ECOHYDROLOGY

Ecohydrol. (2014)
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/eco.1531

Assessing the impact of large-scale water table modifications on

riparian trees: a case study from Australia
Sebastian Pfautsch,1*, Wade Dodson,2 Sally Madden2 and Mark A. Adams1
1

Faculty of Agriculture and Environment, University of Sydney, Level 4, Biomedical Building, Australian Technology Park, Eveleigh, NSW 2015,
Australia
2
Technical Projects/Resource Planning & Development, Rio Tinto Iron Ore, 152-158 St Georges Terrace, Perth, WA 6000, Australia

ABSTRACT
Mining below groundwater tables is increasing globally, yet little is known of how associated large-scale modication of water
tables impact functioning of surrounding ecosystems. We used measurements of foliage density (F) and sapwood-related sap
ow (QS) to assess effects of depth to groundwater on Eucalyptus victrix, a tree species that is common in riparian zones in
central and northern parts of Australia.
Foliage density (F) varied with season and among sites. Of itself, F provided a partial indicator of how trees responded to
falling (more than 10 m) and rising (more than 9 m) water tables. Assessment of QS was highly informative. Across all sites, QS
was least (90130 l m
2 sapwood h
1) where groundwater was naturally deep (30 m) or had fallen substantially over the past 4 years
(from 8 to 19 m). Fastest rates of QS (>245 l m
2 sapwood h
1) were recorded where groundwater had risen to a depth similar
to a site where depth to groundwater remained stable at 67 m. Our analyses of daytime and night-time QS emphasize that
water use by E. victrix is highly plastic and opportunistic. We discuss how empirical analysis of QS, coupled with a sound
understanding of local hydrogeology, can help assess responses in ecosystem function to large-scale modication of
groundwater levels an important issue globally, as well as in Australia. Copyright 2014 John Wiley & Sons, Ltd.
KEY WORDS

tree water use; water abstraction; riparian; phreatophyte; rising water table

Received 14 January 2014; Revised 3 June 2014; Accepted 26 June 2014

INTRODUCTION
Open-cut mining below the water table for industrial
minerals, mineral fuels, and metals requires lowering of the
groundwater level (drawdown) in order to prevent
ooding of mine pits. There is a well-established and
long-standing appreciation of the scale of the drawdown
required to prevent groundwater from entering such mine
pits and its dependence on local hydrogeology. A useful
summary of the approach, including the creation of
cone(s) of depression around abstraction bores (usually
located close to mine sites), is found in Alley et al.
(1999). Spatial extent and base slope of such cones of
depression vary because of a range of factors, including
the following: volume of aquifer(s); transmissivity of
soil and bedrock; amplitude of direct and indirect
recharge; volume and duration of water displacement;
and boreeld design (Alley et al., 1999). Effects of

*Correspondence to: Sebastian Pfautsch, Hawkesbury Institute for the

Environment, University of Western Sydney, Locked Bag 1797, Penrith,
NSW 2751, Australia. E-mail: [email protected]

Current Address: Hawkesbury Institute for the Environment, University

of Western Sydney, Locked Bag 1797, Penrith, NSW 2751, Australia

Copyright 2014 John Wiley & Sons, Ltd.

drawdown on surrounding ecosystems are thus spatially

limited but expected to be greatest immediately adjacent
to the bores. In addition, there are very few completed
studies (refer to Naumburg et al., 2005) of the effects of
discharge of abstracted water. Overall, the combined
phenomena of abstraction, discharge, and local changes
in water tables is fundamentally different to irrigated
agriculture where regional and larger-scale falls in water
tables result from widespread (e.g. hundreds of bores)
water abstraction. The latter situation affects streamow
and aquifer dynamics in many regions of the world
(MacKay, 2006).
In the Pilbara, a mining region in the north of Western
Australia, high-grade channel iron deposits (CIDs) are
increasingly sourced from below water tables as deposits
above water tables become exhausted. These operations
require abstraction of considerable quantities of groundwater
(ranging up to >100 ml day
1 per boreeld). By 2030, the
total amounts of water abstracted and discharged across the
whole region will have tripled (Department of Water, 2010)
as a result of an increasing number of mines, with a high
frequency of below water table operation. These predictions emphasize the need to greatly increase the currently
limited understanding of the effects of local drawdown

S. PFAUTSCH et al.

(and discharge) on surrounding dependent terrestrial

ecosystems (Froend and Sommer, 2010).
The Pilbara is characterized by ephemeral streams and
creeks where depth to groundwater is often shallow but
varies widely according to underlying geology, evapotranspirative water losses, and rates of recharge. The riparian
vegetation of major watercourses is dominated by Eucalyptus
victrix L.A.S. Johnson & K.D. Hill, sometimes interspersed
with Eucalyptus camaldulensis Dehnh. and Melaleuca
argentea W. Fitzg. Research across semiarid landscapes of
Australia and elsewhere has shown that most riparian
eucalypts qualify as facultative phreatophytes (Busch et al.,
1992; Mensforth et al., 1994; Smith et al., 1998; Lamontagne
et al., 2005; Holland et al., 2006; Costelloe et al., 2008). Here,
we dene facultative phreatophytes as species that are
dependent on access to groundwater or water from the
capillary fringe, when soil water originating from surface
recharge becomes limited. Studies elsewhere suggest that
falls in groundwater and associated capillary fringe affect
species composition and productivity of dependent ecosystems (Laine et al., 1995; Murphy et al., 2009; Sommer and
Froend, 2011) and may cause declining tree health (Murray
et al., 2003) or even tree death (Horton et al., 2001; Eamus
and Froend, 2006). Negative effects on ecosystem function
are usually attributed to the limited drought tolerance of
riparian trees (Naumburg et al., 2005; Orellana et al., 2012).
Establishing effects of falling groundwater on facultative
phreatophytes requires both consideration of rates of
drawdown and antecedent groundwater levels (Shafroth
et al., 2000). For example, trees accessing shallow groundwater may be more sensitive to drawdown compared with
trees growing above groundwater at greater depth. The net
result is a range in dependence on groundwater by trees across
a landscape and variable responses to water availability
(Murray et al., 2003). Water use of phreatophytes is
intimately linked to soil water availability. Under conditions
where tree water use (transpiration) generates increasingly
negative water potentials inside the plant, signals from roots
and/or the atmosphere serve to reduce stomatal conductance
(gs), effectively reducing water loss (e.g. Porporato et al.,
2002). Examples of the links among declining soil water
availability, reduced gs, and declining rates of sap ow in tree
stems are provided by Whitehead and Beadle (2004) and
Bovard et al. (2005). Trees can limit water losses by
abscission of foliage (Brda et al., 2006), and for eucalypts,
abscission of foliage has long been linked to acute water
shortages at the end of summer (Attiwill and Adams, 1996).
There is also evidence from riparian poplars of reductions
in gs and abscission of leaves as a result of falling water
tables (refer to Amlin and Rood, 2003 and references
therein), but there is little evidence for other riparian tree
species, including eucalypts.
For the boreeld associated with the study reported here,
hydrological modelling suggests that by 2030, the
Copyright 2014 John Wiley & Sons, Ltd.

