the importance of unsaturated zone biogeochemical processes in determining groundwater composition,...
TRANSCRIPT
The importance of unsaturated zone biogeochemical processes
in determining groundwater composition, southeastern Australia
Matthew Edwards & John Webb
Abstract Analysis of soil, soil water and groundwater inthe Mount William Creek catchment, southeastern Aus-tralia, shows that Mg2+ and Ca2+ within infiltrating rainfallare rapidly depleted by plant uptake and adsorption onclay minerals. Na+ and K+ may exhibit minor enrichmentat shallow depths but are quickly readsorbed, so thatcation/Cl– ratios typical of groundwater are observed insoil water within the upper 200cm of the soil profile for allspecies. The concentrations of K+ and Ca2+ in soil andgroundwater are more depleted than Na+ and Mg2+ due topreferential uptake by vegetation. Removal of organicmatter results in a continuing, long-term export of allmajor cations from the soil profiles. The processes ofbiogeochemical fractionation within the unsaturated zonerapidly modify the cation/Cl– ratios of infiltrating rainfallto values characteristic of seawater. These mechanismsmay have reached steady state, because groundwaters withseawater ion/Cl– ratios are thousands of years old; theexchange sites on the soil clays are probably saturated, socations supplied in rainfall are exported in organic matterand incorporated into recharge infiltrating into the ground-water. Much of the chemical evolution of groundwatertraditionally attributed to processes within the aquifer iscomplete by the time recharge occurs; this evolutionarymodel may have broad application.
Keywords Unsaturated zone . Cation exchange .Plant uptake . Hydrogeochemistry . Australia
Introduction
The role of soil zone reactions in controlling groundwaterchemistry is often regarded in the hydrogeochemicalliterature as subordinate to reactions occurring in thesaturated zone of aquifer systems. Where aquifers arecomposed of reactive minerals such as carbonates, halidesor unstable silicates, interactions between groundwaterand the host lithologies dominate the evolution ofgroundwater composition (e.g. Garrels and Mackenzie1967; Heathcote 1985; Cardenal et al. 1994; Kimblin1995; Rosen and Jones 1998; Stuyfzand 1999; Toth 1999;Rademacher et al. 2001; Dogramaci and Herczeg 2002;Benedetti et al. 2003).
However, in the semi-confined and unconfined alluvialaquifers of the Murray-Darling Basin, southeastern Aus-tralia, it is generally agreed that groundwater solutes arederived from rainfall and that the dominant evolutionarymechanism affecting groundwater chemistry is concentra-tion by evapotranspiration (e.g. Dyson 1983; Arad andEvans 1987; Macumber 1991; Simpson and Herczeg1994; Herczeg et al. 2001; Cartwright et al. 2004).Despite this, high-resolution rainfall chemistry, datasetsfrom this area (e.g. Hutton and Leslie 1958; Blackburnand McLeod 1983; Simpson and Herczeg 1994; Bormann2004) demonstrate that the ratios of most ions to Cl–
within rainfall differ substantially from the groundwaterratios, indicating that processes apart from evapotranspi-ration are affecting the groundwater chemistry.
Studies of many groundwater systems throughoutAustralia (e.g. Lawrence 1975; Arad and Evans 1987;Macumber 1991; Salama et al. 1993; Acworth andJankowski 2001; Herczeg et al. 2001; Cartwright et al.2004) and in other countries (e.g. Kimblin 1995; Elliot etal. 1999; Guler and Thyne 2004) show that groundwatermajor cation/Cl– ratios typically decline rapidly as thegroundwater salinity increases, from values approximatingthose in local rainfall towards values more characteristicof seawater in many cases. To explain these trends,previous workers have cited mechanisms including min-eral weathering (e.g. Jankowski and Acworth 1993;Cartwright et al. 2004), cation exchange (e.g. Kimblin1995; Toth 1999; Acworth and Jankowski 2001; Bennettset al. 2007), dissolution/precipitation reactions (e.g.Stuyfzand 1999; Herczeg et al. 2001; Guler and Thyne2004), the formation of secondary clay minerals (e.g.
Received: 5 April 2008 /Accepted: 23 February 2009Published online: 27 March 2009
© Springer-Verlag 2009
M. Edwards : J. WebbDepartment of Environmental Geoscience,La Trobe University,Kingsbury Drive, Melbourne, Victoria 3086, Australia
M. Edwards ())ENSR Australia,6/417 St Kilda Road, Melbourne, Victoria 3004, Australiae-mail: [email protected].: +61-3-86992199Fax: +61-3-86992122
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
Blake 1989; Salama et al. 1993) and mixing with connateseawater (e.g. Lawrence 1975; Elliot et al. 1999), with thestated or implied assumption that these processes occurbelow the water table. Until recently, only a few studiesdescribed how reactions within the unsaturated zone couldaffect groundwater major element composition (e.g.Spears and Reeves 1975; Drever and Smith 1978; Mossand Edmunds 1992).
However, it is recognised that many soils are a hot spotof geochemical, particularly organic, activity due to thehigh microbial, fungal and root biomass present (Choroveret al. 2007). Microbial metabolic activity concentrated inthe uppermost few millimeters of the soil strongly affectschemical gradients of nutrient species (Garcia-Pichel et al.2003), and detailed studies on soil water composition(White et al. 2002, 2006) have demonstrated the influenceof biological processes on the concentrations of somespecies. This suggests that soil zone processes caninfluence soil water and therefore groundwater composi-tion to a degree that has often been neglected.
This study assesses the evolution of groundwaterchemistry in a small, partially cleared catchment inVictoria, southeastern Australia, through a detailed studyof soil, soil water and groundwater chemistry. It presentsan alternative explanation wherein most change ingroundwater composition is attributed to soil zone process-es, including plant nutrient cycling, previously neglected inthe study of Australian groundwater chemistry.
Site description
The Mount William Creek catchment is located in theupper reaches of the Wimmera River catchment in westernVictoria, southeastern Australia (Fig. 1), and lies on thesouthern margin of the Murray-Darling Basin, whichoccupies a large part of inland Australia. The study areahas a temperate climate with mean annual rainfall of595.5 mm at Moyston, in the south, and 533.5 mm atStawell, in the northeast of the catchment. Pan evaporationexceeds rainfall in all but the winter months, with a meanannual total of 1501.1 mm at Stawell (Bureau ofMeteorology 2003).
The topography is dominated by the Mount Williamand Mount Difficult Ranges of the Grampians-GariwerdNational Park, which form high mountain escarpments inthe west of the study area, and the relatively subdued,lower elevation Ararat Hills and Black Range in the east.Much of the catchment has been cleared for drylandagriculture (largely sheep grazing), excluding somesignificant tracts of native forest vegetation, particularlyin the Grampians-Gariwerd National Park, which extendsalong the entire western boundary.
The area is geologically complex, with a Palaeozoicbasement consisting of Cambrian greenstones and turbi-ditic metasediments, Silurian schists and sandstones and anumber of Devonian-age granitic intrusives (Cayley andTaylor 2001). The basement is overlain in the MountWilliam Creek valley by Cainozoic fluvial sediments and
colluvium, sub-divided into a basal Tertiary sand andgravel unit overlain by Quaternary sandy clay. Thereforethree aquifers are present: Palaeozoic basement, semi-confined Tertiary sediments and Quaternary alluvium.
