fractured bedrock and saprolite hydrogeologic controls on groundwater/surface-water interaction: a...

Fractured bedrock and saprolite hydrogeologic controlson groundwater/surface-water interaction: a conceptual model(Australia)

Edward W. Banks & Craig T. Simmons &

Andrew J. Love & Roger Cranswick &

Adrian D. Werner & Erick A. Bestland & Martin Wood &

Tania Wilson

Abstract Hydrologic conceptual models of groundwa-ter/surface-water interaction in a saprolite-fracturedbedrock geological setting often assume that thesaprolite zone is hydraulically more active than thedeeper bedrock system and ignore the contribution ofdeeper groundwater from the fractured bedrock aquifer.A hydraulic, hydrochemical, and tracer-based studywas conducted at Scott Creek, Mount Lofty Ranges,South Australia, to explore the importance of both thedeeper fractured bedrock aquifer system and theshallow saprolite layer on groundwater/surface-waterinteraction. The results of this study suggest thatgroundwater flow in the deeper fractured bedrock zoneis highly dynamic and is an important groundwaterflow pathway along the hillslope. Deep groundwater istherefore a contributing component in streamflowgeneration at Scott Creek. The findings of this studysuggest that hydrologic conceptual models, which treatthe saprolite-fractured bedrock interface as a no-flowboundary and do not consider the deeper fractured

bedrock in hydrologic analyses, may be overly sim-plistic and inherently misleading in some groundwater/surface-water interaction analyses. The results empha-sise the need to understand the relative importance ofsubsurface flow activity in both of these shallowsaprolite and deeper bedrock compartments as a basisfor developing reliable conceptual hydrologic modelsof these systems.

Keywords Fractured rocks . Saprolite . Groundwater/surface-water relations . Hydrochemistry . Australia

Introduction

Interactions between groundwater and surface water (gw–sw) form one component of the hydrological cycle and arelargely controlled by the effects of physiography (topog-raphy and geology) and climate (Winter et al. 1998).Considerable research on this topic has been undertaken insedimentary aquifer systems (e.g. Beyerle et al. 1999;Schilling et al. 2006; Krause and Bronstert 2007) but veryfew studies are reported for fractured bedrock systems(e.g. Sklash and Farvolden 1979; Haria and Shand 2006;Manning and Caine 2007; Kahn et al. 2008). The latter aresubstantially more complex owing to the geologicalheterogeneity of the fractured-rock aquifer. In addition,the deeper fractured bedrock aquifers are usually overlainby surficial weathered soil and saprolite material. Intui-tively, these different geologic layers and the high level ofheterogeneity in this system is expected to be critical incontrolling both groundwater and surface water responsesand hence the nature of the gw–sw interaction. However,the partitioning of the aquifer system between the soil(weathered material), saprolite (weathered material thatretains structure of the parent rock) and fractured rock(unweathered bedrock) is often simplified. These geologiclayers are usually not considered as explicit geologicfeatures and hydrogeologic controls on gw–sw interaction.Some important questions arise: What is the groundwaterflux through the saprolite zone compared to the deeper

Received: 19 August 2008 /Accepted: 28 May 2009Published online: 21 June 2009

* Springer-Verlag 2009

E. W. Banks ()) :C. T. Simmons :A. J. Love :R. Cranswick :A. D. Werner : E. A. Bestland :M. WoodSchool of Chemistry, Physics and Earth Sciences,Flinders University South Australia,GPO Box 2100, Adelaide, 5001, Australiae-mail: [email protected].: +61-409-097970

T. WilsonDepartment of Primary Industries and Resources of South Australia,Adelaide, Australia

Present Address:Sinclair Knight Merz Pty Ltd.,590 Orrong Road, Armadale, VIC 3143, Australia

Present Address:Australian Water Environments Pty Ltd.,1/198 Greenhill Road, Eastwood, SA 5063, Australia

Hydrogeology Journal (2009) 17: 1969–1989 DOI 10.1007/s10040-009-0490-7

fractured bedrock system? How do these various geologiccontrols influence the nature of gw–sw interaction? Underwhat conditions do these various geologic controls need tobe included in our hydrogeologic conceptual models?What level of simplification is permissible in a hydrologicconceptual model and what are the consequences ofconceptual model choice?

The study of gw–sw connectivity under gaining-streamconditions in fractured-rock systems generally involvesthe application of one of at least three types of conceptualmodels, the most basic forms of which are as classifiedand described in the following (Fig. 1). In the first type,the gw–sw interaction and streamflow response is inter-preted using a single geological system in a bucket-typeapproach, which does not explicitly account for soil,saprolite or fractured bedrock (Fig. 1a). Type 1 conceptualmodels implicitly or explicitly assume that the system ishomogeneous, as is commonly the case in traditionalhydrograph separation methods. In the second type, thesurficial saprolite system dominates the hillslope hydrol-ogy (i.e. there is no groundwater flow in the deeperfractured bedrock aquifer and the saprolite-fracturedbedrock interface is considered an impermeable boundary)and the gw–sw interaction and streamflow response of thegaining-stream is completely interpreted by assumingsubsurface flow occurs in the surficial saprolite layer only(Fig. 1b). Thirdly, both a saprolite layer and fracturedbedrock aquifer system are explicitly considered asgeologic features and hydrogeologic controls on gw–swinteraction (Fig. 1c). The first conceptual model (Fig. 1a)may be regarded as the simplest model, while the third(Fig. 1c) may be regarded as the most complex. A criticalchallenge is that one rarely knows, apriori, which is themost appropriate conceptual model to choose for anygiven field site and what the consequences of conceptualmodel choice (simple or complex) are on the interpretationof gw–sw interaction data and associated analyses.

Numerous studies of gw–sw connectivity have appliedthe first conceptual model (Fig. 1a), particularly in theapplication of hydrograph separation methods (Pinder andJones 1969; e.g. Chapman 1999), where the pathways ofgroundwater discharge are not explicitly determined andoften considered as a ‘blackbox’ within a bucket typeapproach. The assumptions used in the second conceptualmodel (Fig. 1b) can be seen in a wide range of catchment

hillslope hydrological studies (e.g. Kirkby 1988), whichvery often assume that subsurface flow occurs in the soil/saprolite zone only, rather than in the deeper fracturedbedrock groundwater and hence that surficial flows in thesaprolite zone are of far greater concern in the partitioningof streamflow. Gburek and Urban (1990) and Gburek et al.(1999) investigated a layered fractured zone along twohillslope cross-sections and noted that the soil zonerapidly transmitted rainfall to the groundwater and thatlateral flow in the highly fractured zone above lesspermeable bedrock dominated the groundwater flow.

The geological controls on rapid lateral flow that mayoccur along the interface between the saprolite andfractured bedrock are largely unknown and there is stillsurprisingly little information on the importance of thefractured bedrock aquifer in streamflow generation.However, the very few recent studies which haveconsidered the third conceptual model (Fig. 1c; e.g.Shand et al. 2005; Haria and Shand 2006) have shownthat rising water tables in the fractured rock contribute torelatively rapid lateral flows in the saprolite zone and thatdeeper groundwater flows in the fractured bedrock playan active and critical role in streamflow generation.Distinguishing between shallow soil, saprolite andfractured bedrock groundwater contributions to stream-flow generation is difficult and even more complicatedwith the existence of pre-event water. A study by Sklashand Farvolden (1979) of several watersheds characterisedby surficial deposits of 1–20 m thickness overlyingbedrock in Quebec, Canada, noted that groundwater canplay an active and responsive role in streamflowgeneration. The study emphasised that it is not justsurface-water runoff that contributes to peak flow eventsand that groundwater may also control surface waterquality during these events.

Part of the difficulty in assessing these types ofsaprolite-fractured bedrock systems lies in the complexityassociated with the geological heterogeneity inherent to allof fractured rock hydrology. It is well known that the ratesof groundwater flow and connectivity in fractured rockaquifers are difficult to determine, and methods commonlyused for porous media are often not applicable (Cook et al.1996; Love et al. 2002). Multi-tracer approaches havebeen used in fractured rock systems (e.g. Genereux et al.1993; Shand et al. 2007) and have proven invaluable in

Fig. 1 Conceptual models of groundwater/surface-water interaction under gaining stream conditions in a complex saprolite-fracturedbedrock aquifer system: a homogeneous system, b subsurface flow in the surficial saprolite zone only, and c subsurface flow occurs in boththe surficial saprolite layer and deeper fractured bedrock aquifer

1970


constraining contributing sources of solutes, preferentialflow pathways and residence times in these types ofsystems. Groundwater noble gas, age and temperature datawere used by Manning and Caine (2007) in an alpinebedrock catchment to evaluate the groundwater pathwaysand activity of deep fractured bedrock systems, and meanresidence time of groundwater in the catchment.

The movement of water through soils and the saprolitethrough relict fractures and macropores has been shown tobe important in several studies. McKay et al. (2005) foundthat the hydraulic conductivity and groundwater flow in asedimentary rock saprolite was influenced significantly bythe parent lithology, degree of infilling with clays and Fe/Mn oxides, and the physical nature of the macropores.McDonnell (1990) hypothesised that rapid downslopeflow of pre-event water through continuously connectedpipes extending from the hillslopes to the stream wasresponsible for observed runoff in a small headwatercatchment in New Zealand. The importance of rapidtransport pathways via fractures and macropores insaprolite and fractured rock was also shown in a studyby Van der Hoven et al. (2005). Their study evaluatedtemporal variability in tracer data in the soils, saproliteand shallow fractured sedimentary rocks in Tennessee,USA.

The field-based study described in this paper inves-tigates the role of the saprolite-fractured bedrock aquifersystem on gw–sw interactions in the Scott Creekcatchment in the Mount Lofty Ranges, South Australia.The monitoring period for this study was between July2005 and December 2007. It builds upon some earlierhillslope studies in this area by Smettem et al. (1991) andLeaney et al. (1993) which investigated the influence ofmacropores on runoff processes and concluded that lateralsubsurface flows are initiated above the soil–rock interfacerather than the A–B soil horizon boundary once infiltra-tion of the macropore system has been exceeded. It wasimplicitly assumed in those earlier studies that the maincontributors to streamflow were lateral subsurface flow inthe saprolite zone only and surface runoff. The role ofdeeper groundwater from the fractured bedrock zone wasnot considered. As a result, previous conceptual models ofthe saprolite-fractured bedrock interface at Scott Creekwere considered to be no flow boundaries (i.e. to date, thesecond conceptual model (Fig. 1b) as defined previously,has been used as the basis for previous hydrologicanalyses at Scott Creek). Using a hydraulic, hydrochem-ical, and tracer-based approach, the field study reportedhere explores the importance of the deeper fracturedbedrock aquifer system in the Scott Creek catchment todetermine whether it exerts greater influence on gw–swdynamics than has previously been reported. Our resultsprovide conclusive evidence that the deeper fracturedbedrock is hydraulically active and indeed quite possiblymore so than the shallow saprolite system. This suggeststhat, unlike previous studies at this site, the deeperfractured bedrock system should be considered in hydro-logic analyses pertaining to gw–sw connectivity in thissystem.

