doble r et al 2006 spatial relationships between vegetation cover and irrigation

7/27/2019 Doble r Et Al 2006 Spatial Relationships Between Vegetation Cover and Irrigation

1/23

Spatial relationships between vegetation cover and

irrigation-induced groundwater discharge on a

semi-arid floodplain, Australia

Rebecca Doble a,b,*, Craig Simmons a, Ian Jolly b, Glen Walker b

a Flinders University of South Australia, Australiab CSIRO Land and Water, PMB 2, Glen Osmond, South Australia 5064, Australia

Received 16 June 2004; received in revised form 31 January 2006; accepted 3 February 2006

Summary Insight into spatial patterns of groundwater discharge through evapotranspiration is

critical for understanding responses of groundwater dependent ecosystems. In the semi-arid

floodplain environment of the lower River Murray, South Australia, patterns of steady state

groundwater flux may indicate potential areas of anthropogenic salinisation due to adjacentirrigation and aid in management decisions for the protection of indigenous floodplain vegeta-

tion. This paper outlines the use of relationships for evapotranspiration response to groundwa-

ter depth and salt concentration, within the MODFLOW 2000 framework, to model spatial

patterns of groundwater flux. This modelling approach is used to develop relationships between

the steady state net rates evapotranspiration, seepage and baseflow, and floodplain properties

including elevation, geometry, soil and vegetation distribution relevant to the field site.

The model was applied to a key research site on the River Murray, South Australia, to facil-

itate understanding of the factors contributing to the development of spatial patterns of irri-

gation induced groundwater discharge. The inclusion of elevation and vegetation data in the

methodology revealed the effects of microtopography and plant water use on groundwater dis-

charge. Spatial distribution of vegetation had the highest correlation with patterns of ground-

water evapotranspiration. Floodplain elevation, floodplain width and geometry, which related

to vegetation cover, also affected both groundwater discharge and baseflow to the river.c 2006 Elsevier B.V. All rights reserved.

KEYWORDSEvapotranspiration;

Spatial distribution;

Floodplaingeomorphology;

Vegetation;

Numerical modelling;

MODFLOW 2000

Introduction

In arid and semi-arid regions, floodplains are ecologicallyand hydrologically important parts of a catchment. Due totheir proximity to rivers, floodplains and their associated

0022-1694/$ - see front matter c 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.jhydrol.2006.02.007

* Corresponding author. Tel.: +61 8 8303 8705; fax: +61 8 83038750.

E-mail address: [email protected] (R. Doble).

Journal of Hydrology (2006) 329, 7597

a v a i l a b l e a t w w w . s c i e n c ed i r e c t . c o m

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j h y d r o l
mailto:[email protected]:[email protected]


2/23

wetlands have higher biodiversity value than surroundingareas, providing habitat for many aquatic and riparian spe-cies, including vegetation, birds, mammals and fish species.In lowland gaining streams groundwater flowing from thecatchment moves through the floodplain before either flow-ing to the river or being attenuated by evapotranspiration.1

The shallow watertables found within floodplains mean thatevapotranspiration rates are significant, and understanding

the nature of groundwater discharge processes throughevapotranspiration and seepage in this environment is vitalto produce an accurate water balance for the system.

In semi-arid areas where saline regional groundwater dis-charges to streams, such as the lower River Murray in SouthAustralia, evapotranspiration from rising floodplain waterta-bles and altered flow regimes have led to the dieback ofenvironmentally significant riparian vegetation such as redgum (Eucalytpus camaldulensis) and black box (Eucalytpuslargiflorens) (Jolly et al., 1993). The naturally saline regio-nal groundwater tables have risen as a result of higher riverlevels and a three orders of magnitude increase in rechargefrom adjacent irrigation developments, leading to the accu-mulation of salt in the soil and alluvial aquifer through evap-

oconcentration (Jolly, 1996). The hydrology, geomorphologyand vegetation coverage of these floodplains are complex,leading to high spatial variability in evapotranspiration,and variation in vegetation health impacts.

Management of vegetation health is undertaken throughpumping regional groundwater to lower riparian waterta-bles, targeted spear point pumping under key eucalypt com-munities, injection of fresh water under communities, and

localised artificial flooding or environmental irrigation.These management techniques require an understandingof spatial patterns of salinisation risk on a fine scale, andthe true ecological impact of land management and irriga-tion decisions requires knowledge of the proximity of salin-isation risk to vegetation communities. The ability toquantify spatial patterns of groundwater discharge throughmodelling and site characterisation is required in order to

understand the potential impacts from salinisation on vege-tation. To do this, a conceptual model of discharge that ac-counts for the variability in elevation, soils and vegetationcover on floodplains is needed.

Groundwater loss due to evapotranspiration, however, isknown to be the most difficult component of the groundwaterbalance to estimate (Freeze and Cherry, 1979). Much soil sci-ence research of the 19501980s was focused on the under-standing, quantification and representation of capillaryupflow from shallow watertables (Gardner, 1958; Warrick,1988), and more recently transpiration within the contextof environmental science (Thorburn et al., 1995; Mac Nishet al., 2000) and silviculture (Silberstein et al., 1999;Feikema, 2000). Despite this understanding, numerical mod-

elling of groundwater loss has tended to simplify theevapotranspiration process to the point of representing it asa simple sink with no interface definition, or as linearapproximations.

Where patterns of groundwater evapotranspiration rep-resenting salinisation risk at a fine scale are of interest,however, simplified models of groundwateratmosphereflux can limit the understanding of spatial variation and het-erogeneities. A model that incorporates the key factorsinfluencing evapotranspiration, including vegetation type,groundwater salinity, watertable depth and floodingfrequency is required to investigate the origins of spatialpatterns of groundwater discharge and salinisation.

Nomenclature

Parameter Definition (dimensions)a parameter in the K(S) function of (Gardner,

1958) (Ln+1T1)A aAn (Warrick, 1988) (L

n+1T1)An [(p/n)csc(p/n)]

n (Warrick, 1988) ()

Cg groundwater solute concentration (ML3

)Cth threshold soil water solute conc. at which plants

can no longer extract water (ML3)d, Zg depth to groundwater from the soil surface (L)dext extinction depth, below which there is no fur-

ther discharge from the groundwater (L)D cumulative discharge through plant water up-

take (L)Dh confined piezometric head elevation above

ground surface (L)K hydraulic conductivity (LT1)Ksat saturated hydraulic conductivity (LT

1)L aquifer thickness (L)n exponent in the K(S) function of (Gardner,

1958), dependent on soil characteristics ()qlim maximum steady state moisture flux through the

soil (LT3)

q groundwater flux, positive upward (LT3)qmax maximum discharge rate when groundwater is at

the soil surface (LT3)R cumulative recharge to groundwater (L)

t the point in time at which groundwater flux iscalculated (T)

td the length of time between flood inundation (T)tr duration for which the site is inundated (T)U plant groundwater uptake flux (LT1)Zg, d depth of groundwater below the soil surface (L)Zf depth of salt front (plane of evaporation) below

the soil surface (L)c Zf/Zg ()h0 initial water content (L

3L3)h1 water content within capillary zone, approxi-

mately field capacity (L3L3)hFC water content at field capacity (L

3L3)hsat water content under saturated conditions

(L3L3)

1 For the remainder of this paper, any reference to evapotrans-piration refers to capillary upflow from shallow groundwater, ratherthan direct evaporation of rainfall.

76 R. Doble et al.


3/23

The objectives of this study were therefore to modelevapotranspiration at a scale and complexity appropriatefor understanding salinity effects on vegetation health,and to use these results to understand the relative contribu-tion of physical, ecological and hydrological parameters ingoverning groundwater flux distribution.

Specifically, the aims of the paper were to:

Use relationships for evapotranspiration in semi-aridenvironments within a finite difference groundwatermodel in order to understand the long-term patterns ofevapotranspiration and salinisation for the purpose ofdetermining vegetation health in a floodplain environ-ment; and

Develop a conceptual understanding of the hydrologicalrelationships between regional groundwater inflows, veg-etation water use and baseflow to the river for a keyresearch site to understand the factors that governgroundwater flow processes within a floodplainenvironment.

The first part of this study briefly describes the modelling

process undertaken, including the use of a spatially distrib-uted evapotranspiration relationship as an input for MOD-FLOW 2000 in order to facilitate better representation ofgroundwater discharge fluxes. This section provides a back-ground to the modelling, and a more detailed description ofthis methodology is available in Doble et al. (2005).

The second section of this paper applies the model to afloodplain on the River Murray, South Australia in order tounderstand the floodplain hydrological processes. Model re-sults are compared against spatial patterns of dischargefrom remote sensing images and fine scale baseflow datafrom geophysical measurements. A conceptual understand-ing of the relative importance of factors such as floodplain

elevation, geometry, soil and vegetation type on the distri-bution of discharge is then developed for the field site.

Site description

Clarks Floodplain, adjacent to the Bookpurnong IrrigationDistrict (Fig. 1), is located on the lower reaches of the RiverMurray in South Australia (34210S, 140370E). The field siteis located within the semi-arid inland of Australia, with rain-fall varying between 200 and 300 mm a1 and potentialevaporation of approximately 1800 mm a1.