Soil profiles analysed throughout the catchment(Fig. 1) fall between two distinct geological and physio-graphic end-members, represented by sites 27 and 5, bothof which were selected for detailed sampling. Site 27(SP27) is located adjacent to the Grampians-GariwerdNational Park (Fig. 1), in Quaternary transported silty-claysbeneath natural Eucalypt forest. The site lies in a localvalley floor at an elevation of ~190 m relative to AustralianHeight Datum (AHD) (equivalent to mean sea level), andslopes gently to the north and east. The soil composition isdominated mineralogically by kaolinite and quartz withminor anorthite, and consists of medium-grey clay gradingdownwards into light brown silty-sand at 2.0 m depth. Thewater table in the vicinity of SP27 is approximately 10 mbelow the ground surface, such that the surficial Quaternaryalluvium is saturated at this locality.
Site 5 (SP5) lies in cleared farmland on deeplyweathered Silurian schist (Fig. 1). It is located on thecrest of a broad ridge at ~240 m AHD with thesurrounding terrain falling gently to the north and south.The water table lies at a depth of 8.5 m. The soil is a paleyellow silty clay with minor quartz gravel and isdominated mineralogically by kaolinite and quartz withminor illite/muscovite and clinochlore. The farmland atsite 5 is used for sheep grazing; no fertilisers are applied,nor is there any evidence of historical fertiliser usage.
Site hydrogeology
The Palaeozoic basement aquifer incorporates all thePalaeozoic rock units within the catchment, and is largelycomposed of crystalline rocks characterised by fractureporosity. The aquifer outcrops along the western catch-ment boundary as the densely vegetated GrampiansRanges, which consist of metamorphosed sandstones,through the centre of the catchment as a discontinuousridge of metavolcanic greenstone, and extensively in theeast of the catchment as rolling hills of granite and deeplyweathered metamorphic rocks (Fig. 2). Throughout thecentral lowlands and tributary valleys the aquifer isconcealed and confined to semi-confined beneath theCainozoic sedimentary cover, and has a low hydraulicconductivity of the order of 0.8 m/day (Edwards 2006).
The Tertiary Calivil Formation comprises unconsoli-dated sands and gravels, consisting predominantly ofquartz, that lie in palaeodrainage lines beneath the presentMount William Creek floodplain, where deposits canreach 40 m in thickness and up to 6,000 m in width(Fig. 2). It is the dominant aquifer in the catchmentbecause of the large volume of relatively porous sedimentand is also known as the Tertiary sediments aquifer; it isresponsible for much of the down-basin groundwater flowwithin the catchment (Edwards 2006). The aquifer doesnot outcrop and therefore receives recharge from other
1360
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
aquifers. It is heterogenous and anisotropic, due to amoderately well defined fining upwards sequence andstrong lateral variability; hydraulic conductivities areestimated to range from 1.0 m/day up to 100 m/day.
The Quaternary alluvial aquifer consists mostly ofextensive alluvium, which blankets much of the lowerrelief area along Mount William Creek and tributaryvalleys to depths of up to 25 m (Edwards 2006; Fig. 2).The sediments are clay-rich and unconsolidated, withminor sand stringers evident during drilling, both in thisstudy and previous investigations (e.g. Harrison 1993).This aquifer is characterised by a low hydraulic conduc-tivity of the order of 1.0 m/day.
Potentiometric and groundwater geochemical evidencepresented by Edwards (2006) generally indicates a lowdegree of cross-formational interaction throughout themajority of the study area where groundwater movementis dominated by lateral flow; however, vertical flow issignificant in areas of relatively high recharge.
Recharge occurs dominantly from direct rainfall acces-sions and is highly variable throughout the catchment,ranging from 50–100 mm/year in the Grampians Rangeand associated colluvial slopes to 0.2 mm/year in areas ofwell-developed regolith beneath remnant native vegetation(Edwards and Webb 2006). Transient and steady-state soilwater Cl– mass balance modelling (Edwards 2006) show
Fig. 1 Location of the Mount William Creek catchment study area, showing sample sites
1361
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
that recharge at SP5 and SP27 is of the order of 8 and0.2 mm/year, respectively, consistent with the physiogra-phy of the sites.
Sampling and analytical methods
GroundwaterA total of 79 groundwater bores were sampled during thisstudy (Fig. 1), ranging in depth from 5 to 142 m, with atypical depth of ~20 m. Before collection of groundwatersamples, all bores were purged of at least three bore volumesof water using a Grundfos SQ1–80N electric submersiblepump for bores with ≥100 mm diameter casings, and a PVCbailer for bores with smaller diameter casings. Groundwatersamples were collected in 120-ml polyethylene vials follow-ing the stabilisation and measurement of the physicalparameters EC, pH, Eh and temperature; samples wereimmediately refrigerated for transport back to the laboratory.All groundwater samples were filtered through 0.45 µmcellulose nitrate filters prior to analysis.
Ca2+ , Mg2+ and K+ concentrations were determined onacidified samples containing 10% LiCl, using a GBC 933Plus Atomic Absorption Spectrophotometer (AAS) with an
air-acetylene flame. Samples were diluted with deionisedwater tofit the working ranges for Ca2+ ,Mg2+ andK+ of 0.1–4.0, 0.5–2.0 and 0.5–4.0 mg/L, respectively. Na+ wasdetermined on acidified samples using a Sherwood Model410 flame photometer, after dilution to fit the working rangeof 0–30 mg/L. Cl– and other anion concentrations weredetermined on unacidified samples using a PhenomenexA300 anion peak ion chromatograph with a 1.7 mMNaHCO3/1.8 mM Na2CO3 eluent, 25 mM H2SO4 regeneratesolution, and flow rate of 1.5 ml/min. The samples werediluted to fit the working range for Cl– of 0–250 mg/L.Alkalinity was determined by standard titration to pH 4.5with HCl following the method of Rayment and Higginson(1992). Selected groundwater samples were collected andanalysed for 14C concentration. Samples were collected inacid-washed (HNO3) 1-L glass bottles, rinsed three timeswith sample water and filled so that headspace wasnegligible, in order to minimise atmospheric contamination.Analysis was performed using the ANTARES tandemaccelerator at ANSTO, Sydney (Fink et al. 2004). Theaverage error associated with the percent modern carbon(pMC) values here is 0.3%, and the age limit is 50,000±200 years. Age determinations made on the basis of pMCvalues are uncorrected for the addition of radiogenically dead
Fig. 2 Areal extents of the major aquifer systems, Mount William Creek catchment
1362
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
carbon, due to the paucity of carbonate minerals or preservedorganic material within the aquifers (Edwards 2006).
SoilsA total of 36 soil profiles (Fig. 1) were sampled for aminimum analytical suite of soil moisture, EC1:5, pH1:5 andCl–. Soil samples were collected using a Gemco solid flightauger-drilling rig in the winter of 2005. Sub-samples weretaken by hand at the top, middle and bottom of every 50-cm flight length during drilling, and combined to form asingle sample that represents the entire 50-cm interval. Thistechnique ensured that representative samples were collect-ed and that cross contamination was kept to a minimum.Samples were immediately stored in zip-lock plastic bagsand refrigerated for transport back to the laboratory.
