height, as described in Chapter 2. Heat pulse velocity was recorded every half hour from

29th March 2009 until 12th May 2010, except for the period 15th March–12th May 2010

at the impact site, when battery failure resulted in the half hourly measurements being

recorded only between 08:00 and 18:00 on each day (when equipment was powered by

solar panel directly). I was therefore unable to calculate total daily sap flow for this period at

the impact site, but was still able to determine maximum daytime flow rates. The methods

of calculating sap flow velocity (Vs) and the measurement of sapwood cross sectional area

are also described in Chapter 2.

Leaf physiological measurements

Leaf measurements were conducted in March, June and November 2009, and May 2010,

on the five study trees at each site. Leaf water potential (Ψl) was measured with a

Scholander-type pressure chamber, as described in Chapter 2. Leaf 13C was measured in

young, mature leaves collected from the outer canopy, ~5 m height above the ground. Leaf 13C is used as an indication of intrinsic water use efficiency integrated over the life span of

the leaf; RuBisCO preferentially utilises 12CO2, but as stomatal conductance decreases, CO2

within the sub-stomatal cavities and leaf tissues becomes limiting, forcing increased fixation

of 13CO2. Leaf material was dried at 55 °C, ground to a fine powder in a ball mill, and

samples analysed in a Nitrogen-Carbon Analyser Mass Spectrometer (Europa Scientific

Ltd., Crewe, UK), at the West Australian Biogeochemistry Centre. 13C contents are given

as delta values in parts per thousand (‰), relative to the Vienna PeeDee Belemnite

international standard. External error (standard deviation) of the analysis was <0.15‰. For

comparison with the 13C values, stomatal conductance was also measured with a leaf

porometer (Decagon Devices, Pullman, WA, USA) at both sites in November 2009 and

May 2010, as described in Chapter 2. Leaf specific area was determined from the same

samples as used for isotopic analysis; a hole punch was used to cut discs of known size from

ten leaves of each sample, and the discs weighed after drying at 55 °C.


135


Statistical analyses were performed with R version 2.13.0 (R Development Core Team,

2010). Linear mixed-effect models (lme4) were used to assess differences in tree sap flow,

leaf 13C, and xylem sap 2H and 18O. Comparisons were made between the sites, both over

time and over DGW. A random effect of tree individuals was included in the analysis when

it significantly improved the model (as determined by comparing models based on AIC

scores and χ2 log-likelihood tests). The effects of site, days after start of trial, DGW and tree

size on the response variable (sap flow, leaf 13C or isotopic signature) were tested by

comparing nested models differing in a single variable based on AIC scores and χ2 log-

likelihood tests. Effects are reported as significant where the χ2 log-likelihood test yielded a

p value of <0.05, relative to the simpler model.

RESULTS

Water table recession with groundwater abstraction

The amount of groundwater abstraction was less than planned; as a result the range of

groundwater levels observed during the study was mostly within the levels previously

recorded in the aquifer. At the impact site, the water table declined 4.3 m, from 2.4 m

DGW at the start of the trial (March 2009) to 6.7 m DGW in May 2010 (Figure 7.2a).

The water table at the control site also declined during the study period, but only by 1.5 m,

from 0.5 to 2.0 m DGW. The daily rates of decline increased toward the end of the study

period, the maximum recorded was 6 cm day-1 (Figure 7.2b). Cyclonic rainfall just before

the start of the study filled the aquifer to levels that were one of the highest recorded during

the previous ten years (Figure 7.2c). By the end of the study period, the water table at the

impact site had fallen just below the previous ten-year minimum.

Tree water uptake from soil, surface and ground water sources

The δ18O of the ground and soil water remained similar over time at each site (Figure 7.3a).

Groundwater δ18O values differed by 2.3‰ between the upstream control site (at -7.1‰)

and impact site (at -9.4‰), indicative of evaporation from the sandy riverbed as the

groundwater flows downstream underground. Soil water δ18O varied with depth below the

surface, with a similar range of values at both sites, of -13.0 to 3.5‰. Surface pools were

present at the control site at the start of the study, with δ18O of -6.3 to -5.2‰. Volumetric

Chapter Seven

136

soil water content in the upper 4 to 5 m also remained very similar over time at both sites

(Figure 7.3b). At the control site, soil water content ranged from 5% (vol/vol) near the

surface to 25% at 5 m depth, while at the impact site surface soil moisture was less than

2%, increasing to 20% in a high moisture layer at 3.7 m, then returning to 4% by ~4.5 m

depth. At depths below 4.5 m, soil water content decreased over time with the increasing

DGW, with much more pronounced changes over time and depth at the impact site

(Figure 7.3b).

Tree xylem sap δ18O at the control site (mean -6.4‰) was generally similar to that of

the groundwater at that site (-7.1‰), while at the impact site, the sap was slightly more

enriched (mean -7.2‰) compared to groundwater (-9.4‰) (Figure 7.3a). At the end of

the study period at the impact site, δ18O was also significantly more enriched in the sap of

smaller trees, while larger trees had sap isotopic signatures closer to that of groundwater

(p=0.022). There was no significant effect of tree size on sap δ18O at the earlier time points

at the impact site (March, June and November 2009), or at any time point at the control

site. δ2H was also analysed for each water sample, yielding results very similar to those

obtained with δ18O (not shown).

Tree water use

Tree water use reflected a lowering of the water table, even though groundwater levels were

mostly within the historic range (Fig. 7.2). Tree daily Vs did not differ between sites at the

start of the pumping trial (March and April 2009; p=0.95). However, water use at the

impact site decreased significantly over the study period relative to the control site

(p=0.003; Figure 7.4a). Water use at both sites fluctuated over the study period, reflecting

seasonal patterns in VPD (Chapter 2), with lower rates of sap flow during winter (~3 m

day-1 at the control site and ~1.5 m day-1 at the impact site) than in summer (up to 5 m

day-1 at the control site and 3 m day-1 at the impact site) (Figure 7.4a). The proportion of

sap flow occurring during the dark period was consistently slightly lower in trees at the

impact site throughout the study period (on average lower by 4.5%, with nocturnal flow

ranging from 10–41% of total daily flow at the control site, and 7–37% at the impact site),

but there was no detectable effect of groundwater drawdown on nocturnal sap flow rates

(not shown). The maximum rates of sap flow velocity reached each day (Vs-max) were much

more stable over time than the total daily sap flow (Vs) (Figure 7.4b). The Vs-max was lower


137

at

the impact site throughout the study (p=0.002; Figure 7.4b). At the impact site, there was a

significant interaction effect between DGW and tree size on Vs-max, with smaller trees having

lower Vs-max at greater DGW (p<0.001). At the control site there was no detectable effect of

either DGW or tree size on Vs-max (p=0.12 for tree size, p=0.51 for DGW). Effects of tree

Figure 7.3: Soil and ground water sources used by the trees at the Yule River control

(lefthand panels) and impacted (righthand panels) study sites during the pumping trial

period. (a) Isotopic signature (δ18O) of water extracted from tree xylem (upper panels) and from soil (lower panels, open symbols), and of groundwater obtained from bores (lower panels, closed symbols). Values are plotted against tree size (sapwood area) for xylem water, and against depth below the ground surface for source water. (b) Soil water content (SWC) with depth below the ground surface.

CONTROL SITE IMPACT SITE(a)

Sapw

ood

area

(cm

2 )

0

2

4

6

8

10

March 2009July 2009Dec 2009April 2010

tree

sourcewater

xylemD

epth

(m)

(b)

SWC (% (vol))

δ18O (‰) δ18O (‰)

Dep

th (m

)

0 20 40 60 80 0 20 40 60 80 100SWC (% (vol))

1

2

3

4

5

6

7

0

50

100

150

200

250

300March 2009June 2009Nov 2009May 2010

50-5-10-1550-5-10-15

Chapter Seven

138

size were not detectable on Vs at either site. There appeared to be an inflection point in the

Vs-max values at the impact site at ~day 230 (November 2009), after which the Vs-max began

to decline. The inflection point corresponds to DGW of 5.5 m, a decline of 3.1 m since the

start of the study, a depth change that was never reached by the groundwater at the control

site. Toward the end of the study period, when groundwater levels at the impact site were

at their lowest, fluctuations in the groundwater level at that site appear to be reflected in the

tree Vs-max, with a lag time of several days (Figures 7.2, 7.4).

Leaf physiology of M. argentea in relation to lowering of groundwater

In March 2009, before the start of the pumping trial, there were no differences in Ψl

evident between the sites or between trees of different size, but as the trial progressed the

0

5

10

15

20

25

30

35

0

1

2

3

4

5

6

7

ControlImpact

V s (m

day

-1)V s-m

ax (c

m h

-1)

Days after start of pumping

NA MM J J A S O D J F M A

0 100 200 300 400

(a)

(b)

Figure 7.4: Tree sap flux at control and impact sites during the pumping trial study period. (a) Daily sap flux velocity (Vs), and (b) daily maximum sap flux velocity (Vs-max). Values are means (bold lines) and standard errors (pale lines) of five trees. The x-axis displays the time elapsed since the start of the pumping trial study period, as well as letters indicating the months of the year (from March 28th 2009 to May 12th 2010).


139

predawn Ψl decreased in the smallest trees at the impact site (Figure 7.5). There were no

significant changes in the predawn Ψl of the larger trees, or in the midday Ψl of trees of any

size. The whole tree K, estimated from the linear portion of the relationships between Ψl

and Vs, are shown in Figure 7.6 (see also Figure 2.7 in Chapter 2 for examples of the Ψl by

Vs relationships). The K of trees at the impact site decreased over the trial period,

particularly in the smaller trees, while K did not change at the control site. The impact site

trees also had greater K than the control site trees at the beginning of the study period.

The patterns of leaf δ13C signatures over time differed between sites (p=0.006, Figure

7.7a); δ13C increased over time at the impact site (p=0.026), and did not change

significantly at the control site (p=0.11). There was no effect of tree size on δ13C detectable

at either site (p=0.67 control site; p=0.24 impact site). Trees of all sizes at the impact site

therefore appear to have increased their intrinsic water use efficiency progressively over the

study period. Stomatal conductance (gs) measured November 2009 and May 2010 also

revealed differences between sites (Figure 7.7b). The gs of trees at the impact site was lower

throughout the day, and at the mid-morning time points there was also a significant

interaction effect between site and tree size (p=0.0023 Nov 2009, p=0.0481 May 2010). At

the impact site, small trees had lower mid-morning gs than larger trees (p=0.003 Nov 2009,

p=0.034 May 2010), while there was no effect of tree size at the control site at those same

times (p=0.151 and 0.099). There was no detectable effect of tree size at any other time of

day at either site. There was also no detectable difference in leaf specific area between sites,

or over time (data not shown).

DISCUSSION

The DGW in the Yule River borefield appeared to be increased by the abstraction regime;

levels fell by 4.3 m (2.4–6.7 DGW) at the impact site, and by only 1.5 m (0.5–2.0 DGW)

at the control site, over the 13 months of this study. However the impact site has clearly

experienced a history of greater groundwater fluctuation than the control site, so it is

difficult to determine the extent to which the increase in DGW is due to abstraction, as

opposed to ‘natural’ fluctuation patterns. Nonetheless, greater DGW appears to have

mildly impacted physiological functioning of M. argentea trees near the production bores,

compared with trees at the upstream control site, with small trees more strongly affected

Chapter Seven

140

than larger trees. By the end of the study period, the small trees at the impact site had lower

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.20

ControlImpact

-2.5

-2.0

-1.5

-1.0

-0.5

0

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0

100 200 3000

-2.5

-2.0

-1.5

-1.0

-0.5

100 200 3000

Tree sapwood area at 1.2 m height (cm2)

0

-2.5

-2.0

-1.5

-1.0

-0.5

0

-2.5

-2.0

-1.5

-1.0

-0.5

0

Pred

awn

leaf

wat

er p

oten

tial (

ψ; M

Pa) M

idday leaf water potential (ψ

; MPa)

March 2009

June 2009

November 2009

May 2010

March 2009

June 2009

November 2009

May 2010

Figure 7.5: Variation in leaf water potential (ψ) with tree size during the pumping trial study

period. Predawn ψ is shown in the lefthand panels, and midday ψ in the righthand panels; note difference in y-axis scales.


141

(a) Control (b) ImpactK

(ml.

h-1.

cm-2.

MPa

-1)

0

10

20

30

40

50

60

70

0 100 200 300

June 2009 November 2009 May 2010

0 100 200 300 Tree sapwood area at 1.2 m height (cm2)

Figure 7.6: Relationship between whole tree conductance (K) and tree size (sapwood area)

for trees at the (a) control and (b) impact sites. K was estimated from the linear part of the relationship between sapflux and leaf water potential.

0

100

200

300

400

6:00 12:00 18:00

ControlImpact

-34

-33

-32

-31

-30

6:00 12:00 18:000:00 0:00

(b) Nov 2009 (c) May 2010

g s (m

mol

m-2

s-1)

Time of day (h) Time of day (h)

(a)

March2009

June2009

November2009

May2010

δ13C

(‰)

Figure 7.7: (a) Foliage carbon isotope ratio (δ13C) and (b) instantaneous stomatal conduct-

ance (gs) of study trees at the control (closed circles) and impact sites (open circles).

Chapter Seven

142

mid-morning gs, lower predawn water potentials, and sap isotopic signatures that indicated

increased use of soil water and therefore reduced access to the groundwater. In all trees at

the impact site, water use was reduced, intrinsic water use efficiency increased, and whole

tree hydraulic conductance reduced. These observations correspond with the finding that

M. argentea seedlings with lower water availability have lower gs and transpiration rates than

seedlings under ‘optimal’ conditions of partial saturation (Chapter 6). However, the

midday Ψl of trees at the impact site remained at control levels throughout the study, so the

physiological changes I observed appear to have been a mild response only. Changes in leaf

morphological traits associated with increasing water stress, such as reduction in specific

leaf area (e.g. Smith et al., 1991; Chapter 6), were also not detectable during drawdown at

the Yule River. The decreasing whole tree K at the impact site was most likely due to

cavitation of xylem vessels, or resistance at the soil-root interface, or both (Sperry et al.,

1998; Tognetti et al., 2004). In either case, the trees appear to have responded to the lower

K with reduction in gs, which maintained midday Ψl at control levels.

Plants can also respond to water stress with leaf drop to reduce canopy area, thereby

reducing transpiration and increasing the water potential of the remaining leaves (e.g.

Eamus et al., 2000; Cooper et al., 2003). I observed no obvious leaf loss or deterioration of

the canopy over time at the impact site, however it is possible that some leaf area

adjustments also occurred in addition to the reduction in gs. However, the extent of

groundwater recession at the impacted Yule River site was relatively large, at an increase in

DGW of 4.3 m over 13 months; similar and smaller declines in groundwater have resulted

in much more severe effects in other riparian tree species. Populus balsamifera on Willow

Creek in Alberta, Canada, suffered water stress under drawdown of 2.5 m over 2 years, with

widespread early leaf abscission (Amlin & Rood, 2003). On the South Platte River in

Colorado, USA, Populus deltoides suffered crown dieback of up to 30% following acute

groundwater drawdown of 2 m over a period of a few weeks (Cooper et al., 2003). P.

deltoides on Coal Creek in Colorado, USA, showed a threshold response with a decline of

just 1m, from 2.5 to 3.5m DGW; canopy desiccation and branch loss was evident within

weeks, and was associated with higher mortality rates in subsequent years (Scott et al.,

1999). Populus fremontii and Salix gooddingii on the Hassayampa River in Arizona showed


143

a threshold response when groundwater dropped below about 3m depth from the surface,

with increased canopy dieback and mortality at greater depths (Horton et al., 2001b). M.

argentea trees at the Yule River therefore appear to have relatively deep root systems for

phreatophytic riparian trees, with rates of water use remaining unchanged to 5.5 m DGW

(a decline of 3.1 m since the start of the study). As the water table fell beyond 5.5 m, water

use began to decline, primarily in small trees. In large trees there was no change in leaf

water potential and little change in sap flow rates beyond 6.5 m DGW.

The rates of water table decline during abstraction at the Yule River borefield were

also quite high toward the end of the study period. The maximum observed rate was 6 cm

day-1, although that was not maintained for longer than a couple of days at a time, and was

interspersed with declines of 2- 4 cm day-1 and small rises in the water table. Recession rates

of >3 cm day-1 are considered high for many riparian tree species, at least for seedlings, and

can lead to higher mortality rates and reduced growth (Segelquist et al., 1993; Hughes et

al., 1997; Amlin & Rood, 2003; Stella & Battles, 2010). While the tolerance of some

species to more rapid drawdown can be related to their greater drought tolerance (e.g.

Horton & Clark, 2001), in the case of M. argentea which appears to be drought sensitive,

tolerance of drawdown may be due to their rapid root growth rates (Chapter 6), or due to

constitutively deep root systems. Nonetheless, the small M. argentea trees impacted by

drawdown may have been unable to elongate roots rapidly enough, and suffered effects of

mild water stress, while the larger trees’ root systems may have been deep enough or

elongated rapidly enough to compensate. The type of substrate also affects survival and

growth; a given rate of decline is more severe in coarser soils (Mahoney & Rood, 1992;

Hughes et al., 1997). The layers of clay-type soils at the Yule River sites may have mitigated

the decline to some extent, slowing the recession of the non-saturated fringe.

Tree water use patterns in response to abstraction

Changes over time in the daily maximum rates of sap flow appear to be a highly sensitive

indicator of changes in water availability. Toward the end of the study there was a striking

mirroring of the small variations in groundwater depth, and the variations in daily Vs-max.

The lag time of several days could reflect the lag in the rate at which the unsaturated

capilliary fringe zone retreats as the saturated zone declines. Smaller trees had greater

reduction in Vs-max, although all trees at the impact site had slightly reduced water use, and

Chapter Seven

144

increased intrinsic water use efficiency. Larger trees may have reduced their water use by

shortening the plateau period of high water use during the day, rather than through

reduction in Vs-max (Chapter 2).

The large M. argentea trees at the impact site reduced water use and increased water

use efficiency, despite a lack of any other symptoms of water stress, including no change in

predawn Ψl. M. argentea appears to reduce its water use, through partial stomatal closure, in

response to relatively small reductions in water availability (Chapters 4 and 5). Stomata are

generally ‘inefficient’ if water is abundant, with greater losses of water for each incremental

increase in carbon assimilation, and even increasing water use without any change in the

observed assimilation rate (Thomas et al., 1999). For example, Salix babylonia in NSW,

Australia, used substantially more water when standing within the stream, compared with

nearby trees on the river bank (Doody & Benyon, 2010). Eucalyptus camaldulensis over a

shallow aquifer near the Murray River, SE Australia, had slightly increased growth rates

following an experimental flood treatment while Ψl was unchanged, suggesting that the

trees were utilising the additional water, although they were not previously water stressed

(Bacon et al., 1993). Populus fremontii trees on the San Pedro River in Arizona, USA, along

an intermittent reach with ground water 3.5–5 m below the surface had rates of Vs that

were less than half those of trees at a site with shallower ground water, within 3.5 m of the

surface (Gazal et al., 2006). Trees with abundant water are likely to use more, even though

trees at slightly ‘drier’ locations may still have access to an adequate supply.

Effects of abstraction on small trees

The smaller M. argentea trees at the impact site in our study did appear to show early signs

of water stress. Predawn Ψl were reduced, and isotopic signatures indicated greater uptake

of soil water by small trees with groundwater recession, demonstrating their reduced access

to the water table. Rates of sap flow and gs were reduced in all trees at the impact site, but

to a greater extent in small trees. Similar differences among tree size classes under drier

conditions have been observed in other riparian systems. Smith et al. (1991) found that

changes in riparian tree morphology and physiology in the Sierra Nevada resulting from

stream diversion were more pronounced in juveniles than in the mature trees, and

Donovan & Ehleringer (1991) found that Ψl, gs and photosynthesis differed between

juvenile and reproductive plants of several woody species to a much greater extent at a dry


145

location than at a wet location. Mature riparian Populus trees have been shown to survive

quite dramatic changes in hydrological regime, such as river damming and regulation;

rather the problem is often that recruitment of seedlings cannot take place under the new

conditions (Williams & Cooper, 2005). Sustained low groundwater levels can therefore

lead to mortality of young trees, and alter the community structure (Smith et al., 1991).

