Journal of Hydrology 406 (2011) 16–29

Contents lists available at ScienceDirect

Journal of Hydrology

journal homepage: www.elsevier .com/locate / jhydrol

Defining the hyporheic zone in a large tidally influenced river

M.S. Bianchin a,⇑, L. Smith b, R.D. Beckie b

a Lorax Environmental Ltd., 2289 Burrard Street, Vancouver, British Columbia, Canada V6J 3H9b Department of Earth and Ocean Sciences, University of British Columbia, 6339 Stores Road, Vancouver, British Columbia, Canada V6T 1Z4

a r t i c l e i n f o

Article history:Received 30 November 2009Received in revised form 3 March 2011Accepted 24 May 2011Available online 12 June 2011This manuscript was handled by P. Baveye,Editor-in-Chief

Keywords:Hyporheic zoneDelineationTidesFraser RiverGeophysicsGroundwater profiling

0022-1694/$ - see front matter � 2011 Elsevier B.V. Adoi:10.1016/j.jhydrol.2011.05.056

⇑ Corresponding author.E-mail addresses: mbianchin@lorax.ca (M.S. Bian

Smith), rbeckie@eos.ubc.ca (R.D. Beckie).

s u m m a r y

An investigation was conducted to characterize the spatial and temporal distribution of the hyporheiczone of a large tidally influenced river. The field site is located on the Fraser River in British Columbia,Canada, approximately 30 km from its outlet to the ocean. The physical attributes of the riverbed weremapped using geophysical techniques coupled with sediment sampling. The spatial and temporal distri-bution of groundwater composition beneath the riverbed was determined through detailed profiling.Contaminated (fresh) groundwater discharges through a narrow band of the riverbed at a distanceapproximately 88–105 m from the shoreline coinciding with the termination of a massive silty unit. Sal-ine groundwater, as part of a regional flow system, dominates the riverbed sediments from 105 m beyondthe shoreline towards the centre of channel. Three water types occur within the upper 2 m of the riverbedsediments; a result of both mixing of river water, contaminated (fresh) groundwater, and saline ground-water and modification by cation exchange reactions. The interaction of these waters produced distinctzones of Ca–HCO�3 , Na–Cl, and Ca–Cl type waters. The distribution of groundwater solutes indicates thatduring a single tidal cycle, river water penetrates the riverbed to a depth of approximately 15 cm but thelong term effects of tidal pumping of river water into the riverbed is observed to a depth of approximately1 m below the river bed.

� 2011 Elsevier B.V. All rights reserved.

1. Introduction

Surface water bodies such as lakes and rivers can be an integralcomponent of a groundwater flow system. Linkages between thetwo are controlled by the physiography of the landscape, definedby geologic framework and topology, and climate. These linkscan be part of a regional, intermediate or local scale groundwaterflow system. A critically important interaction between groundwa-ter and surface water is hyporheic flow, where water flows to andfro between near-river channel sediments and the active channel.It is at this scale that hyporheic flow paths, which return to thestream in distances less than tens of meters, are distinguished fromthe regional-scale paths that support base flow. Hyporheic pro-cesses are viewed as smaller scale interactions between channelwater and groundwater occurring within larger-scale patterns ofloss and gain of channel water in drainage basins (Harvey andWagner, 2000).

Several definitions have been proposed for hyporheic zones(HZ). For example, Hynes (1974) defined the HZ based on observa-tions of stream organisms and dissolved oxygen, while Triska et al.(1989) defined the HZ as the region where subsurface water con-tains at least 10% surface water. In this paper the definition pro-

ll rights reserved.

chin), lsmith@eos.ubc.ca (L.

posed by White (1993) is adopted: the saturated interstitial areasbeneath the stream bed and into the stream banks that containsome proportion of channel water as a result of hyporheic flow.In a tidally-influenced river, hyporheic flow is dominated by tidalforcing: flow paths are oscillatory (recharging and discharging) un-der high- and low-tide river stages, respectively (Bianchin et al.,2010).

The characteristics of groundwater–surface water interactions(GWSi) on smaller order streams (Anderson et al., 2002; Harveyand Bencala, 1993; Harvey et al., 1996; Kasahara and Wondzell,2003; Landmeyer et al., 2010; Wroblicky et al., 1998), and onephemeral streams (Boulton and Stanley, 1995; Stanley andBoulton, 1995; Valett et al., 1994) are comparatively better under-stood than GWSi of larger, tidally-influenced rivers. Only a fewstudies such as that of Hinkle et al. (2001) on the Willamette Riverin Oregon have been conducted on larger order streams. Moststudies on smaller systems have focused on the influence of river-bed geomorphology on GSWi. Bed-induced perturbations tostream flow result in pressure variations that drive the exchangeof surface water and groundwater at the stream bed (Andersonet al., 2002; Cardenas et al., 2004; Harvey and Bencala, 1993;Kasahara and Wondzell, 2003; Marion et al., 2002; Packman andBrooks, 2001; Storey et al., 2003). The composition of the riverbed sediments also influences the degree to which bed-forminduced hyporheic exchange occurs (Cardenas et al., 2004; Salehinet al., 2004; Storey et al., 2003; Vinson et al., 2001).

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 17

Tidal pumping is also a mechanism for GSWi in rivers adjacent tothe coastal zone (Land and Paull, 2001; Trefry et al., 2007; West-brook et al., 2005). Westbrook et al. (2005) and Trefry et al. (2007)investigated the hypoaigic zone [otherwise known as the subterra-nean estuary (Moore, 1999)] beneath the Canning River, near Perth,Australia. They showed the effect of a seasonally stratified river ongroundwater discharge patterns and discussed implications for con-taminant transport through this mixing zone. However, the fre-quency of their data collection was too low to observe the effect ofthe tides on groundwater – surface water exchange. The effect of ti-dal pumping on contaminant transport has been the focus of severalmodel studies (e.g. (Neeper, 2001; Yim and Mohsen, 1992). Neeper(2001) found that oscillatory flow in the presence of a sorbed phasecontaminant increases the time-average flux of contaminantsthrough increased dispersal, beyond that which would be found insteady flow without sorption processes. Yim and Mohsen (1992)showed that tides could result in the mixing of surface water withgroundwater, diluting contaminants up to a distance of 12 m inlandfrom the surface water–aquifer interface. This large mixing zone re-flects the relatively large dispersivity value (�3 m) used in their sim-ulations. They noted that tidal pumping hastened the migration ofcontaminants to the estuary in comparison to a non-tidal simula-tion. These results were not verified by field data.

The effects of tidal pumping on submarine groundwater dis-charge (SGD) have also been investigated (Burnett et al., 2003;Land and Paull, 2001; Preito and Destouni, 2005; Robinson et al.,2007; Taniguchi, 2002). Preito and Destouni (2005) and Robinsonet al. (2007) have shown that the discharge characteristics ofSGD, that is, size of discharge zone and degree of surface water[ocean water] and groundwater mixing are controlled by the mag-nitude of groundwater flux and amplitude of tidal oscillation. Majiand Smith (2009) emphasize the importance of mixed-water dis-charge occurring within the intertidal zone.

Fig. 1. The field site is located offshore of a wood preservation facility on the north banshore contamination with wood preservatives has led to the development of a dissolved pit eventually discharges to the river.

Little is known of the hyporheic zone of large rivers and, to theknowledge of the authors no such work has been conducted on largetidally-influenced rivers in that part of the estuary beyond the land-ward ingress of saline (ocean) water. Further, one cannot extrapolatethe effect of tidal pumping from SGD studies because density depen-dent flow, wave fetch and slope break modify the groundwater flowpatterns and are not active on a tidally-influenced river.

The transport of groundwater contaminants through the river-bed sediments of a tidally influenced river is complex. Consideringthe many processes on a tidally-influenced river that could drive ex-change, the nature and spatial extent of hyporheic flow can also beexpected to be complex; leading to the following key questions:Where does river water recharge the aquifer or near channel sedi-ments? To what depth does river water penetrate the riverbed? Doesthe depth of penetration vary across the riverbed and if so, why?Based on the observed GWSi patterns on the riverbed, is it possibleto develop a qualitative statement of the process or processes likelyresponsible for driving this exchange? The motivation for this studyis to gain an understanding of how tidal forcing on a large river influ-ences GWSi and ultimately the physical and chemical characteristicsof the hyporheic zone, both spatially and temporally. The objectivesinclude: (1) determining the extent to which river water penetratesthe river bed and the resulting influence on groundwater chemistry,thus delineating the hyporheic zone; and (2) determining the con-trols on how and where groundwater discharge occurs.