groundwater level at the centre of the cone of depression

will have fallen by as much as 180 m (Johnson and Wright,
2001), with much smaller falls at greater distance from the
centre. As groundwater is being abstracted and depth to
groundwater is increasing, there is a prerogative for mine
managers to carefully consider how this water is used.
Clearly, in addition to water for dust suppression and other
industrial procedures, a high priority must be given to the
use of abstracted water to prevent a decline in ecosystem
functioning in the area affected by drawdown. Of particular
concern is maintaining surface ow where that existed
before the commissioning of the mine. For this purpose, a
series of spurs were installed that release groundwater to
areas upstream and downstream of a major surface
expression, known locally as a spring.
Here, we report on the following: (i) a 4-year study of foliage
density (F, a measure of leaf area per unit canopy space) and
(ii) a short-term study of tree water use (using sapwood-related
sap ow (QS) as a surrogate), in the initial stages of the mining
and boreeld operations.
Water availability dictates ecosystem processes, including
plant water uptake (Porporato et al., 2002). For the short-term
(91 days) study, we specically targeted the dry season
when soil water becomes progressively limiting and forces
phreatophytes to increase dependence on water from
deeper sources (Lamontagne et al., 2005). We specically
tested the following hypotheses related to the plasticity of
E. victrix for surviving a broad range of environmental
conditions: (i) A fall in groundwater from moderate to deep
depths would result in trees becoming increasingly water
limited similar to trees with a life history of accessing deep
water sources; (ii) a rise in groundwater from deep to
moderate levels would result in increased water use by trees
compared with those that have a life history of access to
water sources at moderate depth.

MATERIALS AND METHODS

Study area
A detailed description of the hydrogeological characteristics of the study area along Weeli Wolli Creek is provided
as Supplementary Information. Briey, the tree vegetation
within the riparian zone of the Weeli Wolli Creek consists
variously of the following: (i) narrow belts on banks and
elevated locations within the channel and (ii) widely spaced
woodlands on low-lying areas adjacent to the channel.
Trees are irregularly distributed and do not form a
continuous canopy (refer to Figure S3S5). These
characteristics make hydrological scaling of water uxes
problematic. We focused on characterizing the effects of
changing water tables on tree health (as measured by leaf
density) and physiological performance (as measured by
sap velocity).
Ecohydrol. (2014)

RESPONSE OF E. VICTRIX TO VARYING WATER TABLES

Four research sites were established along a 30-km

section of the upper and midsection of the creek where at
least ve trees could be accessed within a reasonable area.
Two sites were established, termed control29 (
23.04lat,
119.18long; 608 m a.s.l.) and control6 (
22.93lat,
119.19long; 577 m a.s.l.) where depth to groundwater
was stable. Location of control29 was outside the cone of
depression, and the groundwater level was naturally deep
(around 29 m; Figure 1, Table I). Depth to groundwater at
control6 was less (6 m) and was articially maintained by
discharge from a series of spurs from a major water
pipeline (Figure 1 and S1, Table I). A third site, termed
drawdown (
22.94lat, 119.17long; 583 m a.s.l.), was
selected immediately above the cone of depression where
depth to groundwater had dropped from 8.3 to 19.2 m since
commencement of drawdown (Figure 1 and S1, Table I).
The fourth site, termed surplus (
22.81lat, 119.29long;
496 m a.s.l.) was established in the midsection of the creek,

15 km downstream of control6. Discharge of excess groundwater further upstream had lifted the water table at this site
from 16 to 7 m (Table I). All measured trees were within
510 m of the main channel of Weeli Wolli Creek.
Nominally, all sample trees were E. victrix. Distinguishing
this species from E. camaldulensis in the eld is difcult,
relying on characteristics of seed capsules that are not
always present.
Regional climate
Climate of the Pilbara is bimodal and generally hot (refer to
also Pfautsch et al., 2011). In the central Pilbara, near the
town of Newman, annual mean maximum air temperature
is 32 C (19962012) and annual precipitation is 315 mm
(19712012; Bureau of Meteorology, 2012). Large
proportions of annual precipitation can fall in a single event
(up to 200 mm or more) originating from decaying summer

Figure 1. Location of research sites along the upper section of Weeli Wolli Creek (control29, drawdown, and Control6) in relation to proposed below
water table (BWT) mine pits. Contour lines indicate extent and depth of drawdown at the end of Nov 2010 compared with pre-mining groundwater
levels. Because of reasons of scale, the map does not show the location of the surplus site (12 km NE of spring site).

Copyright 2014 John Wiley & Sons, Ltd.

Ecohydrol. (2014)

S. PFAUTSCH et al.

Table I. Depth to groundwater for research sites.

