Interaction between shallow groundwater, saline surface

water and contaminant discharge at a seasonally

and tidally forced estuarine boundary

S.J. Westbrooka,b, J.L. Raynerb, G.B. Davisb,*, T.P. Clementa,c, P.L. Bjergd, S.J. Fisherb

aDepartment of Environmental Engineering, University of Western Australia, Nedlands, WA 6907, AustraliabPerth Laboratory, Division of Water Resources, CSIRO Land and Water, Private Bag No. 5, Wembley, WA 6913, Australia

cDepartment of Civil Engineering, Auburn University, Auburn, AL, USAdEnvironment and Resources DTU, Technical University of Denmark, Kongens Lyngby, Denmark

Received 22 August 2003; revised 16 July 2004; accepted 30 July 2004

Abstract

This paper presents findings from a 2-year field investigation of a dissolved hydrocarbon groundwater plume flowing towards

a tidally and seasonally forced estuarine river system in Perth, Western Australia. Samples collected from transects of multiport

wells along the riverbank and into the river, enabled mapping of the fine scale (0.5 m) vertical definition of the hydrocarbon

plume and its longitudinal extent. Spear probing beneath the river sediments and water table, and transient monitoring of

multiport wells (electrical conductivity) was also carried out to define the zone of mixing between river water and groundwater

(the hyporheic zone) and its variability. The results showed that groundwater seepage into the estuarine surface sediments

occurred in a zone less than 10 m from the high tide mark, and that this distance and the hyporheic transition zone were

influenced by tidal fluctuations and infiltration of river water into the sediments. The dissolved BTEXN (benzene, toluene,

ethylbenzene, the xylene isomers and naphthalene) distributions indicated the behaviour of the hydrocarbon plume at the

groundwater/surface water transition zone to be strongly influenced by edge-focussed discharge. Monitoring programs and risk

assessment studies at similar contaminated sites should therefore focus efforts within the intertidal zone where contaminants are

likely to impact the surface water and shallow sediment environments.

Crown Copyright q 2004 Published by Elsevier B.V. All rights reserved.

Keywords: Ground water; Surface water; Hydrocarbons; Discharge; Tides; Seasonal variations

0022-1694/$ - see front matter Crown Copyright q 2004 Published by El

doi:10.1016/j.jhydrol.2004.07.007

* Corresponding author. Tel.: C61 89333 6386; fax: C61 89333

6211.

E-mail address: greg.davis@csiro.au (G.B. Davis).

1. Introduction

As urban and industrial development continues to

expand around the world’s rivers and coastlines, so

does the rate of unintentional release of contaminants

to subsurface and surface waters and the need for

Journal of Hydrology 302 (2005) 255–269

www.elsevier.com/locate/jhydrol

sevier B.V. All rights reserved.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269256

effective assessment of such environments (Winter,

2000). Hydrologists have long known that surface

waters and groundwater are intrinsically linked

systems (e.g. Glover, 1959; Cooper, 1959; Clement

et al., 1996; Simpson et al., 2003). Areas around

streams, rivers, lakes and coastal environments

represent zones of interaction and transition between

the two systems where dissolved constituents such as

pollutants can be diluted, exchanged, transformed or

destroyed. Identifying predominant processes affect-

ing solute exchange across transition zones is there-

fore, critical in assessing contaminant fluxes to the

sediment/water interface, and ultimately in estimating

contaminant exposures for the receiving ecosystems.

Groundwater/surface water interactions in estuar-

ine environments are influenced by a number of

processes forming complex spatially and temporally

variable systems. Density contrasts between the

typically fresh groundwater and saline to brackish

marine and estuarine surface waters leads to mixing

and convective circulation at the groundwater dis-

charge boundary so that the system is characterised by

the intrusion of saltwater into the adjacent coastal

aquifer (Glover, 1959; Cooper, 1959; Reilly and

Goodman, 1985; Ataie-Ashtiani et al., 1999; Simpson

and Clement, 2004). Tidal activity can often induce a

fluctuating water table as well as infiltration of surface

water into sediments, forming a surficial mixing zone

with groundwater discharging from the adjacent

aquifer (Robinson et al., 1998; Ataie-Ashtiani et al.,

1999; Boudreau and Jorgensen, 2001; Acworth and

Dasey, 2003). Although there is still no single

conceptual definition for such a surficial mixing

zone, the terms ‘hyporheic zone’, ‘subsurface estuary’

and ‘groundwater/surface water interface’ or ‘GSI’

are gaining common usage in the scientific literature.

White (1993) conceptually defined the hyporheic zone

as ‘the saturated interstitial area beneath the stream

bed and into the stream banks that contain some

proportion of channel water or that have been altered

by channel water infiltration’. This definition may be

broadened to include rivers, lakes, estuaries and

coastal environments where surface water infiltrates

into the underlying sediments and interacts with

groundwater.

Although numerous studies have addressed

groundwater and solute inputs to surface water bodies

(e.g. Harvey et al., 1987, Gallagher et al., 1996,

Portney et al., 1998, Krabbenhoft et al., 1990, Lorah

and Olsen, 1999, Winter, 2000; Tobias et al., 2001),

few studies to date have examined near-shore

groundwater discharge in detail. Studies of note

however, include those by Robinson et al. (1998);

Robinson and Gallagher (1999); Smith and Turner

(2001); Linderfelt and Turner (2001); Simpson et al.

(2003) and the initial study by Westbrook et al. (2000)

related to the current work.