1.1. Site description

The location of the field site is on the north bank of the FraserRiver, near Vancouver, Canada (Fig. 1). Historic wood treatmentpractices dating back over 70 years have led to a zone of non-aque-ous-phase creosote that penetrates 27 m beneath the ground sur-face. Groundwater flow south across the site towards the Fraser

k of the Fraser River in the Lower Mainland Area of British Columbia, CANADA. Onhase PAH plume that extends approximately 100 m south of the riverbank to where

18 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

River contains dissolved-phase polycyclic aromatic hydrocarbons(PAHs). The groundwater plume extends beneath the river anddischarges some 100 m from the north bank (Bianchin et al., 2006)(see Figs. 1 and 3). The PAH plume is underlain by saline groundwa-ter. Deeper, upward flowing groundwater that represents regionaldischarge has higher dissolved solids and is distinguished fromother groundwater types by its higher chloride content (Fig. 3).The source of saline groundwater may be connate water in the Pleis-tocene-aged sediment immediately underlying the Fraser RiverSand unit (Bridger and Allen, 2006).

In this discharge portion of the aquifer, dissolved and sorbed-phase contaminant concentrations in the upper 1 m of the riverbedare below analytical detection limits (Anthony, 1998; Bianchin,2001; Bieber, 2003; Bianchin, 2010). The attenuation of PAH con-tamination is likely a function of dispersion and biodegradation inthe hyporheic zone. Towards the contaminant source zone on land,the aquifer is anaerobic and characterized by elevated concentra-tions of ferrous iron (Fe+2) and methane (CH4); the by-products ofanaerobic degradation of PAH’s by iron reduction and methanogen-esis (Bianchin et al., 2006). Currently, site management involvesplume containment using a pumping well (DWW-5 shown in Fig. 2).

1.2. Hydrology of the Fraser River estuary

The Fraser River drainage basin has an area of 233,000 km2 anddischarges to the Strait of Georgia, 1370 km from its headwaters.

Fig. 2. Detailed map of the site and the locations of surface geophysical survey lines. Sushown as well as the results of a river bathymetric survey conducted in 1999 (Bianchin

The mean annual flow is 2720 m3 s�1 (Government of Canada,2008). During the winter months the discharge is usually less than1500 m3 s�1. Flow averages above 4000 m3 s�1 during freshet withpeak flow ranging from 5000 to 15,000 m3 s�1. The snow-melt fre-shet occurs from May to mid-July. The range in tidal fluctuationsat the field site, situated approximately 30 km from the Fraser Riv-er’s outlet to the Straight of Georgia, varies between 2 and 3 m.The river has a mean depth of approximately 12 m near the centreof the nearly 1 km wide channel. The maximum inland extent ofsea water in the channel is located approximately 14 km down-stream from the site (Ages, 1979).

1.3. Hydrogeology of the Fraser River Sands aquifer

The hydrostratigraphy of the site is represented in Fig. 3. Thechannel sand unit with a maximum thickness of 27 m (the FraserRiver Sands aquifer) is capped by a low permeability silty-clay layer(�1–5 m thick), interpreted as over bank deposits. Offshore, this unitis 1 to 4 m thick and consists of fine river sediments (silt and finesand) and organic material (bark and logs) to a distance of 100 mfrom the riverbank. The Fraser Sands Aquifer is underlain by a dense,low permeability, sandy-silt unit referred to as Pleistocene-agedsediments, with a vertical thickness not determined at site.

Cores, cone penetrometer testing, and hydraulic testing indicatethat the sand aquifer is homogeneous, with only minor silt string-ers. Transmissivity values obtained from pumping tests at wells

rface geophysical (Seismic (blue solid lines) and GPR (red dashed lines) traces are, 2001).

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 19

DWW-1 and DWW-3 [Fig. 2] are 1.2 � 10�3 and 1.5 � 10�2 m2 s�1,respectively [Golder Associates, per. comm., 1997]. Transmissivityvalues derived from tidal analyses are on the order of 3.6 �10�2–5.3 � 10�2 m2 s�1 (Golder, 1997).

For most of the year, the net groundwater flow at the site is fromthe northern uplands south towards the river. The groundwater flowregime can be separated into two distinct periods defined in terms ofgroundwater gradients; higher gradient and low gradient periods(Zawadzki et al., 2002). Low gradient conditions (5 � 10�4) occurwhen the river stage is high (during spring freshet) and upland re-charge of the aquifer by precipitation is low. This condition typicallyoccurs from May to September. Higher gradient conditions(3 � 10�3) occur when river stage is low and upland recharge is high.This condition occurs from October/November to April, coincidingwith the rainy season (average annual precipitation of 1200 mm).On occasion, for a short period of time during the freshet, a reversalin net gradient has been observed with a net inflow of river waterinto the aquifer (Zawadzki et al., 2002).

The effect of the containment well on the hydraulic gradient hasbeen examined by Zawadzki et al. (2002). The greatest influence oninstantaneous gradients occurs during the mid-tidal cycle when thehydraulic gradient, as a result of tidal pumping, is relatively low. Theeffect of containment well pumping on the hydraulic gradient isnegligible during high and low tides, when gradients are relativelyhigh. On a seasonal time scale, the capture zone of the well is largestduring low gradient (spring to early summer) conditions and issmallest during high gradient (fall to winter) conditions. Groundwa-ter modelling indicates that outside the capture zone, groundwaterflow is essentially perpendicular to the river during the high gradi-ent winter season (Bieber, 2003; Zawadzki et al., 2002).

2. Methods

Data collection began in May 2004 and ended in December2006. The field program consisted of geophysical surveys of theriverbed (including ground penetrating radar (GPR), seismic reflec-tion and, bulk resistivity profiling), groundwater profiling and sed-iment sampling. With the exception of GPR and seismic surveys, allactivities were conducted from aboard HMV Ocean Venture, a 70-foot fishing vessel. This floating platform was kept stationary onthe river using a multipoint anchoring system. The positioning ofthe sampling stations off the platform were determined by triangu-lation with two known fixed points on the river using a BushnellLaser Rangefinder Yardage Pro 500 with an accuracy of ±1 m.

The choice of investigative methods was critical to achievingthe goals of the study as the equipment must be able to withstandthe harsh elements of the Fraser River. Collecting data from theaquifer underlying the Fraser River is challenging because the riveris deep (up to 9 m in the immediate study area), fast flowing(velocities reaching 3 ms�1), and tidally-influenced with diurnalaltering of river stage and direction of river flow. It is also sedimentladen with zero visibility, effectively prohibiting the use of diversto place instruments. In addition, large submerged ‘‘deadhead’’ logscan easily destroy instrumentation. The river bed is littered withindustrial debris, predominantly rocks and sunken logs, makingpenetration by tools difficult and sometimes impossible.

2.1. Geophysical characterization of riverbed geology

A key step in quantifying hyporheic exchange at the site was thedetermination of the lateral extent and continuity of the overbanksilt deposits that cap the Fraser River Sands. ‘‘Windows’’ throughthis silt layer could provide preferential pathways for groundwaterdischarge in the offshore zone, as seen elsewhere (Conant et al.,2004; Nyquist et al., 2008; Slater et al., 2010). Geophysical surveys

have proven useful in delineating zones of GWSi where there is asignificant contrast in sediment porosity and conductivity (Butleret al., 2004; Cross et al., 2008; Lendvay et al., 1998; Naegeliet al., 1996; Ward et al., 2010). The sediment profile was character-ized using ground penetrating radar, a reflection seismic survey,resistivity profiling, coring, and spot inspection using scuba divers.The use of several techniques is advantageous because each meth-od has limitations related to the attenuation of instrument signals.The interpretation of geology beneath the river bed is an iterativeprocess of data synthesis from all methods. An initial reconnais-sance was conducted using water borne geophysical surveys ofthe type described by Haeni (1996). The results of these waterborne surveys served to focus ground truthing and other intrusiveinvestigative techniques.

2.1.1. Ground penetrating radar (GPR)The salt content of the groundwater in the aquifer sands off-

shore was expected to be more conductive than the overlyinglow permeability sediment closer to shore. Therefore, a ground pe-netrating radar (GPR) survey, which is sensitive to this contrast inconductivity, was deployed (Butler et al., 2002; Ward et al., 2010).A detailed account of the application of GPR in hydrogeologic stud-ies can be found in Beres and Haeni (1991).