Site name

Pre-management (m)
28.74
8.28
6.33
16.33

Control29
Drawdown
Control6
Surplus

(0.03)
(0.22)
(0.32)
(0.04)

Current (m)
29.40
19.23
6.23
7.05

(0.02)
(0.13)
(<0.01)
(<0.01)

Change current to
pre-management (m)

0.66

10.95
+0.10
+9.28

Change within study

period (m)

0.42

0.87

0.07

0.22

Pre-management depths represent average depth-to-groundwater between 1998 and 2007. Current depth-to-groundwater was averaged from
2
fortnightly data spanning from 30 June to 27 Nov 2010. Variance (s ) is shown in parentheses.

cyclones. Droughts are frequent during which annual rainfall

can be less than 40 mm year
1. Study sites were all located in
the same broad valley system and subject to the same general
climate (i.e. identical vapour pressure decit (D) and
temperatures within 12 C at any given time). Rainfall
during the wet season preceding that of our study was low
(126 mm between November 2009 and April 2010) relative to
long-term records and insignicant during the rst half of the
preceding dry season (20 mm between May and August
2010). Annual potential evapotranspiration (PET) in the
central Pilbara can reach up to 3700 mm (Luke et al., 1987). A
weather station established close to our drawdown site
recorded air temperature (Tair, C) and relative humidity (rH,
%; HMP43A, Vaisala, Finland), rainfall (mm; TB3/0.2,
Hydrological Services America, USA), solar radiation
(EQ08-E, Middleton Solar, Australia), and wind speed and
direction (Model 051035, RM Young Company, USA) at
10-min intervals. PET (mm) was calculated from these
measurements using a modied PenmanMonteith equation.
Vapour pressure decit was calculated using average Tair and
rH according to Snyder and Shaw (1984):

D

0:6108 


 exp17:27  T  
rH
air
 1

T air 237:3
100
(1)

Canopy monitoring
Long-term patterns of tree health were monitored at all sites
over a 4-year period (20062010), except at control29
(20092010). Canopy monitoring at the latter site was
delayed because of access restrictions. A single sample
point was established and permanently marked underneath
10 tree canopies at three research sites and 12 canopies at
surplus to monitor changes in F. Digital images with an
approximately 15 eld of view were collected at each
point up to four times a year. Foliage cover (CF) and crown
cover (CC) were calculated according to Macfarlane et al.
(2007). An estimate of F within crowns was calculated as
CF:CC. This sampling strategy does not provide an average
measurement of cover across whole stands because only
individual canopies were sampled, omitting gaps between
widely spaced individuals. The measurement objective was
to detect changes in F of tree canopies rather than quantify
Copyright 2014 John Wiley & Sons, Ltd.

stand averages. We used a gap-lling approach to account

for missing images. This approach used the average relative
change in CF and CC among all trees at the relevant site
between two consecutive sample periods to estimate
missing data (<4% of all images were missing from a
total of 1200).
Sap velocity measurements
We recorded heat velocity (Vh) in stems of ve E. victrix at
all sites for 91 days (1 September to 30 November 2010)
except at control6 where data were recorded for 71 days
(1 September to 11 November) using probes that operate
on the principle of the heat ratio method (HRM, Burgess
et al., 2001). All measured trees were located within 10 m of
the riverbank. Physical characteristics of measured trees are
given in Table S1. Additional measurements and procedures
(refer to Pfautsch et al., 2010) allowed conversion of Vh to
sapwood-related sap ow (QS, l (unit sapwood area)
1 (unit
time)
1). At each site, we installed four probe sets in one tree
in four hemispherical directions to detect any circumferential
variation of sap velocity. One probe set was installed at the
southern side of four additional trees. Probe sets, data loggers,
and auxiliary equipment were acquired from ICT International (Armidale, Australia), and probes were shielded to
reduce thermal load. At the end of sap ow measurements, we
extracted two blocks of sapwood from opposite sides of three
measured trees at each site. Sapwood was immediately
analysed for fresh weight and volume (immersion technique).
Once returned to the lab, blocks were dried for 72 h at 105 C
before recording their dry weight. Sapwood density was
calculated as mass per unit volume (g cm
3). In addition, we
extracted wood cores from each of four cardinal positions
from each tree equipped with sap ow sensors for a detailed
analysis of sapwood depth (DS) using light microscopy as
outlined in Pfautsch et al. (2012).
Initial measurements collected across a radial gradient of
DS indicated that maximum QS was concentrated in outer
sapwood. This section of sapwood supports transpiration of
the sunlit canopy and is a strong indicator of availability
and use of water by trees (Fiora and Cescatti, 2005). Given
our focus on effects of drawdown on tree health and
physiological function and the irregular distribution of
these riparian stands, we focused on characterizing patterns
Ecohydrol. (2014)

RESPONSE OF E. VICTRIX TO VARYING WATER TABLES

of QS rather than total volumetric water use on a stand

basis. Consequently, thermistors were positioned at 7 mm
DS in all trees ensuring comparability among trees and
sites. According to the conductive properties of stainless
steel probes (Swanson, 1983), QS reported here represents
an integral of velocities at DS 212 mm.
Statistical analyses
One-way analysis of variance (ANOVA) was used to assess if
sapwood density and moisture content varied signicantly
between sites. Trends in mean monthly peak QS (QSp) were
established relative to the period where QS at control6 (i.e.
benchmark site where trees grow at continuously high
groundwater level) was more or less constant (10001800 h).
As shown previously for this species (Pfautsch et al., 2011),
the relation of QS to D differs markedly between daytime and
night-time. Hence, we used 6-h windows to assess this
relation when average D was at its maximum (QSDmax,
11001600 h) and minimum (QSDmin 01000600 h) for each
site. For these tests, we applied repeated measures (RM)
ANOVA to mean QS (n = ve trees) for each hour of the
identied time window (n = 6) and used Fishers LSD post
hoc test to examine between-subject (month to month)
variation. RM-ANOVA was also used to evaluate if F
differed signicantly among sites. ANOVA was used to test if
F at individual sites changed signicantly from September
2010 to December 2010. Signicance of all tests was given if
P < 0.05. AABEL3 (Gigawiz, Tulsa, USA) was used for
statistical tests.