Robinson et al. (1998) presented results from a

field investigation on unconfined groundwater dis-

charge to estuarine waters at Chesapeake Bay,

Virginia, showing strong tidal and seasonal controls

on fresh groundwater discharge associated with

infiltration of surface water into tidally exposed

sediments. Robinson and Gallagher (1999) further

developed a two-dimensional, field scale, finite-

element model based on density dependent fluid

flow with water table and dynamic tidal boundary

conditions. The model was able to reproduce the

Chesapeake Bay field data on the movement of the

near-shore water table, groundwater salt concen-

trations and groundwater discharge rates and patterns

but was unable to replicate short-term salt fluctuations

in the hyporheic zone due to the wave action of tides

within the intertidal zone (Robinson and Gallagher,

1999). Simpson et al. (2003) performed transport

experiments in a sand tank to study the characteristics

of the seepage-face zone that exists near a ground-

water/surface water interface. Their study concluded

that seepage-face zones, which are dominated by

strong hydraulic gradients, play an important role in

influencing the localized flow and solute transport

processes in shallow unconfined aquifers.

Field observations and mathematical modelling of

density dependent groundwater/surface water inter-

action in the Swan-Canning estuary by Smith and

Turner (2001) showed that density contrasts between

the estuary and adjacent fresh groundwater system are

sufficient to drive mixed-convection cells that give rise

to circulation of river water in the aquifer, providing a

mechanism to transport nutrients between the nutrient-

rich pore fluids in the riverbed sediments and

groundwater. Further, results indicated unconfined

groundwater preferentially discharges into the Swan

River along the outside of river meanders, with very

low discharges or at times saline river water recharge

along the inside of meanders (Smith and Turner, 2001;

Fig. 1. Regional locality map of the Canning river study site in

Perth, Western Australia.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 257

Linderfelt and Turner, 2001). Short-term fluctuations

such as wave-induced displacement, hyporheic flux

and tidal oscillation appeared to cause nutrient release

from the sediment pore fluids, particularly in low flow

summer periods (Linderfelt and Turner, 2001).

Although these findings highlight the complexities

of groundwater discharge at an estuarine boundary, to

the author’s knowledge no studies have investigated

the importance of the near-shore processes in under-

standing the transport of point-source contaminant

plumes across the estuarine transition zone. This

paper presents findings from a 2-year field investi-

gation of a dissolved phase gasoline groundwater

plume, primarily comprised of BTEXN (benzene,

toluene, ethylbenzene, the xylene isomers and

naphthalene), that discharges to an estuarine beach

on the foreshore of the Canning River in Perth,

Western Australia. These compounds are known or

suspected carcinogens and may represent a threat to

the quality of fresh water resources (see, for example

ANZECC, 1992). Interpretation of the data is used to

develop an increased understanding and improved

conceptual model of groundwater/surface water

dynamics, especially for this study site. The model

identifies spatial and temporal characteristics of

groundwater/surface water interactions and examines

the effects of these interactions on contaminant

discharge to the estuarine environment. Ongoing

research at the site uses the conceptual model as a

basis for investigations into geochemical controls on

groundwater quality and contaminant fate in the near-

shore region.

2. Site location and background data

The study site (Figs. 1 and 2) is situated on the

foreshore of the Canning River, approximately 6-km

south-southwest of Perth’s central business district,

Western Australia, and approximately 15 km

upstream from the Indian Ocean. Groundwater at

the site is impacted with dissolved-phase petroleum

hydrocarbons, including the BTEX (benzene, toluene,

ethylbenzene, and xylene isomers) compounds and

naphthalene (N). These impacts originate from a

historical, unintentional, light nonaqueous phase

liquid (LNAPL) release from an underground storage

tank (UST) located approximately 80 m from

the river’s edge. The plume follows an easterly flow

path from the UST source as governed by the local

groundwater hydraulics.

The Canning River forms part of the Swan-

Canning river and estuary system, which is charac-

terised by cyclical density stratification related to

changing surface flow dynamics. The estuary system

is tidal, being open to exchange with the Indian Ocean

and seasonally forced by dominant winter rainfall.

During spring and summer a high-density wedge of

saline ocean water intrudes into the estuary and it’s

rivers, reaching as far as 45 km upstream from the

ocean. High rainfall periods during autumn and winter

give rise to fresh water flushing of the estuary causing

regression of the saline wedge. Smith and Turner

(2001) showed that the seasonal density contrasts

between the estuary and the adjacent fresh water

aquifers were sufficient to drive mixed-convection

cells and saline water into the adjoining surficial

aquifers.

2.1. Hydrogeology of the site

Groundwater of the regional Perth area is domi-

nated by a shallow, unconfined Quaternary ‘superficial

aquifer’ system, overlying confined tertiary and

mesozoic sedimentary formations of the Perth basin

Fig. 2. Plan view of the Canning river field site showing monitoring infrastructure and A–A0, B–B, and C–C0 transects, and approximate plume

outline.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269258

(Davidson, 1995) The superficial aquifer sediments

vary in composition from predominantly clayey in

inland areas to sand and limestone successions in the

coastal plain areas. The sediments have an average

thickness of 20–45 m. Around the Swan-Canning

estuary, the sediments are characterised by sequences

of interbedded sand, silt and clay that directly overly

the confining shale and siltstone of the early Tertiary

Kings Park Formation (Davidson, 1995).