A ground penetrating radar (GPR) survey was carried out onFebruary 7, 2005 using a pulseEkko GPR full bistatic configurationunit (Sensors and Software, Mississauga, Ontario) with 100 MHzfrequency antennae. The antennae were enveloped in thin (2 mil)polyethylene plastic to protect them from river water and theywere placed on either side of a 3 m rubber inflatable dinghy. Sur-veys were conducted along prescribed transects that primarily tra-versed the river channel (Fig. 2). Three shore-perpendicular andfour shore-parallel lines were surveyed at approximately 40 mintervals. The movement of the boat was recorded in real timeusing a mapping grade Trimble global positioning system (G PS)in differential mode.

2.1.2. Seismic reflection surveyingA seismic reflection survey complemented the GPR survey by

providing information on sedimentary structures where EM signalsfrom the GPR system were greatly attenuated. The seismic surveywas conducted on the July 10, 2005. Instrumentation for the reflec-tion seismic survey included a Datasonic Model 1200 SubbottomProfiler, with a NWGS marine hydrophone receiver eel and AppliedAcoustic Engineering geopulse transducer. The survey was con-ducted using an aluminum boat with the receiver mounted onthe forward starboard side at a distance of approximately 0.6 mfrom the hull of the boat. The transducer was positioned on theport side of the stern. Three shore-parallel and 13 shore-perpendic-ular lines at approximately 20 m intervals were surveyed, (seeFig. 2). A Krohn-Hite Amplifier Filter was applied in the field usinga Chesapeake Digital Acquisition system. The data was not postprocessed.

2.1.3. Bulk resistivity profilingBulk resistivity profiles of the riverbed were collected in the

area of groundwater discharge between April 21 and June 23 of2006. The Fraser River sediments contain high gas concentrationswhich impaired the seismic survey and in places the sedimentscontain higher salinity ground water which impaired the GPR sur-vey. The objective of the bulk resistivity profiling was to providedetails of the active zones of GWSi that are not provided by theGPR and seismic surveys. This included vertical lithologicprofiles and profiles of fluid conductivity, essentially mappingout groundwater types. Details of a laboratory study involvingtests of the resistivity probe made on samples of Fraser River sed-iment is available in (Bianchin, 2010).

Fig. 3. Detailed map of the site and the locations of sampling points in plan view (inset upper right) and in cross section along the transect line A-A’ (lower inset).

Table 1The range of fluid conductivity for end members and related bulk resistivity (qt) ofFraser Sand observed beneath the Fraser River offshore of the Braid Street site.

Description of water end member Fluid conductivity (mS/cm) qt (X m)

River water 60–100 500–600Contaminated groundwater 400–600 100–150Saline groundwater 2000–6000 <50

20 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

The results of the laboratory testing suggest that there is suffi-cient resistivity contrast between end member water types at thesite and in the sediment matrix to use bulk resistivity measure-ments to map water salinity and sediment types. Table 1 summa-rizes the salinity of end member water types expected within andbeneath the river accompanied by the bulk resistivity of thesewater types with Fraser River Sand.

Sediment resistivity was profiled using a drive-point in situresistivity probe. The basic principle of operation involves applyinga current through excitation electrodes of one array and measuringthe resulting potential at potential electrodes of another array. Theprobe was built to the specifications described by Rosenbergeret al. (1999) and differs only in the method of deployment, sensorexcitation and data logging. The calibration of the probe produceda configuration factor of 0.3144 m with a standard deviation of±0.0125 m, in good agreement with the theoretical value of0.3181 m determined by Rosenberger et al. (1999). The probewas deployed using drill rods and a pneumatic hammer. The depthof the probe below the river bed was determined by measuring thelength of drill rod above the river bed. The elevation of the riverbed was determined accurately by measuring the length of drillrod above the river bed at the point when the resistance valuesmeasured by the probe differed from that of fresh water (a valuearound 500 X), and from a water gauge surveyed at the site. Theprobe was advanced in 30 cm intervals to depths ranging from 2to 5 m. At each depth 50 measurements were collected with a sam-pling frequency of two measurements per second. The probe wascontrolled using a Campbell Scientific CR10x datalogger in a fourwire half bridge configuration with 1000 X scientific grade refer-ence resistor. The four wire half bridge program provided withthe CR10x [Instruction 9 (Four wire half bridge)] was used to excitethe sensor and make differential voltage measurements, thusavoiding the need to correct for resistance associated with lead

length. The program applied a 250 mV slow integration excitationcorresponding to an alternating current (AC) frequency of 367 Hz,which reduced the effect of ionization (polarization) of the elec-trodes. The probe was also fitted with a thermistor to allow tem-perature corrections to the voltage measurements if needed.

2.2. Core collection

A drive-point piston-sampler (DPPS) (Starr and Ingleton, 1992)fitted with a sample-freezing drive shoe (Murphy and Herkelrath,1996) was used to collect cores of cohesionless sand from the riverbed. This method allows the collection of cores at considerable depthwithout a drilling rig and with a high degree of sample recovery andintegrity. Clear PVC Vacuum tube (50.8 cm OD) in 162.6 cm lengthswere used as core liners. Liquid carbon dioxide (CO2) was used as thefreezing agent. The depth of the core interval below the riverbed wasdetermined following the same technique used for the resistivityprobe. A total of six sediment cores were collected from the riverbedin the zone where fresh groundwater discharges. The first threecores spanned a depth range from 0.3 m below river bed (m.b.r.b.)to 4.6 m.b.r.b. with an average core interval of approximately1.5 m. A second set of three cores were collected approximatelyone-meter away from the first set spanning the same interval.

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 21

2.3. Groundwater sampling

Following Anthony (1998), groundwater samples at depth werecollected using a Waterloo Drive Point Profiler (WDPP) (Pitkinet al., 1994). The WDPP is capable of profiling an aquifer with a res-olution of approximately 30 cm which is suitable for deeper por-tions of the aquifer where conditions are more homogeneous.However, for a HZ estimated to be less than 1.5 m thick a highersampling density is required. A modified version of the WDPPmethod, a multilevel drive point well (MDPW), was deployed tosample the HZ continuously over several tidal cycles. A MDPWwas equipped with six sampling ports spaced at 30 cm. To achievea vertical spacing of 15 cm two MDPWs were used installed 0.5 mapart (laterally) and offset vertically by 15 cm. Details on the con-struction of the MDPW can be found in Bianchin (2001).

Groundwater samples were collected using a peristaltic pump.Samples for cation analyses were preserved in the field by first fil-tering with a 0.45 lm membrane filter followed by pH adjustmentto a pH of 2 using concentrated nitric acid. Samples were stored incoolers under ice-packs and shipped to commercial laboratories lo-cated in Vancouver, British Columbia (see Appendix E (Bianchin,2010)).

3. Results and discussion

3.1. Sediment composition of the riverbed and underlying aquifer

In general, the riverbed sediments may be characterized asthree units: a thin silty fine sand layer containing significant woo-dy debris (wood chips) ubiquitous on the riverbed; a low perme-ability organic-rich silty-clay unit varying in thickness fromabout 5 m near shore, thinning to the sub-meter scale at approxi-mately 100 m offshore; and a massive sand unit making up the Fra-ser Sands aquifer. The silt unit is heterogeneous, containing buriedlogs and other industrial debris (chains, concrete, rocks), and a con-siderable amount of gas, likely CO2, from the degradation of the or-ganic fraction of the silt and anaerobic degradation of naphthalenein the underlying sands.

Fig. 4. Shore perpendicular profile of the river and underlying sediments along GPR surveFigure 2. A: strong reflector representing river/river bed interface; B: river bed multiple;result of gas charged sediments; E: strong reflector at interface of unconsolidated sedim

The seismic and GPR surveys provide important insight into thedistribution and composition of sediments at the riverbed. TransectA–A0 of Fig. 3 tracks closely along the seismic and GPR survey lines0 + 60 and L3, respectively (see Fig. 2) and can be considered repre-sentative of all the survey lines. Seismic line 0 + 60 and GPR line L3are shown in Fig. 4 with a vertical exaggeration of approximately3.3 and 5, respectively. Both lines are approximately 230 m longextending from the north bank of the river south to the northernedge of the Sapperton sand bar, located in the centre of the channel.