Study area and background

The study site Scott Bottom is located near the bottom ofthe Scott Creek Catchment (SCC; Fig. 2). The SCC is arelatively small (27 km2) catchment in the Mount LoftyRanges (MLR), South Australia and is located approxi-mately 30 km southeast of Adelaide. The SCC ischaracterised by a temperate climate, experiencing warmdry summers and wet cool winters. Mean annual evapo-ration is 1,517 mm (Department of Water, Land andBiodiversity Conservation, unpublished data, 2007). Meanannual rainfall ranges from 804 mm at the bottom of thecatchment to 1,009 mm in the upper reaches of thecatchment (based on 1968–2007 data; Department ofWater, Land and Biodiversity Conservation, unpublisheddata, 2007).

The topography of SCC varies from steep slopes togently undulating land. The main watercourse of ScottCreek runs in a north–south direction and minor tributar-ies that contribute to Scott Creek, which tend to be dryduring the summer months, dissect the steep slopedvalleys. The established gauging station on Scott Creekat the study site Scott Bottom, located near the bottom ofthe SCC, has a rectangular stepped V-notch weir withcontinuous streamflow records from 1964 until present.The mean annual flow is approximately 3,710 ML/year(based on 1979–2007 data; Department of Water, Landand Biodiversity Conservation, unpublished data, 2007).Due to increased rainfall from May to October each year,high flows usually occur during this period with maximumflow typically observed around August. Streamflowrecords were instantaneous but for this study data ispresented on a daily interval. A pluviometer at the sitewith a 203 mm throat and 0.2 mm tipping bucket hasrecorded rainfall from 1991 through to present day.

Geology and hydrogeologyThe SCC is characterised by moderate to steep topograph-ic relief with soil (0–3 m thickness) and saprolite (1–20 mthickness) material underlain by fractured bedrock, whichis exposed at the surface in some areas. The geology isstructurally complex, with a diverse range of metamor-phosed sedimentary formations including siltstone, sand-stone and dolomite of the Adelaidean sequences of theAdelaide Geosyncline (Preiss 1987; Fig. 2). The Wool-shed Flat Shale metasediment dominates the area aroundthe study site at Scott Bottom. The eastern side of thecatchment is predominantly Aldgate Sandstone (and to alesser degree Skillogalee Dolomite) and the highertopographic areas on the western side of the catchmentis dominated by Stonyfell Quartzite.

There are approximately 150 bores in the SCC. Whilesome are completed in the shallow alluvial aquifers(<20 m depth), the majority of bores are located in thedeeper fractured Woolshed Flat Shale, Aldgate Sandstoneand Stonyfell Quartzite metasediments (bore depths of upto 190 m depth) due to higher yields and lower salinitiesin the deeper bedrock system. Yields range from 0.02 to

1971


25 L/s, and of the bores that have salinity data records,95% have an electrical conductivity (EC) less than 1,500μS/cm, and 40% have an EC less than 500 μS/cm (James-Smith and Harrington 2002).

Methods

An integrated approach was employed combining physi-cal, hydrogeological, hydrochemical and environmentaltracer measurements made in the field. The hydrogeolog-ical and hydraulic data was used to establish thegroundwater flowpaths, hydraulic responses and relativeflow activity of the shallow saprolite and deeper fracturedbedrock geologic zones over the duration of the studyperiod, between July 2005 and December 2007. Thehydrochemistry and environmental tracers were used toidentify the contributing end members in this flow system,mixing processes and for comparison with the hydraulicdata in order to reinforce our hydraulic conceptual modelof the system.

Piezometer installationForty-two 50 mm PVC piezometers were installed at ScottBottom in July 2005 to complement the existing monitor-ing well network at the site and to monitor thegroundwater processes in the soil, saprolite and fracturedbedrock zones. The existing network includes eight largerdiameter open wells of up to 96 m depth completed in the

fractured bedrock aquifer, which were constructed in2002. The construction details and preliminary hydro-geologic investigation of these wells are described inHarrington (2004), Harrington et al. (2004) and James-Smith and Harrington (2002). The piezometers arearranged in six nests situated along a transect perpendic-ular to the creek valley, on an inferred groundwaterflowpath covering a distance of about 330 m (Fig. 3).There are four nests on the southern side (A, B, C and D)and two nests on the northern side of the creek (E and F),with depths varying from 1.5 to 28.5 m (Cranswick 2005).Piezometer screen intervals were no greater than 3 m inlength to target discrete groundwater flow zones in thesoil, saprolite and fractured bedrock which allows repre-sentative samples from the aquifer to be obtained forgeochemical analyses (Shapiro 2002). The constructiondetails of each piezometer are shown in Table 1. Theground elevation and piezometers were surveyed and thewater table elevations were corrected to a reducedstanding water level (RSWL) relative to the AustralianHeight Datum (mAHD; i.e. the mean sea level around thecoast of the Australian continent). Manual water levelmeasurements in the piezometers were conducted on amonthly basis from July 2005 until December 2007.Stream water level measurements were monitored in ScottCreek in between nests D and E (in line with thepiezometer transect) on a monthly basis and comparedwith the continuous streamflow data from the gaugestation located approximately 50 m upstream of thetransect.

Fig. 2 a Map showing the location of Scott Creek Catchment in the Mount Lofty Ranges, South Australia. b The geology, groundwaterwells and location of Scott Bottom study site in the Scott Creek Catchment. c Location of the nested piezometers, open groundwater wells,pluviometer (rainfall collector is adjacent to pluviometer) and Scott Bottom gauging station at the Scott Bottom study site

1972


Soil and aquifer hydraulic conductivityThe hydraulic conductivity of the soil, saprolite andfractured bedrock was determined in order to estimatethe relative activity of each geologic zone in control-ling the hydrologic fluxes and gw–sw connections.Field-saturated hydraulic conductivity of the shallowsoil zone was determined at the study site using theGuelph permeameter method which has been widelyapplied as a field technique for determining soilhydraulic properties (Elrick and Reynolds 1992;Bagarello and Giordano 1999). Tests were conductednext to the six piezometer nests at 5, 10 and 15 cmdepth below ground. One test was also conducted atthe base of an excavated soil pit on the sandstoneplateau, 200 cm below ground level. When a constantdepth of ponding (H) in mm is maintained in the

borehole, the field-saturated hydraulic conductivity(Kfs) can be determined by:

Kfs ¼ Q

2pH2ln

H

r

� �þ H

2

r2þ 1

!1 2=24

35� 1

8<:

9=; ð1Þ

(Elrick and Reynolds 1992) where r is the radius of theborehole, Q is the steady-state volumetric flow ratecalculated from the cross sectional area (A) of the tubeand measurement of the rate of fall in the reservoir overtime (LT−1). For this particular Guelph design and ourexperiment, a constant head value of 50 mm and radiusof the borehole equal to 28 mm were used.

Single-well aquifer tests were conducted on six of thepiezometers at nests D, E and F to characterise the

Fig. 3 a Summary of the drilling program, depth ranges and description of the A and B horizons (soil zone), saprolite and unweatheredbedrock with photographs of each zone. b Cross-section of the nested piezometer transect from nest F (northern extent) to nest A (southernextent) showing depths of piezometers and inferred lithology at the site. mAHD metres above Australian Height Datum; RSWL reducedstanding water level

1973


hydraulic conductivity of the fractured bedrock zone sothat groundwater fluxes could be estimated at the studysite. Drawdown was measured in the piezometers locatedabove and below the piezometer being pumped to monitorthe vertical connection between the piezometers. Aquifertests are typically more suited to sedimentary systemsbecause the models for interpreting the test data representthe aquifer as a homogeneous and isotropic porousmedium. However, their application to nested piezometersin fractured rock aquifers can provide valuable informa-tion on the vertical variation of hydraulic conductivity andcan be used to derive other physical characteristics of theaquifer (Cook 2003). Where aquifer tests could not beconducted, bail tests were done, using the methoddescribed by Hvorslev (1951).

The Cooper-Jacob straight-line method (Fetter 2001)was used to determine the aquifer transmissivity, where:

T ¼ 2:3Q

4pD h0 � hð Þ ð2Þ

T is the transmissivity [L2T−1], Q is the constantdischarge from the pump [L3T−1] and ∆(h0–h) is thedrawdown per log cycle of time [L] (L is length and T istime). The transmissivity of the aquifer was used to

determine the bulk hydraulic conductivity [LT−1] Kb overthe length of the screen interval, b.

Stream and groundwater samplingAYSI multi-parameter meter was used to measure the pH,electrical conductivity (EC), dissolved oxygen (DO),redox and temperature in the creek and also duringpurging of the piezometers using a flow-through cell.Alkalinity was also measured in the field using a HACHtitration kit. Prior to sampling the piezometers, the staticwater level was measured from top of casing (TOC) usingan electric water level indicator. Samples were collectedafter purging the piezometers and once the physicalparameters had stabilised, indicating that the sample wasrepresentative of the section of the aquifer sampled. Onlythe piezometers that were wet could be sampled and theseare shown in Table 1.

Major ion analyses were conducted on the surfacewater and groundwater samples that were filtered througha 0.45-μm membrane filter. Cations were acidified withnitric acid (1% v/v HNO3) to keep the ions in solution andanalysed by a Spectro CIROS Radial Inductively CoupledPlasma Optical Emission Spectrometer at CSIRO Landand Water Analytical Services, Adelaide, South Australia.