The shallow (top 10 m) hydrogeology at the site

comprises a stratigraphy of well-sorted, medium-

grained, unconsolidated quartz-rich sediments with

thin shelly horizons underlain by a dense clay layer

intersected at around 7.5 m below surface. The

aquifer’s saturated porosity ranges from 0.25 to

0.3 m3 mK3, and the hydraulic conductivity ranges

from 1 to 10 m dayK1. The average depth to water

from the ground surface ranges from 2.02 m in the

near source area at MW02 to 0.94 m at the foreshore

at MW33 (see locations on Fig. 2). The hydraulic

gradient from the source to the river foreshore

varies seasonally between 0.005 and 0.01 m mK1.

Dissolved oxygen concentrations of !1 mg lK1

indicated that the groundwater is typically anaerobic.

2.2. Historical contaminant characterisation data

Environmental assessments of the area since the

early 1990s identified two areas of subsurface

contamination: (1) a northern zone around the source

at the site of the present study, and; (2) a southern

zone situated at a site where underground storage

tanks were previously located. Results from the initial

site assessment showed that the subsurface was

contaminated with hydrocarbon compounds. Soil

contamination and dissolved phase total petroleum

hydrocarbons (TPH) concentrations of O5000 ppm

and O2 ppm, respectively, were reported from

monitoring wells MW-01 and MW-02 situated at the

UST location. Analysis of samples for hydrocarbon

based on chain length indicated that both gasoline and

diesel were possible source fuels.

Up to 37 monitoring wells had been installed at

various localities around the contaminated site

Fig. 3. Schematic diagram showing construction details of a

multiport well used for groundwater sampling at Canning river.

Arrows indicate draw in of water samples through the slotted screen

during sampling.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 259

between 1993 and 1999. Most of these monitoring

wells were installed to depths of 3 m below ground

surface (approximately 1–2 m below the water table).

The wells were screened over their whole depth

intervals. Historic data from periodic sampling

indicated a groundwater plume moving in an easterly

direction away from the source towards the Canning

River (see Fig. 2), with benzene concentrations in

MW33, situated on the river bank, exceeding the

ANZECC (1992) guideline value of 300 mg lK1 in

May 1996, October 1997, December 1998 and March

1999.

In December 1999, an additional three monitoring

wells were installed (MWs 41–43) to depths of 6 m

below ground level and screened over 2.5–6 m depth

intervals. Subsequent sampling of MW43, situated

20 m west of the river bank showed dissolved phase

hydrocarbon concentrations up to 1000 mg lK1 ben-

zene, 955 mg lK1 ethylbenzene and 1600 mg lK1

xylene. The presence of elevated BTEX concen-

trations at depths not previously sampled at the site

suggested that the dissolved phase plume had

migrated downward in the aquifer and highlighted

the uncertainties involved when using shallow, fully

screened wells to monitor groundwater plumes.

3. Methods

3.1. Water quality monitoring

To aid further definition of the plume, multiport

wells (MPs) and spear probing (drive point sampling)

were used to collect groundwater samples for analysis

As shown in Fig. 3, the multiport wells had sampling

screens at 0.5 m depth intervals and comprised a

bundle of individual 2.5 mm internal diameter nylon

tubes, each extending to a different depth. Sampling

screens (or ports) were made of 0.1 m long slotted

stainless steel tubes covered by an inert nylon filter

mesh. The MPs were installed by initially auguring to

the water table using a half split stem auger followed

by insertion of a temporary casing into the hole.

Aquifer material was sludged (bailed) out of the

casing while the casing was pushed deeper until the

required depth was reached. The multiport bundle was

then inserted into the temporary casing, which

was removed to allow the hole to collapse around

the installation. This method was used to minimise the

zone of disturbance around each installation.

After initial investigations using historical moni-

toring data, and pilot multiport and spear probe

sampling, a transect approach was employed to

monitor contaminant distributions. In this approach

the principal axes of the contaminant plume were

designated x, y and z, with x being the longitudinal

direction of flow, y the lateral direction perpendicular

to flow, and z the vertical direction. By orientating

the sampling transects along the principal axes, lateral

(y–z), longitudinal (x–z) and plan (x–y) control

sections were mapped and a three-dimensional picture

of the plume was constructed.

In total, five control sections were installed and

sampled at the site, consisting of three lateral sections

(A–A 0, B–B 0, C–C 0), one longitudinal section (D–D 0)

and one plan section (see Figs. 2 and 4). The lateral

sections, comprised of permanent multiport wells,

were installed to capture the depth distribution of

contaminants across the plume in both the on-shore

and in-river sediments. The longitudinal section

Fig. 4. Near-shore multiport well lateral (B–B 0, C–C 0) and

longitudinal sampling transects (D–D 0) and plan-section sampling

area.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269260

utilised both multiport wells and spear probe sampling

to monitor groundwater quality and contaminant

distribution along the estimated centre line of flow

of the plume as it approached and discharged to the

river. The plan section, which used only spear probe

sampling over a grided area (2 m spaced north-south

grid lines and 1 m east–west sample intervals), was

used to investigate water quality and contaminant

distribution in the area of groundwater discharge near

the sediment/water interface. Samples from the plan

section were taken from 0.20 m below the surface of

the riverbed or 0.20 m below the groundwater table in

the on-shore section of the sampling grid.

Two major monitoring events, designed to provide

snap-shots of the plume distribution during summer

and winter conditions, were carried out in January and

August/September 2001, respectively. The MPs 9–13

were only sampled in August/September 2001. The

monitoring events comprised sampling of all installed

multiport wells for the lateral (y–z) control sections as

well as spear probe sampling of the longitudinal (x–z)

and plan (x–y) control sections. Sampling of selected

multiport wells or control sections was also carried

out intermittently over the investigation in order to

obtain data for analysis of temporal trends.