The results of the seismic survey provide good definition of thesub-bottom. The seismic profile shows a strong, nearly continuousupper reflector (triple polarity of white, black, white banding) repre-senting the river bottom. The continuity of this reflector is brokenfrom UTM 5452430N to 5452400N where diffraction hyperbolasdominate. The hyperbolas likely represent logs or other large debristhat have accumulated in this low point of the channel. Difficulty insediment and groundwater sampling in this area confirms this con-dition. Closer to the shoreline at UTM 5452480, the sub-bottom de-tail is masked by river bottom multiples (repeat reflection of energyat riverbed with longer travel time resulting in the reproduction ofthe riverbed surface at a deeper depth). However, a little further off-shore extending across to the Sapperton sand bar (at 5452250 m)another strong reflector occurs at 0.5 m below riverbed. This strongreflector likely represents the boundary between a layer of looseunconsolidated sediments and underlying sand. The sediments arebe interpreted to be Gyttja, a nutrient-rich sedimentary peat con-sisting mainly of plankton, other plant and animal residues, andmud (Canada, 1976) combined with small-sized woody debris (likebark), and the underlying sand. This Gyttja-type material was ob-served along the entire riverbed during the survey. South of5452400 m, below the Gyttja layer, reflecting surfaces, albeit lessdistinct (an indication of reflection coefficient closer to unity), arehorizontally trending and somewhat uneven, representing horizon-tally trending, inter-bedded sediments.

Of particular interest on the 0 + 60 profile is the area between5452460 m and 5452430 m showing poor sub-bottom penetrationby seismic energy. The acoustic turbidity appears as diffuse andchaotic seismic facies masking nearly all other reflections and most

y line L3 (Plate 1) and seismic survey line 0+60 (Plate 2) lying along Transect A-A’ ofC: diffraction hyperbola (logs and/or larger debris); D: acoustic turbidity likely theents overlying sand of the Fraser Sands Aquifer.

22 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

likely results from scattering of the acoustic energy by interstitialgas bubbles in the sediment (Schubel, 1974). Gas bubbles, easily vis-ible at the surface of the river, were observed during groundwatersampling events when the insertion of drill rods into the low perme-ability sediments (silty unit) provided preferential flow paths for theescaping gases. The area underlying the low permeability sedi-ments, which is some distance from where groundwater dischargesto the river, is a zone of active anaerobic degradation (Bianchin et al.,2006) leading to the production of large amounts of CO2 (g). Acousticturbidity was visible in all survey lines from 0 + 00 to 1 + 20, wheresurvey lines cover the submerged bank. The remaining lines do notextend to the shoreline as that part of the river was inaccessibledue to the presence of moored barges. It is likely that gas-chargedsediments exist further along the shoreline due in part at least togroundwater-PAH contamination which extends approximately200 m up river from transect A–A0 (Fig. 2) (Anthony, 1998; Bianchin,2001). Gas-charged sediments are not observed further offshore be-yond the extent of the low permeability sediment-cover. This likelyindicates that gas bubbles are discharged with groundwater to theriver. A shore parallel seismic profile (survey line 3) is shown inFig. 5. Line 3 lies along the 5452400 Northing line of Fig. 4 which isat the edge of the silty unit capping the sands. Acoustic turbiditydoes not appear on this profile however, multiple hyperbolas occurfrom 509025E to 509100E and are likely due to logs or other debris.Further east, the riverbed is dominated by a strong continuousreflector representative of the central portion of the channel whichis dominated by sand.

While results of the GPR survey are considered to be of mediumquality, they provide valuable insight into the depth of the riverwater–saline groundwater interface beneath the riverbed. GPR sur-vey line L3 which also plots closely to transect A–A0 (Fig. 3) isshown in figure. The depth scale on the plots is based on an EM-wave velocity (V) in water (0.033 m nS�1) whereas in saturatedsand the V = 0.055 m nS�1 (sediment thicknesses are therefore1.7-times that shown in the plot). The river bed along the entiresection is represented by the uppermost and strongest reflector.

This reflector is weaker in the near shore area (on the sub-merged bank) between 5452430N and 5452390N. The lack of aclear reflector represents attenuation of the EM-wave by scatter-ing. This area of the river bed is known to contain considerable

Fig. 5. GPR (upper plate) and seismic (lower plate) profil

amount of debris, mainly logs. The greatest EM-wave penetrationis closest to shore with a maximum depth of approximately 4 m.The reflections in the near shore unit from 5452430N to5452490N, are horizontally trending and hummocky in geometryinterpreted to represent interbedded layers of silt and woody deb-ris. The GPR signal is completely attenuated beyond a depth of1 m.b.r.b. from 5452390 N to 5452250 N. Groundwater salinitywithin this 1 m-deep zone immediately below the riverbed is di-luted as a result of mixing with river water, which thus definesthe depth of the hyporheic zone. This pattern of EM signal attenu-ation at the 1 m-depth level was visible on all survey lines indicat-ing that the hyporheic zone is ubiquitously 1 m in depth where theriverbed is essentially sandy. The shore parallel GPR profile, line L5is shown in Fig. 4. Similar patterns of signal attenuation are visibleup river from 509125E to 5090200E. Further downriver, signalpenetration is improved as the line tracks close to the southernedge of the silt unit. In this area, hummocky structures dominate.

Resistivity profiling was carried out in the area of groundwaterdischarge which coincides with the area where the overlying mas-sive silt unit ends offshore and, where GPR and seismic datayielded relatively poor resolution of the riverbed and subbottommaterials. Fig. 6 shows in situ profiles of bulk resistivity, fluid elec-trical conductivity and core logs for three positions on the river bedin the area where groundwater discharges (locations given inFig. 3). The data reveals that the relatively homogeneous sand unitis represented by uniform bulk resistivity measurements at depths1 m below the river bed whereas, the upper 1 m exhibits consider-able variability. Bulk resistivity values are lower where organicmaterial/wood chips or fine-grained sediments are encounteredand increases with the coarsening downward sequence observedin the core log. The bulk resistivity values obtained for the riverbed sediments agree well with values obtained by others with sim-ilar sediment composition (Campanella and Weemes, 1990; Fukueet al., 2001).

The apparent formation factor for the sand unit is consistentwith the laboratory measurements of the Fraser Sand samples(see Appendix B in (Bianchin, 2010)) with values ranging from 5to 6. Departure of the apparent formation factor from the rangefor the Fraser River Sand denotes a transition in sediment type.From this comparison, it becomes clear that fluid conductivity

es or riverbed along line L5 and line 3, respectively.

Fig. 6. A comparison of bulk resistivity, fluid electrical conductivity, apparent formation factor and core logs at selected sites on the Fraser Riverbed offshore of the BraidStreet site (refer to Figure 2 for locations of sampling sites on riverbed). For reference the elevation of the top of the riverbed is indicated on the bulk resistivity profile.

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 23

measurements are not required to distinguish sediment types gi-ven the strong contrast in bulk conductivity. At the location ofRP17, the core log, fluid conductivity and bulk resistivity indicate

a sharp change in sediment type at �9.5 m.a.s.l between the lowerFraser Sand unit and the overlying silty fine-grained sand and or-ganic layers. This sharp transition in fluid conductivity suggests

24 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

that the upper fine-grained sediments act as a confining unit andthe sand unit at this location is not directly connected to the river.Moving further offshore from RP17, profiles RP15 and RP16 indi-cate that the overlying silt unit is thinning with sandy sedimentsdominating, and with salinity in groundwater increasing.

Fig. 7 illustrates cross section A–A0 of Fig. 3 with six bulk resis-tivity profiles located from 88 m to 108 m offshore. These profilesare representative of the dominant onshore–offshore trend in sed-iment and groundwater observed. The resistivity profiles shown inFig. 7 delineate the river bed zones of different groundwater qual-ity and heterogeneity in the subbottom sediment composition.Detection of the river bottom was apparent as river water (500–600 X) contrasted significantly with the organic-laden river bedsediments (typically +1000 X). Techniques for mapping of a con-taminant plume and salt water interface using bulk resistivitymeasurements used by Campanella and Weemes (1990) are appli-cable here. The 30 X m contour shown on Fig. 7 essentially mapsthe location of saline groundwater. The increase in resistivity to-wards the river bank is expected, as previous investigations (Bian-chin et al., 2006) have shown that the aquifer contains freshgroundwater albeit contaminated with PAHs. Contaminatedgroundwater, as result of higher dissolved ions (in particular iron)due to anaerobic degradative processes (Bianchin et al., 2006) re-duces the resistivity of the groundwater. The 100 X m contourdelineates the extent of the PAH plume in the vicinity of the riverbed.

The shaded zone superimposed on the profile illustrates thelocations where the resistivity indicates a highly heterogeneousriverbed. This shaded zone is interpreted (supported by core logsof Fig. 6 and diver observations (Roschinski, 2007)) to consist ofmostly fine-grained sandy silt containing significant woody mate-rial. Groundwater sampling efforts by others using the WDPP (An-thony, 1998; Bianchin et al., 2006; Bieber, 2003) have shown that alayer of relatively low permeability sediments is found at the riverbottom in the zone from the river bank to 90 m offshore.