averages (refer to Regional climate). Average F was lowest

at control29 (0.65 0.08 (mean 1 standard deviation (SD)).
Trees at all other sites showed F of 0.690.70 (overall 0.09).
Changes in depth to groundwater (refer to Table I) had no
measurable long-term effect on F at drawdown or surplus.
Analyses of data collected in September and December 2010
showed a tendency of declining F across all sites, albeit this
decline was only signicant (0.75 to 0.63, P < 0.001) at
drawdown (Figure 2).
Weather conditions
From September to November 2010, the weather became
generally hotter and drier. Average maximum Tair rose
from 25 C in September to 35 C in November while at the
same time, rH declined from 22 to 10%. D increased
accordingly to reach 7.3 kPa during the hottest day (>41 C)
in November and remained high during night-time where it

RESULTS
Foliage density in tree canopies
Long-term monitoring (October 2006December 2010)
indicated signicant intra-site variation of F, with variation
being greatest at surplus (Table II). Amongst all sampling
intervals and sites, F was largest in February 2007 following
a pronounced wet season. In contrast, F decreased in the wet
season of 2009/2010 where rainfall was below long-term

Figure 2. Variation of foliage density (F) in canopies of Eucalyptus

1
victrix (n = 20 site ) between October 2006 and December 2010. Solid
lines show smoothed averages; error bars show SD.

Table II. Statistics for RM-ANOVA (including BonferroniDunn post hoc) testing differences in foliage density (F) of
Eucalyptus victrix.
Post hoc
Site name

DF F

SI

Control29
Drawdown
Control6
Surplus

18
14
16
18

6
15
15
15

Residual MS F
108
210
240
270

0.02
0.02
0.08
0.06

SI

F-ratio F

SI

Partial 2
for SI

0.03
0.02
0.01
0.03

4.69
5.53
22.05
19.12

10.12
4.82
3.53
9.14

0.29
0.19
0.12
0.29

score total score SepNov Adjusted level

6
10
11
31

No
Yes
No
No

0.0024
0.0004
0.0004
0.0004

All RM-ANOVA indicated that F differed signicantly among SI (P < 0.001). Post hoc scores denote number of signicant differences among all
possible combinations of SI per site. Signicant differences in F during study period are shown separately.
DF, degrees of freedom; MS, mean squares; SI, number of sampling intervals.

Copyright 2014 John Wiley & Sons, Ltd.

Ecohydrol. (2014)

S. PFAUTSCH et al.

regularly exceeded 2 kPa. Maximum D during night-time was

3.9 kPa. During our study, PET totalled close to 600 mm.

Table III. Slope, intercept, and regression coefcient for relation

between peak sap ow (QSp, l m
2 h
1) and daytime hours
(10:0018:00) in Eucalyptus victrix.

Sap ow

Site name

Month

Slope

y-intercept

R2

Control29

Sep
Oct
Nov
Sep
Oct
Nov
Sep
Oct
Nov
Sep
Oct
Nov

2.41

3.00

3.34

3.71

4.50

4.57

1.75

0.56

0.32

1.94
0.34

0.15

136.28
125.32
103.53
154.00
143.78
121.03
209.80
204.35
198.83
243.50
241.52
234.20

0.40
0.89
0.97
0.59
0.84
0.91
0.34
0.31
0.09
0.58
0.23
0.03

Average sapwood density of E. victrix was 0.695 g cm
3

(0.056). Moisture content of sapwood ranged from 32 to
39%. Both parameters did not vary signicantly among sites.
Sap ow within the outer band of sapwood varied
considerably with tree size, ranging from an average of 22
to 224 l m
2 h
1 (52, Figure S2). We calculated mean
monthly diel courses of QS for all sites. One similarity among
sites was that QS never approached zero. Another was a
doubling of QS from minimum rates within 2 h of early
daylight (Figure 3ad). Overall, trees at sites where
groundwater was deep (control29 and drawdown) used less
water compared with trees that grew at sites where
groundwater was closer to the surface (control6 and surplus).
Diel courses of QS varied greatly among sites. For the sites
where the water table was at greatest depth (control29 and
drawdown), QS increased rapidly with the onset of daylight,
in September reaching QSp around midday (control29,
133 l m
2 h
1; drawdown, 148 l m
2 h
1). As the dry season
progressed, QSp declined (in November: control29, 101 lm
2
h
1; drawdown, 111 l m
2 h
1) and was reached earlier each
day (e.g. by 0800 h in November). QS slowed during daytime,
and its relation to D became gradually linear with increasingly
negative slopes (refer to Figure 3a and b, Table III). For sites
where groundwater was closer to the surface (control6 and

Drawdown
Control6
Surplus

Data were averaged for the months of Sep, Oct, and Nov 2010.

surplus), daytime QS was consistently faster throughout the

research period (Figure 3c and d). For these sites, and
following an early and steep increase in QS after dawn, QSp
was reached between 1000 and 1100 h where it remained with
little change for the entire day. On average, QSp was
212 l m
2 h
1 at control6 and >240 l m
2 h
1 at surplus
(Figure 3c and d). There was negligible variation in QSp from
September to November for these sites (Table III).
Hysteresis in D versus QS relationships is shown in
Figure 3eh. Notably, differences in QS between morning

Figure 3. Diel proles of sapwood-related sap ow (QS, panel ad) and the relation of QS to vapour pressure decit (D, panel eh) in Eucalyptus victrix

1
(n = 5 site ). Data were averaged for Sep 2010 (solid line), Oct 2010 (dotted line), and Nov 2010 (dash-dotted line). Circular arrow in (h) indicates
direction of hysteresis.

Copyright 2014 John Wiley & Sons, Ltd.