Implications for vegetation management under increased abstraction

As shown elsewhere, when water is abundant M. argentea consumes large quantities, but the

species is sensitive to small reductions in water availability, even in the absence of water

‘stress’ (Chapters 5 & 6). The maximum daily rates of sap flow reached by the trees appears

to be particularly sensitive, and this parameter has the potential to be used as an indicator

of groundwater levels, or as an ‘early warning signal’ of likely vegetation water stress.

The mature M. argentea trees in this study tolerated a relatively large recession in

groundwater. However, the Yule River site has historically been subjected to large

fluctuation in groundwater levels, both within and between years. The trees may therefore

be unusually tolerant of drawdown, and it is not known whether M. argentea would show

lower tolerance in locations with more consistent historical water table levels.

Due to the more pronounced effect of DGW on smaller trees, it is possible that

increased water abstraction could adversely affect the persistence of M. argentea populations

in the long term. Under increased abstraction, conditions supporting recruitment may

become less frequent, since the ability of seedlings to establish will require recharge events

large enough to replace abstracted water, and refill the aquifer from its reduced levels.

Establishment of young trees may also be impaired by higher rates of drawdown during the

juvenile phase. However, the effects of abstraction on the Yule River riparian ecosystem as a

whole will depend on how widespread the drawdown effects are. If the effects of abstraction

are relatively localised and populations of M. argentea are therefore able to thrive at nearby

unimpacted locations, the total impact is likely to be minor.

146

CHAPTER EIGHT: General discussion: temporal and spatial patterns of

water use by Melaleuca argentea

This thesis demonstrates the patterns of water use by Melaleuca argentea trees at several

scales, from fluxes within different portions of a tree, to seasonal and site differences in tree

water use. Many of the observed patterns and physiological responses are clearly explicable

in terms of the adaptation of M. argentea to the riparian environments of the Pilbara, which

combine harsh atmospheric conditions above ground with highly heterogeneous and

changeable conditions below ground. The water use physiology of this species is adapted to

the sub-zones of highest water availability and highest disturbance within the riparian

environments of northern Australia. High rates of water transport enable high rates of

growth and photosynthesis in the water abundant, high VPD environment. Nonetheless,

the limitation in water use apparent under extreme VPD points toward the fundamental

limits and tradeoffs involved in plant water transport. The plastic root responses of M.

argentea enable rapid adjustment to the disturbance and shifts in resource distribution that

commonly occur around the root system. These compensatory responses include a

previously undescribed mechanism involving rapid increase in the aquaporin levels and

hydraulic conductance of fine roots. The plasticity and waterlogging tolerance of the root

system allow these trees to thrive in the flood and disturbance prone river channels, which

provide resource rich sites in an otherwise harsh environment.

The findings of this research reveal new understanding of the water use physiology

and root system plasticity of M. argentea, which will in turn inform more accurate estimates

of the ecological water requirements of riparian systems in the Pilbara. These results also

considerably expand the relatively few studies made of riparian tree species in arid and

highly dynamic environments elsewhere in the world.

ADAPTATIONS OF M. ARGENTEA

Extreme VPD and temperature conditions

In northwest Australia, atmospheric vapour pressure deficit (VPD) is typically high (greater

than 3 kPa; e.g. Lio et al., 2004; Bobich et al., 2010) and often extreme in the austral

General discussion

147

summer, frequently exceeding 7 kPa, with maximum temperatures often greater than 40 °C

(Chapter 2). Under such conditions, very high rates of transpiration are expected for leaf

cooling and to meet the evaporative demand necessary to maintain photosynthesis and

growth. I observed maximum rates of sap flow in M. argentea in the range of 30–40 cm h-1

and up to 5 m day-1, similar to rates observed in other semi-arid riparian trees including

Populus and Salix species in Arizona, USA (Schaeffer et al., 2000; Gazal et al., 2006). While

these values are indeed relatively high, they are still within the range reported for many

other species, and even trees growing in cooler environments with much lower VPD can at

times reach similar sap velocities (e.g. Zang et al., 1996; Wullschleger et al., 1998). I also

found evidence that the water use of M. argentea did not always fully meet the evaporative

demand of the atmosphere. Tree hydraulic conductance (K) varied over diurnal cycles

during summer, with partial stomatal closure observed during the middle part of the day,

and the relationship between water use and VPD also varied over annual cycles, most likely

due to seasonal variation in tree capacitance and/or K (Chapter 2). The rate and volume of

water transported to the leaf surface therefore appears to place a limit on the water use of

M. argentea, despite abundant water availability to the roots.

At low to moderate evaporative demand (generally defined as VPD of 0–3 kPa),

transpiration usually increases with VPD, both within individual trees over diurnal and

annual cycles, and across climate gradients (e.g. Dragoni et al., 2009). However,

transpiration rates reach an ‘upper limit’ under higher VPD (often at 2–3 kPa), maintained

by reducing stomatal conductance (gs) and often leading to reductions in photosynthetic

rate, as documented in a wide variety of species and environments (e.g. Meinzer et al.,

1993; Granier et al., 1996; Hogg & Hurdle, 1997; Oren et al., 2001; Lio et al., 2004;

Woodruff et al., 2010). High VPD often correlates with drought conditions both

temporally and spatially, and so the reduced gas exchange is often attributable to soil

drought (Chaves et al., 2002). However, the ‘saturation’ of transpiration at high VPD is

also seen when water is not limiting, such as in irrigated orange trees in semi-arid Spain

(Martin-Gorriz et al., 2011), rainforest trees in the Ecuadorian Andes (Motzer et al., 2005),

after rainfall in semi-arid forests of the Loess Plateau region of China (Du et al., 2011), and

in riparian trees in north-west Australia (Pfautsch et al., 2011, Chapter 2). Trees native to

low VPD environments may show transpiration saturation at lower VPD than those native

Chapter Eight

148

to high VPD regions. In Sequoia sempervirens, native to the relatively mild, humid climate

of coastal California, USA, transpiration rate saturated at VPD of less than 1 kPa when

grown under higher VPD conditions with irrigation in the Los Angeles basin, USA (Litvak

et al., 2011). In contrast, transpiration rates of M. argentea in northwest Australia saturated

at VPD of approximately 3.5 kPa (Chapter 2). Nonetheless, 3.5 kPa is still a relatively low

VPD in the context of the extreme atmospheric conditions that commonly occur in the

Pilbara. Transpiration rates therefore appear in general to be limited at high VPD, even in

trees with access to abundant water, and native to environments where high VPD is

common.

Water transport can be limited by constraints on hydraulic architecture and vessel

structure. It would be expected that the optimal strategy in an environment with both high

demand and high availability of water would be to maximise tree transpiration and

hydraulic conductance (K). Indeed, xylem specific K (Ks) may be a key trait that maintains

plant water balance; across climatic zones Ks increases with soil water availability and with

temperature (i.e., with both availability and demand) (Gleason et al., 2012). The increase

in Ks across climate gradients is primarily due to increasing vessel diameter, which may also

increase the vulnerability to embolism (Zanne et al., 2010; Gleason et al., 2012). Many

studies indicate such a tradeoff between K and embolism vulnerability, on both theoretical

and empirical grounds (e.g. Pockman & Sperry, 2000; Wheeler et al., 2005; Hacke et al.,

2006). However, some studies have found contradictory evidence, observing a positive

correlation between transport capacity and cavitation resistance (e.g. Domec et al., 2005;

Bucci et al., 2013). In trees experiencing seasonal drought, Ks and transpiration rates might

be constrained during the wet season by the requirement for higher resistance to embolism

during dry periods, such as in several Eucalyptus species in northern Australian savannas

(Eamus et al., 2000). However, trees growing under conditions of constant high water

availability, such as M. argentea, are not constrained by seasonal water deficit. When

drought does occur, M. argentea adopts a strategy of rapid leaf loss and dormancy,

indicating low cavitation resistance (Chapter 6). The high rates of water use and low

drought tolerance of M. argentea are consistent with a vascular structure enabling high K

but low cavitation resistance.

General discussion

149

The vascular structure of trees is also constrained by mechanical strength

requirements, since an increase in vessel lumen area is associated with structural weakness

(Enquist, 2003). Riparian trees require high wood compression and shearing strengths, as

the environment is frequently subject to high energy flood disturbance (Naiman et al.,

2005; Niklas & Spatz, 2010). Tree hydraulic architecture and xylem structure appear in

general to be consistent with the maximisation of transport efficiency within the constraints

of mechanical strength (McCulloh & Sperry, 2005). In addition, xylem embolism could

still pose risks to transport when water is abundant, during periods of extreme evaporative

demand, if Ks is constrained by strength requirements. A combination of wood mechanical

strength and resistance to cavitation may therefore be the major limiting factors in the rate

of water transport by trees such as M. argentea.

Capacitance is also likely to play a role in the water transport strategy of riparian

trees. High wood and tissue capacitance can allow the use of internal water storage to

prevent highly negative water potentials developing (Meinzer et al., 2009). Capacitance

may thus partly overcome any tradeoff that exists between hydraulic efficiency and safety

(Sperry et al., 2008), though would probably not ameliorate a tradeoff between efficiency

and strength. Low wood density (low dry mass per volume) is associated with high sapwood

capacitance (Hultine et al., 2003b; Gartner & Meinzer, 2005; Scholz et al., 2007), but

lower mechanical strength (Niklas & Spatz, 2010; Zanne et al., 2010). M. argentea wood is

of intermediate density at 0.55 g cm-3, perhaps reflecting a balance between these

conflicting requirements. Both M. argentea and the co-occurring E. camaldulensis in wet

habitats in northwest Australia appear to draw upon internal water stores (Pfautsch et al.,

2011 Chapter 2). Diurnal cycles of dehydration and rehydration were observed in E.

camaldulensis (Pfautsch et al., 2011), and my results with M. argentea trees suggested that

internal stores were depleted and replenished over annual cycles (Chapter 2). Semi-arid

riparian trees on other continents may also utilise capacitance, for example the invasive

Tamarix chinensis in southwestern USA displayed diurnal changes in stem wood water

content (Moore et al., 2008). Many trees in the seasonally dry tropics have high

capacitance, which appears to allow these species to retain their leaves during the dry

season, and to increase their transport capacity (Goldstein et al., 1998; Meinzer et al., 2003;

Borchert & Pockman, 2005). Stem water storage appears to be a common strategy in trees

Chapter Eight

150

exhibiting high rates of water use, reducing their vulnerability to cavitation without

compromising transport efficiency, but the extent of capacitance might be limited by

mechanical strength requirements in some cases.

The tradeoffs and limitations among the various functions of wood tissues (water

transport efficiency, resistance to embolism, water storage capacity, mechanical strength

and resistance to decay) remain only partly resolved, but will be better clarified with further

study of wood biomechanical and hydraulic properties (Chave et al., 2009; Fournier et al.,

2013). Compilation of wood trait databases across species and biomes is currently

underway, with the aim of improving understanding of the inter-relationships among wood

properties and functions (Kattge et al., 2011; Choat et al., 2012). The increasing

availability of technologies such as nuclear magnetic resonance, microtomography and

synchrotron x-ray sources are also allowing in-planta analysis of wood functioning,

including time-lapse imaging of xylem water flow (e.g. Holbrook et al., 2001; Windt et al.,

2006; Kim & Lee, 2010; Page et al., 2012; Brodersen et al., 2013). Similarly, changes in

wood micro- and nano-structures occurring under mechanical stresses can be examined in

detail and in real time using these imaging approaches (Müller, 2009; Fernandes et al.,

2011). These developments will help to resolve questions of the maximum limits to

hydraulic functioning, and species such as M. argentea will provide good examples for

analysis of vascular systems operating near the upper limits of water transport capacity.

Heterogeneous belowground environment

Large and long-lived species in intermittent riparian environments, such as M. argentea,

encounter especially heterogeneous belowground conditions. All plants are exposed to

heterogeneity of numerous properties within their root-zone and canopy, but riparian

environments are particularly heterogeneous at multiple scales (Naiman et al., 2005; Pettit

& Naiman, 2005). Riparian heterogeneity is due primarily to the fluvial processes creating

cycles of disturbance and resource availability, and to the interactions between this resource

heterogeneity and biotic processes such as vegetation type and microbial activity (Naiman et

al., 2005). Highly variable flow patterns, typical of intermittent waterways in semi-arid

regions such as the Pilbara, further exacerbate the degree of spatial and temporal

heterogeneity. The basis of most plant responses to belowground heterogeneity is the

efficient use of available nutrient and water resources. When patches of water and nutrients

General discussion

151

are encountered, roots are able to detect and respond to them, however species appear to

vary widely in the extent to which they exploit patches (e.g. Huang et al., 1994; Gallardo et

al., 1996; Einsmann et al., 1999; Kembel & Cahill, 2005; Hodge, 2006; Neatrour et al.,

2007; Mommer et al., 2011). Semi-arid riparian trees have a high demand for water to

maintain transpiration and growth in the high VPD environment, and therefore are

expected to utilise resource patches. The high degree of root system plasticity in response to

changes in the distribution of water observed in M. argentea (Chapters 4, 5, 6) most likely

constitutes a strategy to maintain high rates of resource capture.

M. argentea displayed rapid root growth and physiological response times within

patches of high water availability, while simultaneously maintaining roots in previously rich

patches for up to several months, even after the water source was no longer present

(Chapters 4 & 5). Heterogeneous resource distribution can theoretically increase plant

performance in responsive species such as M. argentea, by allowing smaller investments in

roots into the rich patches only (foraging precision) (Hutchings & John, 2004; Magyar et

al., 2007). However, high foraging precision is only likely to be advantageous if conditions

change slowly relative to the response time of the roots, and/or if resource distributions are

relatively predictable (Alpert & Simms, 2002; Dunbabin et al., 2004; Hutchings & John,

2004). A permanently accessible water table is a highly predictable element of the habitat in

which M. argentea occurs (Masini, 1988), and accordingly trees appear to readily invest in

large roots which tap this water source, providing the majority of the water to the tree, at

least during drier periods (Chapter 4, 7). Upper soil layers can repeatedly experience pulses

of water and nutrients, and depressions in the creek bed forming temporary pools are likely

to refill with further rainfall. Correspondingly, M. argentea trees maintained surface roots in

the field for several months despite relatively little rainfall, and aquatic root mats were

partially maintained for at least 4 weeks after draining of the pools in glasshouse

experiments (Chapters 4 & 5). At the same time, high root growth rates can be important

in effectively exploiting ephemeral patches (Kembel & Cahill, 2005). Rapid root growth

and physiological responses (such as changes in root hydraulic conductivity) are also likely

to be beneficial in the event of dramatic shifts in resource patterns, such as those following

a destructive flood.

Chapter Eight

152

Water was the focus of the present study, however, nutrient availability will

undoubtedly play a role in the root responses of M. argentea. The relative importance of

water versus nutrients on root responses varies among studies reported in the literature, and

the effects of water and nutrients can also interact (Pregitzer et al., 1993; Coleman, 2007;

Zhou & Shangguan, 2007). In pot experiments, precise foraging for nutrients is commonly

reported, including in riparian trees such as Nyssa aquatica and Fraxinus pennsylvanica

(Blair & Perfecto, 2004; Neatrour et al., 2007). However, root foraging patterns can be

imprecise with respect to water in riparian trees with access to abundant water. For

example, fine root distribution did not closely reflect the distribution of water in Populus

fremontii seedlings grown in containers with vertical heterogeneity of water availability

(Snyder & Williams, 2007). Similarly, in this study when M. argentea seedlings were

planted into split-root containers mimicking access to an aquatic pool, the proportion of

fine roots in the aquatic pool remained unchanged eight weeks after planting, despite the

higher water availability in that compartment (Chapter 5). In contrast, rapid proliferation

of roots into aquatic pools by M. argentea has been observed in the field (personal

observations; Graham, 2001). Natural surface pools can also be nutrient rich, and the lack

of proliferation in the controlled experiment may have been due to the homogeneous

distribution of nutrients between compartments, while water supply was already adequate.

The primary response of riparian trees in water-abundant conditions may be to nutrient

availability, since nutrients are usually in more limited supply. Nonetheless, both water and

nutrient uptake will preferentially occur in the most densely rooted zones, regardless of the

primary response trigger.

My research has demonstrated considerable plasticity of root behavior in M. argentea,

but much remains unknown about how roots detect and signal their responses to soil

resources, especially to water (hydrotropism) (Cassab et al., 2013). My study illustrates how

the lack of a chemical signal (in this case ABA; Appendix C) can itself be an important part

of the signalling process, generating flow-on effects to other parts of the plant. The

apparent absence of a root-to-shoot signal under partial root zone drying in M. argentea led

to demand-driven compensatory responses in the remaining wet part of the root system

(Chapter 5). Many plant responses to water are generated largely through non-chemical,

hydraulic means, for example water potential and cell turgor pressure can directly affect

General discussion

153

stomatal conductance and growth (Brodribb, 2009; Kroeger et al., 2011). Genomic

approaches have revealed a multitude of gene expression responses to hydrostatic pressures

in bacteria and yeast (Ishii et al., 2005; Bravim et al., 2012), demonstrating the potential

for direct, molecular effects of hydrostatic pressure on cell functioning. The water flow

capacity of aquaporin channels also appears to be directly modulated by turgor pressure,

which most likely induces conformational changes (Wan et al., 2004). Further investigation

of gene expression patterns and protein functioning in response to hydrostatic pressure in

plant roots will improve the understanding of root water detection and response at the

molecular level.

Under conditions of spatial moisture heterogeneity, water may be passively

transferred through root systems from wet to dry soil regions (hydraulic redistribution;

HR). The acropetal flows in surface lateral roots of M. argentea during dry periods

demonstrate the transfer of groundwater through the root system into shallow roots

residing in the dry surface soil layers (Chapter 4). HR clearly improves root survival in dry

soil portions (Domec et al., 2004; Bauerle et al., 2008), most likely enabling the almost

instantaneous resumption of water uptake observed following rewetting, such as in pear

(Pyrus communis) roots under alternating drip irrigation (Kang et al., 2003), and in M.

argentea roots both in the field following rainfall (Chapter 4) and in glasshouse experiments

during partial root-zone drying and rewetting (Chapter 5). Although the occurrence of HR

is very widespread, there are a few documented cases of plants not displaying HR despite

moisture gradients across their root system, and so HR may not necessarily be an inevitable

consequence of heterogeneity. Species showing little or no HR activity, such as Quercus

margaretta in South Carolina, USA, appear to rapidly shed drying roots, preventing loss of

water into the dry soil, a mechanism which could itself be beneficial if total water

availability is limited (Ludwig et al., 2003; Espeleta et al., 2004; Prieto et al., 2012).

Differences among species in HR activity may be related to root hydraulic conductivity;

low conductivity may reduce the rate of HR, leading to fine root mortality in dry soils,

which would further prevent HR (Domec et al., 2004; Espeleta et al., 2004). The

substantial HR observed in M. argentea may thus be an indication of high root system

hydraulic conductance, likely to be an adaptive trait given the requirement for high rates of

water transport, and given that water availability is rarely limiting across the root system as

Chapter Eight

154

a whole. HR might also benefit M. argentea by enabling increased nutrient uptake from

surface soil layers, either as a side effect of high root conductance, or as an adaptive trait in

its own right.

The floods that bring pulses of water and nutrients also cause physical disturbance,

through erosion and deposition of sediments along river channels. Riparian trees are

dependent upon this disturbance for recruitment, since their propagules can generally only

establish in bare, moist sediment beds (e.g. Stromberg et al., 2010b). However, established

trees must then also withstand the impacts of burial in sediment and inundation, or

alternatively scouring of sediment and uprooting; both types of disturbance can lead to tree

mortality (Friedman & Auble, 1999; Bornette et al., 2008; Asaeda et al., 2010). For highly

waterlogging tolerant trees such as M. argentea, the greatest threat to survival appears to be

sediment erosion, which may cause the destruction or drying out of part of the root system,

impacting upon the anchorage role of the roots, as well as reducing the capacity for resource

capture (Kazda & Schmid, 2009; Asaeda et al., 2010). M. argentea roots were able to

rapidly compensate for part of the root system drying out by increasing rates of uptake

from other portions of the root system (Chapter 5), a response which may aid recovery

from flood scouring.