Evidence of groundwater discharge to the river is provided bythe conductivity of the groundwater as indicated by the 100 X mcontours, the termination of the overlying silt unit, and the distri-bution of and proximity of the saline groundwater in the river bed.Based on the geophysical observations, groundwater dischargealong this transect occurs between +90 m and 105 m offshore, aspan of less than 15 m. Somewhere between 101 m and 105 m off-shore, the PAH-contaminated groundwater (100 X m contours) isno longer detectable as a result of dilution with river water or sal-ine groundwater. It appears that the resistivity profile at 101 mrepresents a mixing zone between the saline groundwater andcontaminated groundwater as the bulk resistivity profile indicates

RP14

0 200 400

100Ω•m

RP17

0 200 400

RP9

0

riv

Elev

atio

n (m

asl)

Distance 88 90 m 1

plume

-16

-4

-6

-8

-10

-12

-14

Fig. 7. Stratigraphic and resistivity profile of the Fraser River subbottom offshore of the Bby low permeability sediments. X-axis units = ohm�m. RP15 and RP9 yield essentially th

a gradual transition from saline groundwater to contaminatedgroundwater. Further offshore, bulk resistivity profiles do not dis-play mixing between river water and the saline groundwater: thehigh bulk resistivity readings at the riverbed (at 105 m and108 m), due to the thin layer of silty sand and wood chips, maskthe freshwater signal.

The results of the geophysical surveys permit the developmentof a model of the sediment type distribution on and below the riv-erbed (Fig. 2). The area of the main channel is covered by a thinfilm of organic material, which can be interpreted as gyttja. Belowthis layer, sand dominates. This thin layer was represented by highbulk resistivity values, and by a continuous strong reflector on theseismic profiles. From shoreline to about 100 m offshore, a siltlayer caps the Fraser Sands Aquifer. This silt unit, which is inun-dated with logs and other industrial debris, thins towards thechannel. It is at the southern most edge of this silt unit where GWSioccurs.

3.2. Water chemistry of the HZ

3.2.1. Distribution of water typesGroundwater profiling reveals that the variation in groundwa-

ter chemistry within the shallow sediments of the Fraser River iscomplex. Three water types occur within the upper 2 m of the riv-erbed sediments a result of mixing of river water, contaminated(fresh) groundwater and saline groundwater. Detailed solute pro-files, spanning the groundwater discharge area, were obtainedfrom five multilevel drivepoint wells (MW1–MW5) and an upgra-dient profiling station, P3. P3 is examined in this discussion insteadof MW3 as both are in close proximity to each other yet P3 is a dee-per profile (a comparison of fluid conductivity between the twoprofiles is presented in Fig. 6).

The three water types are distinguishable in the Piper plot ofFig. 8. Contaminated groundwater plots as Ca–HCO�3 water, salinegroundwater as Na–Cl water, and ‘mixed’ water occurring at aninterface of two end member waters i.e., river water–salinegroundwater or contaminated groundwater–saline groundwater,plots as Ca–Cl water. River water also plots as Ca–HCO�3 type water.For clarity in Fig. 8 we use the chemistry of water from only twoprofiles, MW4 and MW5, representing the contaminated ground-water discharge and saline groundwater discharge zones, respec-tively. The profiles at MW4 and MW5 are more detailed andconsist of a complete chemical data sets as opposed to MW1 andMW2. The numerical labels on the Piper plots reference to thesame samples that are plotted in the chemistry versus depth pro-files in Fig. 9.

200 400

RP6

0 500 1000

RP16

0 200 400

er bed

saline ground water

from riverbank01 105 m 108

30Ω·m

raid Street site. Shaded area is interpreted to be that part of the riverbed dominatede same result.

Fig. 8. Piper plot of samples collected from the river and groundwater interface. Numbers adjacent to the symbols correlate to those shown in Figure 9 for the profiles of MW4and MW5.

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 25

Fig. 9 includes the chemical profiles of groundwater at 5 pointsalong the cross section A�A0 of Fig. 3 from 88 m to 108 m offshore,spanning the groundwater discharge zone delineated by the geo-physical surveys. MW4 and MW5 are located approximately10 m down river of the A–A0 transect line and show higher ion con-centrations than the adjacent profiles (MW1 and MW2, respec-tively). Small-scale local heterogeneity likely accounts for thesedifferences. For example, a zone of lower permeability would re-duce the interaction between river and groundwater, or con-versely, a high permeability connection to the river would dilutethe groundwater signal. A trend in groundwater composition isdetectable from onshore to offshore. The groundwater compositionat 88 m offshore (Fig. 9a) is representative of the local groundwatermoving from upland recharge zones through the sandy aquifer andcreosote source zone. This groundwater is a Ca–HCO�3 -typegroundwater, with elevated levels of dissolved iron and with chlo-ride concentrations below 0.5 mM. Profiles shown in Fig. 9b and care considered to be in the zone where the contaminated freshgroundwater mixes with the underlying saline groundwater andwith infiltrating river water. From 105 to 108 m offshore (Fig. 9dand e) the groundwater composition below the river bed consistsof a NaCl-type groundwater with chloride concentrations reaching50 mM.

The transition from Ca–Cl to Ca–HCO�3 and back to Ca–Cl atMW4 can be understood considering the positioning of the profilerelative to groundwater flow (see Fig. 3). In the groundwater dis-charge zone, the water-type at a depth of 1.8 m.b.r.b. is Ca–Cl(point 11 of MW4 profile), from 1 to 0.3 m.b.r.b in the contami-nated groundwater zone, the water type is Ca–HCO�3 , and above

0.3 m.b.r.b. the water transitions between Ca–Cl and Ca–HCO�3 -type water (see Fig. 8). At about 1.2 m.b.r.b. the interface betweencontaminated groundwater and saline groundwater occurs as indi-cated by the sharp increase in Cl and Na (Fig. 9b). It is interesting tonote that the salinity (as indicated by chloride concentration) ofthe groundwater increases at the contaminated groundwater–riverwater interface during a high tide. We can only speculate that thesource of chloride is from saline groundwater (we elaborate on thisfurther below).

In the saline groundwater discharge zone, as represented byMW5, the Ca–Cl type water from about 0.5 m.b.r.b. and upwardsis the result of river water mixing with saline groundwater. Indeed,because the data in the Piper plot do no lie on a straight conserva-tive mixing line that connects the two end member waters, theresulting Ca–Cl type water is affected by a non-conservative chem-ical process, most likely ion exchange. This exchange is occurringat the river water–saline groundwater interface, as well as the pre-viously discussed contaminated groundwater–saline groundwaterinterface (MW4). Below, we take advantage of cation enrichmentfrom the ion exchange process, in addition to chloride profiles, todetermine the extent of river water penetration into the riverbed.

3.2.2. Depth of river water penetration and tidal influence on waterchemistry

The distribution of groundwater solutes indicate that during asingle tidal cycle, river water penetrates the riverbed to a depthof approximately 15 cm but the long term effects of tidal pumpingof river water into the riverbed is observed to a depth of approxi-mately 1 m.b.r.b. In the part of the riverbed characterized by the

Fig. 9. Profiles of the major ions in groundwater at specified sampling points along transect A–A0 . a) P3 b)MW4 c) MW1 d)MW5 e)MW2. Black represents sampling thatoccurred at low tide. Pink represents sampling thath occurred at high tide.

26 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

data of MW2 and MW5, the presence of saline groundwater al-lowed use of chloride as a conservative tracer to delineate thedepth of river water movement into the riverbed, but where chlo-

ride results are ambiguous, for example at P3, MW1 and MW4 werely on enriched ion species in groundwater that result from ionexchange. The depletion and enrichment of cations in groundwater

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 27

as a result of cation exchange has been used to map out areas ofgroundwater freshening and saline intrusion in coastal areas, andto differentiate between recent or ‘fresh’ from ‘old’ or matureevents (Andersen et al., 2005; Chapelle and LeRoy, 1983; Mercado,1985; Xue et al., 1993). The affinity of cations for exchange sites(such as clay) during freshening follows the order Ca2+ > Mg2+ > -K+ > Na+. Freshening would produce a chromatographic pattern ofion exchange exemplified by the following reactions (Andersenet al., 2005):

Ca2þ þMg� X2 $Mg2þ þ Ca—X2 ð1Þ

1=2Mg2þ þ Na—X$ Naþ þ 1=2Mg—X ð2Þ

During intrusion of saline water the affinity order and thuschromatographic sequence reverses. Thus, in areas of recent salineintrusion Ca2+ and Mg2+ would be enriched and Na+ depletedwhereas, for recent freshening events zones of enriched Na+ anddepleted in Ca2+ and Mg2+ would be produced. Long term salineintrusion would see domination of the exchange sites and porewater by Na+, with Ca2+ and Mg2+ being flushed out. In the longterm, freshening of Ca2+ and Mg2+ would come to dominate bothexchange sites and pore water, with Ca2+ ultimately dominating.