Soil moisture contents were determined gravimetricallyby weighing sub-samples before and after oven drying at105°C. The EC1:5 and pH1:5 were determined by adding25 ml of deionised water to 5 g of air-dried and sieved soil(<2 mm), and agitating the resultant solution for 1 h. Aftersettling, or centrifuging where necessary, the supernatantwas measured for EC1:5 and pH1:5 using a WTW LF330conductivity meter and probe and a Metrohm 704 pHmeter and probe, respectively. The supernatant solutionsfrom a sub-set of 20 soil samples were analysed for Cl–
using a Phenomenex A300 anion peak ion chromatographto give a correlation between Cl– and EC1:5 (r2=0.98),thus providing Cl–1:5 data for all samples.
The exchangeable cation content of the soils and the soilwater composition (i.e. soluble cations) were determined bystandard techniques (methods 15A1 and 15A2; Raymentand Higginson 1992) on samples from soil profiles SP5 andSP27 only. Initially both soluble and exchangeable cationswere extracted from the soil sample by weighing 5 grams ofair dried and sieved soil into a 125-ml polyethylene vial,adding 100 ml of 1M NH4Cl and mechanically shaking for1 h. The soil extract was then passed through a Whatman 52filter paper, collected in a volumetric flask and analysed forK+ , Ca2+ and Mg2+ using a GBC 933 Plus AtomicAbsorption Spectrophotometer (AAS) with an air-acetyleneflame, and for Na+ using a Sherwood Model 410 flamephotometer. Standards were made up using 1M NH4Cl (sothat matrix compositions were uniform) and samples werediluted to fit the working ranges previously specified forgroundwater analyses.
In the second technique, soluble cations were firstremoved by weighing 5 grams of air dried and sieved soilinto a 30-ml polyethylene vial and adding 25 ml of 60%
aqueous ethanol. This extract was then mechanicallyshaken for 1 h, centrifuged and the supernatant solutiondiscarded. A second 25 ml aliquot of 60% aqueousethanol was added to the soil and the process wasrepeated. A third rinse was carried out using 25 ml of20% aqueous glycerol. Ethanol and glycerol are usedrather than distilled water to avoid modifying thedistribution between soluble and exchangeable species inthe soil. Samples were weighed before and after thisprocedure to account for any entrained solvents, trans-ferred to 125 ml polyethylene vials and exchangeablecations then extracted by adding 100 ml of 1M NH4Cl andmechanically shaking for 1 h. The extract was passedthrough a Whatman 52 filter paper, collected in avolumetric flask and analysed as in the previous proce-dure. Thus the detection limits for Na+, Ca2+, Mg2+ andK+ as both exchangeable and soluble cations in the extractsolutions are 0.1, 0.1, 0.15 and 0.5 mg/L, respectively.The second procedure determines only exchangeablecations; the soluble cation content is obtained by thedifference between the results of the two methods. Extractconcentrations were converted mathematically to 1:5equivalents for ease of comparison to soil Cl– data.
The bulk chemical composition of soil samples fromprofiles SP5 and SP27 were determined by X-rayflorescence (XRF) using 0.75±0.00001 g of sample (ovendried at 40°C until constant mass) mixed with 6.75±0.00001 g of 66:34 lithium tetraborate:lithium metaborateflux (oven-dried at 400°C for 2 h) and fused at 1,050°Cinto a glass button using an automated fusion machine.Glass buttons were analysed using a Siemens SRS 303 ASWavelength Dispersive XRF spectrometer. Elements wereoriginally reported as wt% oxides. The mineralogicalcomposition of the soil samples was determined byqualitative X-ray Diffractometer (XRD) analysis, per-formed from 4–70° for 90 min, using a Siemens D5000X-ray Diffractometer on dried sub-samples after milling to~60 μm. Quantitative XRD analysis was undertaken usingSIROQUANT (Taylor and Hinczak 2001).
Local rainfall composition
Rainfall was not analysed in the present study; however,Victorian rainfall has been studied in detail by Hutton andLeslie (1958) and Bormann (2004), with both studiesincluding sample locations (Beaufort and Ararat) near theMount William Creek catchment (Fig. 1). Comparison ofthe two average weighted datasets (Table 1) reveals
Table 1 Rainfall chemistry for locations proximal to the Mount William Creek catchment study area
Locality Na+
(mg/L)Ca2+
(mg/L)Mg2+
(mg/L)K+
(mg/L)Si(mg/L)
Cl–
(mg/L)Br–
(mg/L)SO4
2–
(mg/L)HCO3
–
(mg/L)
Beauforta 0.9 0.8 0.2 0.1 n/a 1.1 n/a n/a 3.1Araratb 1.9 1.2 0.5 0.9 0.05 4.4 0.01 1.9 n/a
a Hutton and Leslie (1958)b Bormann (2004)n/a denotes a parameter not analysed
1363
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
Tab
le2
Major
elem
entdata
forgrou
ndwatersfrom
theMou
ntWilliam
Creek
catchm
ent,southeastern
Australia
Bore
IDAqu
ifer
screened
Eastin
g(m
)Northing
(m)
Elevatio
n(m
AHD)
Na+
(mg/L)
Ca2
+
(mg/L)
Mg2
+
(mg/L)
K+
(mg/L)
Si
(mg/L)
Cl–
(mg/L)
Br–
(mg/L)
SO42–
(mg/L)
HCO3–
(mg/L)
NO3–
(mg/L)
pMC
(%)
14Cage
(years)
1434
67Quaternary
647,77
95,88
1,88
521
2.00
1,79
413
028
27
20.0
3,10
010
.536
451
7bd
lnm
nm14
3468
Quaternary
647,77
95,88
1,88
521
2.00
2,07
513
531
76
21.3
3,68
411.7
427
590
bdl
nmnm
1434
71Quaternary
649,48
05,87
4,96
123
4.00
2,13
884
251
423
.73,46
311.3
388
474
bdl
nmnm
1434
75Quaternary
650,55
05,87
5,54
823
9.00
2,37
5117
363
1519
.64,40
013
.644
939
3bd
lnm
nm14
3476
Quaternary
652,19
85,88
0,45
722
6.50
1,62
556
303
935
.43,20
17.1
266
0bd
lnm
nm14
3477
Quaternary
652,22
45,88
1,35
122
5.50
2,10
054
397
637
.14,01
66.9
272
10bd
lnm
nm14
3478
Quaternary
647,73
45,87
6,80
422
8.00
695
6816
05
12.8
1,55
62.8
5277
bdl
nmnm
2544
Quaternary
648,51
25,89
8,35
019
0.00
2,85
015
540
312
21.5
5,54
719
.459
517
bdl
nmnm
5314
Quaternary
654,52
75,88
1,08
024
5.00
2,56
373
348
2227
.24,41
614
.965
735
9bd
lnm
nm48
884
Quaternary
644,58
15,89
4,40
319
2.20
1,21
010