Groundwater samples were recovered via suction

directly from the access lines of the multiport

boreholes using glass syringes after purging of the

lines. This strategy was adopted to minimise disturb-

ance of the groundwater and sorptive and/or volatile

losses of the BTEX compounds (Davis et al., 1992,

1999). Groundwater samples were recovered from the

spear probe in a similar manner, using a syringe to

draw the sample through a nylon access tube from a

small-screened interval at the head of the spear.

Groundwater quality parameters measured at each

sample location and depth were dissolved oxygen

(DO), electrical conductivity (EC), pH and redox

potential (or Eh). Groundwater samples taken from

the multiport boreholes and spear probing were

analysed for benzene, ethylbenzene, toluene, m- and

p-xylene, o-xylene, 1,3,5-trimethylbenzene, naphtha-

lene, 1-methylnaphthalene, 2-methylnaphthalene,

m- and p-cresol, o-cresol and phenol. Duplicates

were always taken for at least one or two locations per

sampling period, and blanks were also analysed. The

organic compounds were concentrated from ground-

water samples using micro-solvent extraction tech-

niques (Patterson et al., 1993), and analysed by Gas

Chromatography.

3.2. Water level monitoring

River and groundwater levels were monitored via

four online water level loggers. One logger monitored

river water levels, while the remaining three mon-

itored groundwater levels at increasing distances

away from the river foreshore along the proximal

path of the contaminant plume. The groundwater level

loggers were located in existing monitoring wells

located at 8 m (MW33), 28 m (MW43) and 55 m from

the shoreline.

4. Results and discussion

4.1. On-shore distribution of contaminants

In the on-shore sediments between multiport tran-

sects A–A 0 and B–B 0, the distribution of the dissolved

Fig. 5. Lateral depth-sections through the dissolved plume showing benzene, TEX and naphthalene concentrations in the near-source (A–A 0)

multiport well control sections. AMG, Australian map grid, AHD, Australian height datum, MP, Multiport well. TEXZtoluene, ethylbenzene

and the xylene isomers.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 261

BTEXN was as a thin, lenticular plume in cross-

section, with approximate widths of 20–30 m (Fig. 5).

However, at a finer concentration range, the width

of the benzene plume near the source is much

Table 1

BTEXN concentrations from September 2001, and ratios of TEXN to b

(A–A0) and down-gradient (B–B0) control sections

Depth Concentrations (mg lK1)

B T E X

A–A’

MP09 K2.7 459 3 80 4

MP10 K3.2 82 2 17 6

MP11 K3.6 195 32 4401 13850

MP12 K3.6 233 21 2443 9866

MP13 K3.7 371 127 2625 5707

B–B’

MP03 K1.1 7 40 7 27

MP02 K2.5 109 5 8 27

MP08 K3.9 333 5 2 7

MP01 K1.2 789 10 10 28

MP04 K0.7 103 6 2 9

wider—possibly greater than 40 m wide. Table 1

compares the highest BTEXN (benzene, toluene,

ethylbenzene, xylene and naphthalene) concentra-

tions from near-source (MPs 9–13) and near-shore

enzene concentrations from multiport wells along the near-source

Ratio of organics to benzene

N T:B E:B X:B N:B

8 0.01 0.17 0.01 0.02

126 0.03 0.21 0.07 1.53

806 0.16 22.53 70.90 4.13

470 0.09 10.50 42.42 2.02

321 0.34 7.07 15.38 0.86

1 5.82 0.99 3.86 0.11

78 0.05 0.08 0.25 0.71

4 0.02 0.01 0.02 0.01

242 0.01 0.01 0.04 0.31

28 0.05 0.02 0.09 0.27

Fig. 6. Depth distributions of BTEXN concentrations in near-source (MP-11) and near-shore, down-gradient (MP-1) multiport wells.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269262

(MPs 1–4 and 8) multiport wells and shows ratios of

TEXN organics to benzene concentrations. Fig. 6 also

shows detailed BTEXN distributions with depth from

the centreline multiport wells.

Contaminants in the near-source section (A–A 0,

Fig. 5) were observed to impact the groundwater over

a thin vertical zone of approximately 1 m, at depths

between 5.5 and 6.5 m below ground level (also see

MP-11 data in Fig. 6). The contoured groundwater

hydrocarbon concentration data in section A–A 0

(Fig. 5) was mostly dominated by the xylene,

ethylbenzene and naphthalene compounds, with

concentrations in MP11 (Fig. 6) reaching up

to 13,850, 4401 and 806 mg lK1, respectively, in

September 2001, while benzene was found at lower

concentrations of 195 mg lK1. Benzene, however, was

also observed as a more laterally persistent plume,

extending further to the north where it becomes the

dominant hydrocarbon in MP09 (see Table 1).

Contaminant levels and distributions in the near-

shore section (B–B 0) indicate a down gradient vertical

thickening and lateral narrowing of the plume

combined with shallowing and southward migration

Table 2

Mean concentration and dissolved mass in the A–A0 and B–B 0 sections o

Section B T

A–A0 Mean concentration (mg lK1) 36.4 39.7

Mass (g) 1.5 1.6

B–B 0 Mean concentration (mg lK1) 69.5 4.7

Mass (g) 2.8 0.2

Ratio B–B 0: A–A0 1.91 0.12

of the plume centreline. Increased dispersion due to

fluctuating near-shore groundwater pressure gradients

(see Section 4.3) and possibly increasing vertical

groundwater flow directions may explain the observed

vertical thickening of the plume along transect B–B 0.