The profiles shown in Fig. 9 include measurements at low andhigh tide to demonstrate the effect of the tide and resulting infil-tration of river water into the riverbed. Of the profiles, MW4 andMW5 (Fig. 9b and d, respectively) best demonstrate the effect ofinfiltrating river water on groundwater solute concentrations.The MW5 high-tide chloride concentrations in the first 15–30 cmare lower than their low-tide levels during a single tidal cycle, evi-dence that the conservative solute is diluted with the nearly chlo-ride-free infiltrating river water. The profile of MW4 (Fig. 9b)shows the opposite effect however, chloride concentrations in-crease. The only possible sources for chloride are either: (1) salineriver water; and (2) saline groundwater. The presence of saline riv-er water (from sea water incursion up river) has never been docu-mented this far from the river’s outlet to the sea (Ages, 1979),therefore saline groundwater can be the only reason why chlorideincreases. We believe the increase in chloride could be due thelateral shifting of the fresh/saline groundwater interface; duringa high tide the saline groundwater shifts towards the shorelineand vice versa during a low tide. It could also result from the re-en-try of saline groundwater that discharged during the previous lowtide from up stream saline groundwater sources.

The deeper chloride profile of MW5 suggests that river waterpenetration into the riverbed is deeper than 15 cm. Chloride con-tent increases steadily from about 10 mM at the riverbed to50 mM at 1 m.b.r.b., and generally remains at about 50 mM deeperinto the riverbed. The reduction in salinity is due to mixing of riverwater and saline groundwater by dispersion under an oscillatingflow field (Bianchin et al., 2010).

The most notable effect of river water mixing with groundwaterat MW5 is the calcium and manganese enrichment that occurs to adepth of approximately 0.6 m.b.r.b. Pore water deeper than1 m.b.r.b. is dominated by Na+ (see Fig. 8). At this depth, the ex-change sites are also likely dominated by Na+. However, in theshallow sediments above 1 m.b.r.b Ca2+ appears to dominate theexchange sites. During a high tide the Ca2+ concentration in porewater increases, the opposite effect one would expect in a recentfreshening event. The domination of Ca2+ in these shallow sedi-ments is due to two factors: long term diurnal pumping of riverwater into the riverbed, and the higher affinity of Ca2+ over Na+

for exchange sites. Indeed, during low tide conditions, groundwa-ter discharge would flush Ca2+ in the pore water reducing its con-centration however, the higher affinity of Ca2+ for exchange siteswould make its displacement by Na+ difficult in the short time

span of approximately 12 h (the approximate wavelength of a tidalcycle). This interpretation explains why Ca2+ is not lost to exchangeduring high tides.

As with MW5, the effect of mixing in the shallow sediments atMW4 is visible to a depth of 1 m. However, the Piper plot distribu-tion (Fig. 8) and profiles of Fig. 9b indicate that the chemistry ofthis area is complex likely involving several reaction processesincluding redox and ion exchange. Saline intrusion is occurring inthe shallow sediments to a depth of 0.3 m.b.r.b. The chloride con-tent at the riverbed increases to 1.2 mM during a high tide and re-turns to 0.3 mM chloride on a low tide while the sodium contentremains unchanged. Yet, calcium increases slightly indicating ex-change of calcium by sodium on the exchange sites. Similar ion ex-change patterns have been observed, for example by Xue et al.(1993) with sea-water intrusion in the coastal area of LaizhouBay, China. Xue et al. (1993) noted that CaCl2 waters occurred inareas of fresh saline water intrusion and that elevated Ca2+ oc-curred in these transition zones.

Ion exchange does not appear to be as dominant in the area ofcontaminated groundwater discharge relative to the area wheresaline groundwater discharges (MW5). The concentrations of re-dox sensitive solutes such as dissolved iron and manganese dropsignificantly within 1 m of the riverbed which is likely due to pre-cipitation and/or dilution with river water. In either case, the sol-ute pattern with depth is evidence of river water mixing withgroundwater.

3.2.3. Redox conditions of the hyporheic zoneGroundwater originating from local upland recharge migrates

through an area containing separate phase DNAPL (creosote) be-fore discharging to the river. Anaerobic degradation of dissolvedcreosote (as naphthalene) was confirmed by Bianchin et al.(2006). The terminal electron accepting process (TEAP) in the deepanaerobic portion of the aquifer beneath the river is likely a com-bination of methanogenesis and iron-reduction. This creosote-af-fected groundwater reaches the discharge zone with belowdetection naphthalene, high levels of ferrous iron, methane andinorganic carbon (carbon dioxide). In the area of GWSi ferrous irondecreases gradually from 2.5 mM within the contaminant plume(greater than 1.5 m below river bed) to less than 0.5 mM at the riv-erbed. Methane was not characterized in detail in the shallow sed-iments, however it is ubiquitous throughout the aquifer rangingfrom 0.2 mM to 2 mM (Anthony, 1998; Roschinski, 2007). Dis-solved oxygen decreases with depth from about 2 mM immedi-ately below the riverbed to less than 0.01 mM at 0.6 m belowriverbed at P3 and MW4 (Fig. 9a and b, respectively) and at a depthof only 0.3 m at MW5 (Fig. 9d). The oxygen levels below the riverbed at P3 (Fig. 9a) are suspected to be lower than measured as thesampling point is below the low permeability sediments cappingthe sands and the presence of relatively high levels of ferrous ironwould preclude its presence. The distribution of dissolved oxygenis similar to that observed by Anthony (1998) near or during lowtidal river stage. Oxygen penetration is one third of the approxi-mately 1 m long-term penetration depth of river water into thebed and can be explained by oxygen consumption by processessuch oxidation of organic matter, PAH’s, and methane.

It appears that river water is a source of sulfate. The highestconcentration of sulfate detected was 0.09 mM at the river bedwith detectable levels observed to a depth of 0.6 m below the riverbed (see profiles of MW4 and MW5 of Fig. 9). Nitrate is not consid-ered an important electron acceptor in this system as the FraserRiver water generally contains less than 0.01 mM and, groundwa-ter values are below detection limits (<0.001 mM).

The distribution of reduced chemical species and electronacceptors indicate that the upper 0.5 m of sediments are aerobicbut, during low tide when river water is not flowing into the

28 M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29

sediments the aerobic zone can reduce to 0.2–0.3 m in thickness.With the concentration of sulfate in river water at greater than0.01 mM, and with measurable amounts in the hyporheic zone,sulfate reduction could be another significant process in the atten-uation of groundwater contaminants (Wiedemeier et al., 1999, p.338).

The chemical profiles observed in the riverbed agree well withthe results of temperature profiles and heat transport modelingconducted by Bianchin et al. (2010). The instantaneous time seriesdata of that study indicate that riverbed temperatures, to a depthof 1 m.b.r.b., were affected by the tidally-forced river stage fluctu-ations. Heat transport modeling, using the temperature time seriesdata as calibration targets, reveal that this zone of advective flow israther vigorous with peak instantaneous velocities, at a depth ofless than 0.2 m.b.r.b., can reach 0.45 m/day during either a floodingor ebbing tide. Under these flow conditions, river water can move0.13 m into the river bed during a single high tide; by not account-ing for dispersion this depth of penetration is consideredconservative.

4. Conclusions

Several methodologies were utilized in a multiple line of evi-dence approach to define the spatial extent of the hyporheic zonebeneath the Fraser River.

The hyporheic zone is approximately 1 m deep extending fromapproximately 100 m offshore from the northern shoreline acrossthe channel to the northern edge of the Sapperton sand bar. Geol-ogy controls the location of HZ. A low permeability silt unit capsthe Fraser River Sands aquifer to a distance of 100 m from thenorthern shoreline and interaction between pore water and riverwater is limited and likely diffusion dominated. Beyond the extentof the silt unit where the river bed is sandy, GWSi is dominated byadvective flow.