Plastic and responsive growth patterns appear to be key traits in disturbance-prone

riparian sites, in particular deeply anchored root systems combined with an ability to

rapidly re-sprout both shoots and adventitious roots from stumps (Bornette et al., 2008).

For example, Melaleuca leucadendra in the Burkedin River catchment in NE Australia

displays greater incidence of re-sprouting and clonal growth at higher disturbance locations

than at low disturbance sites (Chong, 2008). The most disturbance prone positions in the

riparian zone appear to be occupied by specialised species that display these highly plastic

strategies. For example, tree species were strongly segregated within the Koike Lake caldera

in Japan; species with life histories well adapted to frequent disturbance such as Neolitsea

aciculata and Quercus gilva occurred within the river channels, while other species were

confined to more stable zones (Ito et al., 2006). M. argentea in the Pilbara region is

confined to the wettest locations closest to the river channels, and possesses a relatively deep

root system, an ability to readily re-sprout following drought, as well as rapid production of

shallow root mats in flooded soils and surface pools (Chapter 5, 6, 7). The ability of M.

General discussion

155

argentea to plastically alter growth and response patterns of both above and below ground

structures is likely to be important for survival in disturbance-prone riparian locations.

ECOSYSTEM SCALE IMPLICATIONS OF WATER USE PATTERNS

Responses of trees to their environment have effects at the ecosystem scale. Transpiration

by riparian trees can be a significant element of the total hydrological budget in some cases,

particularly in dryland regions with high VPD (Scott et al., 2000; Dahm et al., 2002). For

instance, total evapotranspiration (ET) from riparian vegetation was estimated at 20–33%

of total losses from a riparian corridor in the Middle Rio Grande in semi-arid New Mexico,

USA (Dahm et al., 2002). Water fluxes through tree roots can be particularly important in

ecosystem functioning, not only through their influence on total ET, but also through their

effects on soil processes and understorey vegetation (Prieto et al., 2012). Soil moisture

heterogeneity and tree root processes can also have large impacts upon critical ecosystem

parameters such as catchment scale water runoff and soil carbon storage (e.g. Ren &

Henderson-Sellers, 2006; Zheng & Wang, 2007; Ren et al., 2008). Knowledge of the

spatial and temporal patterns of water uptake and root functioning, such as those

investigated in this thesis, can therefore be valuable in land and water management by

providing insights into ecosystem scale processes.

A striking finding of this study was the high variability in water fluxes through M.

argentea trees, at multiple scales. Quantifying and explaining such spatiotemporal

heterogeneity is a central issue in modelling ecosystem scale fluxes, and in the fundamental

understanding of ecosystem functioning (Asbjornsen et al., 2011). While quantifying

hydrological fluxes is critical in heavily populated areas (Schmandt, 2002; Jury & Vaux,

2007), determining aquifer fluxes and budgets in the less intensively managed Pilbara

region is nonetheless important in assessing likely impacts and recovery rates of

hydrological perturbations (Kirkpatrick & Dogramaci, 2010). For Pilbara waterways, the

contribution of tree water use to total hydrological fluxes is likely to vary widely between

locations. In areas such as the Yule River coastal plain, with a shallow aquifer beneath wide

sandy river beds, evaporation is likely to be the major component of ET and transpiration

may be relatively negligible. However, at Marillana Creek in the Hamersley Ranges, ET

accounts for an estimated 20% of losses from the aquifer and the contribution of

Chapter Eight

156

transpiration is likely to be significant though remains unquantified (Kirkpatrick &

Dogramaci, 2010). Temporal variation in sap flow by M. argentea in undisturbed habitat

may be largely predictable from meteorological variables, and over annual scales may also be

sufficiently estimated from short periods of measurements (Chapter 3). However, spatial

variation in flux rates were considerable, for reasons that remain unclear, and more

intensive sampling is required to resolve spatial patterns of water use in M. argentea among

sapwood regions, trees, stands and waterways.

Soil moisture is a key parameter in models of ecosystem water fluxes, and tree

responses to soil moisture and moisture heterogeneity can therefore be critical. Under

heterogeneous soil moisture, M. argentea root systems compensated for partial root zone

drying (Chapter 5), and also transferred water through the root system from wet to dry

regions (Chapter 4); both of these processes are likely to affect ecosystem water use. In

some tree species, moisture heterogeneity such as drying of the surface soil layers is expected

to reduce water use due to stomatal responses to partial root zone drying (Guswa, 2005;

Zheng & Wang, 2007). However, the findings reported in this thesis indicate that water

use by mature M. argentea trees in natural habitat is likely to be unaffected by surface soil

moisture, and to remain high throughout a wide range of conditions, as the root system

readily adjusts according to the spatial water distribution (Chapter 2, 5). Hydraulic lift may

also effect ecosystem net productivity and water use, since transfer of water to the surface

through tree roots can supply water to understorey vegetation, and increase rates of soil

evaporation and nutrient cycling (e.g. Brooks et al., 2002; Peñuelas & Filella, 2003; Scott et

al., 2008; Hawkins et al., 2009; Domec et al., 2010; Armas et al., 2012). For example,

using deuterium as a tracer, water transfer was observed from deep layers into surface soils

by Pinus nigra, where it was taken up by small Quercus ilex trees and sprouts (Peñuelas &

Filella, 2003). Hydraulic lift by a grass in controlled experiments increased the rate of

surface organic matter decomposition and plant nitrogen uptake (Armas et al., 2012).

Through the transfer of deep water to the surface, the presence of M. argentea is likely to

increase ecosystem ET during the early months of dry periods, and may extend the length

of time that water is available in the upper soil layers (Chapter 4). However, the period of

hydraulic lift may be of limited duration, and is therefore unlikely to indefinitely sustain

understorey plants and soil processes throughout dry periods.

General discussion

157

POTENTIAL IMPACTS OF ALTERED WATER REGIMES ON M. ARGENTEA

Riparian trees are adapted to and dependent upon the natural water flow regimes under

which they have developed. In many parts of the world, alterations to flows caused by

human activities such as dam construction, water abstraction or modification of channel

morphologies are frequently deleterious to tree health and ecosystem integrity (Lytle &

Poff, 2004). Consequently, we might expect perturbations to natural flows to pose

significant threats to riparian vegetation in northwest Australia. In parts of the Pilbara, iron

ore deposits are mined from below the water table, which requires abstracting groundwater

that must then be discharged elsewhere (Abbott et al., 2010; see also Chapter 1). The

release of this excess water into what are normally intermittent waterways with only

occasional surface water has raised concern over the effects of prolonged flooding of the

riparian vegetation. Prolonged inundation has occurred in other parts of the world

upstream of dams and caused widespread mortality and lack of regeneration in riparian

species. For example, dam enlargement on the Strymon River in northern Greece led to

extensive loss of waterlogged Salix alba woodlands (Crivelli et al., 1995). However, M.

argentea is highly tolerant of long term waterlogging (Chapter 6), and increased flows are

unlikely to cause major declines in this species. Indeed, greater water availability might in

some cases lead to greater recruitment and establishment, faster growth rates and expansion

of M. argentea populations. Elsewhere, inundation of trees following reservoir construction

in Arkansas, USA, lead to increased growth of the more waterlogging tolerant Quercus

lyrata, with concurrent mortality of less tolerant species including Q. phellos and Q. nuttallii

(King et al., 1998). Although some Melaleuca species can be highly invasive (Lopez-Zamora

et al., 2004), and M. argentea might conceivably invade new areas during riparian

inundation, M. argentea is also very drought sensitive, and these new recruits would be

unlikely to persist once discharge ceased at the completion of mining and groundwater

returned to pre-mining levels. Some other riparian species in the Pilbara may be adversely

affected by inundation, but effects on M. argentea appear to be of little ecological concern.

In contrast, groundwater drawdown resulting from dewatering of iron ore deposits,

and through abstraction for town and industry water supplies (Chapter 1) is likely to cause

declines in health and survival of M. argentea in some cases (Chapters 6 & 7). However,

some mature M. argentea trees appear to have relatively deep and/or responsive root

Chapter Eight

158

systems, which may be partly dependent on site; for instance trees at the Yule River were

able to tolerate groundwater abstraction leading to a greater than 4 m recession over 13

months (Chapter 7). However, whether this is typical of M. argentea trees throughout the

Pilbara is far from clear, since the levels of the Yule River aquifer show considerable natural

fluctuations of greater than 5 m, and so the trees may have been pre-adapted to the large

and rapid recession rates. Likewise, the maximum rates and extent of drawdown that would

tolerable for M. argentea are currently unknown.

The sequence of responses with decreasing water availability that were observed in M.

argentea might be useful in predicting the approach of major drought stress at sites

impacted by drawdown. While M. argentea is highly adaptable provided water is available

to some part of the root system, the species has very limited resistance to a limitation in

water availability across the root system as a whole. If groundwater levels fall below the

range of the root system, an abrupt and rapid decline in tree health would be expected

(Chapter 6). Other riparian tree species subject to groundwater abstraction show similar

threshold-type patterns of decline, for example a groundwater decline of just 1 m at Coal

Creek in Colorado, USA, lead to widespread canopy desiccation and branch loss in Populus

deltoides within weeks, and was associated with higher tree mortality in subsequent years

(Scott et al., 1999). An imminent collapse in tree health can therefore be difficult to

predict, with little visible early indication. However, after foliage loss M. argentea appears to

survive for some time in a dormant state; in a glasshouse experiment droughted seedlings

recovered when re-watered 45 days after initiation of leaf shedding (Chapter 6). The length

of time the trees might survive without water while dormant in the field is unknown, and

warrants further investigation. The physiological parameters of leaf water potential and sap

flow velocity are able to provide quite sensitive indications of early drought stress (Chapter

6, 7). These parameters are more demanding and expensive to monitor than canopy cover

and visual assessments of canopy health. In addition, the values at which major xylem

dysfunction occurs in M. argentea have not yet been ascertained under field conditions, and

in the case of sap flow velocity the threshold values are likely to vary widely among

individuals (Chapter 3). Further calibration under field conditions is therefore required,

before these parameters can be used to accurately monitor for impending ecosystem ‘stress’

and decline.

General discussion

159

Mild levels of inundation and abstraction that are tolerated well by mature trees can

still impact upon recruitment, leading to loss of trees from the impacted area in the longer

term. Of particularly concern for recruitment is the homogenisation of flow regimes and

elimination of the peak flood events, as is often the case on dammed and regulated rivers.

One thoroughly documented example is that of the Populus fremontii and Salix gooddingii

forests along the semi-arid rivers of southwestern North America, where distinct cohorts of

trees clearly correspond to past major floods, and forests are currently in decline along

highly regulated rivers where large floods no longer occur (Stromberg et al., 2010b).

However, such intensive regulation and dam construction is uncommon in the Pilbara, and

so even in impacted waterways, extreme flood events can still occur largely unimpeded.

Tree recruitment is therefore unlikely to be eliminated completely from impacted

waterways. Nonetheless, groundwater abstraction could reduce the frequency of successful

recruitment, by reducing water availability to seedlings during the early establishment phase

(Chapter 7).

Hydrological perturbations can alter the quantities of water used by riparian

vegetation. Stand and ecosystem scale water flux estimates based on values at unimpacted

sites might therefore be inaccurate for impacted areas. Transpiration rates of M. argentea

seedlings were reduced under both drought and stagnant waterlogging of the entire root

system (Chapter 6). Rates of water use also decreased as groundwater declined during

abstraction at the Yule River (Chapter 7). The water use of most other riparian trees also

appears to vary with water availability. Riparian trees in southwestern USA, such as Populus

deltoides and Tamarix ramosissima, can show substantial natural inter-annual variation in

water use, which appears to relate at least partly to the depth to the water table (Devitt et

al., 1998; Dahm et al., 2002). P. fremontii and Salix exigua near a reservoir in Utah, USA,

reduced water use during draining of the reservoir, with rates of water use recovering as the

reservoir re-filled (Hultine et al., 2010). On the Ain River in France, water use by Populus

nigra and Fraxinus excelsior responded to inter-annual variations in river flows (Singer et al.,

2012). Intermittent flooding can increase water use by riparian trees, but longer term

inundation can in some species lead to reductions in water use (Akeroyd et al., 1998;

Parolin & Wittmann, 2010). The apparent sensitivity of the rates of water use by M.

Chapter Eight

160

argentea to water availability suggest that anthropogenic variation in water regimes could

significantly influence the transpiration rate of this species.

CONCLUSION

M. argentea clearly has high water requirements, and is likely to be adversely affected by

reductions in water availability, either through changes in recharge regime due to shifts in

climate patterns, or occurring through groundwater drawdown. My findings illustrate the

need to better characterise the considerable spatial variation in tree water fluxes, if accurate

volumetric estimates of ecosystem water requirements and hydrological fluxes are to be

derived. I also observed a sequence of responses in M. argentea with decreasing water

availability that with further calibration under field conditions, has the potential to provide

sensitive methods to monitor for impending ecosystem water ‘stress’ and decline. Young

trees are likely to be more strongly affected by hydrological perturbations. Accordingly, the

processes of recruitment and establishment should be key considerations in managing

impacts if groundwater drawdown is spatially widespread, regardless of the severity of the

drawdown in terms of water table depth and rate of recession. Water management plans for

the riparian environments of the Pilbara, and of riparian environments more generally,

should carefully consider the often complex responses of dominant riparian tree species.

161

REFERENCES

Aanderud ZT, Richards JH (2009). Hydraulic redistribution may stimulate decomposition. Biogeochemistry 95: 323-333.

Abbott D, Humphreys G, Lynn F (2010). The Pilbara water in mining guideline- Integrating water policy with industry best practise and environmental approval processes. Groundwater 2010: The Challenge of Sustainable Management, Canberra, 31 Oct - 4 Nov 2010.

Abramoff MD, Magelhaes PJ, Ram SJ (2004). Image processing with ImageJ. Biophotonics International 11: 36-42.

Abrams MD (1990). Adaptations and responses to drought in Quercus species of North America. Tree Physiology 7: 227-238.

Abrisqueta JM, Mounzer O, Álvarez S, Conejero W, García-Orellana Y, Tapia LM, Vera J, Abrisqueta I, Ruiz-Sánchez MC (2008). Root dynamics of peach trees submitted to partial rootzone drying and continuous deficit irrigation. Agricultural Water Management 95: 959-967.

Adams HD, Guardiola-Claramonte M, Barron-Gafford GA, Villegas JC, Breshears DD, Zoug CB, Troch PA, Huxman TE (2009). Temperature sensitivity of drought-induced tree mortality portends increased regional die-off under global-change-type drought. Proceedings of the National Academy of Sciences 106: 7063–7066.

Akeroyd MD, Tyerman SD, Walker GR, Jolly ID (1998). Impact of flooding on the water use of semi-arid riparian eucalypts. Journal of Hydrology 206: 104-117.

Alpert P, Simms EL (2002). The relative advantages of plasticity and fixity in different environments: when is it good for a plant to adjust? Evolutionary Ecology 16: 285-297.

Amenu GG, Kumar P (2008). A model for hydraulic redistribution incorporating coupled soil-root moisture transport. Hydrology & Earth System Sciences 4: 3719-3769.

Amlin NA, Rood SB (2001). Inundation tolerances of riparian willows and cottonwoods. Journal of the American Water Resources Association 37: 1709-1720.

Amlin NM, Rood SB (2003). Drought stress and recovery of riparian cottonwoods due to water table alteration along Willow Creek, Alberta. Trees - Structure & Function 17: 351-358.

Antao M, Braimbridge M (2009). Millstream Status Report: A Review of Management and Consolidation of Current Understanding. Department of Water, Government of Western Australia, Perth. Environmental water report no. 9.

Aranda I, Castro L, Pardos M, Gil L, Pardos JA (2005). Effects of the interaction between drought and shade on water relations, gas exchange and morphological traits in cork oak (Quercus suber L.) seedlings. Forest Ecology & Management 210: 117-129.

Aranda I, Forner A, Cuesta B, Valladares F (2012). Species-specific water use by forest tree species: From the tree to the stand. Agricultural Water Management 114: 67-77.

162

Armas C, Kim J, Bleby T, Jackson R (2012). The effect of hydraulic lift on organic matter decomposition, soil nitrogen cycling, and nitrogen acquisition by a grass species. Oecologia 168: 11-22.

Arndt SK, Livesley SJ, Merchant A, Bleby TM, Grierson PF (2008). Quercitol and osmotic adaptation of field-grown Eucalyptus under seasonal drought stress. Plant, Cell & Environment 31: 915–924.

Aroca R, Amodeo G, Fernández-Illescas S, Herman EM, Chaumont F, Chrispeels MJ (2005). The role of aquaporins and membrane damage in chilling and hydrogen peroxide induced changes in the hydraulic conductance of maize roots. Plant Physiology 137: 341-353.

Aroca RR, Ferrante AA, Vernieri PP, Chrispeels MJ (2006). Drought, abscisic acid and transpiration rate effects on the regulation of PIP aquaporin gene expression and abundance in Phaseolus vulgaris plants. Annals of Botany 8: 1301-1310.

Arthington AH, Naiman RJ, McClain ME, Nilsson C (2010). Preserving the biodiversity and ecological services of rivers: new challenges and research opportunities. Freshwater Biology 55: 1-16.

Asaeda T, Gomes PIA, Takeda E (2010). Spatial and temporal tree colonization in a midstream sediment bar and the mechanisms governing tree mortality during a flood event. River Research and Applications 26: 960-976.

Asbjornsen H, Goldsmith GR, Alvarado-Barrientos MS, Rebel K, Van Osch FP, Rietkerk M, Chen J, Gotsch S, Tobón C, Geissert DR, Gómez-Tagle A, Vache K, Dawson TE (2011). Ecohydrological advances and applications in plant-water relations research: a review. Journal of Plant Ecology 4: 3-22.

Bacon PE, Stone C, Binns DL, Leslie DJ, Edwards DW (1993). Relationships between water availability and Eucalyptus camaldulensis growth in a riparian forest. Journal of Hydrology 150: 541-561.

Baimbridge M (2010). Yule River – ecological values and issues, Department of Water, Government of Western Australia.

Baird KJ, Maddock T (2005). Simulating riparian evapotranspiration: a new methodology and application for groundwater models. Journal of Hydrology 312: 176-190.

Baker TR, Affum-Baffoe K, Burslem DFRP, Swaine MD (2002). Phenological differences in tree water use and the timing of tropical forest inventories: conclusions from patterns of dry season diameter change. Forest Ecology & Management 171: 261-274.

Baldocchi DD, Ma S, Rambal S, Misson L, Ourcival JM, Limousin JM, Pereira J, Papale D (2010). On the differential advantages of evergreenness and deciduousness in mediterranean oak woodlands: A flux perspective. Ecological Applications 20: 1583-1597.

Barber M, Jackson S (2011). Water and Indigenous People in the Pilbara, Western Australia: A Preliminary Study. CSIRO: Water for a Healthy Country Flagship.

Baruch Z, Mérida T (1995). Effects of drought and flooding on root anatomy in four tropical forage grasses. International Journal of Plant Sciences 156: 514-521.

References

163

Batchelor CH, Rama Mohan Rao MS, Manohar Rao S (2003). Watershed development: A solution to water shortages in semi-arid India or part of the problem? Land Use and Water Resources Research 3: 1-10.

Bauerle TL, Richards JH, Smart DR, Eissenstat DM (2008). Importance of internal hydraulic redistribution for prolonging the lifespan of roots in dry soil. Plant, Cell & Environment 31: 177-186.

Beard JS (1975). Pilbara. University of Western Australia Press, Perth.

Beaudette PC, Chlup M, Yee J, Emery RJN (2007). Relationships of root conductivity and aquaporin gene expression in Pisum sativum: diurnal patterns and the response to HgCl2 and ABA. Journal of Experimental Botany 58: 1291-1300.

Bell DL, Sultan SE (1999). Dynamic phenotypic plasticity for root growth in Polygonum: a comparative study. American Journal of Botany 86: 807-819.

Benz BR, Rhode JM, Cruzan MB (2007). Aerenchyma development and elevated alcohol dehydrogenase activity as alternative responses to hypoxic soils in the Piriqueta caroliniana complex. American Journal of Botany 94: 542-550.

Berger A, Oren R, Schulze E-D (1994). Element concentrations in the xylem sap of Picea abies (L.) Karst. seedlings extracted by various methods under different environmental conditions. Tree Physiology 14: 111-128.