621
08
8.2
2,51
58.3
414
147
3.8
nmnm
4888
5Quaternary
640,65
15,90
0,76
218
3.44
1,39
514
332
29
19.2
2,98
59.1
419
13bd
lnm
nm48
887
Quaternary
640,52
35,89
8,75
620
3.85
2,87
018
944
212
33.6
5,77
212
.655
926
325
.2nm
nm48
888
Quaternary
641,011
5,89
7,87
519
9.89
803
2487
432
.41,10
22.8
142
486
bdl
nmnm
4888
9Quaternary
640,55
65,90
0,30
020
3.09
1,40
014
732
436
21.7
3,07
37.7
280
99bd
lnm
nm48
890
Quaternary
640,53
25,89
9,27
418
5.28
2,15
065
344
3424
.93,96
26.4
371
329
bdl
nmnm
4889
1Quaternary
640,66
65,89
8,31
818
9.59
1,36
844
159
1640
.32,14
45.2
367
408
bdl
nmnm
5125
0Quaternary
638,90
05,90
0,25
018
1.90
1,91
5111
369
4713
.83,85
98.3
471
32bd
lnm
nm1106
74Quaternary
642,65
05,88
9,60
019
7.07
112
314
47.7
113
0.3
1413
35.6
nmnm
1106
77Quaternary
650,15
05,88
2,20
021
7.88
324
311
326
.033
60.8
7522
8bd
lnm
nm1106
78Quaternary
649,45
05,88
2,00
021
6.56
2,98
021
641
663
22.4
5,62
915
.573
713
010
.6nm
nm1106
80Quaternary
648,90
05,88
2,55
021
3.95
1,94
521
341
629
20.1
4,21
211.1
472
337
bdl
nmnm
1106
81Quaternary
642,55
05,89
3,55
019
4.00
485
4387
97.4
948
2.1
7059
2.4
nmnm
1138
87Quaternary
645,00
05,89
7,48
019
9.00
808
414
88
20.2
1,64
83.0
7436
2.0
nmnm
5125
2Quaternary
636,80
55,89
7,23
819
4.00
651
72
11.1
102
0.6
732
bdl
nmnm
5125
4Quaternary
635,80
05,89
7,20
022
2.60
618
6718
418
16.7
1,47
22.5
6960
2.6
nmnm
5309
Quaternary
654,10
85,88
1,21
023
3.00
3,17
570
478
524
.55,93
117
.870
490
bdl
nmnm
5310
Quaternary
654,09
25,88
1,05
823
7.50
3,46
397
514
1120
.86,48
527
.182
014
7bd
lnm
nm53
11Quaternary
654,09
25,88
0,88
824
3.00
1,07
039
152
1511.4
1,91
46.0
231
181
bdl
nmnm
1434
74Quaternary
652,19
85,88
0,45
722
6.50
2,96
316
446
214
17.7
5,45
517
.856
253
9bd
lnm
nm1121
99Tertiary
646,48
25,87
8,20
422
6.21
558
692
412
.21,15
42.1
162
0.6
67.52
3,15
514
3479
Tertiary
647,45
45,87
7,05
422
6.00
373
5277
925
.084
91.5
2483
bdl
nmnm
1434
73Tertiary
650,55
05,87
5,54
823
9.00
2,18
862
211
433
.13,119
9.7
354
1,02
4bd
lnm
nm48
883
Tertiary
644,58
15,89
4,40
319
2.20
683
126
158
87.1
1,64
74.2
134
112
1.9
32.07
9,14
048
886
Tertiary
640,53
65,89
9,76
418
4.55
515
642
424
.277
71.2
111
118
bdl
12.76
16,540
4890
3Tertiary
645,20
75,88
7,24
120
4.77
409
964
211.6
780
1.4
319
4.1
nmnm
4890
4Tertiary
642,97
15,88
9,07
420
3.85
226
844
46.7
449
1.0
4636
bdl
nmnm
6787
0Tertiary
646,71
65,88
1,98
8211.80
338
7854
918
.976
02.1
6115
50.9
37.67
7,84
01106
76Tertiary
648,75
05,87
2,45
023
6.07
458
4773
513
.382
61.6
4067
bdl
31.40
9,31
01116
93Tertiary
648,42
65,86
5,84
427
0.78
8612
323
11.8
244
0.4
617
0.5
nmnm
1122
00Tertiary
646,48
25,87
8,20
422
6.21
936
162
8.2
190
0.4
717
0.8
nmnm
1130
08Tertiary
644,76
75,88
6,59
820
4.43
8545
145
14.8
156
0.4
117
6bd
l17
.42
14,040
1138
85Tertiary
636,80
05,89
7,14
419
3.63
2,01
013
336
248
7.4
3,63
69.0
440
648
bdl
48.65
5,79
01141
74Tertiary
640,45
65,89
1,80
219
9.61
134
1022
27.5
276
0.6
1135
1.1
60.64
4,02
01141
75Tertiary
640,45
65,89
1,80
219
9.61
307
336
39.3
532
1.5
398
bdl
61.29
3,93
01141
73Tertiary
642,63
45,89
3,44
319
3.98
710
55119
77.0
1,35
44.5
152
118
bdl
68.97
2,98
51138
86Tertiary
645,011
5,89
7,37
019
9.00
925
6215
38
7.4
1,80
55.7
187
56bd
lnm
nm67
750
Tertiary
651,64
35,90
0,01
519
2.50
4,31
326
172
819
17.7
8,02
027
.61,02
520
8bd
lnm
nm48
905
Tertiary
645,17
45,88
6,22
019
9.89
249
938
210
.148
71.0
2218
2.1
75.23
2,28
548
906
Tertiary
642,55
05,88
8,30
020
3.09
212
bdl
324
11.2
399
0.8
251
bdl
nmnm
1364
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
5125
5Basem
ent
634,20
05,89
7,20
027
7.70
401
bdl
bdl
4.8
520.2
48
0.5
nmnm
8218
1Basem
ent
646,12
55,86
6,90
728
0.00
9bd
l1
20.9
14bd
l3
1bd
lnm
nm1121
97Basem
ent
648,75
05,87
2,45
023
6.05
345
7466
1417
.783
71.6
3371
1.3
27.40
1,04
001121
98Basem
ent
646,48
25,87
8,20
422
6.08
7234
155
15.6
131
0.4
214
70.6
nmnm
1122
01Basem
ent
636,79
85,89
7,14
419
3.57
170
724
38.2
353
0.8
1425
bdl
nmnm
1209
69Basem
ent
659,40
05,87
0,90
027
0.00
3,16
021
842
030
21.2
5,80
817
.371
424
8bd
lnm
nm12
0970
Basem
ent
659,50
05,87
0,95
026
1.50
1,46
084
166
720
.12,45
46.2
390
240
8.3
nmnm
1209
71Basem
ent
659,70
05,87
1,00
026
1.50
863
5396
621
.71,44
63.8
183
211
17.4
nmnm
1209
73Basem
ent
657,96
85,86
9,15
627
1.00
2,51
098
312
117
22.1
4,30
611.0
375
416
bdl
nmnm
1209
75Basem
ent
656,116
5,87
4,49
026
2.00
3,00
560
413
7713
.25,19
912
.447
575
4bd
lnm
nm12
0976
Basem
ent
659,40
05,87
3,30
026
8.00
321
818
523
.128
01.0
6825
747
.1nm
nm12
0980
Basem
ent
664,48
05,87
8,80
033
0.00
690
8510
38
16.3
1261
3.1
3840
22.0
nmnm
1209
81Basem
ent
664,48
05,87
8,80
033
0.00
793
3587
1127
.21,311
3.1
7333
6bd
lnm
nm12
0982
Basem
ent
664,60
05,87
9,28
035
5.00
380
839
1523
.662
21.5
3016
511.9
nmnm
1209
84Basem
ent
664,10
05,87
3,10
034
0.00
2,46
014
334
542
18.6
4,29
210
.945
190
910
.6nm
nm14
3469
Basem
ent
644,44
95,87
6,79
327
5.00
462
44
7.2
760.2
33
1.2
nmnm
2543
Basem
ent
648,51
25,89
8,35
019
0.00
2,71
315
839
014
12.0
5,12
115
.759
829
bdl
nmnm
5224
Basem
ent
658,96
75,88
5,90
830
8.00
318
924
148
.545
21.5
8172
1.8
nmnm
5308
Basem
ent
654,10
85,88
1,29
423
8.50
3,55
015
852
533
17.1
6,50
722
.179
030
5bd
lnm
nm53
13Basem
ent
654,52
75,88
1,25
223
5.00
3,10
017
051
734
21.7
5,80
514
.874
056
9bd
lnm
nm12
0978
Basem
ent
660,99
75,87
7,88
928
0.00
294
6816
010
15.3
675
2.3
7153
65.7
nmnm
1209
79Basem
ent
660,99
75,87
7,88
928
0.00
278
1969
621
.035
20.8
4442
514
.6nm
nmBurrum
Basem
ent
664,98
95,86
4,52
633
1.00
2,22
558
444
5342
.44,52
911.4
382
59bd
lnm
nm48
882
Basem
ent
642,89
85,90
3,05
120
3.00
823
2289
1217
.41,35
84.4
179
61bd
lnm
nmMarriot1
Basem
ent
658,48
75,88
6,22
937
3.00
432
11
35.5
280.2
1634
16.8
nmnm
5312
Basem
ent
654,52
75,88
1,25
223
5.00
4,70
046
662
1638
.58,09
425
.61,08
573
bdl
nmnm
6784
9Basem
ent
642,33
45,88
0,31
631
3.89
151
25
6.0
20bd
l4
12bd