The lower River Murray floodplain is characterised by aflat, wide, meandering river within a deep river valley, ex-

cised during the early to middle, and then again in the latePleistocene, and redeposited with coarse-grained quartzsands of the Monoman Formation during higher, post-glacialsea levels between 17,000 and 6000 BP (Twidale et al.,1978). In the late Holocene (approximately 4000 BP), theCoonambidgal Clay, typically consisting of clays and silts,was deposited in the river valley overlying and partly confin-ing the Monoman Formation (Fig. 2). The modern floodplainsurface tends to have rapid variation in soil type due to geo-morphological processes of deposition and erosion that arecontinually occurring. Palaeochannels and backwatersformed from previous river courses are present in manyparts of the floodplain.

The hydrogeology of Clarks Floodplain is typical of theeastern part of the lower River Murray (Jarwal et al.,1996). The Coonambidgal Clay ranges from 3 to 7 m thick,while the Monoman Formation is approximately 7 m thickin this area. The cliffs adjacent the floodplains consist ofa layer of Woorinen Sands over Blanchtown Clay, eachapproximately 2 m thick (not represented in the model),overlying a layer of Loxton Sands up to 35 m in depth. The

whole area is underlain by the Bookpurnong Beds, whichact as an aquitard basement to the shallow aquifer thatencompasses the Monoman Formation and Loxton Sands.Groundwater salinity in the Loxton Sands and Monoman For-mation is in excess of 30,000 mg L1, while irrigation re-charge salinity is typically 5000 mg L1.

The Bookpurnong Irrigation District comprises more than1500 ha of irrigated land, with major irrigation having com-menced in the mid-1960s, and has been increasing sincethat time (Telfer and Overton, 1999). Excess recharge fromthe irrigation area has led to the formation of a groundwatermound, which displaces saline groundwater towards thefloodplain and has led to increased waterlogging and salini-sation, and groundwater seepage at the break of slope adja-

cent to the cliffs. Aerial photographs show that extensivedegradation has taken place between 1972 and 1998, withmost of the vegetation health decline occurring at the backand centre of the floodplain. This is associated with theseepage and presence of a saline backwater that is thoughtto intercept groundwater and concentrate salt further(Fig. 3). Black box (E. largiflorens) and red gum (E. camal-dulensis) tree communities have been most affected bythe salinisation of the floodplain.

Clarks Floodplain is situated below a weir (Lock 4;Fig. 1), and therefore the occurrence of vegetation healthdecline due to weir-induced water logging is minimal. Thispaper therefore focuses specifically on irrigation-induced

salinity.

Simulation of groundwater flux patterns

The intent of the modelling in this section of the paper wasto use a better representation of evapotranspiration pro-cesses within a finite difference groundwater model to de-velop a conceptual understanding of the processes leadingto spatial patterns of evapotranspiration. A more detaileddiscussion of the model development is given in Dobleet al. (2005).

Conceptualisation

The methodology used in this paper involves representingspatial distributions of net groundwater flux by developinga relationship to use within a finite difference groundwaterflow model, in this case MODFLOW 2000. This relationshipuses spatial information for vegetation, soil and groundwa-ter salinity distributions, as well as flooding duration and fre-quency to develop a spatially variable, depth-dependent ETfunction that can be used within MODFLOW. The function,which provides a surface boundary condition to the ground-water model, is similar to the segmented evapotranspirationfunction (ETS1, Banta, 2000) but also incorporates a re-charge component, with the equilibrium depth, rather than

Spatial relationships between vegetation cover and irrigation-induced groundwater discharge 77


4/23

extinction depth describing a net flux of zero. The net rate ofgroundwater flux (recharge subtracted from evapotranspira-

tion) is important for understanding the long-term or ulti-mate rate of salt accumulation, and is a common measureof salinisation potential (Petheram et al., 2003; Slavichet al., 1999). Similarly, the equilibrium depth, where thenet flux is zero is critical for determining whether salinisa-tion will occur. The model is run at steady state, in orderto understand the ultimate salt accumulation potential forfuture riverine and vegetation management planning.

The intent of the study was not produce a model to accu-rately represent field conditions for management purposes,and to calibrate the model to produce a close fit with histor-ical data. Instead, the aim was to conceptualise relation-ships between floodplain groundwater processes, leading

to patterns of evapotranspiration. The modelling thereforerequired, rather than full calibration, only that the model

parameters were within a reasonable range, likely to befound in the field situation being modelled. Spatial andnumerical validation of the model was performed using aseries of geophysical, remote sensing and field measure-ment techniques. A brief description of the spatial valida-tion is given later in the paper in order to justify themodel results.

Previous solutions for groundwater discharge

To determine a threshold for salinisation due to steady stateevaporation of saline groundwater, Warrick (1988) proposedthat the limiting rate of upward flow in a soil was inversely

Figure 1 Clarks Floodplain on the lower River Murray, South Australia (3421 0S, 14037 0E) showing irrigation areas of the

Bookpurnong Irrigation District and flow divisions (numbered 111). Local grid refinement graduates from 200 m by 50 m on the

highland region, to 50 m by 50 m on the floodplain, to 6.25 m 6.25 m at the break of slope in the region indicated by the dashed

line.

78 R. Doble et al.


5/23

proportional to the depth to the watertable but variedaccording to soil properties, as described by:

qlim Anadn. 1

From Eq. (1), the maximum moisture flux or diffuse dis-charge from a soil can be determined for varying groundwa-ter depths. Shallower groundwater will lead to higher

evaporation rates, and therefore greater concentration ofsalts in the surface soil.

The science of predicting both dryland and irrigation in-duced soil salinisation using groundwater discharge rateshas already been established in soil physics, but is less fre-quently described by groundwater hydrologists. Using rela-tionships described by Gardner (1958), Wind (1955) and

Figure 2 Geological cross section of Clarks Floodplain, indicating the excised river valley, Monoman Formation alluvial aquifer,

and Coonambidgal Clay which overlies much of the floodplain. Increased recharge from irrigation taking place on the highland causes

a groundwater mound to form adjacent the floodplain, leading to shallow groundwater and seepage at the break of slope.

Figure 3 Vegetation health on the floodplain is shown in greyscale with black representing dead vegetation, dark grey signifying

vegetation in poor health, light grey representing healthy vegetation and white designating no vegetation present. For ClarksFloodplain, vegetation health decline was observed at the back of the floodplain and the saline backwater, and extends out towards

the centre of the floodplain. Most of the floodplain peninsula areas support healthy vegetation.



6/23

others, Talsma (1963) and Peck (1978) defined thresholdsfor salinisation in terms of groundwater fluxes of103 m d1 for irrigated areas and 104 m d1 for drylandconditions, respectively. This definition relates salinisationto rates of groundwater discharge and accounts for soil typeand groundwater depth.

In numerical models, a linear function, or a constant fluxis often used to represent evapotranspiration. MODFLOW 96

(McDonald and Harbaugh, 1996) assumes a linear evapo-transpiration function where evapotranspiration varies froma maximum rate at the ground surface, to zero when thewatertable is at some depth below the surface (extinctiondepth), such that:

q qmax; d6 0;

q qmax 1 d

dext

; 0 < d< dext;

q 0; dP dext.

2

The linear nature of this relationship may cause discharge tobe under-predicted where the watertable is deeper, andover-predicted where it is shallow, a problem when knowl-

edge of groundwater discharge is required on a fine scale.MODFLOW 2000 includes the module ETS1, which representsevapotranspiration as a piecewise linear function (Banta,2000), which is used to prevent this problem. Setting upan accurate curve, however, requires knowledge of the spa-tially varying floodplain conditions.

Eq. (1) describes unsaturated zone groundwater move-ment at a point scale. Difficulties arise in applying thismodel at a regional, or even local scale to consider spatialdistribution of groundwater discharge. The linkage ofpoint-scale unsaturated models with large-scale saturatedgroundwater flow models has been trialled extensively,but these models tend to have stability issues due to thesensitivity of the Richards Equation to lower boundary fluxes(Boone and Wetzel, 1996). Scaling up data from point scaleunsaturated models to local scale groundwater models re-quires some means of extrapolation (Hatton, 1998) to matchspatial and temporal scales, and the processes that are rel-evant to the system can change between small scales of theunsaturated model and large scales of the saturated model.

The methodology used in this paper employs an equationfor groundwater discharge from plants developed for saline,semi-arid conditions, plus information on recharge rates andflood recurrence patterns, and applies it spatially to theMODFLOW 2000 ETS module using a geographic informationsystem (GIS).

Recharge and dischargeAn analytical model for the uptake of saline groundwater byplants in semi-arid regions was developed by Thorburn et al.(1995), based on unsaturated zone water and solute bal-ances. In this model, discharge of groundwater by vegeta-tion depends on soil water content, groundwater salinity

and the salinity threshold at which vegetation can no longerextract groundwater derived soil water. This relationshipfor instantaneous flux is described by the equation:

U A Zg1 cn1

n 1ACgtdh1Cth

nn1

1 Cgh1 h0

h1Cth

.

3

The cumulative groundwater discharge (D) through plant

water uptake is found by integrating Eq. (3) over the timeperiod 0 to t:

D Zg1 cn1

n 1ACgtdh1Cth

1n1

Zg1 c

!

h1Cth

Cg 1

h0

. 4

For a floodplain environment, this represents the cumula-tive discharge occurring for the continuous period betweenflood cycles, where td represents the length of the dry per-iod between the end of one period of inundation (t = 0) tothe start of the next (t = td). The length of this dry period

is critical for vegetation dependent on groundwater, asevapotranspiration processes will concentrate salt, whichin turn limits plant water uptake. When an overbank flowevent occurs, recharge from flooding will freshen soil andground water, allowing plant water uptake to increase.More frequent flooding (shorter td) will therefore supporthigher volumes of discharge in the long term than wouldbe possible with longer dry periods.