Betsch P, Bonal D, Breda N, Montpied P, Peiffer M, Tuzet A, Granier A (2011). Drought effects on water relations in beech: The contribution of exchangeable water reservoirs. Agricultural & Forest Meteorology 151: 531–543.

Blair BC, Perfecto I (2004). Successional status and root foraging for phosphorus in seven tropical tree species. Canadian Journal of Forest Research 34: 1128-1135.

Bleby TM, Burgess SSO, Adams MA (2004). A validation, comparison and error analysis of two heat-pulse methods for measuring sap flow in Eucalyptus marginata saplings. Functional Plant Biology 31: 645-658.

Bleby TM, McElrone AJ, Jackson RB (2010). Water uptake and hydraulic redistribution across large woody root systems to 20 m depth. Plant, Cell & Environment 33: 2132-2148.

Bobich EG, Barron-Gafford GA, Rascher KG, Murthy R (2010). Effects of drought and changes in vapour pressure deficit on water relations of Populus deltoides growing in ambient and elevated CO2. Tree Physiology 30: 866-875.

Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, White J-SS (2009). Generalized linear mixed models: a practical guide for ecology and evolution. Trends in Ecology & Evolution 24: 127-135.

Bonan GB (2008). Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science 320: 1444-1449.

Bongarten B, Teskey R (1987). Dry weight partitioning and its relationship to productivity in loblolly pine seedlings from seven sources. Forest Science 33: 255-267.

Borchert R, Pockman WT (2005). Water storage capacitance and xylem tension in isolated branches of temperate and tropical trees. Tree Physiology 25: 457-466.

164

Bornette G, Tabacchi E, Hupp C, Puijalon S, Rostan JC (2008). A model of plant strategies in fluvial hydrosystems. Freshwater Biology 53: 1692-1705.

Bosabalidis AM, Kofidis G (2002). Comparative effects of drought stress on leaf anatomy of two olive cultivars. Plant Science 163: 375-379.

Box EO, Fujiwara K (2009). Vegetation types and their broad-scale distribution. In: van der Maarel E ed. Vegetation Ecology. Wiley-Blackwell. pp106-128.

Bravim F, F. da Silva L, T. Souza D, I. Lippman S, R. Broach J, Alberto R. Fernandes A, M.B. Fernandes P (2012). High hydrostatic pressure activates transcription factors involved in Saccharomyces cerevisiae stress tolerance. Current Pharmaceutical Biotechnology 13: 2712-2720.

Brodersen CR, McElrone AJ, Choat B, Lee EF, Shackel KA, Matthews MA (2013). In vivo visualizations of drought-induced embolism spread in Vitis vinifera. Plant Physiology 161: 1820-1829.

Brodribb TJ (2009). Xylem hydraulic physiology: The functional backbone of terrestrial plant productivity. Plant Science 177: 245-251.

Brodribb TJ, Cochard H (2009). Hydraulic failure defines the recovery and point of death in water-stressed conifers. Plant Physiology 149: 575–584.

Brodribb TJ, Feild TS, Jordan GJ (2007). Leaf maximum photosynthetic rate and venation are linked by hydraulics. Plant Physiology 144: 1890-1898.

Brooks JR, Meinzer FC, Coulombe R, Gregg J (2002). Hydraulic redistribution of soil water during summer drought in two contrasting Pacific Northwest coniferous forests. Tree Physiology 22: 1107-1117.

Brooks JR, Meinzer FC, Warren JM, Domec J-C, Coulombe ROB (2006). Hydraulic redistribution in a Douglas-fir forest: lessons from system manipulations. Plant, Cell & Environment 29: 138-150.

Brooksbank K, White D, Veneklaas E, Carter J (2011). Hydraulic redistribution in Eucalyptus kochii subsp. borealis with variable access to fresh groundwater. Trees - Structure & Function 25: 735-744.

Bucci SJ, Scholz FG, Goldstein G, Meinzer FC, Hinojosa JA, Hoffmann WA, Franco AC (2004). Processes preventing nocturnal equilibration between leaf and soil water potential in tropical savanna woody species. Tree Physiology 24: 1119-1127.

Bucci SJ, Scholz FG, Peschiutta ML, Arias NS, Meinzer FC, Goldstein G (2013). The stem xylem of Patagonian shrubs operates far from the point of catastrophic dysfunction and is additionally protected from drought-induced embolism by leaves and roots. Plant, Cell & Environment 36: 2163-2174.

Buckley TN (2005). The control of stomata by water balance. New Phytologist 168: 275-292.

Bunn SE, Thoms MC, Hamilton SK, Capon SJ (2006). Flow variability in dryland rivers: boom, bust and the bits in between. River Research and Applications 22: 179-186.

Bureau of Meteorology (2013). Climate Data Services. Government of Australia. http://www.bom.gov.au/climate/data-services/

References

165

Burgess SSO, Adams MA, Turner NC, Beverly CR, Ong CK, Khan AAH, Bleby TM (2001). An improved heat pulse method to measure low and reverse rates of sap flow in woody plants. Tree Physiology 21: 589-598.

Burgess SSO, Adams MA, Turner NC, Ong CK (1998). The redistribution of soil water by tree root systems. Oecologia 115: 306-311.

Burke M, Jorde K, Buffington JM (2009). Application of a hierarchical framework for assessing environmental impacts of dam operation: Changes in streamflow, bed mobility and recruitment of riparian trees in a western North American river. Journal of Environmental Management 90, Supplement 3: S224-S236.

Busch DE, Smith SD (1995). Mechanisms associated with decline of woody species in riparian ecosystems of the Southwestern U.S. Ecological Monographs 65: 347-370.

Caldwell MM, Richards JH (1989). Hydraulic lift: water efflux from upper roots improves effectiveness of water uptake by deep roots. Oecologia 79: 1-5.

Capon SJ, James CS, Williams L, Quinn GP (2009). Responses to flooding and drying in seedlings of a common Australian desert floodplain shrub: Muehlenbeckia florulenta Meisn. (tangled lignum). Environmental & Experimental Botany 66: 178-185.

Cassab GI, Eapen D, Campos ME (2013). Root hydrotropism: an update. American Journal of Botany 100: 14-24.

Catford JA, Downes BJ, Gippel CJ, Vesk PA (2011). Flow regulation reduces native plant cover and facilitates exotic invasion in riparian wetlands. Journal of Applied Ecology 48: 432-442.

Caylor KK, D'Odorico P, Rodriguez-Iturbe I (2006). On the ecohydrology of structurally heterogeneous semiarid landscapes. Water Resources Research 42: W07424.

Caylor KK, Dragoni D (2009). Decoupling structural and environmental determinants of sap velocity: Part I. Methodological development. Agricultural & Forest Meteorology 149: 559-569.

Centritto M, Wahbi S, Serraj R, Chaves MM (2005). Effects of partial rootzone drying (PRD) on adult olive tree (Olea europaea) in field conditions under arid climate: II. Photosynthetic responses. Agriculture, Ecosystems & Environment 106: 303-311.

Čermák J, Kučera J, Bauerle WL, Phillips N, Hinckley TM (2007). Tree water storage and its diurnal dynamics related to sap flow and changes in stem volume in old-growth Douglas-fir trees. Tree Physiology 27: 181-198.

Čermák J, Nadezhdina N, Meiresonne L, Ceulemans R (2008). Scots pine root distribution derived from radial sap flow patterns in stems of large leaning trees. Plant Soil 305: 61-75.

Chaumont F, Barrieu F, Jung R, Chrispeels MJ (2000). Plasma membrane intrinsic proteins from maize cluster in two sequence subgroups with differential aquaporin activity. Plant Physiology 122: 1025-1034.

Chave J, Coomes D, Jansen S, Lewis SL, Swenson NG, Zanne AE (2009). Towards a worldwide wood economics spectrum. Ecology Letters 12: 351-366.

166

Chaves MM, Pereira JS, Maroco J, Rodrigues ML, Ricardo CPP, Osório ML, Carvalho I, Faria T, Pinheiro C (2002). How plants cope with water stress in the field? Photosynthesis and growth. Annals of Botany 89: 907-916.

Chen L, Zhang Z, Li Z, Tang J, Caldwell P, Zhang W (2011). Biophysical control of whole tree transpiration under an urban environment in Northern China. Journal of Hydrology 402: 388-400.

Choat B, Jansen S, Brodribb TJ, Cochard H, Delzon S, Bhaskar R, Bucci SJ, Feild TS, Gleason SM, Hacke UG, Jacobsen AL, Lens F, Maherali H, Martinez-Vilalta J, Mayr S, Mencuccini M, Mitchell PJ, Nardini A, Pittermann J, Pratt RB, Sperry JS, Westoby M, Wright IJ, Zanne AE (2012). Global convergence in the vulnerability of forests to drought. Nature 491: 752-755.

Chong C (2008). A Riparian Perspective on Species Ecology and Evolution: Melaleuca leucadendra (Myrtaceae). PhD thesis, James Cook University, Australia.

Chu CR, Hsieh C-I, Wu S-Y, Phillips NG (2009). Transient response of sap flow to wind speed. Journal of Experimental Botany 60: 249-255.

Cleverly JR, Dahm CN, Thibault JR, McDonnell DE, Allred Coonrod JE (2006). Riparian ecohydrology: regulation of water flux from the ground to the atmosphere in the Middle Rio Grande, New Mexico. Hydrological Processes 20: 3207-3225.

Cochard H, Delzon S (2013). Misunderstanding sap ascent in trees. Journal of Sigmoidal Plant Hydraulics 1: e0001.

Cohen S, Wheeler J, Holbrook NM (2012). The radial and azimuthal (or tangential) distribution of sap velocity in tree stems - why and can we predict it? Acta Horticulturae 951: 131-137.

Cohen Y, Cohen S, Cantuarias-Aviles T, Schiller G (2008). Variations in the radial gradient of sap velocity in trunks of forest and fruit trees. Plant and Soil 305: 49-59.

Coleman M (2007). Spatial and temporal patterns of root distribution in developing stands of four woody crop species grown with drip irrigation and fertilization. Plant and Soil 299: 195-213.

Cooper DJ, D'Amico DR, Scott ML (2003). Physiological and morphological response patterns of Populus deltoides to alluvial groundwater pumping. Environmental Management 31: 0215-0226.

Cosgrove D (1986). Biophysical control of plant cell growth. Annual Review of Plant Physiology 37: 377-405.

Crivelli A, Grillas P, Lacaze B (1995). Responses of vegetation to a rise in water level at Kerkini Reservoir (1982–1991), a Ramsar site in northern Greece. Environmental Management 19: 417-430.

Croker J, Witte W, Auge R (1998). Stomatal sensitivity of six temperate, deciduous tree species to non-hydraulic root-to-shoot signalling of partial soil drying. Journal of Experimental Botany 49: 761-774.

Dahm CN, Cleverly JR, Allred Coonrod JE, Thibault JR, McDonnell DE, Gilroy DJ (2002). Evapotranspiration at the land/water interface in a semi-arid drainage basin. Freshwater Biology 47: 831-843.

References

167

Davies WJ, Wilkinson S, Loveys B (2002). Stomatal control by chemical signalling and the exploitation of this mechanism to increase water use efficiency in agriculture. New Phytologist 153: 449-460.

Dawson TE, Ehleringer JR (1991). Streamside trees that do not use stream water. Nature 350: 335-337.

Dawson TE, Mambelli S, Plamboeck AH, Templer PH, Tu KP (2002). Stable isotopes in plant ecology. Annual Review of Ecology and Systematics 33: 507-559.

De Simone O, Haase K, ller E, Junk WJ, Gonsior G, Schmidt W (2002a). Impact of root morphology on metabolism and oxygen distribution in roots and rhizosphere from two Central Amazon floodplain tree species. Functional Plant Biology 29: 1025-1035.

De Simone O, Müller E, Junk WJ, Schmidt W (2002b). Adaptations of central Amazon tree species to prolonged flooding: Root morphology and leaf longevity. Plant Biology: 515-522.

Delzon S, Sartore M, Granier A, Loustau D (2004). Radial profiles of sap flow with increasing tree size in maritime pine. Tree Physiology 24: 1285-1293.

Denton M, Ganf GG (1994). Response of juvenile Melaleuca halmaturorum to flooding: Management implications for a seasonal wetland, Bool Lagoon, South Australia. Marine and Freshwater Research 45: 1395-1408.

Department of Regional Development and Lands, Pilbara Development Commission (2011). Pilbara: A Region in Profile.In Government of Western Australia, Perth.

Department of the Environment, Water, Heritage and the Arts (2008). Australia's rivers and catchment condition database. Australia's Natural Lands and Rivers. Government of Australia, Canberra.

Department of Water (2010). Pilbara Regional Water Plan 2010–2030: Supporting Detail.In Government of Western Australia, Perth.

Dettinger MD, Diaz HF (2000). Global characteristics of stream flow seasonality and variability. Journal of Hydrometeorology 1: 289-310.

Devitt DA, Sala A, Smith SD, Cleverly J, Shaulis LK, Hammett R (1998). Bowen Ratio estimates of evapotranspiration for Tamarix ramosissima stands on the Virgin River in southern Nevada. Water Resour. Res. 34: 2407-2414.

Dixon M (2003). Effects of flow pattern on riparian seedling recruitment on sandbars in the Wisconsin River, Wisconsin, USA. Wetlands 23: 125-139.

Dodd IC, Egea G, Davies WJ (2008). Accounting for sap flow from different parts of the root system improves the prediction of xylem ABA concentration in plants grown with heterogeneous soil moisture. Journal of Experimental Botany 59: 4083-4093.

Dodds WK, Clements WH, Gido K, Hilderbrand RH, King RS (2010). Thresholds, breakpoints, and nonlinearity in freshwaters as related to management. Journal of the North American Benthological Society 29: 988-997.

Dogramaci S, Skrzypek G, Dodson W, Grierson PF (2012). Stable isotope and hydrochemical evolution of groundwater in the semi-arid Hamersley Basin of subtropical northwest Australia. Journal of Hydrology 475: 281-293.

168

Domec JC, King JS, Noormets A, Treasure E, Gavazzi MJ, Sun G, McNulty SG (2010). Hydraulic redistribution of soil water by roots affects whole-stand evapotranspiration and net ecosystem carbon exchange. New Phytologist 187: 171-183.

Domec JC, Pruyn ML, Gartner BL (2005). Axial and radial profiles in conductivities, water storage and native embolism in trunks of young and old-growth ponderosa pine trees. Plant, Cell & Environment 28: 1103-1113.

Domec JC, Warren JM, Meinzer FC, Brooks JR, Coulombe R (2004). Native root xylem embolism and stomatal closure in stands of Douglas-fir and ponderosa pine: mitigation by hydraulic redistribution. Oecologia 141: 7-16.

Donovan LA, Ehleringer JR (1991). Ecophysiological differences among juvenile and reproductive plants of several woody species. Oecologia 86: 594-597.

Doody T, Benyon R (2011). Quantifying water savings from willow removal in Australian streams. Journal of Environmental Management 92: 926-935.

Dragoni D, Caylor KK, Schmid HP (2009). Decoupling structural & environmental determinants of sap velocity: Part II. Observational application. Agricultural and Forest Meteorology 149: 570-581.

Du S, Wang Y-L, Kume T, Zhang J-G, Otsuki K, Yamanaka N, Liu G-B (2011). Sapflow characteristics and climatic responses in three forest species in the semiarid Loess Plateau region of China. Agricultural & Forest Meteorology 151: 1-10.

Duff GA, Myers BA, Williams RJ, Eamus D, O'Grady A, Fordyce IR (1997). Seasonal patterns in soil moisture, vapour pressure deficit, tree canopy cover and pre-dawn water potential in a Northern Australian savanna. Australian Journal of Botany 45: 211-224.

Dunbabin V, Rengel Z, Diggle A (2004). Simulating form and function of root systems: efficiency of nitrate uptake is dependent on root system architecture and the spatial and temporal variability of nitrate supply. Functional Ecology 18: 204-211.

Eamus D (1999). Ecophysiological traits of deciduous and evergreen woody species in the seasonally dry tropics. Trends in Ecology & Evolution 14: 11-16.

Eamus D, Froend R, Loomes R, Hose G, Murray B (2006a). A functional methodology for determining the groundwater regime needed to maintain the health of groundwater-dependent vegetation. Australian Journal of Botany 54: 97–114.

Eamus D, Hatton T, Cook P, Colvin C (2006b). Ecohydrology : Vegetation Function, Water and Resource Management. CSIRO Publishing, Melbourne.

Eamus D, O'Grady AP, Hutley L (2000). Dry season conditions determine wet season water use in the wet-dry tropical savannas of northern Australia. Tree Physiology 20: 1219-1226.

Einsmann JC, Jones RH, Pu M, Mitchell AJ (1999). Nutrient foraging traits in 10 co-occurring plant species of contrasting life forms. Journal of Ecology 87: 609-619.

Enquist BJ (2002). Universal scaling in tree and vascular plant allometry: Toward a general quantitative theory linking plant form and function from cells to ecosystems. Tree Physiology 22: 1045-1064.

References

169

Enquist BJ (2003). Cope's Rule and the evolution of long-distance transport in vascular plants: allometric scaling, biomass partitioning and optimization. Plant, Cell & Environment 26: 151-161.

Enstone DE, Peterson CA, Ma F (2003). Root endodermis and exodermis: Structure, function, and responses to the environment. Journal of Plant Growth Regulation 21: 335–351.

Environment Australia (2001). A Directory of Important Wetlands in Australia. Third Edition. Australian Government, Canberra.

Espeleta JF, West JB, Donovan LA (2004). Species-specific patterns of hydraulic lift in co-occurring adult trees and grasses in a sandhill community. Oecologia 138: 341-349.

Evans R (2007). The Effects of Groundwater Pumping on Stream Flow in Australia. Technical Report. Land & Water Australia, Canberra.

Fellman JB, Dogramaci S, Skrzypek G, Dodson W, Grierson PF (2011). Hydrologic control of dissolved organic matter biogeochemistry in pools of a subtropical dryland river. Water Resources Research 47: W06501.

Fernandes AN, Thomas LH, Altaner CM, Callow P, Forsyth VT, Apperley DC, Kennedy CJ, Jarvis MC (2011). Nanostructure of cellulose microfibrils in spruce wood. Proceedings of the National Academy of Sciences 108: E1195-E1203.

Fernandez G (2007). Model Selection in PROC MIXED - A User-friendly SAS® Macro Application. SAS Global Forum, Orlando, Florida.

Fetter K, Van Wilder V, Moshelion M, Chaumont F (2004). Interactions between plasma membrane aquaporins modulate their water channel activity. Plant Cell 16: 215-228.

Fiora A, Cescatti A (2006). Diurnal and seasonal variability in radial distribution of sap flux density: implications for estimating stand transpiration. Tree Physiology 26: 1217-1225.

Flom PL, Cassell DL (2007). Stopping stepwise: Why stepwise and similar selection methods are bad, and what you should use. NESUG Statistics and Data Analysis.

Ford CR, Goranson CE, Mitchell RJ, Will RE, Teskey RO (2004). Diurnal and seasonal variability in the radial distribution of sap flow: predicting total stem flow in Pinus taeda trees. Tree Physiology 24: 951-960.

Ford CR, Hubbard RM, Kloeppel BD, Vose JM (2007). A comparison of sap flux-based evapotranspiration estimates with catchment-scale water balance. Agricultural & Forest Meteorology 145: 176-185.

Fort C, Fauveau ML, Muller F, Label P, Granier A, Dreyer E (1997). Stomatal conductance, growth and root signaling in young oak seedlings subjected to partial soil drying. Tree Physiology 17: 281-289.

Fort C, Muller F, Label P, Granier A, Dreyer E (1998). Stomatal conductance, growth and root signaling in Betula pendula seedlings subjected to partial soil drying. Tree Physiology 18: 769-776.

170

Fournier M, Dlouhá J, Jaouen G, Almeras T (2013). Integrative biomechanics for tree ecology: beyond wood density and strength. Journal of Experimental Botany 64: 4793-4815.

Friedman JM, Auble GT (1999). Mortality of riparian box elder from sediment mobilization and extended inundation. Regulated Rivers: Research & Management 15: 463-476.