Benzene and naphthalene are the dominant hydro-

carbons impacting groundwater, while the TEX

compounds were almost at negligible concentrations

(see Table 1).

Table 2 presents dissolved mass calculations for

the individual and total BTEXN compounds in

sections A–A 0 and B–B 0 for September 2001. The

contaminant mass figures stated are based on an

average section length of 27 m, a saturated aquifer

thickness of 6 m, a control section width of 1 m and an

aquifer porosity of 0.25. Using average contaminant

concentrations calculated from multiport sampling,

the dissolved mass was estimated using the equation

(Wiedemeier et al., 1999)

DMt Z Cavg;tbnLW (1)

where DMt [M] is dissolved mass at time t [T], Cavg,t

[MLK3] is the average concentration at time t, n is

f the groundwater plume

E X N Total BTEXN

264.7 602.7 57.3 1001

10.7 24.4 2.3 40.5

4.6 9.9 53.5 140.4

0.2 0.4 2.2 5.7

0.02 0.02 0.93 0.14

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 263

the aquifer porosity [L3LK3], b is the aquifer

thickness [L], L [L] is the lateral plume length and

W [L] is plume width.

Clearly the near-source plume section is signifi-

cantly higher in total contaminant mass with ethyl-

benzene and xylene comprising the dominant portion.

Ratios to benzene along the plume centreline (i.e.

through MP11/MP12 towards MP08/MP01) showed

that the significant mass loss and probably biodegra-

dation of the non-benzene compounds may have

occurred. Benzene concentrations showed no

enhanced depletion trend along the down gradient

flow path, which is in accord with observations from a

BTEX plume in a similar aquifer formation in Perth

(Davis et al., 1999). Although this data suggests loss

of contaminant mass along the plume flow path, it is

recognised that it is difficult to draw such conclusions

based on ‘snap shot’ sampling alone.

Fig. 7. Pore water electrical conductivity (mS cmK1, shaded contours)

longitudinal transect D–D0 for (A) January and (B) September 2001.

4.2. Groundwater/surface water transition zone

Mapping of electrical conductivity and contami-

nant concentration data from the in-river and near-

shore transects indicated the groundwater/surface

water transition zone at the site to be characterised

by: (1) saline water intrusion (salt wedge) landwards

at the base of the aquifer; (2) shore focussed

groundwater and plume discharge and (3) formation

of a shallow hyporheic zone due to brackish/saline

river water infiltration and mixing with groundwater

below the sediment/water interface. These features

are illustrated in the longitudinal (D–D 0) and the plan

sections from January and September 2001, as shown

in Figs. 7 and 9.

Salinity differences between the groundwater and

river water allow electrical conductivity measure-

ments to be used to map the distribution of water types

and total BTEXN concentration (mg lK1, line contours) from the

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269264

and transition zones in the subsurface, and the

discharge of fresh groundwater towards the sediment/

water interface. At the Canning River field site,

fresh groundwater is characterised by EC values of

!1000 mS cmK1, while river water EC is variable,

generally ranging from 4000 mS cmK1 during winter

months to 47,000 mS cmK1 during summer.

4.2.1. Longitudinal trends (transect D–D 0)

Fig. 7 shows closely contoured electrical conduc-

tivity values along the longitudinal transect D–D 0 for

January and September 2001. Total BTEXN concen-

trations are indicated as line contours overlying the

conductivity data. Saline water intrusion into the

aquifer is clearly indicated by elevated pore water

electrical conductivity underlying the fresh ground-

water. At the salt wedge/groundwater interface,

electrical conductivity values were transitional over

a depth interval of 1.0–1.5 m, steadily increasing up

to 45,000 mS cmK1 in the main body of the saltwater

wedge. With increasing distance along the flow path

towards and into the river sediments, the fresh water

lens was observed to thin above a thickening salt

wedge. Consequent groundwater discharge towards

the sediment surface was observed, resulting in either

the formation of a seepage face at the sediment

surface or mixing with river water in a hyporheic

zone. In this case, we define the hyporheic zone as

the region beneath the sediment/surface water inter-

face (i.e. the river bed) with electrical conductivity

values between that of the onshore groundwater

(i.e. O1000 mS cmK1) and river water (seasonally

variable). Both the seepage face and hyporheic zones

were variable in space and time—as shown by

Fig. 8. Depth distribution of BTEXN concentrations in MP05, MP06 and M

the changes in seepage zone length and EC when

comparing January and September 2001 data

(Fig. 7A and B). Changes in the distribution of the

hyporheic zone are primarily a function of the local

surface water (tidal oscillation) and groundwater

hydraulics (this is discussed later). BTEXN concen-

trations along the transect showed the leading front

of the plume (500 mg lK1) extending a maximum of

only some 2.5–5 m into the river sediments from the

shoreline or 10–12 m east of MP08, with abrupt

lateral decreases in contaminant concentration occur-

ring in the area of hyporheic mixing.

Depth profiles of electrical conductivity in shallow

pore water along transect D–D 0 (as depicted in Fig. 7)

clearly show the higher salinity water of the hyporheic

zone situated above fresh groundwater. The general

trend was for the depth and salinity of the hyporheic

zone to increase with increasing distance away from

the shoreline. Conversely, the thickness of the fresh

groundwater lens steadily decreases and is eventually

pinched out where the salt wedge becomes continuous

with infiltrating river water. The maximum hyporheic

zone thickness recorded was at approximately a depth

of 2 m into the river sediments.