Recharge into the floodplain will be dependent on boththe vertical conductivity of the soil and the period for whichthat part of the floodplain is inundated. This volume of infil-tration, less the water volume that fills the unsaturatedzone to field capacity, gives the net volume of waterrecharging the groundwater. Recharge may be represented

by:

R Ksattr hFC h0Zf hFC h1Zg Zf

if hsat hFCZg > 0. 5

For simplification, the moisture content within the capillaryfringe (h1) may be approximately equal to the moisture con-tent at field capacity (hFC), therefore the second term willcancel. Groundwater recharge in conductive soils will alsobe limited by the total unsaturated pore space that maybe filled during floods. The equation for recharge is there-fore the minimum of the following equations:

Minimum: R Ksattr h1 h0Zf;

R hsat

h1Z

g;

6

where (hsat h1) represents the pore space available to re-ceive recharge water.

Combining recharge and discharge for a cycle of oneflood and one discharge period, an approximation of thenet groundwater flux from plant water use (q) is:

q

minKsattr h1 h0Zf

hsat h1Zg

Zg1 c

n1 n1ACgtd

h1 Cth

1n1

Zg1 c

h1

CthCg

1

h0

tr td

.7

80 R. Doble et al.


7/23

Eq. (7) reflects the long term balance between groundwaterdischarge during dry periods, and recharge during floods.The relationship contains information on evapotranspirationfrom groundwater dependent ecosystems, and flood fre-quency and duration patterns, evaluated as a long termtrend to determine whether areas are dominated by re-charge or discharge, dependent on depth to groundwater.In areas with saline regional groundwater, discharge domi-nance will lead to secondary salinisation.

This equation describes a relationship between ground-water flux and watertable depth that can be used within afinite difference groundwater flow model as a surfaceboundary condition. The function was used to calculate in-put parameters for the ETS1 module using GIS coveragesof elevation, soil and vegetation conditions for each partof the floodplain and each model cell. Data used for thisspatial calculation is shown in Table 1, and durations ofinundation and dry periods (td and tr) provided from floodrecurrence data.

Seepage

At many points along the river valley, the high incoming re-gional groundwater inflows cause the watertable to inter-sect the ground surface at the break of slope, causingdirect seepage of groundwater. Seepage at the break ofslope was modelled with an additional segment of the

evapotranspiration function, such that the maximum ETrate at the soil surface was equal to the potential evapo-transpiration for the region (Table 2). At a point 0.2 m be-low the surface, the evapotranspiration function revertedto the groundwater flux relationship appropriate for the soiland vegetation conditions.

Confining layers

For Clarks Floodplain, and most areas of the lower RiverMurray, the Coonambidgal Clay forms a confining layer overthe Monoman Sands aquifer. The effect of this confininglayer is to increase the pressure head required to forcegroundwater to the surface, or high enough that the capil-

lary fringe reaches the surface for evaporation to occur.This process may be modelled within MODFLOW using a falseevaporation surface above the actual ground surface to rep-resent this additional head requirement. This stretches therechargedischarge curve both upwards above the equilib-rium point, and downward within the recharge section be-low the equilibrium point.

The degree of upward stretching is given by the Darcyequation, which specifies the head required for maximumgroundwater flux at the surface. That is:

q KDh

L. 8

Rearranging and substituting into Eq. (4) for the condition ofthe watertable at the soil surface, and therefore K= Ksatgives:

Dh L

KsatETmax; 9

where recharge is zero, or:

Dh L

Ksat

n 1ACgtdh1Cth

1n1

h1Cth

Cg 1

h0

. 10

For the groundwater flux curves, the maximum rates ofevapotranspiration and recharge remain the same.

Coupling with MODFLOW

The relationships described in the previous sections are ap-plied to MODFLOW 2000 (Harbaugh et al., 2000) through amodified version of the ETS module (Banta, 2000), whichcombines both recharge (RCH) and discharge (ETS) togetheras a single groundwater flux dependent on groundwaterdepth (Fig. 4). The resulting combined rechargedis-charge (CRD) module is a simple additive combination ofthese two functions.

The function applies a net recharge or discharge to themodel, depending on whether the groundwater head isabove the ground (max ET), between the surface and

Table 1 Plant water use and soil parameters from Thorburn et al. (1995) and Jolly et al. (1993)

Parameter Values for each soil, vegetation, groundwater category

Soil: Silty sand Clayey sand Sandy clay Clay Palaeochannel

hsat (v/v) 0.35 0.4 0.45 0.6 0.6

h1 (v/v) 0.2 0.28 0.34 0.4 0.44

h0 (v/v) 0.15 0.2 0.25 0.35 0.35

a (mn d1) 0.02 0.01 0.005 0.001 0.001

n 5 3 2.9 2.5 2.5

A (mn d1) 0.028 0.018 0.009 0.002 0.002

Ksat (m d1) 0.05 0.01 0.005 0.001 0.001

Vegetation: Red gum Black box Lignum Grassland

Cth (g L1) 8 16 6 7

Groundwater: Zone 1, extreme EC Zone 2, high EC Zone 3, moderate EC Zone 4, low EC

Zg (m) 0.5 3 3 3

Zf (m) 0.1 0.4 0.5 0.6

c () 0.2 0.13 0.17 0.2

Cg (g L1) 20 15 8 2

http://-/?-http://-/?-


8/23

equilibrium depth (a) (net ET), between the equilibriumdepth and extinction depth (b) (net recharge) or below theextinction depth (max recharge). The point at which ground-water discharge is equal to zero is termed the equilibriumdepth, that is where recharge and discharge fluxes are equalin the long term. With this method of defining groundwaterflux it is possible to include varying recharge functions suchas zero recharge when groundwater is at the surface.

Vegetation input parameters were applied spatially usinga GIS vegetation survey (Telfer and Overton, 1999), and soiland groundwater parameters using a map of soil types de-

rived from soil logs and geomorphology, and groundwaterchloride concentration data. Each of the three coverageswas divided into categories shown in Table 1 and assignedvalues for parameters used in Eq. (7). A GIS frameworkwas used to calculate four points on the CRD curve, definingmaximum ET, maximum recharge, and equilibrium depth.Spatial distributions of maximum ET, maximum rechargeand equilibrium depth were used as inputs for MODFLOW,which provided the hydrological component and groundwa-ter depth required to calculate long term groundwater flux.

MODFLOW model development

Model parameters and their sources are listed in Table 2.Fig. 5 shows an example cross-section of the model, includ-ing the geological layers used. On the highland, the upper,

unconfined aquifer consisted of Upper Loxton Sands(Kh = 5 m d

1), which was underlain by the Lower LoxtonSands (Kh = 0.2). On the floodplain, the alluvial aquifer with-in the Monoman Formation (Kh = 10) is again underlain byLower Loxton Sands. Overlying the alluvial aquifer is a spa-tially variable semi-confining layer, consisting of heavy Coo-nambidgal Clay (K variable) at the break of slope, to a siltysand deposit near the river. Thicker Coonambidgal Clay wasfound in some low lying palaeochannels. Vertical hydraulicconductivity and other soil properties used for the evapo-transpiration function (Eq. (7)) are shown in Table 1.

The MODFLOW model was discretised into a regular,square grid with cell dimensions ranging from 200 50 m

Table 2 Model parameters

Parameter Source Range Adopted Value

Recharge and discharge parameters

Irrigation recharge AWE (1999) 133243 175 mm a1

Dryland recharge AWE (1999) 0.1 0.1 mm a1

Maximum evapotranspiration

rate on floodplains

Thorburn (1996); Thorburn et al. (1993) 40440 Black box

360720 Red gum

Ranges spatially mm a1

Extinction depth Thorburn (1996), Talsma (1963) 17 Ranges spatially 1.55 mOpen water evaporation rate Bureau of Meteorology 17001800 mm a1 1800 mm a1

Seepage extinction depth 0.10.5 0.2 m

River and elevation parameters

River stage upstream of Lock 4 SA Water 13.2 m 13.2 m

River stage downstream of Lock 4 SA Water 9.810.5 m 10.5 m

Floodplain elevation Overton (2005) 12 m15 m 12 m15 m

Highland elevation Overton (2005), Armstrong et al. (1999) 3540 m 40 m

Aquifer parameters

Upper Loxton Sands Armstrong et al. (1999), AWE (1999),

Jolly et al. (1998), Evans and Kellett

(1989), Barnett and SA Department

of Mines and Energy (1991)

Kh 110 5 m d1

Kv 0.0351 1 m d1

Coonambidgal ClayKh 0.050.1 0.1 m d1

Kv 0.0050.01 Ranges spatially m d1

Monoman Sands

Kh 1035 10 m d1

Kv 0.351 3 m d1

Lower Loxton Sands

Kh 0.054 0.2 m d1

Kv 0.0050.4 0.2 m d1

Recharge

Depth to

groundwater

Discharge Groundwater Flux

(a)

(b)

Figure 4 Combined recharge and discharge in MODFLOW

showing: (a) the equilibrium depth where ET is equal to

recharge, and (b) the extinction depth or maximum recharge,

below which the ET component of flux is zero.