Friedman JM, Lee VJ (2002). Extreme floods, channel change, and riparian forests along ephemeral streams. Ecological Monographs 72: 409-425.

Froend R, Drake P (2006). Defining phreatophyte response to reduced water availability: preliminary investigations on the use of xylem cavitation vulnerability in Banksia woodland species. Australian Journal of Botany 54: 173–179.

Fu P-L, Jiang Y-J, Wang A-Y, Brodribb TJ, Zhang J-L, Zhu S-D, Cao K-F (2012). Stem hydraulic traits and leaf water-stress tolerance are co-ordinated with the leaf phenology of angiosperm trees in an Asian tropical dry karst forest. Annals of Botany 110: 189-199.

Gallardo M, Jackson LE, Thompson RB (1996). Shoot and root physiological responses to localized zones of soil moisture in cultivated and wild lettuce (Lactuca spp.). Plant, Cell & Environment 19: 1169-1178.

Gallardo M, Turner NC, Ludwig C (1994). Water relations, gas exchange and abscisic acid content of Lupinus cosentinii leaves in response to drying different proportions of the root system. Journal of Experimental Botany 45: 909-918.

Galmés J, Pou A, Alsina M, Tomàs M, Medrano H, Flexas J (2007). Aquaporin expression in response to different water stress intensities and recovery in Richter-110 (Vitis sp.): Relationship with ecophysiological status. Planta 226: 671-681.

García-Palacios P, Maestre FT, Bardgett RD, de Kroon H (2012). Plant responses to soil heterogeneity and global environmental change. Journal of Ecology 100: 1303-1314.

Gartner BL, Meinzer FC (2005). Structure-function relationships in sapwood water transport and storage. In: Holbrook NM, Zwieniecki MA eds. Vascular Transport in Plants. Elsevier Academic Press. pp307-331.

Gassmann M, Grenacher B, Rohde B, Vogel J (2009). Quantifying Western blots: Pitfalls of densitometry. Electrophoresis 30: 1845–1855.

Gazal RM, Scott RL, Goodrich DC, Williams DG (2006). Controls on transpiration in a semiarid riparian cottonwood forest. Agricultural & Forest Meteorology 137: 56-67.

Gebauer T, Horna V, Leuschner C (2008). Variability in radial sap flux density patterns and sapwood area among seven co-occurring temperate broad-leaved tree species. Tree Physiology 28: 1821–1830.

George AK, Walker KF, Lewis MM (2005). Population status of eucalypt trees on the River Murray floodplain, South Australia. River Research and Applications 21: 271-282.

Gleason SM, Butler DW, Ziemińska K, Waryszak P, Westoby M (2012). Stem xylem conductivity is key to plant water balance across Australian angiosperm species. Functional Ecology.

References

171

Glenz C, Schlaepfer R, Iorgulescu I, Kienast F (2006). Flooding tolerance of Central European tree and shrub species. Forest Ecology & Management 235: 1-13.

Glinski DS, Karnok KJ, Carrow RN (1993). Comparison of reporting methods for root growth data from transparent-interface measurements. Crop Science 33: 310-314.

Goldhammer DA, Salinas M, Crisosto C, Day KR, Soler M, Moriana A (2002). Effects of regulated deficit irrigation and partial root zone drying on late harvest peach tree performance. Acta Horticulturae 592: 343- 350.

Goldstein G, Andrade JL, Meinzer FC, Holbrook NM, Cavelier J, Jackson P, Celis A (1998). Stem water storage and diurnal patterns of water use in tropical forest canopy trees. Plant, Cell & Environment 21: 397-406.

Gong J-R, Zhang X-S, Huang Y-M, Zhang C-L (2007). The effects of flooding on several hybrid poplar clones in Northern China. Agroforestry Systems 69: 77-88.

Gonzalez-Rodriguez AM, Martin-Olivera A, Morales D, Jimenez MS (2005). Physiological responses of tagasaste to a progressive drought in its native environment on the Canary Islands. Environmental & Experimental Botany 53: 195-204.

Goodrich DC, Scott R, Qi J, Goff B, Unkrich CL, Moran MS, Williams D, Schaeffer S, Snyder K, MacNish R, Maddock T, Pool D, Chehbouni A, Cooper DI, Eichinger WE, Shuttleworth WJ, Kerr Y, Marsett R, Ni W (2000). Seasonal estimates of riparian evapotranspiration using remote and in situ measurements. Agricultural & Forest Meteorology 105: 281-309.

Gowing DJG, Davies WJ, Jones HG (1990). A positive root-sourced signal as an indicator of soil drying in apple, Malus x domestica Borkh. Journal of Experimental Botany 41: 1535-1540.

Graham J (2001). The root hydraulic architecture of Melaleuca argentea. BSc (Honours) Thesis thesis, The University of Western Australia, Perth, Australia, Perth.

Graham MH (2003). Confronting multicollinearity in ecological multiple regression. Ecology 84: 2809-2815.

Granier A, Huc R, Barigah ST (1996). Transpiration of natural rainforest and its dependence on climatic factors. Agricultural & Forest Meteorology 78: 19-29.

Granier A, Loustau D, Bréda N (2000). A generic model of forest canopy conductance dependent on climate, soil water availability and leaf area index. Ann. For. Sci. 57: 755-765.

Greco S, Baldocchi DD (1996). Seasonal variations of CO2 and water vapour exchange rates over a temperate deciduous forest. Global Change Biology 2: 183-197.

Grice AC (2006). The impacts of invasive plant species on the biodiversity of Australian rangelands. The Rangeland Journal 28: 27-35.

Grierson PF, Adams MA (2000). Plant species affect acid phosphatase, ergosterol and microbial P in a Jarrah (Eucalyptus marginata Donn ex Sm.) forest in south-western Australia. Soil Biology and Biochemistry 32: 1817-1827.

Groffman P, Baron J, Blett T, Gold A, Goodman I, Gunderson L, Levinson B, Palmer M, Paerl H, Peterson G, Poff N, Rejeski D, Reynolds J, Turner M,

172

Weathers K, Wiens J (2006). Ecological thresholds: The key to successful environmental management or an important concept with no practical application? Ecosystems 9: 1-13.

Guerfel M, Baccouri O, Boujnah D, Chaïbi W, Zarrouk M (2009). Impacts of water stress on gas exchange, water relations, chlorophyll content and leaf structure in the two main Tunisian olive (Olea europaea L.) cultivars. Scientia Horticulturae 119: 257-263.

Guswa AJ (2005). Soil-moisture limits on plant uptake: An upscaled relationship for water-limited ecosystems. Advances in Water Resources 28: 543-552.

Guyot G, Scoffoni C, Sack L (2012). Combined impacts of irradiance and dehydration on leaf hydraulic conductance: insights into vulnerability and stomatal control. Plant, Cell & Environment 35: 857–871.

Hachez C, Moshelion M, Zelazny E, Cavez D, Chaumont F (2006). Localization and quantification of plasma membrane aquaporin expression in maize primary root: a clue to understanding their role as cellular plumbers. Plant Molecular Biology 62: 305-323.

Hacke U, Sauter JJ (1996). Drought-induced xylem dysfunction in petioles, branches, and roots of Populus balsamifera L. and Alnus glutinosa (L.) Gaertn. Plant Physiology 111: 413-417.

Hacke UG, Sperry JS, Wheeler JK, Castro L (2006). Scaling of angiosperm xylem structure with safety and efficiency. Tree Physiology 26: 689-701.

Haig T (2009). The Pilbara Coast Water Study. Department of Water Hydrogeological Record Series, Government of Western Australia, Perth. Report HG34.

Halse S, Scanlon M, Cocking J, Smith M, Kay W (2007). Factors affecting river health and its assessment over broad geographic ranges: The Western Australian experience. Environmental Monitoring and Assessment 134: 161-175.

Hao X, Chen Y, Li W, Guo B, Zhao R (2010). Hydraulic lift in Populus euphratica Oliv. from the desert riparian vegetation of the Tarim River Basin. Journal of Arid Environments 74: 905–911.

Hatton T, Evans R (1998). Dependence of Ecosystems on Groundwater and its Significance to Australia. Land and Water Resources Research and Development Corportion, Canberra.

Hatton TJ, Moore SJ, Reece PH (1995). Estimating stand transpiration in a Eucalyptus populnea woodland with the heat pulse method: measurement errors and sampling strategies. Tree Physiology 15: 219-227.

Hawkins H, Hettasch H, West AG, Cramer MD (2009). Hydraulic redistribution by Protea ‘Sylvia’ (Proteaceae) facilitates soil water replenishment and water acquisition by an understorey grass and shrub. Functional Plant Biology 36: 752-760.

Heinicke E, Kumar U, Munoz DG (1992). Quantitative dot-blot assay for proteins using enhanced chemiluminescence. Journal of Immunological Methods 152: 227-236.

Henzler T, Waterhouse RN, Smyth AJ, Carvajal M, Cooke DT, Schaffner AR, Steudle E, Clarkson DT (1999). Diurnal variations in hydraulic conductivity and root pressure can be correlated with the expression of putative aquaporins in the roots of Lotus japonicus. Planta 210: 50-60.

References

173

Herrera A, Escala M, Rengifo E (2009). Leaf anatomy changes related to physiological adaptations to flooding in Amazonian tree species. Brazilian Journal of Plant Physiology 21: 301-308.

Hoagland DR, Arnon DI (1950). The water culture method for growing plants without soil. Circular, University of califormia Agricultural Experiment Station, Berkely 347.

Hodge A (2006). Plastic plants and patchy soils. Journal of Experimental Botany 57: 401-411.

Hodge A (2010). Roots: The acquisition of water and nutrients from the heterogeneous soil environment. In: Lüttge UE, Beyschlag W, Büdel B, Francis D eds. Progress in Botany. Springer, Berlin / Heidelberg. pp307-337.

Hogg EH, Hurdle PA (1997). Sap flow in trembling aspen: implications for stomatal responses to vapor pressure deficit. Tree Physiology 17: 501-509.

Holbrook NM, Ahrens ET, Burns MJ, Zwieniecki MA (2001). In vivo observation of cavitation and embolism repair using magnetic resonance imaging. Plant Physiology 126: 27-31.

Horna V, Schuldt B, Brix S, Leuschner C (2011). Environment and tree size controlling stem sap flux in a perhumid tropical forest of Central Sulawesi, Indonesia. Annals of Forest Science 68: 1027-1038.

Horton JL, Clark JL (2001). Water table decline alters growth and survival of Salix gooddingii and Tamarix chinensis seedlings. Forest Ecology & Management 140: 239-247.

Horton JL, Kolb TE, Hart SC (2001a). Leaf gas exchange characteristics differ among Sonoran Desert riparian tree species. Tree Physiology 21: 233-241.

Horton JL, Kolb TE, Hart SC (2001b). Physiological response to groundwater depth varies among species and with river flow regulation. Ecological Applications 11: 1046-1059.

Horton JL, Kolb TE, Hart SC (2001c). Responses of riparian trees to interannual variation in ground water depth in a semi-arid river basin. Plant, Cell & Environment 24: 293-304.

Hose E, Steudle E, Hartung W (2000). Abscisic acid and hydraulic conductivity of maize roots: a study using cell- and root-pressure probes. Planta 211: 874-882.

Hoshika Y, Omasa K, Paoletti E (2011). Both ozone exposure and soil water stress are able to induce stomatal sluggishness. Environmental & Experimental Botany doi: 10.1016/j.envexpbot.2011.12.004.

Howell T, Member A, Dusek D (1995). Comparison of vapour-pressure-deficit calculation methods - Southern High Plains. Journal of Irrigation and Drainage Engineering 121: 191-198.

Huang B, Johnson JW, Nesmith S, Bridges DC (1994). Growth, physiological and anatomical responses of two wheat genotypes to waterlogging and nutrient supply. Journal of Experimental Botany 45: 193-202.

Hubbard RM, Bond BJ, Ryan MG (1999). Evidence that hydraulic conductance limits photosynthesis in old Pinus ponderosa trees. Tree Physiology 19: 165-172.

174

Hughes FMR, Harris T, Richards K, Pautou G, Hames AE, Barsoum N, Girel J, Peiry J-L, Foussadier R (1997). Woody riparian species response to different soil moisture conditions: Laboratory experiments on Alnus incana (L.) Moench. Global Ecology and Biogeography Letters 6: 247-256.

Hukin D, Cochard H, Dreyer E, Thiec DL, Bogeat-Triboulot MB (2005). Cavitation vulnerability in roots and shoots: does Populus euphratica Oliv., a poplar from arid areas of Central Asia, differ from other poplar species? Journal of Experimental Botany 56: 2003-2010.

Hultine KR, Bush SE, Ehleringer JR (2010). Ecophysiology of riparian cottonwood and willow before, during, and after two years of soil water removal. Ecological Applications 20: 347-361.

Hultine KR, Cable WL, Burgess SSO, Williams DG (2003a). Hydraulic redistribution by deep roots of a Chihuahuan Desert phreatophyte. Tree Physiology 23: 353-360.

Hultine KR, Scott RL, Cable WL, Goodrich DC, Williams DG (2004). Hydraulic redistribution by a dominant, warm-desert phreatophyte: seasonal patterns and response to precipitation pulses. Functional Ecology 18: 530-538.

Hultine KR, Williams DG, Burgess SSO, Keefer TO (2003b). Contrasting patterns of hydraulic redistribution in three desert phreatophytes. Oecologia 135: 167-175.

Hutchings MJ, John EA (2004). The effects of environmental heterogeneity on root growth and root/shoot partitioning. Annals of Botany 94: 1-8.

Hutley LB, O’Grady AP, Eamus D (2001). Monsoonal influences on evapotranspiration of savanna vegetation of northern Australia. Oecologia 126: 434-443.

Irvine J, Law BE, Kurpius MR, Anthoni PM, Moore D, Schwarz PA (2004). Age-related changes in ecosystem structure and function and effects on water and carbon exchange in ponderosa pine. Tree Physiology 24: 753-763.

Ishii A, Oshima T, Sato T, Nakasone K, Mori H, Kato C (2005). Analysis of hydrostatic pressure effects on transcription in Escherichia coli by DNA microarray procedure. Extremophiles 9: 65-73.

Ito H, Ito S, Matsui T, Marutani T (2006). Effect of fluvial and geomorphic disturbances on habitat segregation of tree species in a sedimentation-dominated riparian forest in warm-temperate mountainous region in southern Japan. Journal of Forest Research 11: 405-417.

Ivans CY, Leffler AJ, Spaulding U, Stark JM, Ryel RJ, Caldwell MM (2003). Root responses and nitrogen acquisition by Artemisia tridentata and Agropyron desertorum following small summer rainfall events. Oecologia 134: 317-324.

Jackson J, Kern S, Pinder A, Nowicki A, Daniel G (2009a). Resource Condition Report for Significant Western Australian Wetland: Wetlands of the Fortescue River System. Department of Environment and Conservation, Government of Western Australia, Perth.

Jackson MB, Attwood PA (1996). Roots of willow (Salix viminalis L.) show marked tolerance to oxygen shortage in flooded soils and in solution culture. Plant and Soil 187: 37-45.

References

175

Jackson RB, Caldwell MM (1993). Geostatistical patterns of soil heterogeneity around individual perennial plants. Journal of Ecology 81: 683-692.

Jackson RB, Jobbágy EG, Nosetto MD (2009b). Ecohydrology in a human-dominated landscape. Ecohydrology 2: 383-389.

Jang JY, Kim DG, Kim YO, Kim JS, Kang H (2004). An expression analysis of a gene family encoding plasma membrane aquaporins in response to abiotic stresses in Arabidopsis thaliana. Plant Molecular Biology 54: 713-725.

Javot H, Lauvergeat V, Santoni V, Martin-Laurent F, Guclu J, Vinh J, Heyes J, Franck KI, Schaffner AR, Bouchez D, Maurel C (2003). Role of a single aquaporin isoform in root water uptake. Plant Cell 15: 509-522.

Javot H, Maurel C (2002). The role of aquaporins in root water uptake. Annals of Botany 90: 301-313.

Johnson SL (2004). Geology and hydrogeology. In Van Vreeswyk AME, Payne AL, Leighton KA, Hennig P. An Inventory and Condition Survey of the Pilbara Region, Western Australia. Department of Agriculture, Government of Western Australia, Perth.

Johnson SL, Wright AH (2001). Central Pilbara groundwater study. Water and Rivers Commission, Government of Western Australia, Perth. Report HG 8.

Joslin JD, Wolfe MH, Hanson PJ (2000). Effects of altered water regimes on forest root systems. New Phytologist 147: 117-129.

Jury WA, Vaux HJ (2007). The emerging global water crisis: Managing scarcity and conflict between water users. In: Sparks DL ed. Advances in Agronomy. Academic Press. 1-76.

Justin SHFW, Armstrong W (1987). The anatomical characteristics of roots and plant response to soil flooding. New Phytologist 106: 465-495.

Kaldenhoff R, Ribas-Carbo M, Sans JF, Lovisolo C, Heckwolf M, Uehlein N (2008). Aquaporins and plant water balance. Plant, Cell & Environment 31: 658-666.

Kang S, Hu X, Jerie P, Zhang J (2003). The effects of partial rootzone drying on root, trunk sap flow and water balance in an irrigated pear (Pyrus communis L.) orchard. Journal of Hydrology 280: 192-206.

Kang S, Zhang J (2004). Controlled alternate partial root-zone irrigation: its physiological consequences and impact on water use efficiency. Journal of Experimental Botany 55: 2437-2446.

Kattge J, Díaz S, Lavorel S, Prentice IC, Leadley P, Bönisch G, Garnier E, Westoby M, Reich PB, Wright IJ, Cornelissen JHC, Violle C, Harrison SP, Van Bodegom PM, Reichstein M, Enquist BJ, Soudzilovskaia NA, Ackerly DD, Anand M, Atkin O, Bahn M, Baker TR, Baldocchi D, Bekker R, Blanco CC, Blonder B, Bond WJ, Bradstock R, Bunker DE, Casanoves F, Cavender-Bares J, Chambers JQ, Chapin III FS, Chave J, Coomes D, Cornwell WK, Craine JM, Dobrin BH, Duarte L, Durka W, Elser J, Esser G, Estiarte M, Fagan WF, Fang J, Fernández-Méndez F, Fidelis A, Finegan B, Flores O, Ford H, Frank D, Freschet GT, Fyllas NM, Gallagher RV, Green WA, Gutierrez AG, Hickler T, Higgins SI, Hodgson JG, Jalili A, Jansen S, Joly CA, Kerkhoff AJ, Kirkup D, Kitajima K, Kleyer M, Klotz S,

176

Knops JMH, Kramer K, Kühn I, Kurokawa H, Laughlin D, Lee TD, Leishman M, Lens F, Lenz T, Lewis SL, Lloyd J, Llusià J, Louault F, Ma S, Mahecha MD, Manning P, Massad T, Medlyn BE, Messier J, Moles AT, Müller SC, Nadrowski K, Naeem S, Niinemets Ü, Nöllert S, Nüske A, Ogaya R, Oleksyn J, Onipchenko VG, Onoda Y, Ordoñez J, Overbeck G, Ozinga WA, Patiño S, Paula S, Pausas JG, Peñuelas J, Phillips OL, Pillar V, Poorter H, Poorter L, Poschlod P, Prinzing A, Proulx R, Rammig A, Reinsch S, Reu B, Sack L, Salgado-Negret B, Sardans J, Shiodera S, Shipley B, Siefert A, Sosinski E, Soussana JF, Swaine E, Swenson N, Thompson K, Thornton P, Waldram M, Weiher E, White M, White S, Wright SJ, Yguel B, Zaehle S, Zanne AE, Wirth C (2011). TRY – a global database of plant traits. Global Change Biology 17: 2905-2935.

Kazda M, Schmid I (2009). Clustered distribution of tree roots and soil water exploitation. In: Lüttge U, Beyschlag W, Büdel B, Francis D eds. Progress in Botany. Springer, Berlin / Heidelberg. pp223-239.

Kelley G, O’Grady AP, Hutley LB, Eamus D (2007). A comparison of tree water use in two contiguous vegetation communities of the seasonally dry tropics of northern Australia: the importance of site water budget to tree hydraulics. Australian Journal of Botany 55: 700-708.