lnm
nm67
749
Basem
ent
650,40
05,90
0,30
019
1.30
4,28
824
474
423
10.9
8,02
630
.01,10
017
4bd
lnm
nm52
23Basem
ent
658,85
55,88
6,06
831
9.00
415
324
938
.561
61.7
6169
8.6
nmnm
bdlbelowdetectionlim
it(groun
dwater
detectionlim
itforanions
is0.1mg/L,and
is0.1,
0.1,
0.15
and0.5mg/LforNa+
,Ca2
+,M
g2+andK+,respectively);n
mdeno
tesaparameter
not
measured
1365
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
variability imparted by differences in the climatic con-ditions between the two sampling periods. In the 1958study, which was based on monthly analyses over an18-month period, there was above-average rainfall(770.9 mm/year), and the lower rainfall salinity (Table 1)can be attributed to increased rainout effects anddecreased dust activity. By contrast, the 2004 study, whichwas based on monthly analyses over a 12-month period,was conducted during a sustained drought (449.8 mm/year), where reduced rainout and a greater dust entrain-ment produced the higher salinities observed (Table 1).The apparatus used in both studies were standardmeteorological rain gauges that collect wet depositionand any dry deposition falling directly onto them. Thesetwo datasets are likely to encompass most of thecompositional variability in modern rainfall in the area,and rainfall composition will be expressed as a range,where possible, throughout the following discussion.
Groundwater composition
Groundwaters of the Mount William Creek catchmentrange in salinity from very fresh (TDS <100 mg/L) to
relatively saline (TDS ~20,000 mg/L; Table 2). Mostgroundwaters analysed (94%) are dominated by Na+ andCl–, with a small proportion (6%) dominated by Na+ andHCO3
–. The dominance of Na+ and Cl– in the groundwatersolute load increases with increasing salinity, from ~40%by weight in the lowest salinity samples (TDS < 500 mg/L)to ~85% by weight at higher salinities (TDS > 8,000 mg/L). This trend, together with the commensurate decline inthe relative abundance of other cations and bicarbonate,corresponds to a progression from a composition with ion/Cl– ratios reflecting local rainfall to one with ratios moreclosely resembling those of dilute seawater, particularly forCa2+ and Na+ (Fig. 3), and is similar to trends observedelsewhere in southeastern Australia (e.g. Arad and Evans1987; Tickell and Humphrys 1987; Macumber 1991;Cartwright et al. 2004) and throughout the world (e.g.Elliot et al. 1999; Guler and Thyne 2004). This progressionfrom rainwater to seawater ion/Cl– ratios occurs at lowsalinities (TDS < 2,000 mg/L; Fig. 3), such that the vastmajority of groundwaters, which are moderately to verysaline, have approximately the same relative distribution ofdissolved species as seawater. Geochemical calculationsusing PHREEQC (Parkhurst and Appelo 1999) indicatethat groundwaters are typically undersaturated with respect
Fig. 3 Ion/Cl– vs. TDS plots for groundwaters from all aquifers of the Mount William Creek catchment, southeastern Australia
1366
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
to calcite and other common minerals including dolomiteand gypsum.
Soil and soil water chemistry
Ca2+The soluble calcium content of the soil (i.e. within soilwater) is extremely low (below detection limit) at both sites
(Figs. 4a, e and 5), indicating a very rapid attenuation ofthe calcium supplied in rainfall, and this loss must occur inthe upper 50 cm of the profile in both cases. It is due partlyto adsorption onto the exchange sites of clay minerals, asdemonstrated by the much higher concentrations of Ca2+ asexchangeable cations in the soil (Fig. 4a, e). In bothprofiles, the highest exchangeable Ca2+ values occur in theupper 50 cm, complementing the observed loss in solubleCa2+ (Fig. 4); the concentration of exchangeable Ca2+ is
Fig. 4 Soluble and exchangeable cation/Cl ratios with depth for soil profiles SP5 and SP27
1367
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
much higher in SP5 than SP27 (Table 3), consistent withthe higher clay mineral (kaolinite) content of this profile(Table 4). Ca2+ is typical of Group II cations in that it isweakly adsorbed onto clay minerals as outer-spherecomplexes (Sparks 2005), making it susceptible to ex-change for other cations (as discussed in the following).There is some evidence that adsorption of Ca2+ causesdesorption of H+ in the soil profiles, as pH1:5 typicallydeclines with depth (Table 4).
In the upper part of the profile at both sites, Ca2+
storage as exchangeable cations is considerably lower thanstorage of exchangeable Na+ or Mg2+, particularly in SP27(Table 3), even though rainfall delivers Ca2+ and Na+ invery similar quantities, and much more Ca2+ than Mg2+
(Table 1). This indicates that an additional sink for Ca2+
exists within the profile.Calcite precipitation does not provide this sink,
because soil waters in which the Ca2+ loss is observedhave low salinity and are undersaturated with respect tocalcite. No calcite was encountered during sampling, noridentified by XRD, and there is very low total Ca2+ in theprofiles (Table 3). Testing of soils for calcite usinghydrochloric acid yielded negative results in all cases.
Nutrient uptake by vegetation is the most likelymechanism responsible for the Ca2+ removal from theVictorian profiles; this process has been identified assignificant in determining groundwater composition else-where (e.g. Hudson and Golding 1997; Moulton et al.2000; Benedetti et al. 2003), Ca2+ is a macro-nutrient formost plant species: plants take up relatively largequantities of this element directly from the soil solutionand/or from the cation exchange sites of clay minerals(Sutcliffe 1962). In contrast, plant uptake of Na+ is muchless. This mechanism can also explain why Ca2+ isrelatively more depleted than Na+ within groundwatersin the area (Table 5).