82 R. Doble et al.
http://-/?-http://-/?-http://-/?-http://-/?-


9/23

to 6.25 6.25 m, refined about the break of slope, as indi-cated in Fig. 1. The model used constant head cells on theouter boundaries with elevations of 12.5 m, 13.9 m,15.2 m and 12.5 mAHD, starting from the north-western cor-ner and progressing clockwise around the corners of themodel domain. Floodplain elevation data was available inspatial form at vertical intervals of 100 mm on a 30 30 mgrid cell scale (Fig. 6, Overton, 2005). The river boundarywas simulated using the MODFLOW river module. River stagewas set at an elevation of 10.5 m below the weir and 13.2 mabove it. The river is regulated into a series or weir pools,and river stage variation is minimal for the length of the

floodplain. The model was run in steady state for currentirrigation conditions, both for confined and unconfined allu-vial aquifer states.

Vegetation and soil parameters from Table 1 wereadopted from Thorburn, 1996, and extrapolated to sandiersoils using Gardner n parameters provided by Jolly et al.(1993) from field experiments. Groundwater parameterswere taken from field measurements in March 2003. Thespatial distribution of soil, vegetation and groundwaterparameters are shown in Fig. 7ac.

Historical flood data was used to determine the averagereturn period between floods, and the duration of inunda-tion for various river stages. These values were then usedin Eq. (7). Data for River Murray flows into South Australiawas obtained for the years 19572003 (post regulation)

from gauging station GS426200SA (Porter, 2002). A ratingcurve measured at a point downstream of Lock 4 corre-sponding to Clarks Floodplain was used to link flow

Figure 5 Cross-section of the conceptual model showing the partially confined Monoman Sands alluvial aquifer and the highland

Loxton Sands Formation that form the main groundwater flow path toward the river (not to scale). Note the intersection of the

ground surface by the piezometric surface at the break of slope, resulting in seepage.

Figure 6 Floodplain elevation distribution, showing surface elevation varying between 11 and 15 m. Low lying areas are found at

the ends of the floodplain peninsulas and in the centre of the floodplain associated with the saline backwater.

http://-/?-http://-/?-


10/23

Figure 7 Distribution of: (a) groundwater chloride concentrations (g L1) for Clarks Floodplain obtained from field data, (b) soil

type distribution, and (c) surveyed vegetation type, used as spatial inputs for the combined rechargedischarge function. Parameter

values associated with these distributions are shown in Table 1.

84 R. Doble et al.


11/23

frequency data to the spatial coverage of floodplain eleva-tion (Fig. 6). Fig. 8 shows the resulting smoothed relation-ship for average period of inundation and duration

between flooding as a function of floodplain elevation. Thismeans of including flood information allows the long termtrends of the rechargedischarge balance to be investigated.

Spatial distributions of the MODFLOW CRD module in-puts: maximum ET, equilibrium depth and maximum re-charge, were produced using Eq. (7), and the spatial datadescribed above. Fig. 9 shows examples of various re-chargedischarge relationships that cover several differentvegetation, soil and groundwater scenarios on floodplain.The locations at which data for these curves were selectedare indicated in Fig. 7a.

The MODFLOW model was run twice more using uniformevapotranspiration and elevation parameters, the resultsof which were used to determine how much additional infor-mation was gained by considering variation of the floodplainsystem at a finer scale. The first run consisted of a constantfloodplain elevation of 14 m (herein referred to as the uni-form model), and the second with a spatially variable sur-face elevation (variable elevation model, Fig. 6), but withevapotranspiration not defined by vegetation parameters.Both models used a constant maximum ET rate of0.3 mm d1 and an equilibrium depth of 3.5 m, zero flood-plain recharge and a constant vertical hydraulic conductiv-ity (0.01 m d1).

MODFLOW model results

A spatial distribution of net groundwater flux was producedusing the methodology described above (Fig. 10). This wascompared with the flux outputs from the uniform (Fig. 11)and variable elevation (Fig. 12) models. Inclusion of eleva-tion, vegetation, soil type and flooding information withinthe MODFLOW model provided finer-scale predictions fordischarge patterns and revealed a variation in dischargerates across the floodplain. Not only could areas at risk ofsalinisation be identified, but variations in the rate of salin-isation were also quantifiable.

The MODFLOW groundwater flow model predicted highnet discharge of groundwater at the back of the floodplain,near the break of hill slope where the depth to groundwater

was at a minimum (Fig. 10). High groundwater dischargewas also associated with red gum and black box forests at

the ends of floodplain peninsulas and adjacent the river indivisions 1, 7 and 8. Flow divisions used to describe dis-charge locations are numbered in Fig. 1. Localisedelevation, or microtopography, was revealed by high evapo-transpiration rates in the spatial patterns of groundwaterdischarge, both at the break of slope, and associated withthe low-lying palaeochannel that extended along the centr-eline of Clarks floodplain in division 7. Patterns of elevationwere evident within vegetation classes, for example com-paring the elevation profile at the river end of division 3(Fig. 6) with the groundwater discharge patterns inFig. 10. The effect of elevation between vegetation commu-nities was not evident.

Consecutive Inundated and Dry Periodsvs Floodplain Elevation

5000

1000

1500

2000

2500

3000

12 12.5 13 13.5 14 14.5 15

Floodplain Elevation (mAHD)

td

)d(

0

20

40

60

80

100

trd

td tr

Figure 8 Average consecutive days between floods and

during inundation as a function of floodplain elevation. Once

assigned spatially, linked to the floodplain elevation of each

polygon, this information provided the data for td and tr to be

used in Eq. (7).

GW Flux vs Depth

0

1

2

3

4

5

6

7

8

9

10

-4 -3 -2 -1 0 1

GW Flux (mm/d)

)m(htpeD

WG

RG Z4 BB Z3 L Z2

GL Z1 RG, FO Z3

Figure 9 Groundwater flux versus depth relationships forvarious floodplain vegetation, soil and groundwater scenarios,

including: red gums found on clayey sand with low salinity

groundwater (RG Z4), black box community on sandy clay with

moderate salinity groundwater (BB Z3), lignum on heavy clay,

with high salinity groundwater (L Z2), grassland on the heavy

clay flood out at the break of slope, with extremely saline,

shallow groundwater (GL Z1), and red gum communities found

at the river end of the heavy clay flood out, with moderate

salinity groundwater (RG FO Z3).

http://-/?-http://-/?-http://-/?-http://-/?-


12/23

Validation

In light of the objectives of this paper, the validation pro-

cess described below provides a means of establishing thatboth the modelled spatial patterns of evapotranspirationand the rates of groundwater flux lie within a range that isappropriate for the field conditions under investigation.Since the model was not used for detailed floodplain man-agement, a through calibration with historical data was nei-ther necessary nor productive. This validation is thereforeintentionally concise and indicative in order to give confi-dence that the processes representing evapotranspirationare adequate, rather than providing detailed calibrationresults.

Higher rates of salinisation from evapotranspiration atthe break of slope and through the centre of the floodplain

are indicated by dead and unhealthy vegetation in Fig. 3, aswell as observations of seepage and significant accumula-

tion of salt in the soil profile during field investigations.Point scale transpiration measurements from the field

were compared with modelled results. Sap flow measure-ments from Thorburn and Walker (1994), Jolly et al.(1991) and Akeroyd et al. (1998) indicate transpiration rateswithin the lower River Murray region from 70 to 700 mm a1,and 35 to 150 mm a1 for red gum and black box eucalyptspecies, respectively. Evapotranspiration from bare soilson Chowilla Floodplain in the same region of the River Mur-ray was approximated using 2H isotope profiles to be10 mm a1 (Jolly et al., 1993) with a groundwater depthof 45 m. Accounting for recharge, modelled ET rateswere approximately 180360 mm a1 for red gum areas,

Figure 10 Predicted groundwater discharge on Clarks Floodplain, confined model. Note the high rates of discharge at the break of

slope, associated with high plant water use at the ends of the peninsulas, and through the centre of the floodplain associated with a

low-lying palaeochannel.

Figure 11 Predicted groundwater discharge on Clarks Floodplain from the uniformly parameterised MODFLOW model,

incorporating a flat floodplain, constant hydraulic conductivity and vegetation distribution.

86 R. Doble et al.


13/23

40130 mm a1 for black box and 040 mm a1 for the ele-vated grassland, lying within the ranges described. A shal-lower watertable (23.5 m) may explain why the grasslandevaporation rates were slightly higher than that describedby Jolly et al. (1993).

Satellite imagery was used as an indicator of vegetationwater use (Goodrich et al., 2000). Spectral indices such asnormalised difference vegetation index (NDVI) that are de-rived from satellite images can indicate vegetation distribu-tion and health or density changes due to salt accumulation.In groundwater dependent ecosystems, they are also able toprovide a direct indication of groundwater discharge

through transpiration, which may be correlated withgroundwater discharge model outputs. NDVI and other veg-etation water use indices indicate plant water use only,however, and do not account for direct soil evaporation.