Ecohydrol. (2014)

RESPONSE OF E. VICTRIX TO VARYING WATER TABLES

and afternoon for a given D were greater for trees at

control6 and surplus than for trees at control29 and
drawdown (Figure 3eh), suggesting greater exibility in
stomatal opening when water tables are higher. The lower
overall QS and reducing difference (for a given D) between
morning and afternoon at control29 and drawdown suggest
that within-tree water decits may be reducing overall
stomatal opening. When water tables were closer to the
surface (control6 and surplus), maintenance of faster QSp
till late each day (e.g. 1800 h), as signied by the horizontal
beak at the upper end of the hysteresis loop (Figure 3g
and h), reects weaker control by D.
QSDmax at control29 and drawdown declined by around 30%
from September to November (control29: 128 to 86 l m
2 h
1;
drawdown: 139 to 98 l m
2 h
1; Table IVa) but remained
more or less constant at control6 and surplus (Table IVa).
Changes in QSDmax across the research period were signicant
only for control29 and drawdown (Table IVb).
In clear contrast to varying responses in QSDmax among
sites, QSDmin increased signicantly in trees at all sites
(Table IVa and b). Increasing QSDmin was most pronounced
at control6 (+76%) and surplus (+30%), followed by
control29 (+24%) and drawdown (+14%; Table IVa). Mean
D during this interval increased from 1.0 kPa (September, 0.7)
to 2.5 kPa (November, 0.7). QSDmin was the main driver
for increased tree water use from September to November at
control6 and surplus, while decreasing QSDmax caused the
opposite at control29 and drawdown.

DISCUSSION
Our research from semiarid Australia offers insights into how
falling or rising groundwater levels inuence water use of
facultative phreatophytes. There are no comparable studies in
the published literature, as far as we can ascertain, of the
effects of high rates of abstraction (here >96 ml day
1) and a
fast and sustained fall in groundwater level (here >1 cm
day
1 over 4 years) on tree water use and health.
Digital photography (Macfarlane et al., 2007) and visual
assessment of tree crowns (Souter et al., 2010) have both
been used elsewhere to assess leaf density or projected area
and, inter alia, can serve as indices of tree health. Both
techniques have their limitations when assessing widely
spaced riparian stands. Estimation of a leaf area index
cannot fully capture declining health of individual tree
canopies when trees are sparse, while visual assessment
depends on assessor experience and consistency. Using F
is more likely to capture changes in canopy condition of
widely spaced trees as it repeatedly assesses the same
section of a canopy.
Over the 4-year period of study, F varied within
moderately small ranges, despite variations in annual
rainfall of >50% from the long-term annual mean.
Variation in F did not reect depth or access to
groundwater, and equal variability in F was evident across
all sites albeit not during September and December 2010.
Although the decline of F recorded between September

Table IV. (a) Average monthly sapwood-related sap ow (QS, l m
2 h
1) in Eucalyptus victrix (n = 5); (b) related statistical analyses.
Qs (l m
2 h
1)
(a)
Site name
Control29
Drawdown
Control6
Surplus
Control29
Drawdown
Control6
Surplus
(b)
Site name
Control29
Drawdown
Control6
Surplus
Control29
Drawdown
Control6
Surplus

Sep
QSDmax
QSDmax
QSDmax
QSDmax
QSDmin
QSDmin
QSDmin
QSDmin

127.78
138.65
203.94
232.32
32.42
31.21
54.18
79.55

QSDmax
QSDmax
QSDmax
QSDmax
QSDmin
QSDmin
QSDmin
QSDmin

DF
179
179
59
179
179
179
59
179

(29.61)
(34.79)
(43.95)
(57.44)
(8.69)
(8.54)
(19.02)
(30.63)

Oct
110.54
122.16
202.46
242.98
38.23
31.75
78.15
85.73

Nov

(12.67)
(16.53)
(5.54)
(16.11)
(12.72)
(11.26)
(24.90)
(41.38)

85.90
98.18
198.10
233.19
40.28
35.56
95.28
103.31

RM-ANOVA
F-ratio
1.40
1.35
0.97
1.04
1.99
2.30
2.29
3.26

P
<0.01
0.01
>0.50
0.37
<0.01
<0.01
<0.01
<0.01

(11.86)
(13.43)
(5.13)
(13.98)
(17.45)
(14.89)
(34.37)
(49.95)

Nov
Sep

41.88

40.41

4.62
+0.87
+7.86
+4.35
+41.10
+23.76
P (Nov
Sep only)
<0.01
<0.01
<0.01
>0.50
<0.01
<0.01
<0.01
<0.01

QS was recorded 1 Sep30 Nov 2010 and is presented separately for 6-h windows where vapour pressure decit of the atmosphere was at maximum
(QSDmax, 11:0016:00) and minimum (QSDmin, 1:006:00); SD is shown in parenthesis; Nov Sep in (a) indicates change in QS from Sep to Nov.
Results of RM-ANOVA denote between-subject statistics, including degrees of freedom (DF), test statistic (F-ratio), and signicance probability (P);
Nov Sep in (b) shows P according to Fishers LSD test. Signicance was given at P < 0.05 for all tests.
Copyright 2014 John Wiley & Sons, Ltd.

Ecohydrol. (2014)