Table 1 Construction details of the nested piezometers along the hillslope transect and open wells at Scott Bottom

SiteID

Lithology Welldepth

Mid-screendepth

Screenlength

Datesampled

Temp EC DO pH FieldEh

HCO3 Br

m m m oC μS/cm mg/L mV mg/L mg/L

A1 Woolshed Flat Shale 28.5 27.0 3.0 01/08/05 18.9 9,650 2.8 7.2 98 331 11.6A2 Woolshed Flat Shale 20.5 19.0 3.0 01/08/05 7.2 311 36.6A8 Sandy Clay 1.5 1.1 0.8 01/08/05 40 0.4B4 Woolshed Flat Shale 10.5 9.5 2.0 28/07/05 18.5 2,590 2.1 7.2 151 403 2.1B5 Woolshed Flat Shale 7.7 7.2 1.0 28/07/05 18.3 2,590 2.0 7.2 166 401 2.1C4 Woolshed Flat Shale 10.5 10.0 1.0 28/07/05 17.9 2,030 0.1 7.1 43 456 1.6C5 Woolshed Flat Shale 7.5 7.0 1.0 28/07/05 7.3 424 1.7D1 Woolshed Flat Shale 25.5 24.0 3.0 01/08/05 15.5 3,310 0.6 7.3 −81 422 2.5D2 Woolshed Flat Shale 20.5 19.0 3.0 28/07/05 15.5 3,230 0.0 7.2 −74 356 2.4D3 Woolshed Flat Shale 16.1 14.6 3.0 28/07/05 15.4 2,770 3.0 7.4 −166 362 2.1D4 Woolshed Flat Shale 10.5 10.0 1.0 28/07/05 15.5 3,150 0.0 7.2 −49 338 2.4D5 Woolshed Flat Shale 7.5 7.0 1.0 28/07/05 15.1 3,290 1.6 7.3 18 319 2.4D6A Saprolite 4.3 3.8 1.0 28/07/05 14.1 2,350 2.6 7.2 37.5 473 1.3D6B Saprolite 4.5 4.0 1.0 28/07/05 13.8 2,770 2.7 7.2 34 394 1.7D7 Sandy Clay 2.5 2.0 1.0 01/08/05 14.3 4,400 3.5 7.3 120 608 3.2D8 Sandy Clay 1.5 1.1 0.8 28/07/05 3,490 7.9 198 2.8E1 Woolshed Flat Shale 25.5 24.0 3.0 01/08/05 15.6 3,480 1.1 7.5 −18 339 2.4E2 Woolshed Flat Shale 20.5 19.0 3.0 29/07/05 15.5 4,020 0.0 7.2 −43 441 3.1E3 Woolshed Flat Shale 15.5 14.0 3.0 29/07/05 15.3 3,190 3.1 7.4 −60 402 2.6E4 Woolshed Flat Shale 10.5 10.0 1.0 29/07/05 15.6 3,140 0.0 7.2 −43 362 2.3E5 Saprolite 7.5 7.0 1.0 01/08/05 15.1 2,730 2.4 7.4 37 344 2.0E6 Saprolite 4.5 4.0 1.0 29/07/05 14.5 3,650 5.4 7.2 129 328 2.6F2 Woolshed Flat Shale 20.0 18.5 3.0 29/07/05 16.5 3,120 −0.1 7.1 −25 363 2.3F3 Woolshed Flat Shale 15.5 14.0 3.0 29/07/05 16.1 3,090 0.1 7.1 −18 369 2.3F4 Woolshed Flat Shale 10.5 10.0 1.0 29/07/05 16.1 2,790 0.8 7.2 −50 367 2.1F5 Saprolite 7.5 7.0 1.0 29/07/05 16.1 1,006 4.0 7.6 78 255 0.5F6 Saprolite 4.5 4.0 1.0 29/07/05 15.1 340 7.4 7.5 57 127 0.2F7 Saprolite 2.5 2.0 1.0 29/07/05 15.5 842 7.5 7.9 13 340 0.3F8 Clay 1.5 1.1 0.8 14/11/05 15.7 2,554 51 1.8662710650 Woolshed Flat Shale 52.6 26.3 O/H 11/10/06 2,561 0.02 6.8 −112 410 2.0662710655 Woolshed Flat Shale 52.6 26.3 O/H 11/10/06 2,767 0.01 12.8 −33 372 1.9662710653 Woolshed Flat Shale 58.2 29.1 O/H 13/02/07 19.8 1,997 0.5 7.7 −24 420 1.8

Measured physical parameters, major ion chemistry, CFCs and the stable isotope results for the corresponding piezometers, open wells arealso shown. pptv equivalent atmospheric concentration in parts per trillion volume; N/S no sample taken; O/H open hole

1974


Anions were analysed using a Dionex ICS-2500 IonChromatograph.

Groundwater samples were collected for CFC-11(trichlorofluoromethane, CFCl3) and CFC-12 (dichlorodi-fluoromethane, CF2Cl2) analysis, using a N2-pressurisedgas bailer. Analysis of CFCs is by purge and trap gaschromatography. Corresponding analytical errors in ap-parent CFC ages are approximately ±2 years for ages lessthan 20 years, increasing to ±4 years for ages of 30 yearsand higher. The detection limit for both CFCs isapproximately 5 pg kg−1, which equates to an age ofwater dating to approximately the year 1961 (Busenbergand Plummer 1992).

All isotopic concentrations were measured by isotoperatio mass spectrometry using a Europa Geo 20-20 at theCSIRO Land and Water Isotope Analysis Service inAdelaide, South Australia. δ2H and δ18O were analysedby H2O reduction to H2 (for δ2H) by hot Uranium(Dighton et al. 1997) and CO2 equilibrium for δ18O(Socki et al. 1992). The results are reported as a deviationfrom Vienna Standard Mean Ocean Water (vs. VSMOW)in per mil (‰) difference using delta (δ) notation. Theanalytical precision for δ18O and δ2H is ±0.15 and ±1.5‰,respectively. The local meteoric water line (LMWL) forAdelaide, the closest rainfall station located approximately

30 km to the northwest of SCC with reasonable isotopicrecords is δ2H=7.7×δ18O+9.6 (IAEA/WMO 2001). Datafrom the Adelaide rainfall station, provided by theInternational Atomic Energy Agency (IAEA) GlobalNetwork of Isotopes in Precipitation (GNIP) service, werecollected from 1962 and ceased in 1984 (IAEA/WMO2001). The GNIP-derived LMWL for Adelaide is insimilar agreement with the LMWL calculated for BedfordPark (δ2H=7.5×δ18O+11.2) at the western foothills of theMLR (Kayaalp 2001) and the LMWL calculated for ScottBottom (δ2H=6.6×δ18O+6.1) based on several years ofdata between 2004 and 2007. Local rainfall at the ScottBottom site was collected using a 200-mm-wide funnelinserted into a 1.25-L bottle held 1 m above ground in100-mm PVC stormwater pipe next to the pluviometer atthe site. Approximately 10 mm of paraffin oil was addedto each new collection bottle prior to deployment toprevent isotopic fractionation from evaporation of therainfall sample, which was collected on a monthly basis.

Results and discussion

This section discusses the hydraulic data along thehillslope transect and the surface water and groundwater

Cl SO4 Ca K Mg Na TDS δ2H δ180 CFC 11 CFC 11 CFC 12 CFC 12mg/L mg/L mg/L mg/L mg/L mg/L mg/

L‰ relVSMOW

‰ relVSMOW

pptv age pptv Age

3,220 521 529 23 439 994 6,069 −25.0 −4.69 43 1969 90 196810,100 780 1,300 48 1550 2210 16,336 N/S N/S N/S N/S232 30 19 2.6 13 136 474 N/S N/S N/S N/S565 158 128 8 80 303 1,647 −24.2 −4.78 68 1972 196 1974565 156 127 8 79 301 1,639 −24.2 −4.64 76 1972 266 1979479 207 128 8 80 289 1,649 −24.5 −4.85 26 1965 228 1976552 167 126 8 80 306 1,664 −25.7 −4.88 71 1972 293 1980807 305 194 15 115 381 2,242 −25.6 −4.76 <25 <1965 56 <1965748 299 176 10 117 327 2,035 −24.6 −4.79 <25 1965 102 1969646 267 160 12 102 303 1,855 −24.2 −4.68 37 1968 231 1976718 416 199 11 108 351 2,142 −24.6 −4.87 <25 <1965 99 1969715 435 198 11 107 355 2,141 −26.3 −4.99 <25 <1965 108 1969413 247 135 9 85 252 1,614 −26.1 −4.97 100 1974 340 1983523 366 162 8 101 286 1,842 −26.8 −5.06 60 1971 247 19771,062 476 268 16 175 447 3,055 −26.6 −4.97 N/S N/S1,030 261 206 6 174 408 2,287 −21.2 −4.37 N/S N/S750 514 184 17 101 433 2,341 −28.5 −5.47 <25 <1965 263 1979985 322 216 12 151 412 2,542 −23.45 −4.63 123 1976 85 1967827 288 186 18 120 361 2,205 −23.3 −4.65 42 1968 101 1969755 248 168 9 117 316 1,979 −23.7 −4.67 <25 <1965 103 1969635 279 155 13 100 302 1,831 −25.2 −4.93 <25 <1965 77 1967877 343 176 11 136 377 2,251 −23.1 −4.59 86 1973 259 1979692 319 153 13 118 329 1,988 −29.5 −5.35 <25 <1965 <50 <1965696 313 150 13 119 334 1,997 −28.6 −5.53 <25 <1965 <50 <1965537 248 139 12 103 294 1,702 −29.2 −5.3 <25 <1965 <50 <1965155 54 67 3 37 88 660 −23.4 −4.64 106 1975 309 198251 11 25 1 14 37 267 −19.8 −4.41 157 1980 457 199081 35 23 1 31 130 641 −20.9 −4.4 N/S N/S727 124 37 3 57 368 1,369 −19.8 −4.3 N/S N/S582 540 211 13 121 354 2,233 −29.23 −5.46 N/S N/S565 369 192 10 124 303 1,937 −28.49 −5.62 N/S N/S559 267 157 9 131 244 1,789 −30.33 −5.48 N/S N/S

1975


chemistry results from the nested piezometers. It presentsa conceptual model of groundwater flow dynamics andmechanisms of streamflow generation at the study siteScott Bottom. Importantly, the relative importance ofsubsurface flow activity in both of the shallow saproliteand deeper fractured bedrock geologic compartments isassessed.

Hydrogeological characterisationDrilling investigations for 42 shallow piezometers in July2005 at Scott Bottom showed that there were four distinctsoil/rock horizons (Figs. 2 and 3). From the groundsurface to 10–50 cm depth exists an A horizon composedof loose, organic-rich silty clay soils. An abrupt texturalcontrast was observed between the A and B horizons, andis characteristic of a duplex soil (Chittleborough 1992).The B horizon ranged in depth from 10–50 cm to 1.3–2.4 m before a gradual transition to fragmented saprolite.Varying degrees of mottling were observed in the Bhorizon with the presence of biopores decreasing distinct-ly with depth. The saprolite ranged in depth from 1.3–2.4to 6–18 m below ground. Extensive weathering wasobserved in this zone with both visible leached profilesand iron oxide staining. Within the saprolite there was atransition with depth from unconsolidated and highlyporous to progressively less weathered material, whichretained the structure of the parent bedrock. Groundwaterflow in the saprolite zone is expected to occur primarilythrough the permeable material, and from observation ofdrill logs and excavated pits, was extensively fracturedwith a fracture spacing less than 0.1 m. The unweatheredbedrock (Woolshed Flat Shale) below the saprolite was agrey siliceous slate, containing frequent quartz veiningand pyrite precipitates. Acoustic borehole televiewer(BHTV), geophysical and video camera logs of theexisting 96 m deep open hole in the Woolshed Flat Shaleat Scott Bottom identified a structurally complex de-formed and inclined fold sequence.

Groundwater flow in the fractured bedrock is likely tobe dominantly controlled by the fracture network density,geometry, connectivity and mineralization. According tothe BHTV and calliper logs, there are at least 196significant fractures and 5 dominant fracture sets. Thesefracture sets include bedding, bedding-parallel, two mainjoint sets and some random fractures. The dip and dipdirection of these sets are: set A 39/220, set B 55/050, setC 53/292, set D 62/340 and set E 52/144, respectively.These results suggest that there is likely to be some strongbedding controls but no strong structural anisotropy,which is likely to be related to the complexity of thedeformation and folding of the bedrock. Significantconductive fractures and fracture set spacings at intervalsbetween 0.1 and 1 m over the entire depth of the holewere observed (unpublished data 2007). These resultsmatched the outcrop mapping of the Woolshed Flat Shaleunit at a nearby road cutting and mine located less than1 km from Scott Bottom. The average fracture spacing ofthe bedrock was 0.21 m.