Landsat-5 TM images with a 25-m resolution were ob-tained for the 18th February 2002, and the NDVI was calcu-lated using the normalised difference of the red and infrared spectral bands (bands 3 and 4). This data was not usedto create or calibrate the model. Fig. 13 shows the spatialdistribution of NDVI across the floodplain, with darker areasindicating high water use. The image indicates high wateruse at the ends of the floodplain peninsulas and adjacent

Figure 12 Predicted groundwater discharge on Clarks Floodplain from the MODFLOW model with variable floodplain elevation,

incorporating constant hydraulic conductivity and vegetation distribution.

Figure 13 NDVI at Clarks Floodplain on the 18th February 2002, inferring evapotranspiration rates.



14/23

the river. High water use is also evident along the break ofslope, indicating areas where fresh irrigation recharge hasbroken through to the floodplain seepage face and is provid-ing a permanent water supply to waterlogging tolerant veg-etation such as grasses and reeds. There were strongcorrelations in high water use at the ends of the peninsulasand adjacent the river in both model output and NDVI. Thisindicates that the modelled discharge function is providing

an adequate basis for further analysis of spatial patternsof discharge.

NDVI values and long term discharge fluxes from themodel output were compared within different eucalypt veg-etation categories on Clarks Floodplain (Fig. 14, Table 3).Areas where evaporation from annual species or bare soildominated, such as the break of slope and floodplain cen-tre, were excluded from this analysis. Within the tree cate-gories, a trend of increasing modelled discharge rates withhigher NDVI values was observed, suggesting a positive rela-tionship between modelled and indicated discharge.

To test the hydraulics of the model, baseflow to the riverwas tested against field measured and indicated baseflow.

Monitoring of river salinity has estimated that Clarks Flood-plain contributes salt loads of between 30 and 50 t d1 (Por-ter, 2002). This also correlated with modelled salt loadsfrom Clarks Floodplain, which were equivalent to 45 t d1.

Transient electromagnetics at a nanometre scale (Nano-TEM) has been used to measure resistivity in river sedimentsalong the stretch of river between Lock 3 and 4 of the RiverMurray in South Australia (Berens et al., 2004) in order to

identify the conductive anomalies associated with concen-trated saline groundwater inflows. NanoTEM analysis doesnot provide actual values of salt loads or baseflow, but pro-vides a qualitative means of comparing the relative ground-water inputs to a watercourse.

Modelled baseflow showed higher groundwater dischargein areas where the river is close to the highland, and nega-tive baseflow at the ends of the peninsulas, correlatingstrongly with the nanoTEM resistivity patterns. Multiplyingbaseflow rates by a regional groundwater salinity of30,000 mg L1 resulted in values of modelled salt inputs.Modelled salt loads at approximately 250 m intervals werecompared quantitatively with nanoTEM results using athreshold resistivity of 27 X m to distinguish positive and

negative inflows (Fig. 15). Patterns of high conductivity cor-related well with the modelled baseflow.

A map of modelled depth to groundwater is provided inFig. 16, compared with groundwater depth measured fromalluvial groundwater wells. Root mean square error of mod-elled vs field groundwater elevation was calculated to be0.59 m, with the data biased toward the lagoon and breakof slope where the majority of boreholes were located.The greatest variation between field and modelled headswas found between the lagoon and break of slope wherethe change in hydraulic gradient was greatest. Whilst thedepth to groundwater provides an opportunity for a stan-dard model calibration and indicates the underlying ground-

water conditions, it does not provide information on thedistribution of vertical groundwater flux, as there is no pro-vision for spatially variable soil, vegetation or groundwatersalinity conditions. Measuring the potential for salinity asdepth to watertable will be of little benefit if the spatial dis-tribution of salinisation potential is of interest.

Table 3 Classifications for the vegetation survey of Clarks

Floodplain, used in the NDVI analysis (Fig. 14) for model

validation

Classification Vegetation type

1r(f) Healthy red gum forest

1br Healthy mixed black box red gum

woodland, black box dominant

1r Healthy red gum woodland

1r(r) Red gum regeneration

1b Healthy black box woodland

1bl Black box with lignum understory

2r Unhealthy red gum woodland

2r1bc Unhealthy red gum, healthy black box

and cooba

2rl Unhealthy red gum with lignum understory

Comparison of Modelled Discharge and NDVI

R2

= 0.82

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

00 20 40 60 80 100

Discharge (mm/yr)

IVDN

Mallee

1r(f)

1b

1br

2rl

2r

1r

1r(r)1bl

2r1bc

Figure 14 Comparison between modelled long term ground-

water discharge and NDVI for Clarks Floodplain for areas with

eucalypt vegetation. Vegetation classifications are found in

Table 3.

Salt Loads: nanoTEM and Modelled

0

1

2

3

45

6

7

502504506508510512514516518

River Km

)d/t(daoLtlaS

lavr

etniertem052

0

1

2

3

45

6

7

delacSylraeniL

ytivitcudnoC

Modelled Load nanoTEM

Figure 15 Salt load patterns from Clarks Floodplain showing

modelled data and resistivity from the nanoTEM information

using a threshold value of 27 X m to distinguish positive and

negative baseflow, and linearly scaling conductivity to salt

loads.

88 R. Doble et al.


15/23

Spatial patterns of high groundwater discharge

While Fig. 10 shows a distribution of groundwater flux simi-lar to that found from vegetation indices, the simpler uni-form and variable elevation MODFLOW models indicatedhigh discharge associated with the break of slope (Figs. 11and 12), but none of the groundwater flux associated withvegetation water use. Some discharge was predicted onthe low-lying edges of the floodplain for the variable eleva-tion model run, however, and higher evapotranspirationrates are evident around the palaeochannel in the centreof the floodplain. While these simplifications may be suffi-cient for broad regional scale modelling, the lack of distrib-

uted vegetation data meant that representation ofgroundwater discharge was less accurate on a sub-floodplainscale.

Table 4 shows a budget of groundwater flow for eachfloodplain division. As a general trend, groundwater inflowfrom the highland increased for each division from thenorthwest to the southeast depending on the proximity ofthe break of slope to the highest point of the groundwatermound. Seepage only occurred within divisions 510, while

evapotranspiration was present on all divisions, and washighest for divisions 1, 7 and 8. Baseflow was negative forfour of the divisions (1, 2, 7 and 8), indicating flow fromthe river into the alluvial aquifer.

The areas in which the net groundwater flux is rechargingthe aquifer are predicted by the model to be minimal, andoccur only in very low lying areas on the ends of floodplainpeninsulas where sandy deposits occur and vegetation is yetto establish. The dominance of ET within the system is ex-pected to be due to the significantly decreased flood recur-rence interval from river regulation in the lower RiverMurray. Net groundwater recharge from flooding is pre-

dicted to occur predominantly in floodplain divisions 1, 3,7 and 8, characteristically the low-lying and sandy insideof meanders (Table 4).

Table 5 shows the width, amplification factor, averageelevation, average hydraulic conductivity and area of treesfor each floodplain division. The area of trees indicates thearea of floodplain supporting black box and red gumeucalypts, which have high water requirements comparedwith bare ground, lignum or grassland vegetation. The

Table 4 Groundwater balance for each floodplain division

Division Highland inflow (m3 d1) Net recharge (m3 d1) Net ET (m3 d1) Seepage (m3 d1) Baseflow (m3 d1)

1 9.60 7.16 94.05 0.00 73.932 7.22 0.03 26.32 0.00 3.90

3 50.78 4.56 54.92 0.00 6.35

4 79.57 0.71 13.03 0.00 100.31

5 131.84 0.00 14.89 0.85 87.80

6 147.26 0.00 35.18 9.39 141.81

7 164.53 2.41 155.09 20.60 45.97

8 140.01 3.47 116.35 13.15 34.25

9 126.09 0.00 15.65 15.19 136.62

10 239.85 0.00 16.22 23.02 204.64

11 389.72 0.54 6.77 0.00 376.54

Total 1486.48 18.87 548.47 82.20 896.02

Figure 16 Depth to groundwater from modelled data compared with measured groundwater depth from boreholes indicated.



16/23

description of area of trees in the following sections definelow tree area as 010 ha of trees, medium tree area as 1030 ha and high tree area as more than 30 ha of trees.

The spatial patterns of groundwater discharge tended to

be quite robust, and did not change significantly when thegroundwater discharge relationship was altered. The quan-titative rate of groundwater discharge was far more sensi-tive to changes in the discharge function than the spatialpatterns themselves.

Confined and unconfined aquifer responses

The evapotranspiration output for which surface confiningeffects were not modelled (Fig. 18) retains the overall spa-tial effects of the confined case (Fig. 10). Groundwater dis-charge is still predominant at the break of slope, ends ofpeninsulas and in the low-lying palaeochannel, but is higherthan for the confined case. Seepage is especially more pro-nounced at the break of slope, as this area tends to be rep-

resented by a heavier clay with lower vertical conductivity,and is therefore more confining of vertical groundwatermovement. The confining effects of the Coonambidgal Clayhad little impact on the spatial patterns of groundwater dis-charge other than reducing the rates of groundwater seep-age with respect to overall discharge rates. Groundwaterdischarge was lowered overall, but patterns associated withvegetation types and microtopography were retained.

The distribution of groundwater discharge with distancefrom the break of slope (Fig. 17) confirms that while theunconfined scenario gives higher rates of groundwater dis-charge across the entire floodplain, the difference is more

pronounced in the first 200 m from the break of slope whereseepage occurs. In addition to the smaller vertical conduc-tivity in this location, the confining effects from the Coo-nambidgal Clay are most significant where the watertableis close to or above the ground surface, such as at the breakof slope, as the distance of upward fluid flow through claybefore evaporation occurs is greater.