4.2.2. Lateral distributions (transect C–C 0)

Contaminant distributions from the in-river lateral

section C–C 0 (Fig. 8), are consistent with vertical flow

dominated, shore-focussed behaviour of groundwater

discharge as indicated by results from section D–D 0.

Elevated BTEXN levels were only intermittently

recorded from sampling points in the upper 2 m of

sediments indicating the rapid shallowing of the

plume in the intertidal zone.

P07 along transect C–C0 in river sediments during September 2001.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 265

4.2.3. Plan view distributions (rectangular grid)

Grid-based spear-probe sampling of shallow pore

water (0.2 m below the sediment surface, Fig. 10) at the

river discharge boundary showed BTEXN compounds

strongly coincident with the detection of fresh to

slightly brackish pore water. Fig. 4 shows the location

of the rectangular grid used for this sampling. The

distribution of measured electrical conductivity values

Fig. 9. Distribution of electrical conductivity (A and C) and total BTEXN

section, during January (A and B) and August/September (C and D) 2001

below the water table (if over land).

and total BTEXN concentrations for summer (January)

and winter (September) periods are shown in Fig. 9.

The EC data show there were strong transient

variations in discharge patterns that may be attributed

to a combination of seasonal and tidal forcing. Data

from the January sampling (Fig. 9, A and B) are con-

sistent with characteristic summer hydraulic con-

ditions where river salinity approaches a maximum

concentrations (B and D) from spear probe sampling of the plan-

. Samples were recovered from 0.2 m below the river bed or 0.2 m

Fig. 10. River stage (Tide) and near-shore (MW33) water table

elevations (Groundwater) over a 10-day period in September 2001.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269266

and groundwater discharge is low. In this case, ground-

water discharge was restricted to a small zone of

limited lateral extent. Sustained high tidal stages

during the January sampling period are likely to have

compounded this effect, disturbing groundwater press-

ure gradients and stimulating hyporheic mixing below

the sediment/water interface. Conversely, the August/

September sampling period was designed to be repre-

sentative of winter conditions where river salinity was

low and groundwater discharge was relatively high.

Overall, the field data (Fig. 9, C and D) showed a

significant spatial increase of groundwater discharge to

the river sediments, with contaminants travelling fur-

ther into the river before being expressed near the

surface. Low tidal stages during the September

sampling are likely to have contributed to the enhanced

discharge of groundwater across the sediment/water

interface.

Of the hydrocarbon compounds, benzene and

naphthalene concentrations were highest in the spear

probed samples, as was the case for the MP samples

nearer the river. The concentration of benzene ranged

up to 550 mg lK1 and naphthalene ranged up to

304 mg lK1. Typically, however, maximum benzene

and naphthalene concentrations were 100–300 mg lK1,

where detectable.

Fig. 11. River stage and water table elevations (in MW33) over a

24-h period in September 2001. The inset table shows decreasing

electrical conductivity of hyporheic pore water at three depths

(0.05, 0.25 and 1.0 m below the sediment interface) on the out

going tide.

4.3. Tidal controls at the discharge boundary

Monitoring of river stage and on-shore water table

heights showed that near-shore groundwater levels

were strongly responsive to tidal oscillations. Water

table heights measured approximately 8 m landward of

the shoreline in MW33 are shown in Fig. 10 with

simultaneous river stage heights over a 10 day period in

early September 2001. Periods of groundwater flow

reversal, where there is a net movement of river water

into the groundwater system, were initially inferred

during some high tide events when river water levels

exceeded those of the near shore groundwater levels.

However, monitoring at MP04 (also 8 m inland from

the shoreline) over tidal periods showed no change in

groundwater quality, with the salt wedge interface

retaining a stable position. Pore water in river

sediments showed infiltration of river water across

the sediment/water interface to form a hyporheic zone

but also indicated a stable salt wedge interface below

the fresh groundwater lens. This suggests that over the

time scale of a tidal period, the observed changes in

near-shore groundwater levels occurred in response to

changing pressure heads in the intertidal region of

groundwater discharge. Tidal oscillations appeared to

have had little effect on salt wedge intrusion distances

into the aquifer. These observations are consistent with

numerical modelling results published by Ataie-

Ashtiani et al. (1999) for scenarios with aquifer depths

much larger than tidal amplitudes. Flatter beach slopes

are likely to intensify the surface water infiltration

phenomenon associated with tidal fluctuations

(Ataie-Ashtiani et al., 1999). Monitoring probes at 18

and 50 m inland from the shoreline showed little

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 267

response (!0.05 m variation) to the changing river

conditions (data not shown).

Transient sampling of water quality in the river

sediments showed a relationship between tidal stage

and groundwater seepage to the river sediments with

freshening of hyporheic pore water on an outgoing tide

and high groundwater head relative to the river stage.

In Fig. 11, river stage tidal oscillations are compared

with groundwater pressures over a 24 h period in

September 2001. Electrical conductivity profiles taken

over part of the tidal period are shown in the inset table.

Changes in pore water pressure relative to the river

stage taken in the upper 0.5 m of the sediment column

using mini-piezometers indicated vertical pressure

gradients of 0.04 m mK1 in the hyporheic zone.

5. Conceptual model

A generalised conceptual model for the interaction

between shallow groundwater, saline surface water

Fig. 12. Conceptual model for groundwater/surface water interaction at t

respect to: (A) the predominant hydrodynamic processes and; (B) the gen

and contaminant discharge at the Canning River site is

shown in Fig. 12. The model, developed from the field

observations presented above, shows three dominant

hydrodynamic features that affect plume discharge

patterns to river sediments at the site: (1) saline water

intrusion landwards into the aquifer; (2) edge-

focussed groundwater discharge and (3) formation

of a shallow hyporheic zone due to saline river water

infiltration mixing with groundwater below the

sediment/water interface.