For this study the groundwater system has beenpartitioned, according to the change in geology, into threemain groundwater flow zones: soil zone (0 to 1.3–2.4 m),saprolite zone (1.3–2.4 to 6–18 m) and fractured bedrockzone (deeper than 6–18 m). However, it is the relativeactivity of the saprolite zone compared to the fracturedbedrock zone that is of greater interest in this study.

The single-well aquifer tests conducted in the piezom-eters constructed in the fractured bedrock (Woolshed FlatShale) at different rates of constant discharge showed thatthe average bulk hydraulic conductivity was 5 m/day, andranged from 1.5 m/day to 14 m/day. The monitored waterlevels in the piezometers constructed above and below thepumped piezometer only showed minimal drawdownduring the 100-min aquifer tests. The lack of drawdownsuggests that the vertical connection between piezometersvia the fracture network is limited and/or the horizontalconnection of the fracture network is extensive and thatthere is a high horizontal to vertical hydraulic conductivityanisotropy ratio. Changing the rate of constant dischargefrom ca. 13–50 m3/day (maximum pumping rate) for thetests did not significantly affect the degree of connectionbetween the pumped and monitored piezometers. For thetests with a higher discharge rate, the calculated hydraulicconductivity was less than one order of magnitudedifferent when compared to the tests conducted with alower discharge rate. This suggests that the rate ofdischarge has little effect on the estimated hydraulicconductivity of the fractured bedrock but that the testwas limited by the maximum pumping rate (and hencesampling spatial scale), which could be employed here.Investigations in porous and heterogeneous carbonaterocks by Schulze-Makuch and Cherkauer (1998) foundthat increases in apparent hydraulic conductivity aredependent on the scale of the aquifer test and not themethod of measurement. They also noted that the scaledependency is caused by heterogeneities and connected-ness within the aquifer. As discussed earlier, the highfracture density and orientation of the major fracture setswithin the Woolshed Flat Shale unit implies that there islaterally extensive, well-connected fractures along bed-ding and bedding parallel. In the vertical direction,fracture connectivity is likely to be limited to two majorjoint sets. The fracture density of the fractured bedrockzone (average=0.21 m) and the high measured hydraulicconductivity from the aquifer tests is typical of a highlyconductive aquifer with the groundwater flow controlledby well connected and conductive fractures (Gburek et al.1999). Direct connectivity between the fractured bedrockzone and Scott Creek at Scott Bottom is also evident.During the drilling of the piezometers into the fracturedrock at nest D and during development of the well, asteady line of rising air bubbles in Scott Creek wasobserved indicating that there is a direct connectionbetween the fractured rock aquifer and the creek.

Investigations by Mortimer (2009) and Skinner andHeinson (2004) in the Clare Valley (approximately110 km to the north of Scott Bottom and in the sameregional geology) found that near vertical structures and

1976


bedding planes dominate the hydraulically conductivefractures and that near-horizontal structures are critical tofracture network connection. Mortimer (2009) also iden-tified that groundwater flow in fractured rock aquifers canbe affected by in-situ stress fields at depths less than200 m. Whilst the palaeo-stress regimes play the primaryrole in creating the original properties of the fractured rock(orientation, fracture density, length, etc) and hence thedominant influence on groundwater flow, the more recentin-situ stress fields tend to alter fracture hydraulic aperturedistributions and fracture network connectivity. Thesurface processes of weathering, erosion and unloadingalso contributes to spatial heterogeneity in the fracturedrock and the depth dependent changes in fracture densityhave a significant influence on groundwater flow systems.

Single well aquifer tests were attempted on three of thepiezometers that were located in the saprolite zone.However, the piezometers ran dry even at very lowpumping rates. As an alternative, bail tests were conductedon these piezometers using the method described byHvorslev (1951). The hydraulic conductivity in thesaprolite zone ranged from 0.04–2.5 m/day and a meanof 1.5 m/day. Whilst it indicates considerable heterogene-ity in the saprolite, it is generally less than the hydraulicconductivity values determined for the fractured bedrockzone.

Soil-saturated hydraulic conductivity was estimated toimprove our understanding of the rainfall infiltration ratesand the physical properties of the soil zone at the studysite. The saturated hydraulic conductivity measurementsusing the Guelph permeameter in the soil zone between 5and 200 cm depth at Scott Bottom were much lower thanthe single well aquifer tests of the fractured bedrock.Hydraulic conductivities ranged from 0.003 to 0.33 m/day,a mean of 0.045 m/day and the highest conductivityrecorded at 200 cm depth. If these measurements reflecteffective hydraulic conductivity of surface layer materialsat the site, then only very slow groundwater recharge isexpected through the unsaturated soil zone. This isqualitatively reinforced by the observations of overlandflow at the site during significant rainfall events andfurther suggests that recharge (rates not known) must beoccurring via preferential pathways. It is interesting tocomment on the range of measured hydraulic conductivityvalues in the soil zone. A study by McKay et al. (2005)conducted in sedimentary rock saprolite in Tennessee,USA, showed that variations in hydraulic conductivity didnot always correspond with changes in lithology, but acomplex interaction of several factors, including parentbedrock lithology, the nature of macropores and degree ofinfilling with pedogenic clays and Fe/Mn oxides. Prefer-ential weathering processes are thought to occur in highlyfractured zones or in areas exposed to increased chemicalor physical weathering. The potential origin and processescontrolling the generation of the soils at Scott Bottom isdiscussed by Chittleborough (1992). The strong texturecontrast between surface and subsurface horizons areimportant in controlling the resultant hydraulic propertiesof interest in soil flow analyses.

Water movement in the soil zone appears to becontrolled by infiltration rates through the A-horizon andthe rate at which macropores are filled. A previous studyby Leaney et al. (1993) concluded that flow throughmacropores was the major mechanism for infiltration andthroughflow (horizontal movement of water through theunsaturated zone) at a site near Scott Bottom. Their resultsshowed that the major component of throughflow at thesite was precipitation associated with current storm/rainfall event rather than pre-existing water in the soilzone regardless of the magnitude of the rainfall event orseason. Over the course of their study, it was observed thatrainfall intensity often exceeded the infiltration rate of thesoil zone during the winter months resulting in overlandflow. It was not clear from drilling whether observedbiopores or macropores of any other type, penetratethrough the B-horizon to the saprolite. It would beexpected, however, since the hillslope was only clearedof its native vegetation approximately 80 years ago, thatremnant biopores from native vegetation might still beactive (Harris 1976).

Groundwater/surface-water interactionThe hydrogeological characterisation demonstrated thephysical differences between the soil and saprolite zoneswith the fractured bedrock groundwater flow zone. Thefollowing section describes the direction and relativegroundwater flow rates. Groundwater level time seriesdata was collected over the study period to establish thehydraulic responses in each of the aquifer zones to rainfallevents and hydraulic gradients at the site. Groundwaterlevel data from the six piezometer nests along the hillslopetransect and rainfall collected from 15 July 2005 until 31December 2007 are shown as time series plots in Fig. 4.Water levels measured at each of the nests are showncorrected to a RSWL relative to mAHD.

At nest A (the southern upslope end of the transect),only the two deepest piezometers A1 (28.5 m) and A2(20.5 m) were wet for the duration of the study period(Fig. 4a). The six shallower piezometers remained dryexcept for piezometer A7 (2.5 m) and A8 (1.5 m) whenthere was a significant late spring rainfall event inNovember 2005 and May through July of 2007. Thevertical hydraulic gradient between the deeper piezom-eters A2 and A1 was in a downward direction indicatingprocesses of groundwater recharge. The aquifer responseto rainfall at nest A in the fractured bedrock zone wasconsiderably dampened with a lag time between event andwater-table/water-level rise in the order of weeks tomonths. The lag in response was in part due to the depthto the piezometric surface being greater than 20 m.

Similar to nest A, the shallower piezometers at nest Bwere dry throughout the study period except for the twodeepest piezometers B4 (10.5 m) and B5 (7.65 m), locatedin the fractured bedrock. The vertical hydraulic gradientbetween these piezometers was in a downward directionfor the duration of the study period (Fig. 4b). At nest C,there was also a downward vertical hydraulic gradient

1977


between piezometers C4 (10.5 m) and C5 (7.5 m), locatedin the fractured bedrock. Piezometer C6 (4.5 m), locatedin the saprolite, was only wet during the higher rainfallevents. The water level in C6 was very similar to C4 andC5 indicating that there is a strong hydraulic connectionbetween the saprolite and fractured bedrock zone at thislocation (Fig. 4c). An interesting feature of the hydraulicdata at nest C was the larger piezometric surface responseto rainfall and greater seasonal variation between headlevels than at nests A and B. This is indicative ofsignificant recharge processes most likely associated withrainfall events in combination with low aquifer storage.

Figure 4d shows a large aquifer response to rainfallevents at all depths at nest D, located on the southern sideof the creek. There was a large vertical hydraulic gradientupwards between the deepest piezometer D1 (25.5 m) andpiezometer D6A at 4.5 m depth. Between the shallowestpiezometer D8 (1.5 m) and D6A there was a verticalhydraulic gradient downwards. This indicates a conver-gence of shallow and deep groundwater within thesaprolite zone. The piezometric surface elevations in nest

D were all higher than the stage height in Scott Creekindicating that Scott Creek is gaining along this reach.

At nest E, on the northern side of Scott Creek, anupward vertical hydraulic gradient was observed betweenthe deepest piezometer E1 (25.5 m) and E6 (4.5 m)(Fig. 4e). The shallower piezometers E7 (2.5 m) and E8(1.5 m) remained dry throughout the study period. Similarto nest D, the piezometric surface elevations in nest Ewere all higher than the stage height in Scott Creekindicating that Scott Creek is gaining along this reach. Theaquifer response to rainfall at nest E was rapid and large atall depths. The observed trends in the water levels in eachof the piezometers at nest E show three zones of hydraulicconnectivity. These zones may relate to the physicalcharacteristics and boundaries of the soil deposits,saprolite and the fractured bedrock at this location.