Kembel SW, Cahill JF (2005). Plant phenotypic plasticity belowground: a phylogenetic perspective on root foraging trade-offs. The American Naturalist 166: 216-230.

Khalil AAM, Grace J (1993). Does xylem sap ABA control the stomatal behaviour of water-stressed sycamore (Acer pseudoplatanus L.) seedlings? Journal of Experimental Botany 44: 1127-1134.

Kiley D, Schneider R (2005). Riparian roots through time, space and disturbance. Plant and Soil 269: 259-272.

Kim HK, Lee SJ (2010). Synchrotron X-ray imaging for nondestructive monitoring of sap flow dynamics through xylem vessel elements in rice leaves. New Phytologist 188: 1085-1098.

King SL, Allen JA, McCoy JW (1998). Long-term effects of a lock and dam and greentree reservoir management on a bottomland hardwood forest. Forest Ecology & Management 112: 213-226.

Kingsford RT (2000). Ecological impacts of dams, water diversions and river management on floodplain wetlands in Australia. Austral Ecology 25: 109-127.

Kingsford RT, Lemly AD, Thompson JR (2006). Impacts of dams, river mangement and diversions on desert rivers. In: Kingsford RT ed. Ecology of Desert Rivers. Cambridge University Press. pp203-247.

Kirkpatrick G, Dogramaci S (2010). Yandicoogina Water Balance; Pre and Post Mining Hydraulics and Hydrochemistry. Rio Tinto, Perth, Australia.

Knipfer T, Das D, Steudle E (2007). During measurements of root hydraulics with pressure probes, the contribution of unstirred layers is minimized in the pressure relaxation mode: comparison with pressure clamp and high-pressure flowmeter. Plant, Cell & Environment 30: 845-860.

References

177

Knipfer T, Fricke W (2010). Root pressure and a solute reflection coefficient close to unity exclude a purely apoplastic pathway of radial water transport in barley (Hordeum vulgare). New Phytologist 187: 159-170.

Kogawara S, Yamanoshita T, Norisada M, Masumori M, Kojima K (2006). Photosynthesis and photoassimilate transport during root hypoxia in Melaleuca cajuputi, a flood-tolerant species, and in Eucalyptus camaldulensis, a moderately flood-tolerant species. Tree Physiology 26: 1413-1423.

Kozlowski TT (1997). Responses of woody plants to flooding and salinity. Tree Physiology 17: 490. Online Monograph, DOI: 410.1093/treephys/1017.1097.1490.

Kranjcec J, Mahoney JM, Rood SB (1998). The responses of three riparian cottonwood species to water table decline. Forest Ecology & Management 110: 77-87.

Kroeger JH, Zerzour R, Geitmann A (2011). Regulator or driving force? The role of turgor pressure in oscillatory plant cell growth. PLoS ONE 6: e18549.

Kumagai T, Aoki S, Nagasawa H, Mabuchi T, Kubota K, Inoue S, Utsumi Y, Otsuki K (2005). Effects of tree-to-tree and radial variations on sap flow estimates of transpiration in Japanese cedar. Agricultural & Forest Meteorology 135: 110-116.

Kume T, Tsuruta K, Komatsu H, Kumagai T, Higashi N, Shinohara Y, Otsuki K (2010). Effects of sample size on sap flux-based stand-scale transpiration estimates. Tree Physiology 30: 129-138.

Kunert N, Schwendenmann L, Hölscher D (2010). Seasonal dynamics of tree sap flux and water use in nine species in Panamanian forest plantations. Agricultural & Forest Meteorology 150: 411-419.

Larned ST, Datry T, Arscott DB, Tockner K (2010). Emerging concepts in temporary-river ecology. Freshwater Biology 55: 717-738.

Lawlor DW (1973). Growth and water absorption of wheat with parts of the roots at different water potentials. New Phytologist 72: 297-305.

Lee JE, Oliveira RS, Dawson TE, Fung I (2005). Root functioning modifies seasonal climate. Proceedings of the National Academy of Sciences 102: 17576-17581.

Leighton KA (2004). Climate.In Van Vreeswyk AME, Payne AL, Leighton KA, Hennig P. An Inventory and Condition Survey of the Pilbara Region, Western Australia. Department of Agriculture, Government of Western Australia, Perth. 19-38.

Leiva MJ, Fernandez-Ales R (1998). Variability in seedling water status during drought within a Quercus ilex subsp. ballota population, and its relation to seedling morphology. Forest Ecology & Management 111: 147-156.

Leng C, Lin Y, Wahba G (2006). A note on the lasso and related procedures in model selection. Statistica Sinica 16: 1273-1284.

Li S, Pezeshki SR, Goodwin S, Shields FD (2004). Physiological responses of black willow (Salix nigra) cuttings to a range of soil moisture regimes. Photosynthetica 42: 585-590.

178

Lio A, Fukasawa H, Nose Y, Kakubari Y (2004). Stomatal closure induced by high vapor pressure deficit limited midday photosynthesis at the canopy top of Fagus crenata; Blume on Naeba mountain in Japan. Trees - Structure & Function 18: 510-517.

Lite SJ, Bagstad KJ, Stromberg JC (2005). Riparian plant species richness along lateral and longitudinal gradients of water stress and flood disturbance, San Pedro River, Arizona, USA. Journal of Arid Environments 63: 785-813.

Lite SJ, Stromberg JC (2005). Surface water and ground-water thresholds for maintaining Populus - Salix forests, San Pedro River, Arizona. Biological Conservation 125: 153-167.

Litvak E, McCarthy HR, Pataki DE (2011). Water relations of coast redwood planted in the semi-arid climate of southern California. Plant, Cell & Environment 34: 1384-1400.

Liu L, McDonald AJS, Stadenberg I, Davies WJ (2001). Stomatal and leaf growth responses to partial drying of root tips in willow. Tree Physiology 21: 765-770.

Lockhart C, Austin DF, Aumen NG (1999). Water level effects on growth of Melaleuca seedlings from Lake Okeechobee (Florida, USA) littoral zone. Environmental Management 23: 507-518.

Lockhart CS (1996). Aquatic heterophylly as a survival strategy in Melaleuca quinquenervia (Myrtaceae). Canadian Journal of Botany 74: 243-246.

López-Bernal Á, Alcántara E, Testi L, Villalobos FJ (2010). Spatial sap flow and xylem anatomical characteristics in olive trees under different irrigation regimes. Tree Physiology 30: 1536-1544.

Lopez-Zamora I, Comerford NB, Muchovej RM (2004). Root development and competitive ability of the invasive species Melaleuca quinquenervia (Cav.) S.T. Blake in the South Florida flatwoods. Plant and Soil 263: 239-247.

Loustau D, Berbigier P, Roumagnac P, Arruda-Pacheco C, David JS, Ferreira MI, Pereira JS, Tavares R (1996). Transpiration of a 64-year-old maritime pine stand in Portugal. 1. Seasonal course of water flux through maritime pine. Oecologia 107: 33–42.

Lubczynski M (2009). The hydrogeological role of trees in water-limited environments. Hydrogeology Journal 17: 247-259.

Ludwig F, Dawson TE, Kroon H, Berendse F, Prins HHT (2003). Hydraulic lift in Acacia tortilis trees on an East African savanna. Oecologia 134: 293–300.

Luu D-T, Maurel C (2005). Aquaporins in a challenging environment: molecular gears for adjusting plant water status. Plant, Cell & Environment 28: 85-96.

Lytle DA, Poff NL (2004). Adaptation to natural flow regimes. Trends in Ecology & Evolution 19: 94-100.

Macfarlane C, Hoffman M, Eamus D, Kerp N, Higginson S, McMurtrie R, Adams M (2007). Estimation of leaf area index in eucalypt forest using digital photography. Agricultural and Forest Meteorology 143: 176–188

Magnani F, Borghetti M (1995). Interpretation of seasonal changes of xylem embolism and plant hydraulic resistance in Fagus sylvatica. Plant, Cell & Environment 18: 689-696.

References

179

Magyar G, Kun Á, Oborny B, Stuefer JF (2007). Importance of plasticity and decision-making strategies for plant resource acquisition in spatio-temporally variable environments. New Phytologist 174: 182-193.

Mahdieh M, Mostajeran A, Horie T, Katsuhara M (2008). Drought stress alters water relations and expression of PIP-type aquaporin genes in Nicotiana tabacum plants. Plant & Cell Physiology 49: 801-813.

Mahoney J, Rood S (1998). Streamflow requirements for cottonwood seedling recruitment—An integrative model. Wetlands 18: 634-645.

Mahoney JM, Rood SB (1992). Response of a hybrid poplar to water table decline in different substrates. Forest Ecology & Management 54: 141-156.

Mariaux J-B, Bockel C, Salamini F, Bartels D (1998). Desiccation- and abscisic acid-responsive genes encoding major intrinsic proteins (MIPs) from the resurrection plant Craterostigma plantagineum. Plant Molecular Biology 38: 1089-1099.

Martin-Gorriz B, Egea G, Nortes PA, Baille A, González-Real MM, Ruiz-Salleres I, Verhoef A (2011). Effects of high temperature and vapour pressure deficit on net ecosystem exchange and energy balance of an irrigated orange orchard in a semi- arid climate (Southern Spain). ISHS Acta Horticulturae 922: 149-156.

Martı ́nez-Ramos M, Alvarez-Buylla ER (1998). How old are tropical rain forest trees? Trends in Plant Science 3: 400-405.

Martre P, Morillon R, Barrieu F, North GB, Nobel PS, Chrispeels MJ (2002). Plasma membrane aquaporins play a significant role during recovery from water deficit. Plant Physiology 130: 1-10.

Masini RJ (1988). Inland Waters of the Pilbara, Western Australia. Part 1. A Report of a Field Study Carried Out in March–April, 1983. Environmental Protection Authority Technical Series. Government of Western Australia, Perth.

Maurel C, Simonneau T, Sutka M (2010). The significance of roots as hydraulic rheostats. Journal of Experimental Botany 61: 3191-3198.

Maurel C, Verdoucq L, Luu D-T, Santoni V (2008). Plant aquaporins: Membrane channels with multiple integrated functions. Annual Review of Plant Biology 59: 595-624.

Mays L (2009). Integrated Urban Water Management : Arid and Semi-Arid Regions. Taylor & Francis, Leiden and UNESCO Publishing, Paris.

McCulloh KA, Sperry JS (2005). Patterns in hydraulic architecture and their implications for transport efficiency. Tree Physiology 25: 257-267.

McDowell N, Barnard H, Bond B, Hinckley T, Hubbard R, Ishii H, Köstner B, Magnani F, Marshall J, Meinzer F, Phillips N, Ryan M, Whitehead D (2002). The relationship between tree height and leaf area: sapwood area ratio. Oecologia 132: 12-20.

McDowell N, Pockman WT, Allen CD, Breshears DD, Cobb N, Kolb T, Plaut J, Sperry J, West A, Williams DG, Yepez EA (2008). Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought? New Phytologist 178: 719-739.

180

McDowell NG (2011). Mechanisms linking drought, hydraulics, carbon metabolism, and vegetation mortality. Plant Physiology 155: 1051-1059.

McElrone AJ, Bichler J, Pockman WT, Addington RN, Linder CR, Jackson RB (2007). Aquaporin-mediated changes in hydraulic conductivity of deep tree roots accessed via caves. Plant, Cell & Environment 30: 1411-1421.

McIntyre RES, Adams MA, Ford DJ, Grierson PF (2009). Rewetting and litter addition influence mineralisation and microbial communities in soils from a semi-arid intermittent stream. Soil Biology and Biochemistry 41: 92-101.

McJannet D (2008). Water table and transpiration dynamics in a seasonally inundated Melaleuca quinquenervia forest, north Queensland, Australia. Hydrological Processes 22: 3079-3090.

McKenzie NL, van Leeuwen S, Pinder AM (2009). Introduction to the Pilbara biodiversity survey, 2002–2007. In: George AS, McKenzie NL, Doughty P eds. A Biodiversity Survey of the Pilbara Region of Western Australia, 2002 – 2007. Records of the Western Australian Museum. pp3–89.

Meinzer FC (2003). Functional convergence in plant responses to the environment. Oecologia 134: 1-11.

Meinzer FC, Andrade JL, Goldstein G, Holbrook NM, Cavelier J, Jackson P (1997a). Control of transpiration from the upper canopy of a tropical forest: the role of stomatal, boundary layer and hydraulic architecture components. Plant, Cell & Environment 20: 1242-1252.

Meinzer FC, Bond BJ, Warren JM, Woodruff DR (2005). Does water transport scale universally with tree size? Functional Ecology 19: 558-565.

Meinzer FC, Brooks JR, Bucci S, Goldstein G, Scholz FG, Warren JM (2004a). Converging patterns of uptake and hydraulic redistribution of soil water in contrasting woody vegetation types. Tree Physiology 24: 919-928.

Meinzer FC, Goldstein G, Holbrook NM, Jackson P, Cavelier J (1993). Stomatal and environmental control of transpiration in a lowland tropical forest tree. Plant, Cell & Environment 16: 429-436.

Meinzer FC, Hinckley TM, Ceulemans R (1997b). Apparent responses of stomata to transpiration and humidity in a hybrid poplar canopy. Plant, Cell & Environment 20: 1301-1308.

Meinzer FC, James SA, Goldstein G (2004b). Dynamics of transpiration, sap flow and use of stored water in tropical forest canopy trees. Tree Physiology 24: 901-909.

Meinzer FC, James SA, Goldstein G, Woodruff D (2003). Whole-tree water transport scales with sapwood capacitance in tropical forest canopy trees. Plant, Cell & Environment.

Meinzer FC, Johnson DM, Lachenbruch B, McCulloh KA, Woodruff DR (2009). Xylem hydraulic safety margins in woody plants: coordination of stomatal control of xylem tension with hydraulic capacitance. Functional Ecology 23: 922-930.

References

181

Meinzer FC, Woodruff DR, Eissenstat DM, Lin HS, Adams TS, McCulloh KA (2013). Above- and belowground controls on water use by trees of different wood types in an eastern US deciduous forest. Tree Physiology 33: 345-356.

Merritt DM, Poff NL (2010). Shifting dominance of riparian Populus and Tamarix along gradients of flow alteration in western North American rivers. Ecological Applications 20: 135-152.

Millikin Ishikawa C, Bledsoe CS (2000). Seasonal and diurnal patterns of soil water potential in the rhizosphere of blue oaks: evidence for hydraulic lift. Oecologia 125: 459-465.

Mingo D, Theobald J, Bacon M, Davies W, Dodd I (2004). Biomass allocation in tomato (Lycopersicon esculentum) plants grown under partial rootzone drying: enhancement of root growth. Functional Plant Biology 31: 971–978.

Miranda I, Gominho J, Pereira H (2009). Variation of heartwood and sapwood in 18-year-old Eucalyptus globulus trees grown with different spacings. Trees - Structure & Function 23: 367-372.

Mitchell PJ, O'Grady AP, Tissue DT, White DA, Ottenschlaeger ML, Pinkard EA (2013). Drought response strategies define the relative contributions of hydraulic dysfunction and carbohydrate depletion during tree mortality. New Phytologist 197: 862-872.

Miyazawa Y, Tateishi M, Komatsu H, Kumagai To, Otsuki K (2011). Are measurements from excised leaves suitable for modeling diurnal patterns of gas exchange of intact leaves? Hydrological Processes 25: 2924-2930.

Mommer L, Visser EJW, van Ruijven J, de Caluwe H, Pierik R, de Kroon H (2011). Contrasting root behaviour in two grass species: a test of functionality in dynamic heterogeneous conditions. Plant and Soil 344: 347-360.

Moore GW, Cleverly JR, Owens MK (2008). Nocturnal transpiration in riparian Tamarix thickets authenticated by sap flux, eddy covariance and leaf gas exchange measurements. Tree Physiology 28: 521-528.

Motzer T, Munz N, Küppers M, Schmitt D, Anhuf D (2005). Stomatal conductance, transpiration and sap flow of tropical montane rain forest trees in the southern Ecuadorian Andes. Tree Physiology 25: 1283–1293.

Müller M (2009). Synchrotron radiation x-ray scattering techniques for studying the micro- and nanostructure of wood and their relation to the mechanical properties. Materials Science Forum 599: 107-125.

Myers R (1983). Site susceptibility to invasion by the exotic tree Melaleuca quinquenervia in southern Florida. Journal of Applied Ecology 20: 645–658.

Myster R (2009). Plant communities of Western Amazonia. The Botanical Review 75: 271-291.

Nadezhdina N, Čermák J, Ceulemans R (2002). Radial patterns of sap flow in woody stems of dominant and understory species: scaling errors associated with positioning of sensors. Tree Physiology 22: 907-918.

182

Naidu BP, Jones GP, Paleg LG, Poljakoff-Mayber A (1987). Proline analogues in Melaleuca Species: Response of Melaleuca lanceolata and M. uncinata to water stress and salinity. Australian Journal of Plant Physiology 14: 669-677.

Naidu BP, Paleg LG, Jones GP (2000). Accumulation of proline analogues and adaptation of Melaleuca species to diverse environments in Australia. Australian Journal of Botany 48: 611-620.

Naiman RJ, Bechtold JS, Drake DC, Latterell JJ, O'Keefe TC, Balian EV (2005). Origins, patterns, and importance of heterogeneity in riparian systems. In: Lovett G, Jones CG, Turner MG, Weathers KC eds. Ecosystem Function in Heterogeneous Landscapes. Springer, New York, USA. pp279.

Naiman RJ, Decamps H (1997). The ecology of interfaces: Riparian zones. Annual Review of Ecology and Systematics 28: 621-658.

Nakai A, Yurugi Y, Kisanuki H (2009). Growth responses of Salix gracilistyla cuttings to a range of substrate moisture and oxygen availability. Ecological Research 24: 1057-1065.

Nardini A, Tyree MT, Salleo S (2001). Xylem cavitation in the leaf of Prunus laurocerasus and its impact on leaf hydraulics. Plant Physiology 125: 1700-1709.

Neales TF, Masia A, Zhang J, Davies WJ (1989). The effects of partially drying part of the root system of Helianthus annuus on the abscisic acid content of the roots, xylem sap and leaves. Journal of Experimental Botany 40: 1113-1120.

Neatrour MA, Jones RH, Golladay SW (2007). Response of three floodplain tree species to spatial heterogeneity in soil oxygen and nutrients. Journal of Ecology 95: 1274–1283.

Neuhaus A, Turner DW, Colmer TD, Kuo J, Eastham J (2007). Drying half the root-zone of potted avocado (Persea americana Mill., cv. Hass) trees avoids the symptoms of water deficit that occur under complete root-zone drying. The Journal of Horticultural Science and Biotechnology 82: 679-689.

Neumann RB, Cardon ZG (2012). The magnitude of hydraulic redistribution by plant roots: a review and synthesis of empirical and modeling studies. New Phytologist 194: 337-352.

Newman BD, Wilcox BP, Archer SR, Breshears DD, Dahm CN, Duffy CJ, McDowell NG, Phillips FM, Scanlon BR, Vivoni ER (2006). Ecohydrology of water-limited environments: A scientific vision. Water Resources Research 42: W06302.

Niinemets Ü, Valladares F (2006). Tolerance to shade, drought, and waterlogging of temperate Northern Hemisphere trees and shrubs. Ecological Monographs 76: 521-547.

Niklas KJ, Spatz H-C (2010). Worldwide correlations of mechanical properties and green wood density. American Journal of Botany 97: 1587-1594.

Nobel PS, Cui MY (1992). Hydraulic conductances of the soil, the root soil air gap, and the root - changes for desert succulents in drying soil. Journal of Experimental Botany 43: 319-326.

North GB, Nobel PS (1998). Water uptake and structural plasticity along roots of a desert succulent during prolonged drought. Plant, Cell & Environment 21: 705–713.

References

183

North GB, Nobel PS (2000). Heterogeneity in water availability alters cellular development and hydraulic conductivity along roots of a desert succulent. Annals of Botany 85: 247-255.

Nuttle WK (2002). Is ecohydrology one idea or many? Hydrological Sciences Journal 47: 805-807.

O'Brien JJ, Orerrauer SF, Clark DB (2004). Whole tree xylem sap flow responses to multiple environmental variables in a wet tropical forest. Plant, Cell & Environment 27: 551-567.