S. PFAUTSCH et al.

and November 2010 in trees at the drawdown was signicant,

ongoing monitoring suggests that trees recovered completely
in the following wet season (data not presented). Additional
information, say as provided by QS, is critical to assessing
the effects of water availability for riparian ecosystems as
studied here.
The control6 site provides a clear example of water use
dynamics of E. victrix with more or less unlimited access to
water. The dynamics were similar to those reported
recently by Pfautsch et al. (2011) for sites where depth
to groundwater was around 12 m Similar patterns (i.e. a
steep increase during early morning, followed by a plateau
during most of the day and rapid decline with nightfall)
have been reported from other semiarid environments
(e.g. Meinzer et al., 1999; Bucci et al., 2008; OGrady et al.,
2009). Under well-watered conditions, this pattern remains
stable regardless of season (OGrady et al., 1999; Eamus
et al., 2000; Pfautsch et al., 2011) and is indicative of a
dominant effect of D (Scott et al., 2004; Huang et al., 2009).
Limitation of gs as D increases reduces risks of damage
(e.g. via cavitation) to vascular systems. Our data not only
from control6 but also from surplus support this regulatory
mechanism. Stomatal limitation of transpiration during
daytime with increasing D, and looser regulation during
night-time when QS tracks D, has been reported for E.
victrix (Pfautsch et al., 2011) and other eucalypts under
natural conditions (Buckley et al., 2011) as well as in
common gardens (Phillips et al., 2010).
Groundwater at moderate depth (67 m) resulted in
stable QSDmax. Where groundwater was deep, QSDmax
declined signicantly during the dry season (by 30 2%).
On the other hand, QSDmin increased during this period
across all sites, albeit to a smaller degree at control29 (24%)
and drawdown (14%) when compared with surplus (30%)
and control6 (76%). This observation is in line with previous
description of relations between stomatal regulation and D.
During daytime, QSDmax declines when soil water availability
becomes limiting and remains constant when sufcient soil
moisture can be accessed. As D increased steadily throughout
the research period, including at night, increasing QSDmin is
likely a result of increasing night-time transpiration where
the magnitude of increasing QSDmin reects the availability
of soil moisture.
Supercially, trees functioned similarly at sites with deep
groundwater, irrespective of abstraction effects. Differences
between these sites in rates of sap ow were consistently
small and arguably not signicant. However, this observation is in broad agreement with the life history concept of
roots, which states that the impact of groundwater
drawdown will affect trees growing over historically
shallow groundwater will more negatively compared with
trees that had developed over historically deeper levels of
groundwater (Scott et al., 1999; Shafroth et al., 2000).
QSDmin was consistently less at drawdown compared with
Copyright 2014 John Wiley & Sons, Ltd.

control29. At the other extreme, where groundwater had

risen as the result of discharge (surplus), E. victrix
transpired at faster rates compared with a site where depth
to groundwater remained at 6 m (control6). Taken together,
our observations support both of our hypotheses and
underline the capacity of E. victrix to survive a wide range
of ecohydrological settings.
It is unclear if roots of E. victrix access groundwater as
deep as 29 m. Across biomes, average maximum rooting
depth of trees is around 7 1.2 m (Canadell et al., 1996).
However, tree species from semiarid environments have
been shown to access groundwater at depths below 30 m
(Zencich et al., 2002), and tap roots of Eucalyptus
marginata have been found at depths beyond that (Dell
et al., 1983). Generally, dimorphic root systems, as well as
capacity to reach water sources deep in soil proles or the
regolith, are well described for phreatophytes (Ehleringer
et al., 1991; Dawson and Pate, 1996). Even so, factors such
as low oxygen concentrations in deep soil layers, impeding
soil layers, and increasing resistance to water transport in
roots often severely limit transport of large volumes of
water over long path lengths (Tyree and Ewers, 1991;
Canadell et al., 1996). The capacity of root growth to
match rates at which groundwater levels fall can also
severely constrain tree water access (Zencich et al., 2002).
However, the capacity of another Australian phreatophyte
(Banksia spp.) to adjust growth dynamics of roots in
accordance to uctuation of groundwater tables has been
documented recently (Canham et al., 2012). Nevertheless,
some reports of maximum rates of root growth of
315 mm day
1 by arid zone species (reviewed in
Naumburg et al., 2005) do not take account of soil depth
(i.e. can trees grow roots as quickly at 20 m depth as at 1 m
depth?), and maximum rates for riparian species in
semiarid environments are unknown. Large investments
of carbon needed for root growth at great depth may not be
available because of low gs imposed on trees by water
limitation (Reich, 2002; Naumburg et al., 2005). Then
again, falling groundwater levels increase the volume in
soil available for storage of precipitation and capillary rise,
which has been shown to increase water use of some plants
(Jackson et al., 2000). This advantage may be of limited
use for plants growing in semiarid environments where
annual precipitation is low.
Where groundwater had risen from 16 to 7 m depth
(surplus), QS was generally faster compared with where
groundwater remained at 6 m (control6). Previous research
suggests that plant productivity might increase with rising
groundwater (assuming no anoxia; e.g. Groeneveld and
Crowley, 1988; Naumburg et al., 2005). Anoxic conditions
were not a factor for trees at surplus, as many metres of soil
remained unsaturated. If rising groundwater tables lead to
increasing productivity in E. victrix remains a question for
future research.
Ecohydrol. (2014)

RESPONSE OF E. VICTRIX TO VARYING WATER TABLES

E. victrix displayed opportunistic physiognomy in using

water, a strategy also reported for facultative phreatophytes
from the wet tropics of Australia (OGrady et al., 2006)
and the USA (Hultine et al., 2010) but not for some other
species (Amlin and Rood, 2003). In a global context,
riparian ecosystems have emerged as intense political,
economic, social and ecological battlegrounds over limited
water resources (sensu Cleverly et al., 2006). Here, we
have shown how anthropogenic modication of water
tables can affect water use of eucalypts that dominate these
zones. We demonstrated a remarkable capacity of E. victrix
to sustain a signicant short-term decline in depth to
groundwater. Similar investigations will be required in
other semiarid environments where open-cut mining affects
groundwater and dependent tree species. Our results
suggest repeated (even if intermittent) monitoring of
selected parameters like tree water use will be required.
We agree with MacKay (2006) that negative impacts of
dewatering can only be minimized once sound understanding
of local hydrogeology is developed. Models describing plant
responses to either rising or falling water tables have been
developed (Shafroth et al., 2000; Naumburg et al., 2005). To
date, these models do not consider situations common to
open-cut mining where water tables may drop signicantly
over short time spans. Improving our understanding of how
dewatering and associated discharge impact nearby ecosystems is now a pressing issue, as open-cut mining below water
tables is increasing in Australia and elsewhere.

ACKNOWLEDGEMENTS

This research was funded by the Australian Research

Council Linkage Project 0776626, which included support
by Rio Tinto Iron Ore (RTIO). The authors would like to
thank Pauline Grierson (University of Western Australia),
Nicole Gregory (RTIO), Shawan Dogramaci (RTIO), Craig
Macfarlane (Commonwealth Scientic and Industrial
Research Organisation) and Neil Murdoch (University of
Sydney) for their assistance during the project.