At nest F, the vertical hydraulic gradient observedbetween the deeper piezometers, F2 (20 m), F3 (15.5 m)and F4 (10.5 m), and the four shallower piezometers, F5(7.5 m), F6 (4.5 m), F7 (2.5 m) and F8 (1.5 m) wasupwards suggesting groundwater discharge from the

Fig. 4 Manual water level data (corrected to RSWL mAHD) from the piezometer nests a A, b B, c C, d D, e E and f F at Scott Bottomfrom 15 July 2005 until 31 December 2007. Depths of piezometers are shown in the legend. Discontinuous hydrographs of somepiezometers are a result of the water table falling below the base of screen. Automated rainfall data from the pluviometer and creek gaugeheight between nest D and E adjusted from the gauge station at the study site are also shown

1978


Tab

le2

Measuredph

ysical

parameters,major

ionchem

istryandthestable

isotop

eresults

forthesurfacewater

andlocalrainfallsamples

atScottBottom

Site

IDDate

sampled

Temp

EC

DO

pHField

Eh

HCO3

Br

Cl

SO4

Ca

KMg

Na

TDS

δ2H

δ180

°CμS/cm

mg/L

mV

mg/L

mg/L

mg/L

mg/L

mg/L

mg/L

mg/L

mg/L

mg/L

‰rel

VSMOW

‰rel

VSMOW

ScottCreek

02/08/05

9.0

816

12.8

7.8

138

145

0.5

184

4030

530

9552

9−2

3.2

−4.72

ScottCreek

14/11/05

442

7.4

930.1

7516

164

1553

271

−17.0

−3.79

ScottCreek

03/05/06

N/S

N/S

N/S

N/S

N/S

N/S

N/S

N/S

N/S

−23.6

−4.37

ScottCreek

15/11/06

15.0

1353

1.1

8.0

1127

20.9

290

6665

662

172

934

−20.6

−3.65

ScottCreek

18/12/06

18.2

1844

5.5

8.2

34411

1.2

384

9991

879

204

1277

−16.3

−3.29

ScottCreek

13/02/07

19.8

1998

0.5

7.7

−24

378

1.4

421

117

998

8924

113

54−2

0.2

−3.38

ScottCreek

30/08/07

12.3

948

10.4

7.8

125

192

0.5

176

4438

537

9859

3−2

3.0

−4.65

ScottCreek

15/03/07

16.6

2430

0.2

7.56

−636

81.7

495

165

123

9117

271

1549

−18.5

−3.08

ScottCreek

20/06/07

8.1

855

14.7

892

148

0.5

173

7537

.54.5

40.7

110

589

−26.1

−4.67

Upstream

15/03/07

20.1

2425

0.8

7.9

1135

81.6

478

162

122

9113

267

1511

−20.0

−2.9

Dow

nstream

15/03/07

16.1

2482

0.3

7.3

−124

408

1.8

547

153

119

1010

930

516

52−1

6.5

−3.3

Upstream

20/06/07

8.0

850

9.8

7.9

9614

40.5

173

7837

541

108

586

−27.7

−4.6

Dow

nstream

20/06/07

7.9

872

11.3

8.0

8014

80.5

176

7838

441

112

598

−25.7

−4.8

Mon

thly

rain

Jun-04

−1.7

−0.04

Mon

thly

rain

Oct-04

−21.2

−4.49

Mon

thly

rain

Aug

-04

−26.0

−5.32

Mon

thly

rain

Feb-05

−9.0

−1.60

Mon

thly

rain

Mar-05

−19.0

−4.04

Mon

thly

rain

Apr-05

−13.9

−3.08

Mon

thly

rain

May-05

−3.9

−1.85

Mon

thly

rain

Jun-05

−24.5

−5.52

Mon

thly

rain

Jul-06

−28.8

−4.84

Mon

thly

rain

Aug

-06

−8.3

−2.37

Mon

thly

rain

Sep-06

−15.6

−3.38

Mon

thly

rain

Oct-06

5.5

−0.21

Mon

thly

rain

Jan-07

−20.9

−3.68

Mon

thly

rain

Mar-07

−20.5

−4.13

Mon

thly

rain

Apr-07

−52.5

−7.43

Mon

thly

rain

May-07

−35.2

−5.16

Mon

thly

rain

Jun-07

−12.2

−2.43

Mon

thly

rain

Jul-07

−18.8

−4.01

Mon

thly

rain

Aug

-07

−13.4

−3.10

Rainfallevent

Jun-07

−13.0

−3.13

Rainfallevent

Jul-07

−24.6

−4.98

Rainfallevent

Jul-07

−18.3

−3.59

N/S

nosampletaken

1979


fractured bedrock aquifer towards the overlying perchedaquifer (Fig. 4f). However, measured field parameters(EC, pH and temp) and observations of boggy soilconditions and permanent puddles of surface water forthe duration of the study period indicated that groundwa-ter discharge from the fractured rock zone is occurring atthe ground surface, less than 10 m below nest F via anatural spring. The variations in the potentiometric surfaceof the fractured bedrock aquifer system at nest F is smallwith a dampened response to rainfall events and suggeststhat this aquifer may be semi-confined in this area of thesite. The hydraulic head data in piezometers F2, F3 and F4are almost identical indicating that they are hydrogeolog-ically well connected. However, the data also suggests thisgroup of F piezometers are hydraulically disconnectedfrom the piezometers above. According to the drill holelogs of nest F, the transition zone between the saproliteand fractured bedrock occurs at about 10 m depth, whichis in between the screen depth of piezometers F4 (10.5 m)and F5 (7.5 m). This transition zone is likely to form thebase of the shallow perched aquifer above. On severaloccasions (October and November 2005), in response toheavy rainfall events at Scott Bottom, the water level ofthe perched aquifer was higher than the potentiometricsurface of the fractured bedrock aquifer. This indicatedpossible recharge from the perched aquifer to the fracturedbedrock aquifer. Frequent monitoring (days to weeks) ofthe water levels of the perched aquifer at nest F (F5, F6,F7 and F8) showed rapid responses to rainfall events withthe water table rising to within 0.2 m of the groundsurface and falling significantly over a period of weeks.

Hydraulic conceptual modelObservations of the hydraulic data from each of thepiezometer nests along the hillslope transect show severalinteresting features which are described here as a basis forformulating the hydraulic conceptual model. At any pointin time the hydraulic data shows the movement ofgroundwater from the elevated areas at the northern andsouthern extents of the transect towards the creek at thevalley bottom, i.e. head gradients always point from theaquifer toward Scott Creek. This demonstrates that ScottCreek is a gaining stream. The groundwater flux to thecreek is determined by the hydrogeological characteristicsof the interface between the saprolite and bedrock with thecreek bottom. Observations of fracture orientation anddensity and outcropping in the creek bed suggest thatthere is exchange between the fractured bedrock and thecreek. Over the period of the study, the higher groundwa-ter levels in July–September (winter) compared to thelower groundwater levels in March (baseflow onlyconditions) at nests D and E relative to the creek standingwater level suggests that groundwater discharge to thecreek would be greatest during July-September as a resultof the larger hydraulic gradient. Baseflow conditions inScott Creek typically occur from the beginning ofNovember through to May and there are very fewrecorded occasions when flow has ceased completely.

In summary, the piezometer hydrographs of nests C, D,E and F show that there is a seasonal response of thepotentiometric surface to groundwater recharge in the soilzone, saprolite zone and also in the fractured bedrockzone. This indicates that the fractured bedrock groundwa-ter system is dynamic and hydraulically active. Hydrau-lically active refers to the responsiveness of the aquifersystem to groundwater recharge and flow processes. Thewater level elevations in each piezometer are generallyhigher in July–September and have a lag time of severalweeks after the rainfall event. More importantly, there aresignificant water level fluctuations in the fracturedbedrock zone, similar to the soil and saprolite zones asobserved in piezometric information. The limited hydrau-lic gradient data of the saprolite zone restricts any robustquantitative comparison between groundwater dischargevolumes and flow velocities through the soil, saprolite andfractured bedrock zones. However, there is a discernibledifference between the hydraulic conductivities of the soil,saprolite and fractured bedrock zones, which can be usedto infer the relative activity of these groundwater zones. Inparticular, it is noted that hydraulic conductivity is aparameter, which can vary by many orders of magnitude(e.g. Love et al. 2002), but that field scale hydraulicgradients are typically less variable than the variability inK. Thus, hydraulic conductivity is a good proxy forgroundwater velocity. For the surficial soil layer, thehydraulic conductivity ranges from 0.003–0.33 m/day, forthe saprolite system hydraulic conductivity ranges from0.04–2.5 m/day and in the deeper fractured bedrock it wasfound to range from 1.5–14 m/day (similar to a fine-medium sand in a sedimentary system).

The measured horizontal hydraulic gradient in thefractured bedrock (10−2) at the site is higher than whatmay be considered a typical field gradient (say approxi-mately 10−3) and comparable to the range in verticalhydraulic gradients between each of the groundwater flowzones at each of the nests (ca. 10−1–10−3). Combining therelatively high hydraulic conductivity data for the frac-tured bedrock aquifer with an average horizontal gradientof 10−2, points to a highly active groundwater flow systemin the bedrock aquifer. In contrast, the lower hydraulicconductivity of the saprolite and soil zone with a similarhorizontal hydraulic gradient (based only on the gradientbetween nest F and E in the saprolite zone) suggests thatthe saprolite zone (and also the soil layer) is lesstransmissive than the fractured bedrock zone laterally.When these basic calculations are extrapolated to computevolumetric contributions by including layer thickness perunit width (i.e. average thickness of soil=1.85 m andsaprolite=10 m), it is expected that the saprolite (andindeed soil) zones would be smaller contributors yetagain. This is because they not only have smallerhorizontal hydraulic conductivity and hydraulic gradients(and hence horizontal flow rates) but their effectivecontributing thicknesses are also significantly smaller thanthe deeper fractured bedrock aquifer (approximately100 m). The calculated discharge of the soil, saproliteand fractured bedrock groundwater flow zones using the

1980


range of measured hydraulic conductivities, gradients andaverage thicknesses is 10−5–10−3, 10−3–10−1 and 10–100 m3/day, respectively. Whilst it is apparent that thehydraulic conductivity, groundwater flow speed, andgroundwater volumetric discharge from the deeper fracturedbedrock are expected to be larger than those which are likelyto be encountered in the shallower saprolite and soil zones,further data would be required to verify these estimates.

Groundwater chemistry variationThe hydrogeological conceptual model has indicated anactive fractured bedrock aquifer system at the study site.In the following sections, some of the hydrochemical andisotope results are compared to the hydraulic model. Thegroundwater samples collected between the 28 July and 2August 2005 at Scott Bottom show chemical variabilitybetween the soil, saprolite and fractured bedrock zonesand highlight the spatial differences along the hillslopetransect (Fig. 5 and Table 1). The plots in Fig. 5 show thesurface water and groundwater major ion (Ca2+, Mg2+,Na+, K+, SO4

2− and HCO3−) to chloride ratios versus

chloride compared to the respective ion/chloride ratios oflocal rainfall collected at Scott Bottom. It is assumed thatchloride behaves conservatively and therefore variationsin ionic ratios are presumed to indicate addition or loss ofother major ions via flow through the system. Thegroundwater samples from the six piezometer nests andgroundwater samples from the open wells at the study siteare presented in the plots relative to the lithology type atthe sample depth. Surface water samples from Scott Creektaken at different times of the year are also shown forcomparison (Table 2).

The plots of the major ion/chloride ratios versuschloride for the groundwater samples show broad lineartrends of increasing individual ion concentrations relativeto increased chloride concentration (Fig. 5). The trends inCa2+, Mg2+, Na+ and HCO3

− lie slightly above the rainfalldilution line implying some degree of water–rock interac-tion and the weathering of primary silicate minerals. Otherstudies (e.g. Poulsen et al. 2006) have indicated that themajority (75%) of dissolved salts in the Mount LoftyRanges are of marine origin (rainfall) and only a minorcomponent (25%) comprised of terrestrial sources such asrock mineral weathering.