Modelling the surface confining layers may not be neces-sary in situations with deeper groundwater or coarser soils,but should be considered in situations where seepage orhigh discharge rates from very shallow groundwater are ofinterest.

Interpretations of model results

The remaining sections of this paper are directed towardthe development of relationships between evapotranspira-tion and factors such as floodplain geometry, elevationand distribution of vegetation, soil parameters.

Surface elevation is a key parameter in determining wheresalt accumulation will occur, as the greatest rates of ground-

Table 5 Floodplain characteristics used in the analysis, varying with division

Division Floodplain width (m) Amplification factor () Area trees (ha) Mean height above river

at break of slope (m)

Kv average (m d1)

1 1492 12.36 39.22 3.91 0.0085

2 496 2.00 14.64 4.13 0.0034

3 830 3.90 16.44 3.99 0.0042

4 461 2.00 4.78 3.96 0.0048

5 494 1.11 3.60 3.42 0.00336 626 0.79 8.44 3.13 0.0040

7 2059 11.17 52.44 2.90 0.0045

8 1895 9.33 36.68 3.41 0.0028

9 544 1.27 2.04 3.35 0.0018

10 261 0.93 2.59 3.56 0.0028

11 119 1.00 2.54 3.70 0.0017

Distribution of groundwater discharge withdistance from break of slope

0

50

100

150

200

250

0 020 0

04 060

080 0

01

021

00 041

0 061

0

Distance from break of slope (m)

m(egrahcsiD

3

)d/

Confined Unconfined

Figure 17 Distribution of groundwater discharge with distance from break of slope for confined and unconfined models. The

greatest difference occurs in the seepage area close to the break of slope.

90 R. Doble et al.


17/23

water discharge by evapotranspiration will be located in lowlying areas where the depth to watertable is a minimum. Mic-rotopography, or small-scale variation in elevation, definesboth the areas of groundwater discharge and the frequencyof flood inundation. As ecosystem communities are linkedwith microtopography, it also defines the impact on biodiver-sity in the region (Cramer and Hobbs, 2002).

Soil type (Talsma, 1963; Warrick, 1988), proximity to theriver and vegetation characteristics (Holland, 2002) will alsoaffect the extent of salt accumulation. The episodic leach-ing of salt from the soil profile by flood-induced rechargefurther complicates the establishment of salt stores. Flood-plain geometry will affect distribution of groundwater flowto a river or water body, as observed in the embaymentsof Green Bay and Lake Michigan (Cherkauer and McKere-ghan, 1991), and meanders of the Swan River, Perth (Linder-felt and Turner, 2001). The sinuosity of the river, andassociated geometry of the floodplain is also likely to deter-mine the distribution of groundwater discharge from afloodplain environment.

It was hypothesised that each of these parameters wouldaffect the distribution of long term groundwater flux andtherefore salt accumulation. The order of importance ofthese parameters for defining discharge patterns, however,was unknown.

Analysis

The floodplain was divided into a series of flow divisions(Fig. 1) in order to compare characteristics of the floodplainwithin each division. The flow lines were defined as the flowpath that groundwater took from a set of equally spacedpoints along the edge of the floodplain to the river, andwere determined by generating a set of equipotential con-tours with groundwater flow being perpendicular to theseand toward the river. This minimised groundwater flow be-tween flow divisions. Zonebudget (Harbaugh, 1990) wasused to perform water balance calculations for each ofthese divisions.

The average rate of evapotranspiration for a division,calculated at steady state, was compared with average ele-vation above the river, floodplain width and vertical hydrau-lic conductivity. Amplification or divergence of the division,that is, the length of river boundary divided by the width ofthe division at the break of slope was also compared withevapotranspiration and baseflow, as was the proportion ofhealthy and total trees found in each division. These resultswere compared with those from the uniform model. Resultsfrom the elevation only example were almost identical tothe uniform model, and are therefore not shown.

Dependence of seepage on local elevation

Constriction of groundwater flow as it moves from theunconfined highland into the partially confined floodplainaquifer led to groundwater heads at the break of slopeexceeding the floodplain elevation and caused seepage tooccur. Seepage was greatest in the area of the palaeochan-nel, which corresponded to the points of lowest elevationalong the cliff. These predictions match with observedseepage at the site.

There was very little correlation between the fraction ofinflow from the highland expressed as seepage and themean floodplain elevation for each of the flow divisions(R2 = 0.12). Comparison with the mean elevation of a

200 m wide strip of floodplain at the break of slope wheregroundwater heads are highest however, gave an inversecorrelation (R2 = 0.61; Fig. 19). No grouping of divisions withhigh tree areas or wide floodplain (points 1, 7 and 8) werefound, suggesting that, as expected, vegetation type hadfar less effect on seepage than elevation or geomorphology.Seepage is an elevation controlled process, and is most af-fected by low depths to watertable, whether by low surfaceelevation, or high groundwater table.

While the volume of seepage is dependent on inflowrates, seepage only appears in divisions with break of slopeelevations lower than 3.7 m above the river stage. Apartfrom changes in hydraulic head, and therefore depth to

Figure 18 Predicted groundwater discharge on Clarks Floodplain, unconfined model.



18/23

groundwater, no other geometric factor significantly af-fected the seepage rate from the floodplain.

Effects of vegetation coverage and floodplain width

The spatial distribution of vegetation across the floodplainhad the greatest overarching effect on patterns of ground-water discharge. The area of high water using trees withineach floodplain division correlated very closely with evapo-transpiration (R2 = 0.96; Fig. 20). While the maximum rateof groundwater discharge was found at the break of slope,associated with shallow groundwater and seepage, the totalvolume of discharge from the eucalypt communities wasalso high due to the large areas of vegetation with highwater requirements and deep extinction depths.

Wider sections of floodplain result in a longer groundwa-ter travel time within the floodplain. As the hydraulic headis constrained at both the break of slope and the river,longer flow paths reduce the hydraulic gradient and ground-water velocity. This consequently subjects groundwater toevapotranspiration for longer periods of time. Larger flood-plain widths also provide a greater area over which evapo-transpiration may take place. Fig. 21 shows a strongrelationship between the rate of evapotranspiration from

a division and the floodplain width (R2 = 0.96) despite vary-

ing groundwater inflows from the highland. However, thearea of trees within a division was also found to increasewith floodplain width, indicating some interdependence.

Results from the uniform MODFLOW model showed norelationship with either tree area, which was expecteddue to the lack of vegetation parameters, nor floodplainwidth, despite the inclusion of floodplain spatial geometrywithin the uniform model. The lack of relationship betweenfloodplain width and evapotranspiration in the uniformmodel confirms that ET is more dependent on vegetationdistribution than floodplain width alone, and that inclusionof the spatial patterns of vegetation is fundamental forstudies of groundwater discharge on a sub-floodplain scale.

Dependence of groundwater discharge on

floodplain elevation and soil type

Warrick (1988), Gardner and Fireman (1958) and Thorburnet al. (1992) show that the relationship between watertabledepth and evaporation in soils is an inverse power relation-ship. It was hypothesised that given a relatively level piezo-metric surface, floodplain elevation, which varies morerapidly in a spatial context than groundwater elevation, islikely to be the primary variable for depth to watertable.Relative elevation of the floodplain is therefore thought togovern distribution of groundwater discharge on a spatialscale.

Fig. 22 shows evapotranspiration rates plotted againstfloodplain height relative to the river level. While thereare no strong overall trends, there are sub-groups withinthe data set defined by the area of trees present withinthe division. The divisions with a high tree area exclusivelyrepresent the group with the highest ET rates. The secondgroup, below this, is represented by divisions with a lowto medium area of trees, and shows a trend of increasingET with decreasing elevation (R2 = 0.64).

The location of low-lying regions within a division willalso dictate the relative effect of surface elevation. Low-lying areas that coincide with the break of slope (divisions59) or eucalypt communities (ends of divisions 1, 3, 7

Seepage/Highland Inflow - Mean Elevation

Above River at Break of Slope

R2

= 0.61

0.0

0.1

0.2

2.7 2.9 3.1 3.3 3.5 3.7 3.9 4.1 4.3

Mean elevation above river at break of slope (m)

Seepage/Infow

8

7

1

Figure 19 Seepage/highland inflow vs mean elevation above

river at break of slope, showing highly treed divisions in each

subset. Vegetation coverage has little significance for ground-

water seepage. No relationship can be derived between local

elevation and seepage from the uniform model results, as no

elevation variation was included within the model.

ET vs Area of trees

R2

= 0.96

0

50

100

150

200

0 10 20 30 40 50 60

Tree Area (ha)

m(TE

3

)d/

Figure 20 Evapotranspiration vs area of trees. There is a

strong positive trend for the modified modelling results (),

even for divisions represented by points 1 and 2, which lie below

the trend line, and have extremely low inflow. There is no

relationship between ET and tree area for the uniform MOD-

FLOW model ( ).

ET - Floodplain Width

R2

= 0.96

0.0

50.0

100.0

150.0

200.0

0 500 1000 1500 2000 2500

Floodplain Width (m)

m(TE

3

)d/

High tree

area

Increasing tree

area

Figure 21 Evapotranspiration vs floodplain width. There is a

strong correlation for the modified model results (), despite

the highly variable inflows for each of the divisions. The

relationship between ET and floodplain width for the uniform

MODFLOW model ( ) is much less defined, and suggests that the

ET is more dependent on vegetation than floodplain width

alone.