O'Grady AP, Cook PG, Eamus D, Duguid A, Wischusen JDH, Fass T, Worldege D (2009). Convergence of tree water use within an arid-zone woodland. Oecologia 160: 643–655.

O'Grady AP, Eamus D, Cook PG, Lamontagne S (2006a). Comparative water use by the riparian trees Melaleuca argentea and Corymbia bella in the wet-dry tropics of northern Australia. Tree Physiology 26: 219-228.

O'Grady AP, Eamus D, Cook PG, Lamontagne S (2006b). Groundwater use by riparian vegetation in the wet–dry tropics of northern Australia. Australian Journal of Botany 54: 145-154.

O'Grady AP, Eamus D, Hutley LB (1999). Transpiration increases during the dry season: patterns of tree water use in eucalypt open-forests of northern Australia. Tree Physiology 19: 591-597.

O'Grady AP, Worledge D, Battaglia M (2008). Constraints on transpiration of Eucalyptus globulus in southern Tasmania, Australia. Agricultural & Forest Meteorology 148: 453-465.

Oliveira RS (2013). Can hydraulic traits be used to predict sensitivity of drought-prone forests to crown decline and tree mortality? Plant and Soil 364: 1-3.

Oliveira RS, Dawson TE, Burgess SSO, Nepstad DC (2005). Hydraulic redistribution in three Amazonian trees. Oecologia 145: 354-363.

Oren R, Phillips N, Ewers BE, Pataki DE, Megonigal JP (1999). Sap-flux-scaled transpiration responses to light, vapor pressure deficit, and leaf area reduction in a flooded Taxodium distichum forest. Tree Physiology 19: 337-347.

Oren R, Sperry JS, Ewers BE, Pataki DE, Phillips N, Megonigal JP (2001). Sensitivity of mean canopy stomatal conductance to vapor pressure deficit in a flooded Taxodium distichum L. forest: hydraulic and non-hydraulic effects. Oecologia 126: 21-29.

Osunubi O, Davies W (1981). Root growth and water relations of oak and birch seedlings. Oecologia 51: 343-350.

Page GF, Liu J, Grierson PF (2011). Three-dimensional xylem networks and phyllode properties of co-occurring Acacia. Plant, Cell & Environment 34: 2149-2158.

Pallardy S, Rhoads J (1993). Morphological adaptations to drought in seedlings of deciduous angiosperms. Canadian Journal of Forest Research 23: 1766- 1774.

Pang J, Wang Y, Lambers H, Tibbett M, Siddique KHM, Ryan MH (2013). Commensalism in an agroecosystem: hydraulic redistribution by deep-rooted legumes

184

improves survival of a droughted shallow-rooted legume companion. Physiologia Plantarum: DOI: 10.1111/ppl.12020.

Parolin P, Wittmann F (2010). Struggle in the flood: tree responses to flooding stress in four tropical floodplain systems. AoB Plants doi:10.1093/aobpla/plq003.

Peñuelas J, Filella I (2003). Deuterium labelling of roots provides evidence of deep water access and hydraulic lift by Pinus nigra in a Mediterranean forest of NE Spain. Environmental & Experimental Botany 49: 201-208.

Pettit NE, Froend RH (2001). Variability in flood disturbance and the impact on riparian tree recruitment in two contrasting river systems. Wetlands Ecology and Management 9: 13-25.

Pettit NE, Naiman RJ (2005). Flood-deposited wood debris and its contribution to heterogeneity and regeneration in a semi-arid riparian landscape. Oecologia 145: 434-444.

Petts G (2005). Water in history. In: Lehr JH, Keeley J eds. Water Encyclopedia. John Wiley & Sons. 726–731.

Pezeshki S, Anderson P (1996). Responses of three bottomland species with different flood tolerance capabilities to various flooding regimes. Wetlands Ecology and Management 4: 245-256.

Pezeshki S, Anderson P, Shields FJ (1998). Effects of soil moisture regimes on growth and survival of black willow (Salix nigra) posts (cuttings). Wetlands 18: 460-470.

Pezeshki SR (2001). Wetland plant responses to soil flooding. Environmental & Experimental Botany 46: 299-312.

Pfautsch S, Bleby TM, Rennenberg H, Adams MA (2010). Sap flow measurements reveal influence of temperature and stand structure on water use of Eucalyptus regnans forests. Forest Ecology & Management 259: 1190-1199.

Pfautsch S, Keitel C, Turnbull TL, Braimbridge MJ, Wright TE, Simpson RR, O’Brien JA, Adams MA (2011). Diurnal patterns of water use in Eucalyptus victrix indicate pronounced desiccation–rehydration cycles despite unlimited water supply. Tree Physiology 31: 1041-1051.

Phillips NG, Lewis JD, Logan BA, Tissue DT (2010). Inter- and intra-specific variation in nocturnal water transport in Eucalyptus. Tree Physiology 30: 586-596.

Phillips NG, Ryan MG, Bond BJ, McDowell NG, Hinckley TM, Čermák J (2003). Reliance on stored water increases with tree size in three species in the Pacific Northwest. Tree Physiology 23: 237-245.

Phillips NG, Scholz FG, Bucci SJ, Goldstein G, Meinzer FC (2009). Using branch and basal trunk sap flow measurements to estimate whole-plant water capacitance: Comment on Burgess and Dawson (2008). Plant and Soil 315: 315-324.

Pimenta J, Bianchini E, Medri M (1998). Adaptations to flooding by tropical trees: morphological and anatomical modifications. Oecologia Brasiliensis 6: 157-176.

Pinheiro C, Chaves MM Photosynthesis and drought: can we make metabolic connections from available data? Journal of Experimental Botany 62: 869-882.

References

185

Pockman WT, Sperry JS (2000). Vulnerability to xylem cavitation and the distribution of Sonoran Desert vegetation. American Journal of Botany 87: 1287-1299.

Poff NL, Olden JD, Merritt DM, Pepin DM (2007). Homogenization of regional river dynamics by dams and global biodiversity implications. Proceedings of the National Academy of Sciences 104: 5732-5737.

Poff NL, Zimmerman JKH (2010). Ecological responses to altered flow regimes: a literature review to inform the science and management of environmental flows. Freshwater Biology 55: 194-205.

Polasky S, Carpenter SR, Folke C, Keeler B (2011). Decision-making under great uncertainty: environmental management in an era of global change. Trends in Ecology & Evolution 26: 398-404.

Postaire O, Tournaire-Roux C, Grondin A, Boursiac Y, Morillon R, Schaffner AR, Maurel C (2010). A PIP1 aquaporin contributes to hydrostatic pressure-induced water transport in both the root and rosette of Arabidopsis. Plant Physiology 152: 1418-1430.

Powles JE, Buckley TN, Nicotra AB, Farquhar GD (2006). Dynamics of stomatal water relations following leaf excision. Plant, Cell & Environment 29: 981-992.

Pregitzer KS, Hendrick RL, Fogel R (1993). The demography of fine roots in response to patches of water and nitrogen. New Phytologist 125: 575-580.

Prieto I, Armas C, Pugnaire FI (2012). Water release through plant roots: new insights into its consequences at the plant and ecosystem level. New Phytologist 193: 830-841.

Prieto I, Kikvidze Z, Pugnaire F (2010). Hydraulic lift: soil processes and transpiration in the Mediterranean leguminous shrub Retama sphaerocarpa (L.) Boiss. Plant and Soil 329: 447-456.

R Development Core Team (2010). R: A language and environment for statistical computing R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.

Ren D, Henderson-Sellers A (2006). An analytical hydrological model for the study of scaling issues in land surface modeling. Earth Interactions 10: 20.

Ren D, Leslie LM, Karoly DJ (2008). Sensitivity of an ecological model to soil moisture simulations from two different hydrological models. Meteorology and Atmospheric Physics 100: 87-99.

Richards JH, Caldwell MM (1987). Hydraulic lift: Substantial nocturnal water transport between soil layers by Artemisia tridentata roots. Oecologia 73: 486-489.

Ridolfi L, D'Odorico P, Laio F (2006). Effect of vegetation-water table feedbacks on the stability and resilience of plant ecosystems. Water Resources Research 42: W01201.

Rodriguez-Iturbe I (2000). Ecohydrology: A hydrologic perspective of climate-soil-vegetation dynamics. Water Resources Research 36: 3-9.

Rood SB, Braatne JH, Hughes FMR (2003). Ecophysiology of riparian cottonwoods: stream flow dependency, water relations and restoration. Tree Physiology 23: 1113-1124.

Rood SB, Patiño S, Coombs K, Tyree MT (2000). Branch sacrifice: cavitation-associated drought adaptation of riparian cottonwoods. Trees - Structure & Function 14: 248-257.

186

Ryel R, Caldwell M, Yoder C, Or D, Leffler A (2002). Hydraulic redistribution in a stand of Artemisia tridentata: evaluation of benefits to transpiration assessed with a simulation model. Oecologia 130: 173-184.

Ryel RJ, Ivans CY, Peek MS, Leffler AJ (2008). Functional differences in soil water pools: a new perspective on plant water use in water-limited ecosystems. In: Lüttge U, Beyschlag W, Murata J eds. Progress in Botany. Springer, Berlin / Heidelberg. 397–422.

Sabo JL, Sponseller R, Dixon M, Gade K, Harms T, Heffernan J, Jani A, Katz G, Soykan C, Watts J, Welter J (2005). Riparian zones increase regional species richness by harboring different, not more, species. Ecology 86: 56-62.

Sala A, Piper F, Hoch G (2010). Physiological mechanisms of drought-induced tree mortality are far from being resolved. New Phytologist 186: 274-281.

Salleo S, Nardini A, Lo Gullo MA, Ghirardelli LA (2002). Changes in stem and leaf hydraulics preceding leaf shedding in Castanea Sativa L. Biologia Plantarum 45: 227-234.

Santiago LS, Mulkey SS (2003). A test of gas exchange measurements on excised canopy branches of ten tropical tree species. Photosynthetica 41: 343-347.

Savage JA, Cavender-Bares JM (2011). Contrasting drought survival strategies of sympatric willows (genus: Salix): consequences for coexistence and habitat specialization. Tree Physiology 31: 604-614.

Saveyn A, Steppe K, Lemeur R (2008). Spatial variability of xylem sap flow in mature beech (Fagus sylvatica) and its diurnal dynamics in relation to microclimate. Botany 86: 1440-1448.

Scanlon BR, Keese KE, Flint AL, Flint LE, Gaye CB, Edmunds WM, Simmers I (2006). Global synthesis of groundwater recharge in semiarid and arid regions. Hydrological Processes 20: 3335–3370.

Schaeffer SM, Williams DG, Goodrich DC (2000). Transpiration of cottonwood/willow forest estimated from sap flux. Agricultural & Forest Meteorology 105: 257-270.

Schmandt J (2002). Bi-national water issues in the Rio Grande/Rı ́o Bravo basin. Water Policy 4: 137-155.

Scholz FG, Bucci SJ, Goldstein G, Meinzer FC, Franco AC (2002). Hydraulic redistribution of soil water by neotropical savanna trees. Tree Physiology 22: 603-612.

Scholz FG, Bucci SJ, Goldstein G, Meinzer FC, Franco AC, Miralles-Wilhelm F (2007). Biophysical properties and functional significance of stem water storage tissues in Neotropical savanna trees. Plant, Cell & Environment 30: 236-248.

Scholz FG, Phillips NG, Bucci SJ, Meinzer FC, Goldstein G (2011). Hydraulic capacitance: Biophysics and functional significance of internal water sources in relation to tree size In: Meinzer FC, Lachenbruch B, Dawson TE eds. Size- and Age-Related Changes in Tree Structure and Function. Springer Netherlands. 341-361.

Schwinning S, Sala OE (2004). Hierarchy of responses to resource pulses in arid and semi-arid ecosystems. Oecologia 141: 211-220.

References

187

Scott ML, Auble GT, Friedman JM (1997). Flood dependency of cottonwood establishment along the Missouri River, Montana, USA. Ecological Applications 7: 677-690.

Scott ML, Shafroth PB, Auble GT (1999). Responses of riparian cottonwoods to alluvial water table declines. Environmental Management 23: 347-358.

Scott RL, Cable WL, Hultine KR (2008). The ecohydrologic significance of hydraulic redistribution in a semiarid savanna. Water Resources Research 44: W02440.

Scott RL, Edwards EA, Shuttleworth WJ, Huxman TE, Watts C, Goodrich DC (2004). Interannual and seasonal variation in fluxes of water and carbon dioxide from a riparian woodland ecosystem. Agricultural & Forest Meteorology 122: 65-84.

Scott RL, Shuttleworth WJ, Goodrich DC, Maddock T (2000). The water use of two dominant vegetation communities in a semiarid riparian ecosystem. Agricultural & Forest Meteorology 105: 241-256.

Segelquist CA, Scott ML, Auble GT (1993). Establishment of Populus deltoides under simulated alluvial groundwater declines. American Midland Naturalist 130: 274-285.

Sena Gomes AR, Kozlowski TT (1980). Responses of Melaleuca quinquenervia seedlings to flooding. Physiologia Plantarum 49: 373-377.

Sheldon F, Bunn SE, Hughes JM, Arthington AH, Balcombe SR, Fellows CS (2010). Ecological roles and threats to aquatic refugia in arid landscapes: dryland river waterholes. Marine and Freshwater Research 61: 885-895.

Shen Y, Chen Y (2010). Global perspective on hydrology, water balance, and water resources management in arid basins. Hydrological Processes 24: 129-135.

Shinohara Y, Tsuruta K, Ogura A, Noto F, Komatsu H, Otsuki K, Maruyama T (2013). Azimuthal and radial variations in sap flux density and effects on stand-scale transpiration estimates in a Japanese cedar forest. Tree Physiology.

Siefritz F, Tyree MT, Lovisolo C, Schubert A, Kaldenhoff R (2002). PIP1 plasma membrane aquaporins in tobacco: From cellular effects to function in plants. Plant Cell 14: 869-876.

Singer MB, Stella JC, Dufour S, Piégay H, Wilson RJS, Johnstone L (2012). Contrasting water-uptake and growth responses to drought in co-occurring riparian tree species. Ecohydrology 6: 402-412.

Small EE, McConnell JR (2008). Comparison of soil moisture and meteorological controls on pine and spruce transpiration. Ecohydrology 1: 205-214.

Smith SD, Wellington AB, Nachlinger JL, Fox CA (1991). Functional responses of riparian vegetation to streamflow diversion in the eastern Sierra Nevada. Ecological Applications 1: 89-97.

Snyder KA, Williams DG (2007). Root allocation and water uptake patterns in riparian tree saplings: Responses to irrigation and defoliation. Forest Ecology & Management 246: 222-231.

Sojka RE (1988). Measurement of root porosity (volume of root air space). Environmental & Experimental Botany 28: 275-280.

188

Sperry JS, Adler FR, Campbell GS, Comstock JP (1998). Limitation of plant water use by rhizosphere and xylem conductance: results from a model. Plant, Cell & Environment 21: 347-359.

Sperry JS, Hacke UG, Oren R, Comstock JP (2002). Water deficits and hydraulic limits to leaf water supply. Plant, Cell & Environment 25: 251-263.

Sperry JS, Ikeda T (1997). Xylem cavitation in roots and stems of Douglas-fir and white fir. Tree Physiology 17: 275-280.

Sperry JS, Meinzer FC, McCulloh KA (2008). Safety and efficiency conflicts in hydraulic architecture: scaling from tissues to trees. Plant, Cell & Environment 31: 632–645.

Sperry JS, Stiller V, Hacke UG (2003). Xylem hydraulics and the soil-plant-atmosphere continuum. Agronomy Journal 95: 1362-1370.

Stella JC, Battles JJ (2010). How do riparian woody seedlings survive seasonal drought? Oecologia 164: 579–590.

Steudle E (1994). Water transport across roots. Plant and Soil 167: 79-90.

Steudle E (2000). Water uptake by plant roots: an integration of views. Plant and Soil 226: 45–56.

Steudle E, Frensch J (1996). Water transport in plants: Role of the apoplast. Plant and Soil 187: 67-79.

Steudle E, Peterson CA (1998). How does water get through roots? Journal of Experimental Botany 49: 775-788.

Stromberg JC (2001). Restoration of riparian vegetation in the south-western United States: importance of flow regimes and fluvial dynamism. Journal of Arid Environments 49: 17-34.

Stromberg JC, Lite SJ, Dixon MD (2010a). Effects of stream flow patterns on riparian vegetation of a semiarid river: Implications for a changing climate. River Research and Applications 26: 712-729.

Stromberg JC, Tluczek MGF, Hazelton AF, Ajami H (2010b). A century of riparian forest expansion following extreme disturbance: Spatio-temporal change in Populus/Salix/Tamarix forests along the Upper San Pedro River, Arizona, USA. Forest Ecology & Management 259: 1181-1189.

Tateishi M, Kumagai To, Suyama Y, Hiura T (2010). Differences in transpiration characteristics of Japanese beech trees, Fagus crenata, in Japan. Tree Physiology 30: 748-760.

Tateishi M, Kumagai To, Utsumi Y, Umebayashi T, Shiiba Y, Inoue K, Kaji K, Cho K, Otsuki K (2008). Spatial variations in xylem sap flux density in evergreen oak trees with radial-porous wood: comparisons with anatomical observations. Trees - Structure & Function 22: 23-30.

Thomas DS, Eamus D, Bell D (1999). Optimization theory of stomatal behaviour II. Stomatal responses of several tree species of north Australia to changes in light, soil and atmospheric water content and temperature. Journal of Experimental Botany 50: 393-400.

References

189

Tiegs S, O'leary J, Pohl M, Munill C (2005). Flood disturbance and riparian species diversity on the Colorado River Delta. Biodiversity & Conservation 14: 1175-1194.

Tognetti R, d'Andria R, Morelli G, Calandrelli D, Fragnito F (2004). Irrigation effects on daily and seasonal variations of trunk sap flow and leaf water relations in olive trees. Plant and Soil 263: 249-264.

Tournaire-Roux C, Sutka M, Javot H, Gout E, Gerbeau P, Luu D-T, Bligny R, Maurel C (2003). Cytosolic pH regulates root water transport during anoxic stress through gating of aquaporins. Nature 425: 393-397.

Tsuruta K, Kume T, Komatsu H, Higashi N, Umebayashi T, Kumagai To, Otsuki K (2010). Azimuthal variations of sap flux density within Japanese cypress xylem trunks and their effects on tree transpiration estimates. Journal of Forest Research 15: 398-403.

Tyree M (2003). Hydraulic limits on tree performance: transpiration, carbon gain and growth of trees. Trees - Structure & Function 17: 95-100.

Tyree MT, Kolb KJ, Rood SB, Patiño S (1994). Vulnerability to drought-induced cavitation of riparian cottonwoods in Alberta: a possible factor in the decline of the ecosystem? Tree Physiology 14: 455-466.

Tyree MT, Patiño S, Bennink J, Alexander J (1995). Dynamic measurements of roots hydraulic conductance using a high-pressure flowmeter in the laboratory and field. Journal of Experimental Botany 46: 83-94.

Van de Wal BAE, Guyot A, Lovelock CE, Lockington DA, Steppe K (2013). Influence of spatial variation in sap flux density on estimates of whole-tree water use in Avicennia marina. Acta Horticulturae 991: 101-106.

Van Nieuwstadt MGL, Sheil D (2005). Drought, fire and tree survival in a Borneo rain forest, East Kalimantan, Indonesia. Journal of Ecology 93: 191-201.

Van Splunder I, Voesenek LACJ, De Vries XJA, Blom CWPM, Coops H (1996). Morphological responses of seedlings of four species of Salicaceae to drought. Canadian Journal of Botany 74: 1988-1995.

Van Vreeswyk AME, Payne AL, Leighton KA, Hennig P (2004). An Inventory and Condition Survey of the Pilbara Region, Western Australia. Department of Agriculture, Government of Western Australia, Perth.

Vandeleur RK, Mayo G, Shelden MC, Gilliham M, Kaiser BN, Tyerman SD (2009). The role of plasma membrane intrinsic protein aquaporins in water transport through roots: Diurnal and drought stress responses reveal different strategies between isohydric and anisohydric cultivars of grapevine. Plant Physiology 149: 445-460.