It would be expected that the degree of chemicalweathering along the inferred groundwater flowpaths fromthe upslope area towards the creek (discharge zone) at thevalley bottom would increase because of the longer waterresidence time in the aquifer. However, the hydrochemicaldata suggests that there are more complex hydrogeochem-ical interactions taking place and that merging groundwa-ter flowpaths and interaction between the groundwaterflow zones is influencing the groundwater composition. Itis worth noting that there is greater chemical variation(and in some cases higher concentrations) in the samplesfrom the soil and saprolite zones compared to thefractured rock zone. This is shown more clearly in thedepth profile plots of nest D, E and F in Fig. 6. This may

be a result of the spatial variability between atmosphericand weathering (clay soils and weathered fracturedbedrock metasediments) derived sources, evaporativeprocesses and variable rates of flushing of pore watersvia discrete flow paths during recharge events. The morepositive deuterium values close to the soil surface arecharacteristic of an evaporative process, however, in theδ2H versus δ18O plot (Fig. 7) there is little indication ofevaporative enrichment in all three groundwater zones andtherefore it cannot be considered as a dominant factor tothe higher solute concentrations. Whilst Scott CreekCatchment has a large proportion of native vegetation,there are areas that have been cleared in the last 80 yearsincluding parts of Scott Bottom, which may have alsoaffected the movement of salt in the landscape. Theprocesses of salt mobilisation, increased recharge andstream salinisation are well documented in Australia (e.g.Peck and Hurle 1973; Williamson et al. 1987). Vegetationclearance is likely to have resulted in increased recharge tothe aquifer system and therefore the soil and saprolite zonesmay still be undergoing a transition to re-equilibrium.

The samples taken from the saprolite zone at nests Dand E have a similar isotopic composition to thefractured bedrock zone, which may be a result ofgroundwater mixing before final groundwater dischargeoccurs to the creek from this zone, which was shown inthe hydraulic data. In comparison, the groundwatersamples F5, F6 and F7 from the saprolite zone at nest Fhave low dissolved solutes similar to rainfall and a morepositive deuterium value thought to be a result ofseasonal groundwater recharge events. The soil andsaprolite groundwater samples are distinctly different tothe samples taken from the piezometers at greater depth(F2, F3 and F4) located in the fractured bedrock, whichhave higher major ion concentrations, and more depletedisotopic compositions.

The dissolved oxygen and redox potential measure-ments from nests D, E and F show the groundwatersvarying from oxidising to reducing conditions with depth,which coincide with the boundary between the saproliteand fractured bedrock zones at about 10 m (Fig. 6).Similar changes of other physical parameters, includingtemperature and pH profiles, with increasing depth at nestsD and E may be a result of the piezometer screensintersecting consistent fracture networks and hence similarpreferential pathways of groundwater flow. The samplesfrom the fractured bedrock at each of the nests plot in acluster indicating that the groundwater in this zone isrelatively well mixed (Fig. 5). The samples with thehigher chloride concentrations are from the deeperpiezometers and the piezometers located at the valleybottom suggesting a greater groundwater residence time.The similarity of the major ion (Ca2+, Mg2+, Na+, Cl− andHCO3

−) profiles at nests D, E and F illustrate that thegroundwater is relatively well mixed within the fracturedbedrock, and also reflects the upward hydraulic headgradients at these nests as a groundwater discharge zone(Fig. 6). The deuterium composition of the fracturedbedrock zone at nest F is considerably more depleted than

1981


in the fractured bedrock zone at nests D and E, whichsuggests a different source of recharge for these ground-waters. This source may be deeper groundwater flow froman intermediate flow system that has recharged the aquiferat a higher elevation. These different chemical signaturesreinforce the hydraulic model described earlier, whichindicated a perched separated shallow aquifer overlyingthe deeper fractured bedrock aquifer at this location.

The creek is a mixed sample of groundwaters, soilwaters and surface runoff and the chemistry at any given

point in the creek is indicative of processes that representthe groundwater pathways, residence time and mixingwithin the entire upstream catchment. The range ofchloride concentrations and variability of the ion/chlorideratios of the samples from Scott Bottom clearly shows thatthe creek is comprised of a mix of different waters andthat dilution and evaporative processes influence itstemporal variability (Fig. 5). The surface-water samplestaken upstream and downstream of Scott Bottom show anincrease in concentration downstream implying that the

Fig. 5 Composite diagrams of major ion (Ca2+, Mg2+, Na+, K+, SO42− and HCO3

−) to chloride ratios versus chloride of surface water(August 2005 to June 2007) and groundwater collected between 28 July and 2 August 2005 at Scott Bottom

1982


creek is gaining along this section. Whilst the source tothe creek is not definitive in the hydrochemical data, thereis evidence to suggest that there is a groundwater sourcefrom the fractured bedrock zone.

Origin and ages of groundwaterThe isotopic ratios (δ2H and δ18O) of rainfall, surfacewater and groundwater samples from the Scott Bottom siteare plotted in Fig. 7, relative to the Adelaide LMWL and

Fig. 6 Temperature (Temp), dissolved oxygen, redox potential, pH, major ion and deuterium profiles of groundwater collected between 28July and 2 August 2005 at nests D, E and F along the hillslope transect. The vertical error bars represent the length of the screened interval

1983


the calculated Scott Bottom LMWL, to investigate thepotential source of the waters and in assisting thedevelopment of the hydrogeological conceptual model.The isotopic ratios of the groundwater samples from thenested piezometers range between −29.5 and −19.6‰ forδ2H, and between −5.5 and −3.6‰ for δ18O. The majorityof samples plot closer to the Scott Bottom LMWL thanthe Adelaide LMWL and indicates that despite the limitedisotopic rainfall data set at Scott Bottom it shows thelikely origin of the groundwaters. The δ2H and δ18O ofthese samples are similar or more positive than theweighted average rainfall for Scott Bottom (δ2H=−24‰and δ18O=−4.5‰). The more positive ratios of thesesamples are thought to represent a mixture of seasonallocalised recharge events and/or small amounts of evap-oration that may occur at high humidity. The threegroundwater samples from the open wells at the studysite (δ2H range between −29.2 and 30.3‰) and samplesfrom piezometers E1, F2, F3 and F4 plot close to the ScottBottom LMWL and have a more depleted isotopiccomposition than the weighted average rainfall for ScottBottom, which is indicative of diffuse recharge duringcooler autumn and winter rainfall events, and/or altitudeeffects. Altitude effects on isotopic composition is atemperature-related affect and leads to a reduction in bothδ18O (∼0.15–0.5‰ per 100 m of increase in altitude) andδ2H (∼1–4‰ per 100 m of increase in altitude) values

(Clark and Fritz 1997). The elevation of Scott Bottom is250 m above seal level, about 200 m higher than theAdelaide GNIP station and located on the eastern side ofthe Mount Lofty Ranges. The surrounding hills adjacentto Scott Bottom are at an elevation of about 320 m.Therefore, the most depleted groundwater isotopic valuesmentioned in the preceding can most likely be attributedto altitude effects and represent a deeper intermediategroundwater flow system at the study site. Additionally,however, observations of higher rainfall and increasedwater table elevations during autumn and winter at ScottBottom also provide evidence of seasonal recharge effects.Cooler rainfall would result in more depleted groundwaterisotopic values relative to mean Adelaide rainfall, andtherefore recharge to the deeper fractured bedrock zonemay only occur during these periods.

The variability of the isotopic composition of thegroundwater samples from the soil and saprolite zones isa result of seasonal recharge events of the currentclimate. This is comparatively different to the majorityof the groundwater samples from the fractured bedrockzone, which plot in a relatively tight cluster and representan isotopically well-mixed system where the seasonalsignature has been lost to dispersive mixing. Nosignificant deviation of the isotopic composition of thegroundwater samples from the Scott Bottom LMWL wasobserved and suggests that there has been minimal

Fig. 7 δ2H versus δ18O for local rainfall, Scott Creek and groundwater samples from the nested piezometers and open wells at ScottBottom over the monitoring period. The LMWL for Adelaide is δ2H=7.7 δ18O+9.6 and the LMWL for Scott Bottom is δ2H=6.6 δ18O+6.1. The mean weighted rainfall for Adelaide is δ2H=−26‰ VSMOWand δ18O=−4.7‰ VSMOWand the mean weighted rainfall for ScottBottom is δ2H=−24‰VSMOW and δ18O=−4.6‰ VSMOW

1984


isotopic fractionation by evaporative process prior torainfall infiltration to the three groundwater flow zones.Similarly, the water samples collected during the studyperiod from Scott Creek at Scott Bottom have isotopiccompositions close to the LMWL. However, samplesfrom Scott Creek collected during baseflow conditions(November–May) show signs of isotopic enrichment byevaporation, whilst samples collected during the winterperiod plot closer to the line and near the cluster ofgroundwater samples from the fractured bedrock zoneand the weighted average rainfall at Scott Bottom (Fig. 7;Table 2). The intersection of an evaporation trend line ofthe Scott Creek samples with the Scott Bottom LMWLconfirms the likely sources to Scott Creek. The bulkmonthly rainfall samples from the Scott Bottom siteshow seasonal isotopic variations as well as isotopicallydifferent rainfall events. The samples from the wintermonths generally fall close to the LMWL and below theaverage Adelaide rainfall, whilst the δ2H and δ18O fromthe summer months are generally more positive and plotnear the LMWL and above the average rainfall as a resultof warmer temperatures and greater evaporation duringsummer rainfall events (Coplen et al. 1999).

To summarise, the stable isotope data suggests thatthere are distinct differences between the soil, saproliteand fractured bedrock groundwater flow zones, and thatthere has been minimal evaporation of the water in eachof these zones. Samples from some of the deepestpiezometers (i.e. nests E and F) in the fractured bedrockzone appear to only be recharged during the coolerwinter months after significant rainfall events and athigher elevation than Scott Bottom. These groundwatersamples are likely to be sourced from a much deeperintermediate flow system because they have isotopiccompositions similar to the deep open wells. Themajority of the samples from the fractured bedrock zonehave a similar composition to the mean-weighted ScottBottom rainfall and represent an isotopically well-mixedsystem. The samples from the soil and saprolite zonesshow much greater variability and are influenced byseasonal recharge events. Surface-water samples at ScottBottom, particularly those sampled during the summermonths, have evolved from an isotopic compositionsimilar to groundwater from the fractured bedrock zoneas well as the soil and saprolite zones.