The uptake of nutrients by vegetation is also supportedby the depletion of the groundwater HCO3
–/Cl– ratios inthe more saline groundwaters in the study area (Fig. 3e).This most likely reflects the addition of saline rechargefrom the unsaturated zone that has been depleted inHCO3
– by plant uptake; bicarbonate is a macronutrient for
all plants (Sutcliffe 1962), and no other carbon sinks suchas calcite precipitation, exist in the soil profiles (seeprevious discussion). It is unlikely that significant plantuptake of bicarbonate occurs directly from the saturatedzone due to the water table depths in the areas investigated(>8 m).
K+The concentration of K+ in soil water is only detectable inthe upper 50 cm of the profiles at both sites (Table 3;Fig. 4). The soluble K+/Cl– ratio at the top of the profileapproximates that of rainfall in SP27 and exceeds that ofrainfall in SP5, indicating an additional source of K+ inSP5. This may represent the input of K+ from thebreakdown of organic material and/or desorption of K+
from clays by cation exchange with Ca2+ and/or Mg2+.Because soluble Ca2+ is strongly depleted in the shallowsubsurface (<50 cm; Figs. 4 and 5), the latter explanationmay be more likely; however, both processes are probablyoccurring to some extent. There is also a considerableaccumulated storage of K+ on cation exchange sites in theupper profile (Fig. 4; Table 3). The elevated levels ofsoluble and exchangeable K+ are not derived from theweathering of primary K+ -bearing minerals, because apartfrom some illite, none are present in the upper profiles.
K+ is rapidly removed from the soil water as it migratesdownwards, and at depths of less than 100 cm the levelsfall below the limits of detection, the soluble K+/Cl– ratiosapproximating those of average groundwater (Fig. 4b, f).As with Ca2+ , the removal of K+ is partly the result ofadsorption to the cation exchange sites of clay minerals(Table 3), as shown by the elevated ratio of exchangeableK+ to Cl– compared to the soluble and rainfall K+/Cl–
ratios, particularly in the upper profile (Fig. 4b, f). Theelevated exchangeable K+/Cl– ratio may also be due torecycling from vegetation, which has been shown to causea surface enrichment of essential elements like Ca2+ andK+ that cannot be explained by abiotic processes (Jobbagyand Jackson 2004).
The amount of K+ stored on clay minerals in the upperprofiles is relatively low compared to Na+ and Mg2+
Fig. 5 Total cation concentrations in soil water (soluble cations). Where solute concentrations are below detectable limits, they are plottedat the detection limit
1368
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
(Table 3). This partly reflects the smaller amounts of K+
delivered to the soil in rainfall (Table 1), but is probablyalso due to plant uptake, as discussed for Ca2+ ; K+ is alsoa macronutrient for most plant species (Sutcliffe 1962).This is consistent with the very strong depletion of bothCa2+ and K+ relative to other major cations in localgroundwater (Table 5). Conversion of kaolinite to illitewill remove K+ , but is unlikely to be occurring in thepresent case due to the low K+/H+ ratios of the soil water(Table 3). Some illite is present in the profile of SP5(Table 4), but is most probably derived from the basementschist.
Na+Groundwater Na+ concentrations are not depleted relativeto rainfall to the same extent as the other cations. Theaverage groundwater Na+/Cl– mass ratio is 0.57, whilethat of local rainfall ranges from 0.44 (Bormann 2004) to0.87 (Hutton and Leslie 1958). This lower degree ofdepletion is shown by the fact that the soluble Na+/Cl–
ratios approximate or exceed those of rainfall at mostsampling depths (Fig. 4c, g). However, there is a minoroverall declining trend in soluble Na+/Cl– ratios withdepth at both sites, so that the soil water Na+/Cl– ratiostowards the base of the profiles are indistinguishable fromthe groundwater ratio (Fig. 4c, g). This minor depletion ofNa+ relative to Cl– is controlled by cation adsorption toexchange sites on clay minerals, as clearly demonstratedby the elevated exchangeable Na+ contents (Fig. 4c, g;Table 3).
In SP27, and to a lesser extent SP5, some of thesoluble Na+/Cl– ratios exceed those of local rainfall,probably due to desorption of Na+ from the exchangeablesites on clay mineral surfaces. Na+-bearing minerals werenot detected by XRD in the soil profiles at either location,and indeed, the amount of total Na+ present in bothprofiles (as determined by XRF) is approximately equal tothe sum of soluble and exchangeable Na+ (Table 3). Thedesorption of Na+ is most likely controlled by cationexchange for Ca2+ and/or Mg2+, as previously discussedfor K+.
Plant uptake appears to play no significant role in thedistribution of Na+ in the soils of the study area, in contrastto Ca2+ and K+. This is probably because of thecomparatively low uptake of Na+ by plants (Sutcliffe1962) and the lower proportion of this species in thebiomass; White et al. (2006) found that grassland biomassin California has an order of magnitude less Na+ than Ca2+
and K+ in most cases.
Mg2+Groundwaters and soil waters are depleted in Mg2+
relative to local rainfall; the soil waters acquire Mg2+/Cl– ratios similar to those of groundwater within the top50 cm of the soil profile (Fig. 4d, h). The very low Mg2+
content of soil water (below detectable limits in somecases) is accompanied by very high storage of exchange-T
able
3Abu
ndance
ofsoluble,
exchangeable
andtotalbu
lksoilcatio
ns(w
eigh
tpercentdrysoil)
Cl
KCa
Na
Mg
Soil
profi
leSam
ple
interval
(cm)
Water
content
(wt%
)
Solub
lewt%
(x10
–3)
Solub
lewt%
(x10
–3)
Exchang
eable
wt%
(x10
–3)
Total
wt%
a
(x10
–3)
Solub
lewt%
(x10
–3)
Exchang
eable
wt%
(x10
–3)
Total
wt%
(x10
–3)
Solub
lewt%
(x10
–3)
Exchang
eable
wt%
(x10
–3)
Total
wt%
(x10
–3)
Solub
lewt%
(x10
–3)
Exchang
eable
wt%
(x10
–3)
Total
wt%
(x10
–3)
50–
5023
.910
.83.2
17.7
493.9
bdl
43.0
71.5
8.6
28.5
44.5
3.0
80.7
265.3
550
–100
20.9
28.1
bdl
16.6
585.2
bdl
28.6
50.1
11.7
52.8
48.2
1.1
113.8
379.