Hydrodynamic observations at the Canning River

site were consistent with classical seawater intrusion

models (Glover, 1959; Cooper, 1959; Reilly and

Goodman, 1985; Ataie-Ashtiani et al., 1999)—so, as

such, the conceptual model in its simple form as

depicted in Fig. 12 is not new. However, the work

here has help constrain the magnitude of the extent of

the groundwater discharge zone into the river (the

order of 5–10 m), its transient behaviour, and the

variability induced in concentrations due to tidal and

seasonal influences.

he estuarine discharge boundary showing the hyporheic zone with

eral behaviour of the contaminant plume.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269268

With increasing distance towards the shoreline,

groundwater movement above the saltwedge becomes

dominated by vertical flow paths and increased

velocities, with groundwater discharge to surface

water occurring close to the shoreline (Bear, 1979).

Even neglecting density effects, studies have shown

that unconfined groundwater discharges preferentially

at the shoreline of a surface water body (Duffy and

Al-Hassan, 1988; Diersch, 1988). This phenomenon is

commonly referred to as edge-focussed discharge

(Smith and Turner, 2001). Mapping of the dissolved

BTEXN distributions indicated the behaviour of the

hydrocarbon plume at the groundwater/surface water

transition zone to be strongly influenced by the edge-

focussed discharge, with the leading front of the plume

extending only some 2.5–5 m into the river sediments

from the shoreline. Monitoring programs and risk

assessments studies at similar contaminated sites

should therefore, focus efforts within the intertidal

zone where contaminants are likely to impact the

surface environment.

The field observations from this study also show

that the hyporheic zone, although often ignored in

classical seawater intrusion models, can significantly

influence the spatial and temporal discharge patterns

of groundwater at the sediment/water interface.

Neglect of this zone and its dynamics may introduce

significant uncertainties when attempting to estimate

groundwater and solute fluxes across the interface.

Further, results indicated that the hyporheic zone

plays an important role in controlling near-shore

groundwater levels. During increasing tide levels,

estuarine surface water infiltrates into the sediments

and mixes with groundwater to form the hyporheic

zone that effectively acts to retard groundwater

discharge across the sediment/water interface. During

the ebb period, the surface water level recedes past the

groundwater discharge zone. This allows the for-

mation of a seepage face after a lag period during

which the infiltrated saline surface water is either

diluted with groundwater discharge and/or drained

down slope. As a result, the intertidal hyporheic zone

acts as a fluctuating head discharge boundary.

6. Conclusions

Three-dimensional mapping of water quality and a

dissolved petroleum groundwater plume has

improved our understanding and conceptual model

of groundwater and plume behaviour at an estuarine

river discharge boundary. Results indicated that

shore-focused discharge strongly influences the area

of contaminated groundwater seepage to the estuarine

environment. Electrical conductivity measurements

show that groundwater seepage to the estuarine

surface sediments occurs in a zone less than 10 m

from the high tide mark, and mostly less than 5 m

from the shoreline. Tidal activity plays an important

role, inducing a fluctuating head boundary and the

formation of a hyporheic zone due to infiltration of

saline surface water in the intertidal sediments. Given

that the presence of the hyporheic zone can influence

flow, models not properly accounting for its presence

will give a higher degree of uncertainty when

estimating groundwater and solute fluxes across the

sediment/water interface. Seasonal and tidal fluctu-

ations of the interface location plays a role in re-

distributing hydrocarbon concentrations, governing

the lateral extent of discharge into the river, and the

proximity of high hydrocarbon concentrations to the

sediment/surface water interface.

These observations and conceptual model, along

with other results obtained in the study, will be used to

more fully model the exchange processes and provide

greater confidence in the bio-attenuation and transport

processes that may occur where groundwater dis-

charges to such surface water environments.

Acknowledgements

BP Soil and Groundwater Centre of Expertise

and the Shell Company of Australia part-funded

the project. We thank Nathan Innes, Blair Robertson,

David Briegel, Robert Woodbury and many others for

their assistance with installations, sampling and

interpretation.

References

Acworth, R.I., Dasey, G.R., 2003. Mapping of the hyporheic zone

around a tidal creek using a combination of borehole logging,

borehole electrical tomography and cross-creek electrical

imaging, New South Wales, Australia. Hydrogeology Journal.

11, 368–377.

S.J. Westbrook et al. / Journal of Hydrology 302 (2005) 255–269 269

ANZECC, 1992. Australian water quality guidelines for fresh and

marine waters, National Water Quality Management Strategy.

Australian and New Zealand Environmental Conservation

Council, Canberra.

Ataie-Ashtiani, B., Volker, R.E., Lockington, D.A., 1999. Tidal

effects on sea water intrusion in unconfined aquifers. Journal of

Hydrology 216, 17–31.

Bear, J., 1979. Hydraulics of Groundwater. McGraw-Hill, New York.

Boudreau, B.B., Jorgensen, B.B. (Eds.), 2001. The Benthic

Boundary Layer: Transport Processes and Biogeochemistry.

Oxford University Press, Oxford.

Clement, T.P., Wise, W.R., Molz, F.J., Wen, M., 1996. A

comparison of modeling approaches for steady-state unconfined

flow. Journal of Hydrology 181, 189–209.

Cooper, H.H., 1959. A hypothesis concerning the dynamic balance

of fresh water and salt water in a coastal aquifer. Journal of

Geophysical Research 64, 461–467.