CFCs were used to determine the apparent age ofgroundwater and to provide information on the ground-water flow processes, including depth of circulation andvertical connectivity. CFCs have been used as ageindicators for groundwater studies since about 1979(Szabo et al. 1996). Measurable concentrations of CFCs(apparent groundwater age less than 40 years old) werefound in the fractured bedrock zone to depths of 20 mbelow the water table at the upper and lower elevationsof the hillslope (Table 1). Detectable CFC concentra-tions indicate that there is active groundwater flow inthe fractured bedrock zone. Although the screenintervals are short, there may also be mixing of oldergroundwaters with post-1950s groundwater from the

shallower flow zones. The majority of the CFC-11groundwater ages were older than the CFC-12 ages,which suggests that some retardation of CFC-11 hasoccurred in the unsaturated zone or degradation bymicrobial activity (Busenberg and Plummer 1992).Observed fracture spacings of the fractured bedrockzone at Scott Bottom are typically small and certainlyless than 1 m and as a result CFC concentrations in thefracture are expected to be in equilibrium with the porewater in the matrix (Cook et al. 1996; Cook et al.2005). Therefore, the apparent groundwater ages can beused to indicate the minimum depth of circulation andmay also be used to estimate groundwater recharge(Love et al. 2002).

CFC concentrations of samples taken from F2, F3 andF4 at nest F were below detection limit (5 pg kg−1)indicating groundwater ages older than 40 years. Thissupports the results from the hydraulic and isotope data,which showed that the groundwater from the fracturedbedrock zone at nest F is from a deeper intermediate flowsystem. Comparatively, CFC concentrations of sample F5(7.5 m depth) and F6 (4.5 m depth) in the saprolite zone,represent groundwater that is relatively young (<20 years).The presence of young groundwater at this depth supportsthe hydraulic and hydrochemical evidence of the relativelyfresh, perched aquifer at nest F.

Despite the upward hydraulic gradients in the fracturedbedrock zone at nests D and E, the presence of CFCs inthe fractured bedrock zone suggest that groundwater fromthe overlying soil and saprolite zones may be mixing withthe groundwater below or that there is a source of younggroundwater in the fractured bedrock up gradient of nestsD and E that is moving laterally towards the stream. Thehydraulics and hydrochemical data indicated that there islikely to be more dominant downward vertical groundwa-ter flow in the soil and saprolite zones. Whereas in thefractured bedrock, the data indicated a relatively well-mixed system where the dominant groundwater flow pathsare horizontal and converge with groundwater flowpathsof the overlying flow zones prior to discharge to the creekor in the case of nest F at the ground surface.

Therefore, the study data has demonstrated that thefractured bedrock zone may play a more active role instreamflow generation at Scott Bottom than has previouslybeen documented by other authors working at this site (e.g.Leaney et al. 1993) and internationally in analogousgeologic systems (e.g. McDonnell 1990; Manning andCaine 2007; Shand et al. 2007). These chemistry and tracerdata reinforce the hydraulic data. They are stronglysuggestive of the fact that the implicit or explicitassumption of a no-flow boundary condition at thesaprolite-fractured bedrock zone interface must be carefullyevaluated in the conceptual model of groundwater flow andgw–sw interaction in these types of geologic systems.

Groundwater rechargeIn the following, a chloride mass balance (CMB)technique is used to compare recharge rates at various

1985


locations throughout the site and in the different geologiccompartments. This is useful as a further indicator ofgroundwater dynamics at the site. Evidence of activerecharge in different geologic layers will provide anothermethod for defining the hydraulic conceptual model (e.g.,Is there any evidence of active recharge in the deeperfractured bedrock aquifer?). The CMB technique has beenused successfully in sedimentary aquifer systems (e.g.Allison and Hughes 1978) and has been suggested themost reliable technique for determining recharge rates tofractured rock aquifers systems (Cook 2003). However,the recharge rate determined from CMB should beconsidered as a minimum rate because of the addition ofother sources of chloride, which may occur, by rockweathering. Changes in environmental conditions (i.e.land clearing) will also impact on the equilibrium ofchloride in the fractures with the rock matrix and may takea significant amount of time for the diffusion of salts fromthe matrix into the fractures to re-equilibrate. Thesubsequent steady state mass balance equation can beused to estimate recharge (R):

R ¼ PCp

Cgwð3Þ

R is the estimated recharge [LT−1], P is the averageannual rainfall [LT−1], Cp is the chloride concentration ofrainfall [ML−1], and Cgw is the chloride concentration ofgroundwater [ML−1].

According to the long-term rainfall record, the averageannual rainfall at Scott Bottom is about 800 mm. Theaverage chloride concentration in rainfall from gaugestations in close proximity to SCC is about 9 mg/L(Department of Water, Land and Biodiversity Conserva-tion, unpublished data, 2007). Using the measuredchloride concentrations from the soil (4 samples), saprolite(6 samples) and fractured bedrock (19 samples) zones, theestimated recharge rate to the groundwater system in each

of these zones is 7–31, 8–141, and 0.7–15 mm/year,respectively. The average rate of 18 mm/year (2% ofrainfall) from all of these three zones is within the range ofestimated recharge rates of 26 natural catchments insouthern Australia documented by Scanlon et al. (2006)and similar to the estimated range of recharge rates (16–111 mm/year) in the western MLR (Green and Zulfic2008). There is significant variability in recharge over alarge spatial scale and the range of rates from all three ofthese zones represents only 0.1 to 18% of the long-termaverage annual precipitation. As discussed earlier, therewas greater variability in the hydrochemistry data in thesoil and saprolite zones, and therefore the applicability ofthis method to these zones may not be appropriate as thesystem is unlikely to be in a state of equilibrium. Moreimportantly, the recharge rates calculated from chloridedata in the deeper fractured bedrock system indicate thatthere is effective and significant recharge to this zone, anobservation that is entirely consistent with the conceptualmodel that recognises a dynamic and active fracturedbedrock system.

Summary and conclusions

Hydrological conceptual models of groundwater/surface-water (gw–sw) interaction in a saprolite-fractured bedrockgeological setting often do not consider the contribution ofgroundwater from below the bedrock interface (consider-ing it a no flow or impermeable boundary). More recentstudies have begun to address this issue (Shand et al.2005; Haria and Shand 2006; Manning and Caine 2007).The purpose of this investigation was to determine therelative importance and contribution of both the soil-saprolite and fractured bedrock aquifer systems and theirinfluence on gw–sw interaction. From the hydrogeologyand hydrochemical results presented in this study, a

Fig. 8 Conceptual model of the study site Scott Bottom. Arrows indicate direction of inferred groundwater flow

1986


conceptual model describing the significance of thefractured bedrock zone and its influence on the interac-tions between the groundwater system and Scott Creek atScott Bottom was developed (Fig. 8). This study providesan important demonstration of how geologic controlsgreatly influence groundwater flows.

The results presented here suggest that the relativeactivity of the groundwater in the fractured bedrock zoneis indeed not inactive and plays just as important a role asthat in the shallower soil and saprolite zones. Groundwaterflow in the fractured bedrock zone is dynamic and may bean important flow pathway along the hillslope via a well-connected fracture network. The seasonal variation of thewater levels in the nested piezometers and higherelevations during winter verify active groundwater re-charge and discharge occurring along the hillslope andthat groundwater movement is from the higher parts of thecatchment to the valley bottom. Groundwater flow in thefractured bedrock aquifer appears to be significant. It ishypothesised that the hydraulic properties and high degreeof connection between the soil, saprolite and fracturedbedrock zones is a dominant control on the majorcontribution of groundwater to Scott Creek. The hydraulicand hydrochemical data suggest that the groundwater inthe fractured bedrock zone is relatively well mixed andthat there is some mixing of this water with the shallowergroundwater in the soil and saprolite zones where thegroundwater flow paths converge at the valley bottomprior to discharge to the creek. The data also shows thatthere is a hydraulic disconnection between the soil andsaprolite zones with the deeper fractured bedrock in theupslope areas of the study site as a result of the increasingthickness of the soil and saprolite zones. Perched aquifersare established in the shallow soil zone and persistthroughout the year. These shallow aquifer systems mayplay an important role in overland flow and throughflowto the creek as well as a ‘solute’ mixing zone of shallowand deep groundwater.

Previous studies (Smettem et al. 1991; Leaney et al.1993) at the Scott Creek field site have assumed thatsubsurface flow occurs in the shallow saprolite zone only.In striking contrast, the results of this study are importantbecause they clearly show that groundwater flow in thedeeper fractured bedrock zone is highly dynamic and animportant groundwater flow pathway along the hillslopeand readily exchanges with the creek. As a result, deepgroundwater from the fractured bedrock aquifer istherefore expected to be a significant component instreamflow generation at Scott Creek. These are importantand new observations at the Scott Creek site, whichprofoundly alter the hydrogeologic conceptualisation ofthis system. Our results also reinforce the findings of alimited number of previous studies in similar geologicsettings (e.g. Shand et al. 2005; Haria and Shand 2006;Manning and Caine 2007). This finding will influence theway in which gw–sw interaction is analysed at the ScottCreek site and will inform the development of quantitativenumerical models of the system. More generally, however,this study suggests that hydrologic conceptual models that

do not consider the deeper fractured bedrock in hydrologicanalyses and hence treat the saprolite-fractured bedrockinterface as a no flow boundary may be overly simplisticand inherently misleading in some groundwater/surface-water interaction analyses such as the Scott Creek sitepresented here. The choice of conceptual model employedin the gw–sw interaction assessment (homogeneous,saprolite only, saprolite and deeper fractured bedrock) iscritical. What is particularly problematic is that the choiceof conceptual model is often made implicitly withoutcareful justification. Our results emphasise the need tounderstand the relative importance of subsurface flowactivity in both of these shallow saprolite and deeperbedrock compartments as a basis for developing reliableconceptual hydrologic models of these systems. Thesechoices are in turn crucial for understanding what level ofsimplification is permissible in conceptual model definition.Understanding fractured bedrock and saprolite hydrogeo-logic controls are clearly important for accurate quantitativeflux analyses, streamflow interpretation and a range of othermatters that arise in gw–sw interaction studies.

Acknowledgements The authors would like to thank M. Pichler, P.Kretschmer, E. Kwantes, L. Mortimer, S. Mercer and R. Baird forhelp with fieldwork and data collection. We gratefully acknowledgethe Department of Water, Land and Biodiversity Conservation andthe National Water Initiative Mount Lofty Ranges program forfunding the drilling, piezometer installation and instrumentation atthe Scott Creek study site. A Flinders University Program Grantalso supported field investigation and sampling activities. Themanuscript was improved by the constructive comments from S.Van der Hoven, V. Heilweil and M. Lenczewski.