95
100–
150
11.7
41.9
bdl
7.4
639.1
bdl
16.2
35.8
24.9
60.4
48.2
bdl
84.0
319.6
515
0–20
011.1
26.1
bdl
3.6
1153
.7bd
l7.7
28.6
13.1
31.4
29.9
bdl
17.4
313.6
270–
5012
.713
.62.0
5.2
240.7
bdl
2.3
28.6
19.4
41.7
92.8
11.4
55.4
78.4
2750
–100
13.2
69.4
bdl
5.5
195.1
bdl
2.3
28.6
47.9
59.6
141.0
8.9
61.6
66.3
2710
0–15
015
.112
0.1
bdl
6.8
182.6
bdl
2.3
28.6
70.2
83.7
137.3
7.3
90.7
162.8
2715
0–20
06.0
47.5
bdl
3.2
199.2
bdl
2.9
21.5
30.5
29.2
26.0
1.0
30.9
bdl
bdlbelow
detectionlim
itaBulkchem
ical
compo
sitio
nas
determ
ined
byXRFanalysis
1369
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
able Mg2+ throughout the profiles (Table 3), so the loss ofsoluble Mg2+ is at least partially the result of adsorption toclay mineral surfaces in the shallow soil. In SP27, the totalMg2+ content of the soil approximates the sum of solubleplus exchangeable Mg2+; there is no other significantsource of Mg2+ within the soil (Table 3). The very largestore of Mg2+ on exchange sites has built up despite therelatively low amounts of this ion in rainfall. This reflectsthe preference for divalent cations by clays in these soilprofiles.
There is no direct evidence of the influence of plantuptake on the distribution of Mg2+ in the soils of the studyarea, but studies elsewhere have demonstrated thatbiological cycling can remove Mg2+ from the systemalong with Ca2+ and K+ (White et al. 2002), despite thecomparatively low uptake of Mg2+ by plants (Sutcliffe1962). Considerations of the long-term effect of unsatu-rated zone processes on groundwater composition in thestudy area (see the following) also suggest that uptake ofMg2+ by vegetation is occurring.
Discussion
Is the removal of ions in the unsaturated zonea steady-state situation?The results of this study clearly show that ions supplied inrainfall are being removed in the unsaturated zone,particularly Ca2+ , Mg2+ and K+ , both by cation exchangeon clays and uptake by plants. As previously discussed,these processes cause the higher salinity groundwaters(>2,000 mg/L TDS) in the study area to have ion/Cl–
ratios typical of seawater; groundwaters sampled have 14C
ages ranging from ~2,500 up to ~16,500 years (Table 2).This implies that the cation-removal processes in theunsaturated zone have been operating at a more or lessconstant rate for thousands of years. Despite 14C agedeterminations being derived from groundwater in the semi-confined Calivil Formation, these waters have the samerelative ion concentrations as those waters in the overlyingalluvium, due in large part to the fact that recharge occurs viathe overlying unit.
The exchangeable cation content of the soils in thestudy area is less than might be expected from this timeframe. Assuming a bulk soil density of 1.6 tons/m3
(typical of soils in this part of Australia; Dyson andJenkin 1981), 1 m2 of the upper 50 cm of the soil profilesat both sites contains ~800 kg of soil with 35–50 moles ofexchangeable cations (K+ , Ca2+ , Na+ and Mg2+; fromTable 3); assuming all exchangeable cations were derivedfrom infiltrating pore waters, this represents 500–900 years’ supply of cations in rainfall (from Table 1).The exchangeable cation content of the entire 200 cmthickness of both profiles is 160–190 moles/m2, represent-ing 2,000–3,500 years’ supply of cations in rainfall. Sincethe unsaturated zone processes responsible for the seawa-ter-type ion/Cl– ratios in the higher salinity groundwatersappear to have been operating for many thousands ofyears, this indicates that the concentration of exchangeablebase cations in the soil profiles may have reached a steadystate, implying that continuing net removal of base cationsby cation exchange is no longer occurring. The totalexchangeable cation content of the kaolinite in the upper50 cm of soil profile SP5 (0.08 mol kg–1; from Tables 3and 4) lies within the range of experimental cationexchange capacities for kaolinites in southeastern Aus-
Table 4 Soil profile pH and clay mineralogy
Soil profile Sample interval (cm) pH1:5 % quartz % kaolinite % illite/muscovite
5 0–50 6.46 23.3 76.7 -5 50–100 6.64 14.8 85.2 -5 100–150 5.56 19.8 55.5 24.65 150–200 5.92 35.1 24.2 40.627 0–50 5.72 75.6 8 16.427 50–100 5.16 87.9 4.8 7.427 100–150 5.00 84.9 8.5 6.627 150–200 5.25 96.2 2.5 1.3
Table 5 Depletion of cation species in groundwater relative to a theoretical composition derived by the concentration of local rainfall,using Cl– as the reference ion
Na+ Ca2+ Mg2+ K+ Cl–
Average rainfall (mg/L)a 1.3 0.9 0.3 0.4 2.3Highest salinity groundwater (mg/L)b 4,700 46 662 16 8,094Normalised rainfall (mg/L)c 4,562 3,271 1,146 1,391 8,094Depletion (%) –3d 99 42 99 0
Shows strong depletion of Ca2+ and K+ in Mount William Creek catchment groundwaters relative to other cationsa Volume weighted average composition from Table 1b Cation and Cl– analysis of highest salinity groundwater (Bore 5312)c Average rainfall concentrations normalised to Cl– of highest salinity groundwaterd Negative value indicates enrichment of Na+ , due to small errors
1370
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
tralia (0.028–0.244 mol kg–1; Ma and Eggleton 1999), sothe clays in the upper parts of the soil profiles in the studyarea may be saturated with respect to base cation storage.
However, continuing, long-term removal of ions byplant uptake is occurring. Ions taken up by vegetationfrom both the soil solution and exchange sites are returnedto the soil through the continual breakdown of organicmatter, and this may be responsible for the occasionalgroundwater cation/Cl– ratios exceeding those in localrainfall (Fig. 3), as shown elsewhere (Jobbagy andJackson 2004). However, this return of ions to the soil isdisrupted by the net erosional loss of particulate organicmaterial through overland transport (sheetflow and rivers,e.g. Hopmans et al. 1987; White et al. 2002); White et al.(2006) reported high Ca2+ concentrations in surface watersdraining grassland terraces in coastal California, whichthey attributed to seasonal biomass cycling. In addition,organic material is lost as smoke during bushfires (e.g.Stewart and Flinn 1985), and, in farmed areas, throughremoval of agricultural products (animals and crops).Significant amounts of both Ca2+ and K+ can also be lostby accumulation on plant leaves during photosynthesisand dispersion by wind (Artaxo and Orsini 1987; de Mello2001). Therefore the removal of organic matter will resultin a continuing, long-term export of ions, particularly Ca2+
and K+, from soil profiles, and the ions most stronglydepleted in Mount William Creek groundwaters are Ca2+
and K+ (Table 5).It is therefore possible that a steady state scenario has
become established at the study site; the soil clays areprobably saturated with respect to base cation storage, andthe cations supplied in rainfall are partly exported inorganic matter and partly incorporated into rechargeinfiltrating downwards into the groundwater. These pro-cesses are apparently sufficient to maintain the patternsobserved.
Relationship between unsaturated zone processesand groundwater salinity and ion/Cl– ratios acrossthe catchmentThe results of this study demonstrate that, in the studyarea, much of the chemical evolution of the groundwatersnormally attributed to groundwater-aquifer interactions infact occurs prior to recharge. How do these processescause the variability in the major ion/Cl– ratios across thecatchment and why, once the salinity exceeds ~2,000 mg/L, is there a further downgradient increase in salinitywithin the aquifers in the study area without any change inion/Cl– ratios (Fig. 3)?