92 R. Doble et al.


19/23

and 8) are likely to produce higher discharge rates than ifthey coincide with grassland or low water using lignum

(Muehlenbeckia cunninghamii) covered areas further fromthe break of slope (centre of division 8). Within each vege-tation class, elevation is also a factoring determining pat-terns of groundwater discharge. The effects of localisedelevation patterns (Fig. 6) are visible within different vege-tation areas, rather than between them, but as elevation isan indicator of flooding frequency, vegetation distributionand floodplain elevation are not completely independentvariables.

The weaker dependence on elevation than expected ispotentially as a result of the higher recharge rates associ-ated with lower lying areas. Although lower elevation leadsto higher rates of evapotranspiration, it is partly offset byincreased recharge through more frequent inundation, so

the sensitivity of net flux to elevation is lessened.Groundwater discharge and average vertical hydraulic

conductivity (Kv) of a division appear to have a limited rela-tionship for Clarks Floodplain (R2 = 0.12 for increasing dis-charge with increasing Kv) as shown in Fig. 23. Separatingtwo groups of data points, high and low tree area, revealsa slight positive trend in discharge with increasing Kv, but

this overall trend is overwhelmed by the effects of vegeta-tion type on ET.

Effects of floodplain amplification on baseflow

The ratio between river frontage of a floodplain division toits width at the break of slope, or amplification factor,strongly influenced where baseflow to the river occurred.

High amplification factors were associated with negativebaseflow rates (Fig. 24), and within the divisions with lowareas of trees, baseflow increased with both higher ground-water inflows and lower amplification rates. Divisions withlong river frontages coincided with a wider floodplain andgreater area of trees, facilitating high evapotranspirationrates adjacent the river. To achieve this evaporation rate,river water infiltrates into the alluvial aquifer, facilitatedby the lower groundwater heads under the eucalypt commu-nities. This behaviour has been recorded in riparian environ-ments in the San Pedro River, described by Mac Nish et al.(2000) and Stromberg et al. (1996). While there was a nota-ble correlation between baseflow and division amplificationor river frontage, the area of trees present on a floodplaindivision was also represented in data clusters. The interde-pendent nature of river frontage and vegetation coveragemade these parameters difficult to distinguish, and sug-gested that baseflow was defined by a combination of rivergeomorphology and ecology, rather than any singleparameter.

Fig. 24 shows that the uniform model results were similarto the more detailed modelling for floodplain divisions withlow amplification factors and low tree cover. Baseflow fol-lowed the same trend, increasing with rising highland in-flows. For divisions with high amplification factors,however, the lack of spatial variation in the ET functionmeant the uniform model still predicted positive baseflow.

Baseflow was shown to be influenced firstly by groundwaterinflows to the floodplain, then by amplification factor andtherefore also by proximity to the highland. Spatially vari-able ET rates were only important for areas further from

ET - Floodplain Height Above River

0.0

50.0

100.0

150.0

200.0

3 3.2 3.4 3.6 3.8 4

Mean Height Above River (m)

m(TE

3

)d/

High tree

area

Med-low

tree area

Figure 22 Evapotranspiration vs floodplain elevation above

the river. Note the cluster of points within tree area classes of

high, and low and medium. No relationship can be derived

between elevation and evapotranspiration from the uniform

model results, as no elevation variation was included within the

model.

ET and Seepage - Average Kv

0.0

50.0

100.0

150.0

200.0

0 0.002 0.004 0.006 0.008 0.01

Average Kv (m)

m(egapeeS+TE

3

)d/

High tree

area

Low-med

tree area

Figure 23 Evapotranspiration and seepage vs averaged Kv for

each floodplain division. There is a slight positive trend, but the

high tree area class divisions overwhelm the results. No

relationship can be derived between vertical conductivity and

evapotranspiration from the uniform model results, as no

conductivity variation was included within the model.

Baseflow - Amplification Factor

-100.0

0.0

100.0

200.0

300.0

400.0

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

Amplification Factor

m(wolfesaB

3

)d/

High tree

areaMed tree

area

Increasing

highland

inflow

Figure 24 Baseflow vs division amplification factor. Note

again the separate tree area classes, and the effect of inflow

within the low tree area class within the modified model result

results (). Uniform model results ( ) show a similar trend for

baseflow in divisions with low amplification factors, but

baseflow in the divisions with high amplification factors and

high tree areas is positive in contrast to the negative baseflow

of the modified model.



20/23

the highland. This indicates that while patterns of ET aresensitive to spatially variable model inputs, baseflow pat-terns are relatively robust.

Regression

A linear multiple regression was undertaken, comparing themodelled evapotranspiration with highland inflow, flood-

plain width, area of trees, amplification factor, average Kvand mean height of floodplain section above the river level.This indicated the relative contributing factors toevapotranspiration, and allowed a series of variables to becompared with evapotranspiration simultaneously. A rela-tionship for evapotranspiration was derived with an R2 valueof 0.998. In order of significance, the parameters contribut-ing to evapotranspiration were the area of trees, highlandinflow, floodplain width and mean height above the flood-plain, with P< 0.05 indicating significance (Table 6). Ampli-fication factor and average Kv had little effect on patternsof evapotranspiration. Baseflow, floodplain width and treearea were highly interdependent, with wider floodplain divi-sions also having high tree coverage and negative baseflow.Conversely, narrow divisions tended to have low tree cover-age and higher, positive baseflow.

Results from the regression supported the conclusionsdrawn from the figures above.

Linking discharge with salt accumulation

Net flux output from the model can also be used to under-stand the spatial patterns of salt accumulation. Salt accu-mulation rates were predicted by multiplying the coverageof groundwater discharge rates (Fig. 10) with a spatialmap of local groundwater salinity determined from fieldstudies. High rates of salt accumulation are predicted at

the break of slope, and also on the outsides of the bend indivisions 37 (Fig. 25). Although Fig. 10 shows a high dis-charge rate from red gum and black box vegetation at theends of the peninsulas, long-term salt accumulation is min-imal (Fig. 25). Local groundwater salinity at the ends of pen-insulas is low, as most saline regional groundwater has beendischarged as baseflow at much narrower parts of the flood-plain. High water use from vegetation has lowered ground-water tables and led to an influx of fresh water into thealluvial aquifer from the river.

Vegetation health mapping (Telfer and Overton, 1999)shows a concentration of dead trees or trees in poor healtharound the backwater on Clarks Floodplain (Fig. 3), corre-sponding with areas of high salt accumulation (Fig. 25). Poorvegetation health is also evident in the combined red gumand black box stands around the outside of the bend corre-lating with higher salt accumulation within divisions 47.High rates of salt accumulation indicate areas where vege-tation health or water bodies are under threat from salinisa-tion, and focus where remediation and management effortsshould be directed. Methods such as direct irrigation duringdrought, targeted spear-point pumping of groundwater be-low affected vegetation and flood augmentation may beused to reduce the salt impact on floodplain ecosystems.

Interrelationships between vegetation and

groundwater hydrology

The analysis section of this paper has treated vegetationdistribution as an independent variable in order to ascertain

Figure 25 Modelled salt accumulation on Clarks Floodplain. High salinity at the break of slope from regional groundwater inflows

causes salt accumulation to be highest in this area, while low salinity at the ends of the peninsulas due to fresh inflows from the river

lowers the rate of salinisation.

Table 6 Regression results for evapotranspiration

Parameter P-value

Highland inflow (m3 d1) 0.015

Floodplain width (m) 0.020

Area trees (ha) 0.003

Amplification factor () 0.325

Mean height above river (m) 0.041Kv average (m d

1) 0.904

94 R. Doble et al.


21/23

causal factors for groundwater flux distribution, when inreality it has complex relationships with geomorphology,groundwater hydrology and salinity. Vegetation with highwater requirements such as the red gum eucalypt is morelikely to be found in areas with access to fresh groundwater,such as the inside of meanders or other sandy deposits withgood connection to the river. High rates of evapotranspira-tion are able to occur as a result of the available fresh

water.Floodplain width, tree area and amplification factor

were found to be highly correlated, with correlation coeffi-cients ranging between 0.93 and 0.96, with wider floodplaindivisions having high tree coverage and high amplificationdue to the presence of meanders. Correlation was less pro-nounced for floodplain elevation, soil type and highlandinflow.

The value of these results lies in determining the mostcritical parameters to include in groundwater models toadequately represent groundwater discharge processes ona fine spatial scale. In finite difference numerical models,varying the evapotranspiration function spatially accordingto vegetation distribution will give a more appropriate rep-

resentation of groundwater flow processes than uniformparameters, or variation of evapotranspiration with soilor other factors. For full ecological analyses, however,the lack of feedback on groundwater salinity from saltaccumulation will be a disadvantage for temporal investi-gations of vegetation health responses using thismethodology.

Conclusions

This paper has provided a discussion of the importance ofgeometry, geomorphology and elevation in determining

the spatial distribution of discharge from a floodplain. Itused a quasi-three-dimensional MODFLOW model to confirmthe critical parameters required to represent groundwaterdischarge patterns affecting vegetation health, and com-pared these results with a uniform MODFLOW model usingno spatial distribution of vegetation or elevationparameters.