Davidson, W.A., 1995. Hydrogeology and groundwater resources of

the Perth Region, Western Australia. Bulletin 142. Geological

Survey of Western Australia, Perth, WA.

Davis, G.B., Barber, C., Briegel, D., Power, T.R., Patterson, B.M.,

1992. Sampling Groundwater Quality for Inorganics and

Organics: Some Old and New Ideas, Drill ’92. Australian

Drilling Industry Assoiation Conferece, Perth October 1992

1992 pp. 24.1.–.9.

Davis, G.B., Barber, C., Power, T.R., Thierrin, J., Patterson, B.M.,

Rayner, J.L., Wu, Q., 1999. The variability and intrinsic

remediation of a BTEX plume in anaerobic sulphate-rich

groundwater. Journal of Contaminant Hydrology 36, 265–290.

Diersch, H.J., 1988. Finite element modelling of recirculating

density-driven saltwater intrusion processes in groundwater.

Advances in Water Research 11, 25–43.

Duffy, C.J., Al-Hassan, S., 1988. Groundwater circulation in a

closed desert basin: topographical scaling and climatic forcing.

Water Resources Research 24 (10), 1675–1688.

Gallagher, D.L., Dietrich, A.M., Reay, W.G., Hayes, M.C.,

Simmons Jr.., G.M., 1996. Ground water discharge of

agricultural pesticides and nutrients to estuarine surface water.

Ground Water Monitoring and Remediation 16 (1), 118–129.

Glover, R.E., 1959. The pattern of fresh-water flow in a coastal

aquifer. Journal of Geophysical Research 64, 457–459.

Harvey, J.W., German, P.F., Odum, W.E., 1987. Geomorphological

control of subsurface hydrology in the creekbank zone of tidal

marshes. Estuarine and Coastal Shelf Science 25, 677–691.

Krabbenhoft, D.P., Bowser, C.L., Anderson, M.P., Valley, J.W.,

1990. Estimating groundwater exchange with lakes: 1. The

stable isotope mass balance method. Water Resources Research

26 (10), 2445–2453.

Linderfelt, W.R., Turner, J.V., 2001. Interaction between shallow

groundwater, saline surface water and nutrient discharge in

a seasonal estuary: the swan-canning system. Hydrological

Processes 15, 2631–2653.

Lorah, M.M., Olsen, L.D., 1999. Natural attenuation of chlorinated

volatile organic compounds in a freshwater tidal wetland: field

evidence of anaerobic biodegradation. Water Resources

Research 35 (12), 3811–3827.

Patterson, B.M., Power, T.R., Barber, C., 1993. Comparison of two

integrated methods for the collection and analysis of volatile

organic compounds in ground water. Ground Water Monitoring

and Remediation 13 (3), 118–123.

Portney, J.W., Nowicki, B.L., Roman, C.T., Urish, D.W., 1998. The

discharge of nitrate-contaminated groundwater from developed

shoreline to marsh-fringed estuary. Water Resources Research

34 (11), 3095–3104.

Reilly, T.E., Goodman, A.S., 1985. Quantitative analysis of

salt water–freshwater relationships in groundwater systems—a

historical perspective. Journal of Hydrology 80, 125–160.

Robinson, M.A., Gallagher, D.L., 1999. A model of groundwater

discharge from an unconfined coastal aquifer. Ground Water 37

(1), 80–87.

Robinson, M.A., Gallagher, D.L., Reay, W.G., 1998. Field

observations of tidal and seasonal variations in ground water

discharge to estuarine surface waters. Ground Water Monitoring

and Remediation 18 (1), 83–92.

Simpson, M.J., Clement, T.P., 2004. Improving the worthiness of

the Henry problem as a benchmark for density-dependent

groundwater flow models. Water Resources Research 40 (1),

W01504 doi:10.1029/2003WR002199.

Simpson, M.J., Clement, T.P., Gallop, T.A., 2003. Laboratory and

numerical investigation of flow and transport near a seepageface

boundary. Ground Water 41 (5), 690–700.

Smith, A.J., Turner, J.V., 2001. Density-dependent surface water–

groundwater interaction and nutrient discharge in the Swan-

Canning Estuary. Hydrolological Processes 15, 2595–2616.

Tobias, C.R., Harvey, J.W., Anderson, I.C., 2001. Quantifying

groundwater discharge through fringing wetlands to estuaries:

seasonal variability, methods comparison, and implications for

wetland-estuary exchange. Limnology and Oceanography 46

(3), 604–615.

Westbrook, S., Davis, G.B., Rayner, J.L., Fisher, S.J.,

Clement, T.P., 2000. Initial site characterisation of a dissolved

hydrocarbon groundwater plume discharging to a surface water

environment, in: Johnston, C.D. (Ed.), Contaminated Site

Remediation: From Source Zones to Ecosystems Proceedings

2000 Contaminated Site Remediation Conference, Melbourne,

4–8 December 2000, pp. 189–196.

White, D.S., 1993. Perspectives on defining and delineating

hyporheic zones. Journal of North American Benthological

Society 12 (1), 61–69.

Wiedemeier, T.H., Rifai, H.S., Newell, C.J., Wilson, J.T., 1999.

Natural Attenuation of Fuels and Chlorinated Solvents in the

Subsurface. Wiley, New York.

Winter, T.C., 2000. Interaction of Ground Water and Surface Water,

Proceedings of the Ground-Water/Surface-Water Interactions

Workshop. US Environmental Protection Agency. EPA/542/R-

00/007. July 2000 2000.