The water chemistry of the hyporheic zone consists of three dis-tinct water types: locally recharged contaminated groundwater(Ca–HCO�3 ) discharging through a narrow 10 m band on the river-bed where the overlying silt unit terminates; saline groundwater(Na–Cl) found in the remainder of the riverbed and the FraserSands aquifer offshore; and Ca–Cl water which is ubiquitous acrossthat part of the channel where GWSi occurs. This latter water typeis present from the riverbed to a depth of approximately0.3 m.b.r.b., and results from the interaction of river water andthe two end member groundwater types. In the contaminatedgroundwater discharge zone, saline intrusion accounts for the for-mation of Ca–Cl water. Within the saline groundwater zone, fresh-ening of the aquifer accounts for the formation of the Ca–Cl water.

Within the saline groundwater zone of the riverbed, dilutionand ion exchange appear to be dominant reactions. Within the con-taminated groundwater discharge zone, dilution and redox reac-tions likely dominate. Ion exchange also occurs but, not to theextent as in the saline groundwater zone further offshore. Duringa single high tide event river water may travel 0.30 m into the riv-erbed as indicated by a comparison of groundwater profiles col-lected during low and high-tide river stages. Dissolved oxygenwas observed to a depth of approximately 0.3 m.b.r.b. during hightide and to a depth of 0.2 m.b.r.b. during low tide conditions.Therefore, only the upper 0.3 m of the hyporheic zone is consid-ered aerobic.

The existence of a 1 m-thick hyporheic zone beneath the FraserRiver is a significant finding with implications for contaminant fateand transport, and determining the impact of contaminants on riv-er ecology. In a parallel study involving quantifying flow ratesthrough this 1 m hyporheic zone Bianchin et al. (2010) concludedthat the average residence time for groundwater solutes in the

hyporheic zone was on the order of 58 days. This highly reactivezone may have the potential to significantly attenuate redox-sensi-tive contaminants in water.

The data collected here for the lower Fraser River are probablybroadly representative of river delta environments elsewhere. In-deed, deltaic environments can be expected to have near surfacesandy aquifers with groundwaters made reducing by the high nat-urally occurring organic matter content. The geological faciesfound in deltas suggest that one can expect that groundwater dis-charge will occur where the surface aquifer outcrops beneath theriver, that the depth of penetration of river water will be on the or-der of one meter (depth modulated by the local tidal amplitudes),and that towards the center of the channel older, more salinegroundwater of deeper origin will discharge. These findings canbe used by decision makers to understand the processes that canaffect solutes as they move through the hyporheic zone.

Acknowledgements

We thank the four anonymous reviewers for their thoughtfuland constructive comments. We are also grateful to Jennifer Owenand Amrit Laly of Lorax Environmental for assistance in revisingthe figures. The first author is indebted to Lorax Environmentalfor their support during revisions of this manuscript. We also rec-ognize the contribution of Rob Luzitano and Dick Sylwester of Gol-der Associates for providing GPR and seismic equipment and theirassistance in conducting the surveys. This research was supportedby a National Sciences and Engineering Research Council of CanadaStrategic Grant awarded to L. Smith and R. Beckie.

References

Ages, A., 1979. The salinity intrusion in the Fraser River: salinity, temperature andcurrent observations, 1976, 1977. Pacific Marine Science Report 79-14. In:Institute of Ocean Sciences, S., British Columbia.

Andersen, M.S., Nyvang, V., Jakobsen, R., Postma, D., 2005. Geochemical processesand solute transport at the seawater/freshwater interface of a sandy aquifer.Geochim. Cosmochim. Acta 69 (16), 3979–3994.

Anderson, J.K., Wondzell, S.M., Gooseff, M.N., 2002. Stream geomorphology, watersurface slope, and implications for patterns in hyporheic exchange. In:Anonymous (Ed.), Geological Society of America, Cordilleran Section, 98thAnnual Meeting. Geological Society of America (GSA), Boulder, CO, UnitedStates.

Anthony, T., 1998. An Investigation of the Natural Attenuation of a DissolvedCreosote and a Pentachlorophenol Plume. M.Sc. Thesis, University of Waterloo,Waterloo, ON, 235 pp.

Beres, M.J., Haeni, F.P., 1991. Application of ground-penetrating – radar methods inhydrogeologic studies. Ground Water 29 (3), 375–386.

Bianchin, M., 2001. A Field Investigation into the Fate and Transport of Naphthalenein a Tidally Forced Anaerobic Aquifer. M.Sc. Thesis, University of BritishColumbia, Vancouver, Canada, 220 pp.

Bianchin, M., 2010. A Field Investigation Characterizing the Hyporheic Zone of aTidally-Influenced River. Ph.D. Thesis, University of British Columbia,Vancouver, Canada, 308 pp.

Bianchin, M., Smith, L., Barker, J.F., Beckie, R.D., 2006. Anaerobic degradation ofnaphthalene in a fluvial aquifer: a radiotracer study. J. Contam. Hydrol. 84 (3–4), 178–196.

Bianchin, M., Smith, L., Beckie, R., 2010. Quantifying hyporheic exchange in a macro-tidal estuary using temperature time series. Water Resour. Res. 46. Art. No.W07507.

Bieber, C., 2003. Field Sampling and Modeling of Creosote-Derived Contaminationin a Tidally Forced Aquifer. M.Sc. Thesis, University of British Columbia,Vancouver, BC, 202 pp.

Boulton, A.J., Stanley, E.H., 1995. Hyporheic processes during flooding and drying ina Sonoran desert stream. 2. Faunal dynamics. Arch. Hydrobiol. 134 (1), 27–52.

Bridger, D., Allen, D., 2006. An investigation into the effects of diffusion on salinitydistribution beneath the Fraser River Delta. Hydrogeol. J. 14 (8), 1423–1442.

Burnett, W.C., Bokuniewicz, H., Huettel, M., Moore, W.S., Taniguchi, M., 2003.Groundwater and pore water inputs to the coastal zone. Biogeochemistry 66,3–33.

Butler, J.J.J., Healey, J.M., McCall, G.W., Garnett, E.J., Loheide, P.I., 2002. Hydraulictests with direct-push equipment. Ground Water 40 (1), 25–36.

Butler, K.E., Nadeau, J.-C., Parrott, R., Daigle, A., 2004. Delineating recharge to a rivervalley aquifer by riverine seismic and EM methods. J. Environ. Eng. Geophys. 9(2), 95–109.

M.S. Bianchin et al. / Journal of Hydrology 406 (2011) 16–29 29

Campanella, R., Weemes, L., 1990. Development and use of an electrical resistivitycone for groundwater contamination studies. Can. Geotech. J. 27, 557–567.

Canada, 1976. Glossary of Terms in Soil Science. Department of Agriculture, 1459,pp. 44. <http://sis.agr.gc.ca/cansis/references/1976rba.html>.

Cardenas, M.B., Wilson, J.L., Zlotnik, V.A., 2004. Impact of heterogeneity, bed forms,and stream curvature on subchannel hyporheic exchange. Water Resour. Res. 40.

Chapelle, F.H., LeRoy, L.K., 1983. Aqueous geochemistry and the exchangeablecation composition of glauconite in the aquia aquifer, Maryland. Ground Water21 (3), 343–352.

Conant Jr., B., Cherry, J.A., Gillham, R.W., 2004. A PCE groundwater plumedischarging to a river: influence of the streambed and near-river zone oncontaminant distributions. J. Contam. Hydrol. 73, 249–279.

Cross, V., Bratton, J.F., Crusius, J., Colman, J.A., McCobb, T.D., 2008. SubmarineHydrogeological Data from Cape Cod National Seashore, USGS Open-File Report2006-1169. <http://woodshole.er.usgs.gov/pubs/of2006-1169/index.htm>.

Fukue, M., Minato, T., Matsumoto, M., Horibe, H., Taya, N., 2001. Use of a resistivitycone for detecting contaminated soil layers. Eng. Geol. 60, 361–369.

Golder, A.L., 1997. Final Report on Site Characterization. 25 Braid Street WoodPreserving Facility, Coquitlam, BC, Coquitlam.

Government of Canada, E.C., Water Survey of Canada, 2008. Archived HydrometricData, Station Name: Fraser River at Port Mann Pumping Station ID: 08MH126[DATA FILE]. <http://scitech.pyr.ec.gc.ca/waterweb>.

Haeni, F.P., 1996. Use of ground-penetrating radar and continuous seismic-reflection profiling on surface-water bodies in environmental and engineeringstudies. J. Environ. Eng. Geophys. 1 (1), 27–35.

Harvey, J.W., Bencala, K.E., 1993. The effect of streambed topography on surface-subsurface water exchange in mountain catchments. Water Resour. Res. 29 (1),89–98.