The key findings from this study are that:

To assess the magnitude of the impact on floodplain bio-diversity and ecological value, groundwater modellingneeds to take into account spatial patterns of groundwa-ter discharge and rates of salinisation on a fine scale. Auniform model with no spatial variation or recharge

information was not able to accurately represent spatialdistributions of evapotranspiration.

Progressively more information on the spatial patterns ofdischarge was gained when moving from a uniformparameter floodplain model to variation of elevation,to variation of vegetation and soil parameters. Bettercorrelations between model results and field measure-ments were achieved when vegetation parameters wereincluded in the modelling.

Due to the large range of vegetation present on the flood-plain studied, the distribution of vegetation type had themost significant effect on the spatial distribution andmagnitude of groundwater discharge between groundwa-

ter flow divisions. Floodplain elevation and width had alesser effect, but the distribution of vegetation relatedto both of these parameters.

Localised elevation is the most critical parameter indetermining the proportion of inflow to a floodplainexpressed as seepage and is significant in determiningspatial patterns of evapotranspiration. Geomorphic fea-tures such as backwaters, lagoons and palaeochannel,

which are commonly found in meandering river systems,can produce elevation changes that significantly effectthe spatial distribution of discharge.

The complex meandering patterns of the river definedwhere baseflow occurs on a fine scale. Higher proportionsof baseflow are associated with areas of floodplain thatare close to the break of slope, and where river frontage,or division amplification is low. Negative baseflow wasfound on the inside of river bends where eucalypt com-munities are present, and where division amplificationis high. The uniformly parameterised model was able torepresent baseflow with some degree of accuracy, butcould not predict the negative baseflow associated withhigh vegetation water use.

The uniformly parameterised and more complex distrib-uted models produced similar baseflow distributiondespite very different spatial patterns of evapotranspira-tion. While many groundwater models are calibrated orvalidated against baseflow calculations, further compari-son with spatial indicators of groundwater discharge, suchas vegetation water use or salt accumulation maps arerequired when vertical groundwater flux is of interest.

Although long-term patterns of net groundwater dis-charge were found to depend on vegetation distribution,elevation, soil type and river geometry, the relationship be-tween these interdependent parameters is complex, and

based on the geomorphology and ecology of the riverine sys-tem. The value of these results, however, lies in determin-ing the most critical parameters to include in groundwatermodels to adequately represent spatial groundwater dis-charge processes.

Acknowledgements

The first author acknowledges the support of Land andWater Australia and the Centre for Groundwater Studies.The work was also supported through the project AssessingCurrent and Future Impacts of Land Management InducedGroundwater Discharge on Floodplain Health, which is a

partnership between the River Murray Catchment WaterManagement Board (RMCWMB), CSIRO Land and Water, theSouth Australian Department of Environment and Heritage,and South Australian Department for Water, Land and Biodi-versity Conservation (DWLBC). The project was fundedthrough the Natural Heritage Trust and by the project part-ners, and managed by RMCWMB.

References

Akeroyd, M.D., Tyerman, S.D., Walker, G.R., Jolly, I.D., 1998.Impact of flooding on the water use of semi-arid riparianeucalypts. Journal of Hydrology 206, 104117.



22/23

Armstrong, D., Yan, W., Barnett, S.R., 1999. Loxton irrigation area groundwater modelling of groundwater/river interaction.Report for the Department of Primary Industries and Resources,Adelaide, South Australia.

AWE, 1999. Bookpurnong and lock 4 preliminary land and watermanagement plan. Background report and appendices. Reportfor the Bookpurnong and Lock 4, Environmental Group Inc.

Banta, E.R., 2000. MODFLOW-2000, the US geological surveymodular ground-water model-documentation of packages for

simulating evapotranspiration with a segmented function (ETS1)and drains with return flow (DRT1). US Geological Survey Open-File Report 00-466, 127p.

Barnett, S.R., SA Department of Mines and Energy, 1991. RenmarkHydrogeological Map (1:250,000 Scale). Bureau of MineralResources, Geology and Geophysics, Canberra, Australia.

Berens, V., Hatch, M., Wilson, T., Hill, T., 2004. River-borneNanoTEM survey for location of salt accession to the RiverMurray at Loxton. In: Proceedings of the ASEG 17th GeophysicalConference and Exhibition, Sydney, Australia.

Boone, A., Wetzel, P.J., 1996. Issues related to low resolutionmodeling of soil moisture: experience with the PLACE model.Global and Planetary Change 13, 161181.

Cherkauer, D.S., McKereghan, P.F., 1991. Ground-water dischargeto lakes: focusing in embayments. Ground Water 29 (1), 7280.

Cramer, V.A., Hobbs, R.J., 2002. Ecological consequences ofaltered hydrological regimes in fragmented ecosystems insouthern Australia: Impacts and possible managementresponses. Austral Ecology 27, 546564.

Doble, R.C., Walker, G.R., Simmons, C.T., 2005. Understandingspatial patterns of discharge in semi-arid regions using arechargedischarge balance to determine vegetation health.CSIRO Land and Water Technical Report 13/05.

Evans, W.R., Kellett, J.R., 1989. The hydrogeology of the MurrayBasin, southeastern Australia. BMR Journal of Australian Geologyand Geophysics 11, 147166.

Feikema, P.M., 2000. Tree growth and water relations in eucalyptplantations over shallow saline groundwater. PhD thesis for theDepartment of Civil Engineering, Monash University, MelbourneAustralia.

Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Inc.,New Jersey, USA.

Gardner, W.R., Fireman, M., 1958. Laboratory studies of evapora-tion from soil columns in the presence of a water table. SoilScience 85, 244249.

Gardner, W.R., 1958. Some steady state solutions of the unsatu-rated moisture flow equation with application to evaporationfrom a water table. Soil Science 85, 228232.

Goodrich, D.C., Scott, R., Qi, J., Goff, B., Unkrich, C.L., Moran,M.S., Williams, D., Schaeffer, S., Snyder, K., MacNish, R.,Maddock, T., Pool, D., Chehbouni, A., Cooper, D.I., Eichinger,W.e., Shuttleworth, W.J., Kerr, Y., Ni, W., 2000. Seasonalestimates of riparian evapotranspiration using remote and in situmeasurements. Agricultural and Forest Meteorology 105, 281309.

Harbaugh, A.W., 1990. A computer program for calculating subre-gional water budgets using results from the US geological surveymodular three-dimensional ground-water flow model. US Geo-logical Survey Open-File Report 90-392.

Harbaugh, A.W., Banta, E.R., Hill, M.C., McDonald, M.G., 2000.MODFLOW-2000, The US Geological Survey ModularGround-water Model User Guide to Modularization Conceptsand the Ground-water Flow Process. US Geological Society,Reston, VA.

Hatton, T., 1998. Catchment scale recharge modelling. In: Zhang,L., Walker, G.R. (Eds.), Part 1 of Studies in CatchmentHydrology: The Basics of Recharge and Discharge. CSIRO, Vic.,Australia. Available from: .

Holland, K.L., 2002. Blackbox (Eucalyptus largiflorens) tree healthand groundwater discharge on the lower river Murray floodplain,South Australia: linking spatial and temporal scales. PhD thesis,Flinders University of South Australia.

Jarwal, S.D., Walker, G.R., Jolly, I.D., 1996. General site descrip-tion. In: Walker, G.R., Jolly, I.D., Jarwal, S.D. (Eds.), Salt andWater Movement in the Chowilla Floodplain. CSIRO Division ofWater Resources, pp. 16309 (Chapter 2).

Jolly, I.D., 1996. The effects of river management on the hydrology

and hydroecology of arid and semi-arid floodplains. In: Ander-son, M.G., Walling, D.E., Bates, P.D. (Eds.), Floodplain Pro-cesses. Wiley, Chichester, New York, pp. 577609.

Jolly, I.D., Narayan, K.A., Armstrong, D., Walker, G.R., 1998. Theimpact of flooding on modelling salt transport processes tostreams. Environmental Modelling and Software 13, 87104.

Jolly, I.D., Walker, G.R., Thorburn, P.J., 1993. Salt accumulation insemi-arid floodplain soils with implications for forest health.Journal of Hydrology 150, 589614.

Jolly, I.D., Walker, G.R., McEwan, K.L., Thorburn, P.J., Hatton,T.J., Bennett, J., Hacker, J.M., 1991. The role of floodplainvegetation in salt accession to the River Murray in SouthAustralia. In: Proceedings of the International Hydrology andWater Resources Symposium, Perth. The Institution of Engi-neers, Australia.

Linderfelt, W.R., Turner, J.V., 2001. Interaction between shallowgroundwater, saline surface water and nutrient discharge in aseasonal estuary: the Swan-Canning system. Hydrological Pro-cesses 15, 26312653.

Mac Nish, R.D., Unkrich, C.L., Smythe, E., Goodrich, D.C.,Maddock, T., 2000. Comparison of riparian evapotranspirationestimates based on a water balance approach and sap flowmeasurements. Agricultural and Forest Meteorology 105, 271279.

McDonald, M.G., Harbaugh, A.G., 1996. Users documentation forMODFLOW-96, An update to the US Geological Survey modularfinite-difference ground-water flow model. Scientific Softwaregroup, Washington, DC, USA.

Overton, I.C., 2005. Modelling floodplain inundation on a r

doble r et al 2006 spatial relationships between vegetation cover and irrigation

Documents