REFERENCES
Alley WM, Reilly TE, Franke OL. 1999. Sustainability of ground-water
resources. U.S. Geological Survey Circular 1186, Denver, Colorado.
Amlin NM, Rood SB. 2003. Drought stress and recovery of riparian
cottonwoods due to water table alteration along Willow Creek, Alberta.
Trees 17: 351358.
Attiwill PM, Adams MA. 1996. Nutrition of Eucalyptus. CSIRO
Publishing: Melbourne, Australia.
Bovard BD, Curtis PS, Vogel CS, Su B-H, Schmid HP. 2005.
Environmental controls on sap ow in a northern hardwood forest.
Tree Physiology 25: 3138.
Brda N, Huc R, Granier A, Dreyer E. 2006. Temperate forest trees and
stands under severe drought: a review of ecophysiological responses,
adaptation processes and long-term consequences. Annals of Forest
Science 63: 625644.

Copyright 2014 John Wiley & Sons, Ltd.

Bucci SJ, Scholz FG, Goldstein G, Hoffmann WA, Meinzer FC, Franco AC,
Giambelluca T, Miralles-Wilhelm F. 2008. Controls on stand transpiration
and soil water utilization along a tree density gradient in a Neotropical
savanna. Agricultural and Forest Meteorology 148: 839849.
Buckley TN, Turnbull TL, Pfautsch S, Adams MA. 2011. Nocturnal
water loss in mature subalpine Eucalyptus delegatensis tall open
forests and adjacent E. pauciora woodlands. Ecology and Evolution
1: 435450.
Bureau of Meteorology, Australian Government. Website accessed
05.10.2012: http://www.bom.gov.au/climate/averages/tables/
cw_007176.shtml
Burgess SSO, Adams MA, Turner NC, Beverly CR, Ong CK, Kahn AAH,
Bleby TM. 2001. An improved heat pulse method to measure low and
reverse rates of sap ow in woody plants. Tree Physiology 21: 589598.
Busch DE, Ingraham NL, Smith SD. 1992. Water uptake in woody
riparian phreatophytes of the southwestern United States: a stable
isotope study. Ecological Applications 2: 450459.
Canadell J, Jackson RB, Ehleringer JR, Mooney HA, Sala OE, Schulze ED.
1996. Maximum rooting depth of vegetation types at the global scale.
Oecologia 108: 583595.
Canham CA, Froend RH, Stock WD. 2012. Dynamics of phreatophyte
root growth relative to a seasonally uctuating water table in a
Mediterranean-type environment. Oecologia 170: 909916.
Cleverly JR, Dahm CN, Thibault JR, McDonnell DE, Allred Coonrod JE.
2006. Riparian ecohydrology: Regulation of water ux from the ground
to the atmosphere in the Middle Rio Grande, New Mexico.
Hydrological Processes 20: 32073225.
Costelloe JF, Payne E, Woodrow IE, Irvine EC, Western AW, Leaney FW.
2008. Water sources accessed by arid zone riparian trees in highly saline
environments, Australia. Oecologia 156: 4352.
Dawson TE, Pate JS. 1996. Seasonal water uptake and movement in root
systems of Australian phreatophytic plants of dimorphic root morphology: a stable isotope investigation. Oecologia 107: 1320.
Dell B, Bartle J, Tacey W. 1983. Root occupation and root channels of
jarrah forest subsoils. Australian Journal of Botany 31: 615627.
Department of Water. 2010. Pilbara Regional Water Plan 20102030
Supporting detail. Department of Water: Perth, Western Australia.
Eamus D, Froend B. 2006. Groundwater-dependent ecosystems: the
where, what and why of GDEs. Australian Journal of Botany
54: 9196.
Eamus D, OGrady AP, Hutley L. 2000. Dry season conditions determine
wet season water use in dry-wet tropical savannas of northern Australia.
Tree Physiology 20: 12191226.
Ehleringer JR, Philips SL, Schuster WFS, Sandquist DR. 1991.
Differential utilization of summer rains by desert plants: implications
for competition and climate change. Oecologia 88: 430434.
Fiora A, Cescatti A. 2005. Diurnal and seasonal variability in radial
distribution of sap ux density: implications for estimating stand
transpiration. Tree Physiology 26: 12171225.
Froend R, Sommer B. 2010. Phreatophytic vegetation response to climatic
and abstraction-induced groundwater drawdown: examples of long-term
spatial and temporal variability in community response. Ecological
Engineering 36: 11911200.
Groeneveld DP, Crowley DE. 1988. Root system response to ooding in
three desert shrub species. Functional Ecology 2: 491497.
Holland KL, Tyerman SD, Mensforth LJ, Walker GR. 2006. Tree water
sources over shallow, saline groundwater in the lower River Murray,
south-eastern Australia: implications for groundwater recharge mechanisms. Australian Journal of Botany 54: 193205.
Horton JL, Kolb TE, Hart SC. 2001. Responses of riparian trees to
interannual variation in groundwater depth in a semi-arid river basin.
Plant, Cell and Environment 24: 293304.
Huang Y, Zhao P, Zhang Z, Li X, He C, Zhang R. 2009. Transpiration of
Cyclobalanopsis glauca (syn. Quercus glauca) stand measured by sapow method in a karst rocky terrain during dry season. Ecological
Research 24: 791801.
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: 347361.
Jackson RB, Sperry JS, Dawson TE. 2000. Root water uptake and
transport: using physiological processes in global predictions. Trends in
Plant Science 5: 482488.