References

Allison GB, Hughes MW (1978) The use of environmental chlorideand tritium to estimate total recharge to an unconfined aquifer.Aust J Soil Res 16(2):181–195. doi:10.1071/SR9780181

Bagarello V, Giordano G (1999) Comparison of procedures toestimate steady flow rate in field measurement of saturatedhydraulic conductivity by the Guelph permeameter method. JAgric Eng Res 74(1):63–71

Beyerle U, Aeschbach-Hertig W, Hofer M, Imboden DM, Baur H,Kipfer R (1999) Infiltration of river water to a shallow aquiferinvestigated with 3H/3He, noble gases and CFCs. J Hydrol 220(3–4):169–185

Busenberg E, Plummer LN (1992) Use of chlorofluorocarbons(CCl3F and CCl2F2) as hydrologic tracers and age-dating tools:the alluvium and terrace system of central Oklahoma. WaterResour Res 28(9):2257–2283

Chapman T (1999) A comparison of algorithms for stream flowrecession and baseflow separation. Hydrol Process 13:701–704

Chittleborough DJ (1992) Formation and pedology of duplex soils.Aust J Exp Agric 32:815–825

Clark ID, Fritz P (1997) Environmental isotopes in hydrogeology.CRC, New York

Cook PG (2003) A guide to regional groundwater flow in fracturedrock aquifers. CSIRO Land and Water, Clayton South,Australia, 108 pp

Cook PG, Solomon DK, Sandford WE, Busenberg E, Plummer LN,Poreda RJ (1996) Inferring shallow groundwater flow insaprolite and fractured rock using environmental tracers. WaterResour Res 32(6):1501–1509

Cook PG, Love AJ, Robinson NI, Simmons CT (2005) Groundwa-ter ages in fractured rock aquifers. J Hydrol 308(1–4):284–301

1987


http://dx.doi.org/10.1071/SR9780181

Coplen TB, Herczeg AL, Barnes CJ (1999) Isotope engineering:using stable isotopes of the water molecule to solve practicalproblems. In: Cook PG, Herczeg AL (eds) Environmentaltracers in subsurface hydrology. Kluwer, Boston, pp 79–110

Cranswick RH (2005) Hillslope geological controls on surfacewater–groundwater interaction: evidence of recharge to afractured rock aquifer. Honours Thesis, Flinders UniversitySouth Australia, Adelaide

Dighton JC, Leaney FW, Herczeg AL, Allison GB, Hughes MW(1997) A rapid and robust method for the preparation ofisotopically unaltered hydrogen gas from water for stableisotope mass spectrometry, 6th Australian-New Zealand Envi-ronmental Isotope Conference, Wellington, New Zealand, 2–4April 1997

Elrick DE, Reynolds WD (1992) Infiltration from constant headwell permeameters and infiltrometers. In: Topp GC, ReynoldsWD, Green RE (eds) Advances in measurement of soil physicalproperties: bringing theory into practice. Soil Science Society ofAmerica, Madison, WI, pp 1–24

Fetter CW (2001) Applied hydrogeology. Prentice Hall, UpperSaddle River, NJ

Gburek WJ, Urban JB (1990) The shallow weathered fracture layerin the near-stream zone. Ground Water 28(6):875–883

Gburek WJ, Folmar J, Urban JB (1999) Field data and ground watermodeling in a layered fractured aquifer. Ground Water 37(2):175–184

Genereux DP, Hemond HF, Mulholland PJ (1993) Use of radon-222and calcium as tracers in a three-end-member mixing model forstreamflow generation on the West Fork of Walker BranchWatershed. J Hydrol 142(1–4):167–211

Green G, Zulfic D (2008) Summary of groundwater rechargeestimates for the catchments of the Western Mount LoftyRanges Prescribed Water Resources Area. DWLBC Technicalnote 2008/16, Government of South Australia, through Depart-ment of Water, Land and Biodiversity Conservation, Adelaide

Haria AH, Shand P (2006) Near-stream soil water-groundwatercoupling in the headwaters of the Afon Hafren, Wales:implications for surface water quality. J Hydrol 331:567–579

Harrington GA (2004) Hydrogeological Investigation of the MountLofty Ranges, Progress Report 3: Borehole water and formationcharacteristics at the Scott Bottom research site, Scott CreekCatchment. Report DWLBC 2004/03, Department of Water,Land and Biodiversity Conservation, Adelaide

Harrington GA, James-Smith JM, Wohling D, Van Den Akker J(2004) Hydrogeological investigations of the Mount LoftyRanges, progress report 5: Drilling phases 2.1 to 2.3: researchand monitoring wells at Scott Creek, Balhannah, WillungaFault, Lobethal, Eden Valley and Ashbourne. Report DWLBC2004/04, Department of Water Land and Biodiversity Conser-vation, Adelaide

Harris C (1976) Vegetation clearance in South Australia. Report ofInterdepartmental Committee on Vegetation Clearance, SADepartment of Environment and Planning, Adelaide

Hvorslev MJ (1951) Time lag and soil permeability in groundwaterobservations. Bulletin 36, U.S. Army Corps Engineers Water-ways Experimentation Station, Vicksburg, MS

International Atomic Energy Agency and World MeteorologicalOrganisation (IAEA/WMO) (2001) Global Network of Isotopesin Precipitation (GNIP) database. http://isohis.iaea.org. CitedMarch 2006

James-Smith JM, Harrington GA (2002) Hydrogeological investi-gation of the Mount Lofty Ranges, Progress Report 1: hydro-geology and drilling phase 1 for Scott Creek Catchment. ReportDWLBC 2002/17, Department of Water, Land and BiodiversityConservation, Adelaide

Kahn K, Ge S, Caine J, Manning A (2008) Characterization of theshallow groundwater system in an alpine watershed: HandcartGulch Colorado USA. Hydrogeol J 16(1):103–121.doi:10.1007/s10040-007-0225-6

Kayaalp AS (2001) Application of rainfall chemistry and isotopedata to hydro-meteorological modelling. PhD Thesis, FlindersUniversity of South Australia, Australia

Kirkby M (1988) Hillslope runoff processes and models. J Hydrol100(1–3):315–339

Krause S, Bronstert A (2007) The impact of groundwater-surfacewater interactions on the water balance of a mesoscale lowlandriver catchment in northeastern Germany. Hydrol Process 21(2):169–184

Leaney FW, Smettem KRJ, Chittleborough DJ (1993) Estimating thecontribution of preferential flow to subsurface runoff from ahillslope using deuterium and chloride. J Hydrol 147(1–4):83–103

Love AJ, Cook PG, Harrington GA, Simmons CT (2002)Groundwater Flow in the Clare Valley. Report BookDWR02.03.0002, Department for Water Resources, Adelaide

Manning AH, Caine JS (2007) Groundwater noble gas, age, andtemperature signatures in an Alpine watershed: valuable tools inconceptual model development. Water Resour Res 43, W04404.doi:10.1029/2006WR005349

McDonnell JJ (1990) A rationale for old water discharge throughmacropores in a steep, humid catchment. Water Resour Res 26(11):2821–2832

McKay LD, Driese SG, Smith KH, Vepraskas MJ (2005) Hydro-geology and pedology of saprolite formed from sedimentaryrock, eastern Tennessee, USA. Geoderma 126(1–2):27–45

Mortimer L (2009) In-situ stress field affects on groundwater flowin shallow fractured rock aquifers. PhD Thesis, FlindersUniversity South Australia, Australia

Peck AJ, Hurle DH (1973) Chloride balance of some farmed andforested catchments in Southwestern Australia. Water ResourRes 9(3):648–657

Pinder GF, Jones JF (1969) Determination of the ground-watercomponent of peak discharge from the chemistry of total runoff.Water Resour Res 5(2):438–445

Poulsen DL, Simmons CT, Le Galle La Salle C, Cox JW (2006)Assessing catchment-scale spatial and temporal patterns ofgroundwater and stream salinity. Hydrogeol J 14:1339–1359.doi:10.1007/s10040-006-0065-9

Preiss WVC (1987) The Adelaide Geosyncline-late Proterozoicstratigraphy, sedimentation, palaeontology and tectonics. Bulle-tin (Geological Survey of South Australia) 53:43–72

Scanlon BR, Keese KE, Flint AL, Flint LE, Gaye CB, EdmundsWM, Simmers I (2006) Global synthesis of groundwaterrecharge in semiarid and arid regions. Hydrol Process 20(15):3335–3370. doi:10.1002/hyp.6335

Schilling KE, Li Z, Zhang Y-K (2006) Groundwater-surface waterinteraction in the riparian zone of an incised channel, WalnutCreek, Iowa. J Hydrol 327(1–2):140–150

Schulze-Makuch D, Cherkauer DS (1998) Variations in hydraulicconductivity with scale of measurement during aquifer tests inheterogeneous, porous carbonate rocks. Hydrogeol J 6(2):204–215

Shand P, Haria AH, Neal C, Griffiths KJ, Gooddy DC, Dixon AJ,Hill T, Buckley DK, Cunningham JE (2005) Hydrochemicalheterogeneity in an upland catchment: further characterisationof the spatial, temporal and depth variations in soils, streamsand groundwaters of the Plynlimon forested catchment, Wales.Hydrol Earth Syst Sci 9(6):611–634

Shand P, Darbyshire DPF, Gooddy D, Haria A (2007) 87Sr/86Sr asan indicator of flowpaths and weathering rates in the Plynlimonexperimental catchments, Wales, UK. Chem Geol 236(3–4):247–265

Shapiro AM (2002) Cautions and suggestions for geochemicalsampling in fractured rock. Ground Water Monit Remediat 22(3):151–164. doi:10.1111/j.1745-6592.2002.tb00764.x

Skinner D, Heinson G (2004) A comparison of electrical andelectromagnetic methods for the detection of hydraulic path-ways in a fractured rock aquifer, Clare Valley, South Australia.Hydrogeol J 12(5):576–590. doi:10.1007/s10040-004-0356-y

Sklash MG, Farvolden RN (1979) The role of groundwater in stormrunoff. J Hydrol 43(1–4):45–65

Smettem KRJ, Chittleborough DJ, Richards BG, Leaney FW (1991)The influence of macropores on runoff generation from ahillslope soil with a contrasting textural class. J Hydrol 122(1–4):235–251

1988


http://isohis.iaea.org

http://dx.doi.org/10.1007/s10040-007-0225-6

http://dx.doi.org/10.1029/2006WR005349

http://dx.doi.org/10.1007/s10040-006-0065-9

http://dx.doi.org/10.1002/hyp.6335

http://dx.doi.org/10.1111/j.1745-6592.2002.tb00764.x

http://dx.doi.org/10.1007/s10040-004-0356-y

Socki RA, Karlsson HR, Gibson EK (1992) Extraction techniquefor the determination of oxygen-18 in water using preevacuatedglass vials. Anal Chem 64(7):829–831. doi:10.1021/ac00031a026

Szabo Z, Rice DE, Plummer LN, Busenberg E, Drenkard S,Schlosser P (1996) Age dating of shallow groundwater withchlorofluorcarbons, tritium/helium 3, and flow path analysis,southern New Jersey coastal plain. Water Resour Res 32(4):1023–1038

Van der Hoven SJ, Solomon DK, Moline GR (2005) Natural spatialand temporal variations in groundwater chemistry in fractured,sedimentary rocks: scale and implications for solute transport.Appl Geochem 20(5):861–873

Williamson DR, Stokes RA, Ruprecht JK (1987) Response of inputand output of water and chloride to clearing for agriculture. JHydrol 94(1–2):1–28

Winter TC, Harvey JW, Franke OL, Alley WM (1998) Groundwaterand surfacewater: a single resource. US Geol Surv Circ 1139

1989


http://dx.doi.org/10.1021/ac00031a026

http://dx.doi.org/10.1021/ac00031a026

fractured bedrock and saprolite hydrogeologic controls on groundwater/surface-water interaction: a...

Documents