Harvey, J.W., Wagner, B.J., 2000. Quantifying hydrologic interactions betweenstreams and their subsurface hyporheic zones. In: Jones Jeremy, B., MulhollandPatrick, J. (Eds.), Streams and ground waters. Academic Press, San Diego, CA,United States.

Harvey, J.W., Wagner, B.J., Bencala, K.E., 1996. Evaluating the reliability of thestream tracer approach to characterize stream-subsurface water exchange.Water Resour. Res. 32 (8), 2441–2451.

Hinkle, S.R. et al., 2001. Linking hyporheic flow and nitrogen cycling near theWillamette River; a large river in Oregon, USA. J. Hydrol. 244 (3–4), 157–180.

Hynes, H.B.N., 1974. Further studies on the distribution of stream animals withinthe substratum. Limnol. Oceanogr. 19 (1), 92–99.

Kasahara, T., Wondzell, S.M., 2003. Geomorphic controls on hyporheic exchangeflow in mountain streams. Water Resour. Res. 39 (1), 1005–1029.

Land, L.A., Paull, C.K., 2001. Thermal gradients as a tool for estimating groundwateradvective rates in a coastal estuary: while Oak River, North Carolina. USA J.Hydrol. 248, 198–215.

Landmeyer, J.E., Bradley, P.M., Trego, D.A., Hale, K.G., Haas, J.E., 2010. MTBE, TBA,and TAME attenuation in diverse hyporheic zones. Ground Water 48 (1), 30–41.

Lendvay, J.M. et al., 1998. Geophysical characterization, redox zonation, andcontaminant distribution at a groundwater/surface water interface. WaterResour. Res. 34 (12), 3545–3559.

Maji, R., Smith, L., 2009. Quantitative analysis of seabed mixing and intertidal zonedischarge in coastal aquifers. Water Resour. Res. 45 (W11401). doi:10.1029/2008WR007532.0.

Marion, A., Bellinello, M., Guymer, I., Packman, A.I., 2002. Effect of bed formgeometry on the penetration of nonreactive solutes into a streambed. WaterResour. Res. 38 (10).

Mercado, A., 1985. The use of hydrogeochemical patterns in carbonate sand andsandstone aquifers to identify intrusion and flushing of saline water. GroundWater 23 (5), 635–645.

Moore, W.S., 1999. The subterranean estuary: a reaction zone of ground water andsea water. Mar. Chem. 65, 111–125.

Murphy, F., Herkelrath, W., 1996. A sample-freezing drive shoe for a wireline-pistoncore sampler. Ground Water Monit. Remediat. 16 (3), 86–90.

Naegeli, M., Huggenberger, P., Uehlinger, U., 1996. Ground penetrating radar forassessing sediment structures in the hyporheic zone of a prealpine river. J. N.Am. Benthol. Soc. 15 (3), 353–366.

Neeper, D.A., 2001. A model of oscillatory transport in granular soils, withapplication to barometric pumping and earth tides. J. Contam. Hydrol. 48,237–252.

Nyquist, J.E., Freyer, P.A., Toran, L., 2008. Stream bottom resistivity tomography tomap ground water discharge. Ground water 46 (4), 561–569.

Packman, A.I., Brooks, N.H., 2001. Hyporheic exchange of solutes and colloids withmoving bed forms. Water Resour. Res. 37 (10), 2591–2605.

Pitkin, S., Ingleton, R.A., Cherry, J.A., Anon., 1994. Use of a Drive Point SamplingDevice for Detailed Characterization of a PCE Plume in a Sand Aquifer at a DryCleaning Facility. Ground Water Management. vol. 18. Water Well Journal Pub.Co., Dublin, pp. 395–413.

Preito, C., Destouni, G., 2005. Quantifying hydrological and tidal influences ongroundwater discharges into coastal waters. Water Resour. Res. 41 (W12427).doi:10.1029/2004WR003920.

Robinson, C., Li, L., Barry, D.A., 2007. Effect of tidal forcing on a subterraneanestuary. Adv. Water Res. 30, 851–865.

Roschinski, T., 2007. Geochemistry in The Hyporheic Zone of the Lower Fraser River.M.Sc. Thesis, The University of British Columbia, Vancouver, Canada, 95 pp.

Rosenberger, A., Weidelt, P., Spindeldreher, C., Heesemann, B., Villinger, H., 1999.Design and application of a new free fall in situ resistivity probe for marine deepwater sediments. Mar. Geol. 160, 327–337.

Salehin, M., Packman, A.I., Paradis, M., 2004. Hyporheic exchange withheterogeneous streambeds: laboratory experiments and modeling. WaterResour. Res. 40.

Schubel, J., 1974. Gas bubbles and the acoustically impenetrable, or turbid,character of some estuarine sediments. In: Kaplan, I. (Ed.), Marine Science.Plenum Press, New York, pp. 275–298.

Slater, L.D. et al., 2010. Use of electrical imaging and distributed temperaturesensing methods to characterize surface water–groundwater exchangeregulating uranium transport at the Hanford 300 Area, Washington. WaterResour. Res. 46.

Stanley, E.H., Boulton, A.J., 1995. Hyporheic processes during flooding and drying ina sonoran desert stream. 1. Hydrologic and chemical dynamics. Arch. Hydrobiol.134 (1), 1–26.

Starr, R.C., Ingleton, R.A., 1992. A new method for collecting core samples without adrilling rig. Ground Water Monit. Remediat. 12 (1), 91–95.

Storey, R., KWF, H., Williams, D.D., 2003. Factors controlling riffle-scale hyporheicexchange flows and their seasonal changes in a gaining stream: a three-dimensional groundwater flow model. Water Resour. Res. 39 (2), 1034–1051.

Taniguchi, M., 2002. Tidal effects of submarine groundwater discharge into theocean. Geophys. Res. Lett. 29 (12). doi:10.1029/2002GL014987.

Trefry, M.G., Svensson, T.J.A., Davis, G.B., 2007. Hypoaigic influences ongroundwater flux to a seasonally saline river. J. Hydrol. 335, 330–353.

Triska, F.J., Kennedy, V.C., Avanzino, R.J., Zellweger, G.W., Bencala, K.E., 1989.Retention and transport of nutrients in a third-order stream in NorthwesternCalifornia: hyporheic processes. Ecology 70, 1893–1905.

Valett, H.M., Fisher, S.G., Grimm, N.B., Camill, P., 1994. Vertical hydrologic exchangeand ecological stability of a desert stream ecosystem. Ecology 75 (2), 548–560.

Vinson, D.S., Block, S.E., Pershall, A.D., Littlepage, D.M., Crossey, L.J., 2001. High-resolution vertical examination at the water table; relationship betweenreactive solids and redox chemistry of shallow groundwater along the RioGrande, New Mexico. In: Anon., (Ed.), Geological Society of America, 2001Annual Meeting. Geological Society of America (GSA). Boulder, CO, UnitedStates.

Ward, A.S., Gooseff, M.N., Singha, K., 2010. Imaging hyporheic zone solute transportusing electrical resistivity. Hydrol. Process. 24 (7), 948–953.

Westbrook, S.J. et al., 2005. Interaction between shallow groundwater, salinesurface water and contaminant discharge at a seasonally and tidally forcedestuarine boundary. J. Hydrol. 302 (1–4), 255–269.

White, D.S., 1993. Perspectives on defining and delineating hyporheic zones. J. N.Am. Benthol. Soc. 12 (1), 61–69.

Wiedemeier, T.H., Rifai, H.S., Newell, C.J., Wilson, J.T., 1999. Natural Attenuation ofFuels and Chlorinated Solvents in the Subsurface. John Wiley & Sons, New York,NY, United States (USA), 617 pp.

Wroblicky, G.J., Campana, M.E., Valett, H.M., Dahm, C.N., 1998. Seasonal variation insurface-subsurface water exchange and lateral hyporheic area of two stream-aquifer systems. Water Resour. Res. 34 (3), 317–328.

Xue, Y. et al., 1993. Sea-water intrusion in the coastal area of Laizhou Bay, China: 1.Distribution of sea-water instrusion and its hydrochemical characteristics.Ground Water 31 (4), 532–537.

Yim, C.S., Mohsen, M., 1992. Simulation of tidal effects on contaminant tranpsort inporous media. Ground Water 30 (1), 78–86.

Zawadzki, W., Chorley, D.W., Patrick, G., 2002. Capture-zone design in an aquiferinfluenced by cyclic fluctuations in hydraulic gradients. Hydrogeol. J. 10, 601–609.