Ecohydrol. (2014)

S. PFAUTSCH et al.
Johnson SL, Wright AH. 2001. Central Pilbara Groundwater study.
Water and Rivers Commission, Hydrogeological Record Series,
Report HG8.
Laine J, Vasander H, Laiho R. 1995. Long-term effects of water level
drawdown on the vegetation of drained pine mires in southern Finland.
Journal of Applied Ecology 32: 785802.
Lamontagne S, Cook PG, OGrady A, Eamus D. 2005. Groundwater use
by vegetation in a tropical savanna riparian zone (Daly River,
Australia). Journal of Hydrology 310: 280293.
Luke GJ, Burke KL, OBrian TM. 1987. Evaporation data for Western
Australia. Resource Management Technical Report No. 65, Department
of Agriculture, Government of Western Australia.
Macfarlane C, Hoffmann M, Eamus D, Kerp N, Higginson S, McMurtie R,
Adams MA. 2007. Estimation of leaf area index in eucalypt forest using
digital photography. Agricultural and Forest Meteorology 143: 176188.
MacKay H. 2006. Protection and management of groundwater-dependent
ecosystems: emerging challenges and potential approaches for policy
and management. Australian Journal of Botany 54: 231237.
Meinzer FC, Goldstein G, Franco AC, Bustamante M, Igler E, Jackson P,
Caldas L, Rundel PW. 1999. Atmospheric and hydraulic limitations on
transpiration in Brazilian cerrado woody species. Functional Ecology
13: 273282.
Mensforth LJ, Thorburn TJ, Tyerman SD, Walker GR. 1994. Sources of
water used by riparian Eucalyptus camaldulensis overlying highly
saline groundwater. Oecologia 100: 2128.
Murphy M, Laiho R, Moore TR. 2009. Effects of water table drawdown
on root production and aboveground biomass in a boreal bog.
Ecosystems 12: 12681282.
Murray BR, Zeppel MJB, Hose GC, Eamus D. 2003. Groundwaterdependent ecosystems in Australia: its more than just water for rivers.
Ecological Management and Restoration 4: 110113.
Naumburg E, Mata-Gonzalez R, Hunter RG, McLendon L, Martin DW.
2005. Phreatophytic vegetation and groundwater uctuations: a review
of current research and application of ecosystem response modeling
with an emphasis on great basin vegetation. Environmental Management
35: 726740.
OGrady 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: 591597.
OGrady AP, Eamus D, Cook PG, Lamontagne S. 2006. Groundwater use
by riparian vegetation in the wet-dry tropics of northern Australia.
Australian Journal of Botany 54: 145154.
OGrady AP, Cook PG, Eamus D, Duguit A, Wischusen JDH, Fass T,
Worldege D. 2009. Convergence of tree water use within an arid-zone
woodland. Oecologia 160: 643655.
Orellana FP, Verma SP, Loheide II, Daly E. 2012. Monitoring and
modeling water-vegetation interactions in groundwater-dependent
ecosystems. Reviews of Geophysics 50: RG3003. DOI: 10.1029/
2011RG000383.
Pfautsch S, Bleby TM, Rennenberg H, Adams MA. 2010. Sap ow
measurements reveal inuence of temperature and stand structure on
water use of Eucalyptus regnans forests. Forest Ecology and
Management 259: 11901199.
Pfautsch S, Keitel C, Turnbull TL, Braimbridge MJ, Wright TE, Simpson
RR, OBrien JA, Adams MA. 2011. Diurnal patterns in water use of

Copyright 2014 John Wiley & Sons, Ltd.

Eucalyptus victrix indicate pronounced desiccation-rehydration cycles

despite unlimited water supply. Tree Physiology 31: 10411051.
Pfautsch S, Macfarlane C, Ebdon N, Meder R. 2012. Assessing sapwood
depth and wood properties in Eucalyptus and Corymbia spp. Using
visual methods and near infrared spectroscopy (NIR). Trees 26:
963974.
Phillips NG, Lewis JD, Logan BA, Tissue DT. 2010. Inter- and intra-specic
variation in nocturnal water transport in Eucalyptus. Tree Physiology
30: 586596.
Porporato A, DOdorico P, Laio F, Ridol L, Rodriguez-Iturbe I. 2002.
Ecohydrology of water-controlled ecosystems. Advances in Water
Resources 25: 13351348.
Reich PB. 2002. Rootshoot relationships: optimality in acclimation and
allocation or the emperors new clothes? In Plant Roots: The Hidden
Half, Waisel Y, Eshel A, Kafka U (eds). Marcel Dekker: New York;
205220.
Scott ML, Shafroth PB, Auble GT. 1999. Responses of riparian
cottonwoods to alluvial water table declines. Environmental Management
23: 347358.
Scott RL, Edwards EA, Shuttleworth WJ, Huxman TE, Watts C, Goodrich DC.
2004. Interannual and seasonal variation in uxes of water and carbon
dioxide from a riparian woodland ecosystem. Agricultural and Forest
Meteorology 122: 6584.
Shafroth PB, Stromberg JC, Patten DC. 2000. Woody riparian vegetation
response to different alluvial water table regimes. Western North
American Naturalist 60: 6676.
Smith SD, Devitt DA, Sala A, Cleverly JR, Busch DE. 1998. Water
relations of riparian plants from warm desert regions. Wetlands 18:
687696.
Snyder RL, Shaw RH. 1984. Converting humidity expressions with
computers and calculators. Leaet 21372, Cooperative Extension, Div.
Agric. Nat. Res., University of California.
Sommer B, Froend R. 2011. Resilience of phreatophytic vegetation to
groundwater drawdown: is recovery possible under a drying climate?
Ecohydrology 4: 6782.
Souter NJ, Cunningham S, Little S, Wallace T, McCarthy B, Henderson M.
2010. Evaluation of a visual assessment method for tree condition of
eucalypt oodplain forests. Ecological Management and Restoration 11:
210214.
Swanson RH. 1983. Numerical and experimental analyses of implantedprobe heat pulse velocity theory. Dissertation, University of Alberta.
Tyree MT, Ewers FW. 1991. The hydraulic architecture of trees and other
woody plants. New Phytologist 119: 345360.
Whitehead D, Beadle CL. 2004. Physiological regulation of productivity
and water use in Eucalyptus: a review. Forest Ecology and
Management 193: 113140.
Zencich SJ, Froend RH, Turner JV, Gailitis V. 2002. Inuence of
groundwater depth on the seasonal sources of water accessed by Banksia
tree species on a shallow, sandy costal aquifer. Oecologia 131: 819.

SUPPORTING INFORMATION
Additional supporting information may be found in the online
version at the publishers web site.

Ecohydrol. (2014)