Vertessy RA, Benyon RG, O’Sullivan SK, Gribben. PR (1995). Relationship between stem diameter, sapwood area, leaf area and transpiration in a young mountain ash forest. Tree Physiology 15: 559-568.

Vertessy RA, Hatton TJ, Reece P, O'Sullivan SK, Benyon RG (1997). Estimating stand water use of large mountain ash trees and validation of the sap flow measurement technique. Tree Physiology 17: 747-756.

Voesenek LACJ, Benschop JJ, Bou J, Cox MCH, Groeneveld HW, Millenaar FF, Vreeburg RAM, Peeters AJM (2003). Interactions between plant hormones regulate

190

submergence-induced shoot elongation in the flooding-tolerant dicot Rumex palustris. Annals of Botany 91: 205-211.

Vysotskaya LB, Arkhipova TN, Timergalina LN, Dedov AV, Veselov SY, Kudoyarova GR (2004). Effect of partial root excision on transpiration, root hydraulic conductance and leaf growth in wheat seedlings. Plant Physiology & Biochemistry 42: 251-255.

Wakrim R, Wahbi S, Tahi H, Aganchich B, Serraj R (2005). Comparative effects of partial root drying (PRD) and regulated deficit irrigation (RDI) on water relations and water use efficiency in common bean (Phaseolus vulgaris L.). Agriculture, Ecosystems & Environment 106: 275-287.

Walker KF (2006). Serial weirs, cumulative effects: the Lower River Murray, Australia. In: Kingsford RT ed. Ecology of Desert Rivers. Cambridge University Press. pp248-279.

Wan X, Steudle E, Hartung W (2004). Gating of water channels (aquaporins) in cortical cells of young corn roots by mechanical stimuli (pressure pulses): effects of ABA and of HgCl2. Journal of Experimental Botany 55: 411-422.

Warren JM, Meinzer FC, Brooks JR, Domec J-C, Coulombe R (2007). Hydraulic redistribution of soil water in two old-growth coniferous forests: quantifying patterns and controls. New Phytologist 173: 753-765.

Weih M, Bonosi L, Ghelardini L, Rönnberg-Wästljung AC (2011). Optimizing nitrogen economy under drought: increased leaf nitrogen is an acclimation to water stress in willow (Salix spp.). Annals of Botany 108: 1347-1353.

Wheeler JK, Huggett BA, Tofte AN, Rockwell FE, Holbrook NM (2013). Cutting xylem under tension or supersaturated with gas can generate PLC and the appearance of rapid recovery from embolism. Plant, Cell & Environment 36: 1938-1949.

Wheeler JK, Sperry JS, Hacke UG, Hoang N (2005). Inter-vessel pitting and cavitation in woody Rosaceae and other vesselled plants: a basis for a safety versus efficiency trade-off in xylem transport. Plant, Cell & Environment 28: 800-812.

Williams CA, Cooper DJ (2005). Mechanisms of riparian cottonwood decline along regulated rivers. Ecosystems 8: 382-395.

Williams M, Bond BJ, Ryan MG (2001). Evaluating different soil and plant hydraulic constraints on tree function using a model and sap flow data from ponderosa pine. Plant, Cell & Environment 24: 679-690.

Windt CW, Vergeldt FJ, De Jager PA, Van As H (2006). MRI of long-distance water transport: a comparison of the phloem and xylem flow characteristics and dynamics in poplar, castor bean, tomato and tobacco. Plant, Cell & Environment 29: 1715-1729.

Woodruff DR, Meinzer FC, McCulloh KA (2010). Height-related trends in stomatal sensitivity to leaf-to-air vapour pressure deficit in a tall conifer. Journal of Experimental Botany 61: 203-210.

Wullschleger SD, Meinzer FC, Vertessy RA (1998). A review of whole-plant water use studies in tree. Tree Physiology 18: 499-512.

Xu Z, Zhou G (2011). Responses of photosynthetic capacity to soil moisture gradient in perennial rhizome grass and perennial bunchgrass. BMC Plant Biology 11: 21. DOI: 10.1186/1471-2229-1111-1121.

References

191

Yamanoshita T, Masumori M, Yagi H, Kojima K (2005). Effects of flooding on downstream processes of glycolysis and fermentation in roots of Melaleuca cajuputi seedlings. Journal of Forest Research 10: 199–204.

Yin D, Chen S, Chen F, Guan Z, Fang W (2010). Morpho-anatomical and physiological responses of two Dendranthema species to waterlogging. Environmental & Experimental Botany 68: 122-130.

Young JP, Horton RF (1985). Heterophylly in Ranunculus flabellaris Raf.: The effect of abscisic acid. Annals of Botany 55: 899-902.

Young WJ, Kingsford RT (2006). Flow variability in large unregulated dryland rivers. In: Kingsford RT ed. Ecology of Desert Rivers. Cambridge University Press. pp11-46.

Yu T, Feng Q, Si J, Xi H, Li Z, Chen A (2013). Hydraulic redistribution of soil water by roots of two desert riparian phreatophytes in northwest China’s extremely arid region. Plant and Soil: 1-12.

Zang D, Beadle CL, White DA (1996). Variation of sapflow velocity in Eucalyptus globulus with position in sapwood and use of a correction coefficient. Tree Physiology 16: 697-703.

Zanne AE, Westoby M, Falster DS, Ackerly DD, Loarie SR, Arnold SEJ, Coomes DA (2010). Angiosperm wood structure: Global patterns in vessel anatomy and their relation to wood density and potential conductivity. American Journal of Botany 97: 207-215.

Zekri M, Parsons LR (1990). Response of split-root sour orange seedlings to NaCl and polyethylene glycol stresses. Journal of Experimental Botany 41: 35-40.

Zeppel M, Eamus D (2008). Coordination of leaf area, sapwood area and canopy conductance leads to species convergence of tree water use in a remnant evergreen woodland. Australian Journal of Botany 56: 97-108.

Zeppel M, Macinnis-Ng C, Ford C, Eamus D (2008). The response of sap flow to pulses of rain in a temperate Australian woodland. Plant and Soil 305: 121-130.

Zeppel MJB, Murray BR, Barton C, Eamus D (2004). Seasonal responses of xylem sap velocity to VPD and solar radiation during drought in a stand of native trees in temperate Australia. Functional Plant Biology 31: 461-470.

Zhang J, Schurr U, Davies WJ (1987). Control of stomatal behaviour by abscisic acid which apparently originates in the roots. Journal of Experimental Botany 38: 1174-1181.

Zheng C, Wang Q (2012). Water-use response to climate factors at whole tree and branch scale for a dominant desert species in central Asia: Haloxylon ammodendron. Ecohydrology: 10.1002/eco.1321.

Zheng Z, Wang G (2007). Modeling the dynamic root water uptake and its hydrological impact at the Reserva Jaru site in Amazonia. Journal of Geophysical Research 112: G04012.

Zhou Z, Shangguan Z (2007). Vertical distribution of fine roots in relation to soil factors in Pinus tabulaeformis Carr. forest of the Loess Plateau of China. Plant and Soil 291: 119-129.

192

Zweifel R, Häsler R (2001). Dynamics of water storage in mature subalpine Picea abies: temporal and spatial patterns of change in stem radius. Tree Physiology 21: 561-569.

193

APPENDIX A: Verification of methods used to measure stomatal conductance in the

mid-canopy of mature trees

Introduction

Stomatal conductance (gs) was measured throughout this thesis using a leaf porometer.

When conducting field measurements on mature trees, it was not possible to access leaves

within the mid-canopy directly, so the viability of taking measurements from severed

foliage was tested, to determine whether leaves sampled from the mid-canopy with pole

clippers could provide a reasonable estimate of intact gs. While measurements from severed

branches can be acceptable, the response of stomata to excision varies among plant species,

and the utility of such measurements should therefore be tested for each species (Santiago

& Mulkey, 2003; Miyazawa et al., 2011). In addition, once M. argentea leaves are severed

from the canopy it can be difficult to identify the abaxial and adaxial surfaces, since the

appearance of leaves is very similar on each surface, and mature leaves frequently attach to

the stem in an almost vertical orientation. Therefore, differences in gs and stomatal density

between the two leaf surfaces were also tested, to assess the importance of standardising or

accounting for leaf surface during measurement of gs.

Materials & methods

Testing took place on six seedlings collected from Weeli Wolli Creek, upstream from

Marillana Creek, and growing in pots in a glasshouse at the University of Western

Australia. The gs measurements were taken with a leaf porometer (Decagon Devices,

Pullman, WA, USA), at midday, on sunlit, young, mature leaves, with air temperatures

recorded at the leaf surfaces during measurements of 28–32 °C. Stomatal density was

measured from leaf surface images taken through a microscope (method detailed in

Chapter 6, example image in Appendix D).

Results & Discussion

Figure A-1 shows the comparison between repeated measurements taken on leaves attached

to the plants (control), on excised single leaves, and on leaves attached to excised

branchlets. Stomatal conductance declined slightly with repeated measurement in the

control leaves. Leaves on excised branchlets gave similar values to the control, when

194

measurements were conducted within minutes of severing from the plant. However, when

single leaves were severed, their gs increased during the five minutes after severing that was

examined in this experiment. A transient increase in gs (the ‘wrong-way’ response) is

commonly observed following perturbations of hydraulic supply or demand, including leaf

severing, and is most likely due to a decline in epidermal and guard cell turgor passively

causing a widening of the stomatal aperture (Buckley, 2005; Powles et al., 2006). Following

the transient ‘wrong-way’ response, the gs then declines in the expected ‘right-way’

response, attributed to active guard cell closure through osmo-regulation. Severing

individual leaves in our test experiment clearly induced the stomatal ‘wrong-way’ response,

and the duration of our experiment was probably too short to observe the eventual decline

in gs, which may not occur until more than 10 minutes after severing (e.g. Hoshika et al.,

2011). Severing of branchlets, however, appears to have buffered the effect on stomata,

yielding gs values similar to attached leaves.

-200

-150

-100

-50

0

50

100

150

200control branchlet excised leaf excised

Chan

ge in

gs s

ince

tim

e ze

ro (m

mol

m-2

s-1)

Time (minutes)

1:00 2:00 3:00 4:00 5:00 6:000:00

Figure A-1: Stomatal conductance (gs) of leaves measured repeatedly at time intervals,

while either remaining attached to the plant (control), or after severing of from the plant at

time zero. Leaves were severed either at the petiole (leaf excised) or as a branchlet (branchlet excised). Values of gs at each time point were subtracted from the pre-severing value for that leaf (for severed treatments), or from the first measurement made for that leaf (control treat-ment). Values are mean ± standard error of four to six leaves, each leaf from a different plant.

Appendices

195

The gs and stomatal density were slightly higher on the adaxial than the abaxial leaf

surfaces in four out of the five leaves tested. However the differences were small; much

smaller than the differences observed among the individual ‘equivalent’ leaves, and indeed

non-significant in paired t-tests (p > 0.05; Table A-1).

Table A-1: Difference in (a) stomatal conductance (gs; mmol m-2 s-1) and (b) stomatal density (stomata mm-2) between adaxial and abaxial leaf surfaces of Melaleuca argentea seedlings.

(a) gs

Leaf*

1 2 3 4 5

adaxial 373.2 484.8 313.0 287.9 432.9 abaxial 244.0 483.9 392.7 284.6 347.8

difference 129.2 0.9 -79.7 3.3 85.1

(b) stomatal density 1 2 3 4 5 adaxial 444.3 308.9 499.0 355.1 360.5 abaxial 502.5 305.2 473.6 354.5 356.6

difference -58.2 3.7 25.4 0.6 3.9 *Each leaf was from a different plant, and a different set of leaves were used in parts (a) and (b).

Conclusions

All field measurements of gs presented in this thesis were therefore conducted by severing

branchlets from the canopy, and completing the measurement as quickly as possible after

cutting; generally within 60 seconds. Measurements were performed on either side of a leaf,

which aided in minimising the delay between cutting and measurement, since time was not

spent attempting to identify the surfaces.

196

APPENDIX B: Supplementary information for Chapter Five

The contents of this appendix were included as Online Supporting Information with the

published version of Chapter 5.

Table B-1: Moisture release properties of the river sand used in this study. Values are means ± standard error of three replicates.

Applied pressure (MPa) Soil water content (% v/v)

0.001 33.9 ± 0.4 0.005 21.0 ± 0.6 0.01 1.52 ± 0.09 0.03 1.44 ± 0.04 0.1 1.32 ± 0.06 1.5 1.1 ± 0.03

0

20

40

60

80

100

SWC

(% fi

eld

capa

city

)

Control PRD

CRD PRD-RW

0 2 4 6 8 10 12 14Time after treatment (days)

Figure B-1: Water content of the soil

compartments during treatment.

Plants were subjected to partial root zone drying (PRD), PRD with restricted water (PRD-RW), complete root drying (CRD) or remained fully watered as controls. Treatments were initiated at time 0 by draining the aquatic com-partment. Values are means ± standard error of five to six plants.

Dry mass (g)

R2= 0.87

0

200

400

600

800

1000

1200

0 0.5 1.0 1.5

Surf

ace

area

(cm

2)

2.0

Figure B-2: Relationship between fine

root surface area and dry mass for

Melaleuca argentea seedlings. Fine roots were sampled from 12 plants, each point is a sample from a single plant. Due to the strong linear correla-tion, dry mass was used as a proxy for fine root surface area in the calculation of root hydraulic conductance (Lp).

Appendices

197

Method of measuring of plant water uptake from each split-pot compartment

Plant water uptake was measured as the change in pot weight over the time intervals

between scheduled watering. During each watering, the soil compartments were watered to

capacity, and the aquatic compartments filled to a consistent level so that water use since

the previous watering could be calculated by subtraction from the fully hydrated weights.

In order to accurately refill aquatic compartments, a hole was drilled in each compartment

3 cm from the top, and sealed over with waterproof tape. The pots were then filled to

above the level of the hole, allowed to drain for 2 min, and the holes re-sealed. Through

this method the refilling of the aquatic compartments was reproducible to within 5 ml (2-

4% of daily plant uptake from this compartment). Other methods, such as refilling to a

line marked in the pots, resulted in levels of error that were at times ~20% of the daily

plant water uptake from the compartment. At each measurement time, each split-pot unit

was weighed just before scheduled watering, then the aquatic compartments refilled as

described, and the pots weighed again. Water uptake from the soil compartments could

then be calculated by subtracting the weights after aquatic refilling, from the weight of the

fully hydrated pot. Water uptake from the aquatic compartment was calculated by

subtracting the weights before aquatic refilling from the weights after refilling. Rates of

evaporation were also accounted for, by subtracting the water loss over the same time

period from pots without plants. Water uptake was measured in all plants over several days

just prior to the start of treatments, and water use during the treatment periods expressed as

a change relative to the pre-treatment water use, in order to normalise variation due to

differences in plant sizes.

198

30

30

45

6645

66

97

97

(a) PIP1 (b) PIP2

1 2 1 2 3kDa

Figure B-3: Specificity of PIP antibodies against Melaleuca argentea fine root total protein

samples. Proteins were extracted from fine roots collected from the hydrated zone of plants subjected partial root-zone drying, and from control plants. Total protein (12–15 μg) was sepa-rated by SDS-PAGE, blotted onto nitrocellulose membrane, and incubated with antibodies against (a) PIP1 or (b) PIP2 subfamily proteins. The positions of the size marker bands are shown beside the images. The multiple bands in each lane appear to represent the mono-, di- and trimeric forms of PIP proteins, as demonstrated by the single trimer-sized band when the reducing agent dithiothreitol was added to a sample at a 100-fold lower concentration than usual (panel (a), lane 2). Expected size of the monomers is 23–31 kDa (Maurel et al., 2008). Lanes in panel (b) are three different root protein samples.

Appendices

199

0

100

200

0

200

400

600

800

Rewet PRD-RW

ControlPRD-RW

22 24 26 28 30 32 34Time after PRD treatment (days)

Wat

er u

ptak

e (m

l)

(a) Aquatic roots

(b) Soil roots

Figure B-4: Recovery of root water uptake following re-wetting of the dried root portion

after 28 days of partial root-zone drying (PRD). Saplings of Melaleuca argentea had roots divided between an aquatic compartment and a soil compartment. The PRD treatment was applied by draining the aquatic compartment while the soil compartment remained watered, although total water supply was restricted, inducing water stress and a reduction in total water use by treated plants relative to controls. Water uptake from (a) aquatic compartments and (b) soil compartments is shown for control (untreated) plants, plants under PRD with restricted water (PRD-RW), and plants under PRD-RW treatment for 28 days followed by re-wetting of the dried aquatic compartment (PRD-RW re-wet; re-wetting point is indicated by arrows). Values are means ± standard error of four to six plants. The rapid resumption of water uptake by the aquatic roots is evidence that some fine roots survived the PRD treatment.

200

APPENDIX C: Tissue ABA content during partial root zone drying (Chapter Five)

Measurements of abscisic acid (ABA) were completed subsequent to the publication of

Chapter 5.

1 4 14

0

200

400

600

800

0

50

100

150

200

250

100

200

300

400

Control

020406080

100120140

020406080

100120140

Control 24 h 48 h

1 2Days after treatment

0

20

40

60

80

100

Treatment

Days after treatment

Days after treatment

PRDDrought

(a) Leaves

(b) Aquatic roots

(c) Soil roots

0

1

2

3

4 (d) Xylem sap

(e) Leaves

(f) Aquatic roots

(g) Soil roots

ABA

(ng

g-1 d

ry m

ass)

ABA

(ng

g-1 f

resh

mas

s)

ABA

(ng g-1 dry mass)

ABA

(ng g-1 fresh mass)

0

20

40

60

(h) Xylem sap80

1 4 140

Appendices

201

Figure previous page: Abscisic acid (ABA) content of leaves, roots and xylem sap in Melaleuca argentea seedlings under fully watered (control), partial root zone drying (PRD) and drought treatments. Seedlings had roots divided between an aquatic compartment and a soil compartment. The PRD treatment was applied by draining the aquatic compartment while the soil compartment remained watered, and the drought treatment applied by draining the aquatic compartment and also ceasing watering of the soil compartment. ABA was measured by enzyme-linked immunosorbent assay, in freeze dried material from a subset of plants from experiments 1 (panels d−g) and 3 (panels a−c); full descriptions of experiments are provided in the main text of Chapter 5. Values are means ± standard error of one to three plants, note differing axis scales in each panel.

Plants under PRD treatment showed no significant change in ABA levels relative to controls in any examined tissue (a−d). Under drought treatment, little or no change in ABA levels were observed during the first few days of treatment, but marked increases in ABA were seen in leaves and sap by the14th day of drought (e, h). Results from roots of droughted plants were less clear, but drought did appear to induce smaller increases in root ABA, at least at some timepoints (f, g). PRD in M. argentea therefore appears not to elicit an ABA root-to-shoot signal from the drying portion of the roots, in contrast with findings in many other plant species. However, M. argentea still clearly possesses the capacity to generate ABA in response to water stress, as elevated levels were detected during drying of the entire root system. See also the discussion of expected signalling mechanisms in the main text of Chapter 5.

202

Examples of micrographs used for measurement of leaf and root morphological and

anatomical features. (a) Leaf transverse section showing midvein vascular bundle (VB), palli-sade mesophyll (PM), spongy mesophyll (SM) and epidermis (EP). (b) Leaf surface epifluores-cent image under ultraviolet illumination; stomatal guard cells (GC) and trichomes (T) fluoresce blue, oil glands (OG) appear black. Transverse sections of roots under brightfield illumination, illustrating (c) absence and (d) presence of air spaces (A) within the cortex. The stele (ST) and endodermis (EN) are also indicated. Epifluorescent images of root transverse sections under ultraviolet light, illustrating (e) endodermis with two rings of cells with suber-ised walls and (f) endodermis with unsuberised regions. Lignin and suberin in xylem and endodermal cell walls fluoresce blue. All images are fresh, unstained tissue, sectioned by hand. Scale bars 100 μm.

(a)

(d)

(e) (f)

VB

PM

SM

EP

T

GCOG

ST

STST

EN

EN

EN

A

PM

(b)

(c)

ST

EN

APPENDIX D: Examples of micrographs used for anatomical measurements in

Chapter Six

Documents

Patterns of water use by the riparian tree Melaleuca argentea in …€¦ · Melaleuca New Phytologist 192: 664–675. I was the major contributor to this paper, and completed the