The elevated parts of the study area have relatively thinsoil cover and are characterised by rapid recharge (asshown by groundwater Cl– mass balance and hydrographdata; Edwards 2006; Edwards and Webb 2006). Becauseof the limited time that this recharge spends in theunsaturated zone, it undergoes relatively little evapotrans-piration and interaction with plants and clays, so that thebase cation/Cl– ratios are often only slightly or moderatelydepleted compared to rainfall. This is clearly shown by the
Ca2+/Cl– and K+/Cl– plots (Fig. 3). In contrast, over thelow relief remainder of the catchment, where soil andregolith profiles are thicker and better developed (e.g. SP5and SP27), recharge is slow and diffuse (Edwards 2006;Edwards and Webb 2006), allowing sufficient time forevapotranspiration to increase the salinity and for unsat-urated zone processes to modify the ion/Cl– ratios so thatthey resemble those of dilute seawater (Figs. 3 and 4).
There is a concomitant downflow increase in ground-water salinity and decrease in ion/Cl– ratios across MountWilliam Creek catchment up to salinites of ~2,000 mg/LTDS (the value at which seawater ion/Cl– ratios becomeestablished; Fig. 3). This is caused by mixing of moresaline diffuse recharge that has undergone evapotranspi-ration and ion depletion in the unsaturated zone withfresher, less depleted groundwater recharged in the higherelevation parts of the catchment. Mass balance mixingcalculations using Na+ and Cl– demonstrate the amount ofmixing. Adding dilute groundwater with relatively highion/Cl ratios (<200 mg/L TDS; Table 2) to the very salinesoil water at the base of profiles SP5 and SP27 (1,180 and2,750 mg/L Na+ and 2,350 and 4,280 mg/L Cl–
respectively; from Table 3) shows that groundwater with~2,000 mg/L TDS contains 25–70% soil water; the rangelargely reflects the variability in salinity of the dilutegroundwater compositions used in the calculation.
There is a further downgradient increase in groundwatersalinity from 2,000 to 20,000 mg/L without any significantchange in ion/Cl– ratios (Fig. 3), due to the continuingaddition down the groundwater flow path of saline rechargewith seawater ion/Cl– ratios. This will progressivelyincrease the groundwater salinity but not alter its ion/Cl–
ratios, because these already match the recharge values.Groundwater studies elsewhere in southeastern Australia
have demonstrated the same general process. Bennetts et al.(2006) used major ion and stable isotope data to show thata downflow increase in groundwater salinity within anunconfined aquifer is due to the continued addition alongthe flow path of saline diffuse recharge that has undergoneevapotranspiration in the unsaturated zone.
This explanation of groundwater chemical trends doesnot require that cation exchange and/or mineral precipitationoccur within the saturated zone of the aquifer. There is noevidence for these processes in the Mount William Creekcatchment: no valence co-dependent cation enrichment anddepletion is observed in chemical data (e.g. ion ratio plots,Fig. 3), and groundwaters are typically undersaturated withrespect to calcite and other common minerals, particularly atthe low salinities where ion depletion is observed.
Mixing between two sources of recharge in this contexthas been proposed before (e.g. Tickell and Humphrys1987; Love et al. 1994; Cartwright et al. 2004; Bennetts etal. 2006); however, this study differs from previous workon a number of points. Firstly, it explicitly emphasises theunsaturated zone as the medium in which groundwaterchemical characteristics are imparted. Secondly, it assignsthe often-quoted process of cation adsorption and ex-change to a specific location in the flow-path, usingempirically-derived evidence. Thirdly, it provides evi-
1371
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
dence that plant nutrient cycling affects soil water andgroundwater composition.
Effect of unsaturated zone processeson groundwater compositionwithin the Murray-Darling Basin and elsewhereThe groundwater chemistry trends in the study area(Fig. 3) have been observed by numerous authors acrossthe Murray-Darling Basin, of which the Mount WilliamCreek catchment is a part (e.g. Lawrence 1975; Arad andEvans 1987; Tickell and Humphrys 1987; Macumber1991; Herczeg et al. 2001; Cartwright et al. 2004).Groundwaters in this basin have a distinct oceaniccharacter that is achieved by the time salinites reach~2,000 mg/L in terms of down basin evolution of thecomposition. The trends responsible for this compositionhave been interpreted in terms of mineral weathering inhigher elevation recharge areas of exposed/thinly coveredbedrock to produce the high cation/Cl– and HCO3
–/Cl–
ratios in low salinity waters, and some combination ofcation exchange, mineral precipitation and/or evaporation/transpiration within shallow unconfined/semi-confinedaquifers to produce the depleted base cation/Cl– ratios inhigh salinity groundwater. These mechanisms have alsobeen implicated in inverse modelling simulations (e.g.Herczeg et al. 2001). However, the present study showsthat processes within the unsaturated zone can exertultimate control on the groundwater composition, andwater-rock interaction within the aquifer may play little orno part. In fact, much of the chemical evolution of thegroundwaters normally attributed to groundwater-aquiferinteractions occurs prior to recharge. Furthermore, the lowsalinity groundwaters in the Murray-Darling Basin withhigh base cation/Cl– ratios are commonly viewed ascation-enriched due to mineral weathering, but the presentstudy demonstrates that they are probably, in fact, cation-depleted compared to local rainfall.
The general trends from rainfall chemistry to ground-water chemistry observed in the study area (Fig. 3) arecommon phenomena elsewhere in Australia and through-out the world (e.g. Salama et al. 1993; Kimblin 1995;Elliot et al. 1999; Guler and Thyne 2004). In addition, theprocesses occurring in the soils of the Mount WilliamCreek catchment have been identified in detailed studieson soil composition elsewhere, e.g. in the United States(Fryar et al. 2001; White et al. 2002, 2006). Combinedwith the non-unique nature of the physiographic andhydrogeological setting of the study area, this suggeststhat unsaturated zone processes may be responsible for theevolution of groundwater chemistry in shallow aquifers insimilar contexts world-wide.
Conclusions
Cation/Cl– ratios in the groundwaters of the MountWilliam Creek catchment are depleted compared torainfall, and this is caused both by plant uptake and
adsorption to the exchangeable sites of clay minerals inthe shallow sub-surface. In the case of Ca2+ , Mg2+ andK+, this occurs with remarkable rapidity, such that cation/Cl– ratios are at or below those in groundwater within theupper 50 cm of the soil profile; for Na+ , cation/Cl– ratiosapproximating those in groundwater are still achievedwithin the upper 200 cm.
Enrichment of the major cations within groundwater,above those ratios observed in rainfall, is very rare, and maybe due to return of ions through breakdown of organicmatter and desorption of monovalent ions consistent with acation exchange process favouring divalent cations.
The concentrations of exchangeable Na+ and Mg2+ onclay mineral surfaces are substantially greater than theconcentrations of K+ and Ca2+ . This does not correlate withthe original concentration of these species in rainfall, nor doesit correlate with any valence-dependent cation exchangepreferences. Both K+ and Ca2+ are important plant nutrients;groundwater data show that both K+ and Ca2+ are moredepleted than Na+ and Mg2+ , indicating that plant uptake ofK+ and Ca2+ exerts substantial control over groundwatercompositions. Removal of organic matter through erosionalloss and other mechanisms will result in a continuing, long-term export of these ions from the soil profiles.
These processes of biogeochemical fractionation meanthat recharge through the unsaturated zone is rapidlymodified so that ion/Cl– ratios are depleted relative torainfall but characteristic of seawater. In the higherelevation parts of the catchment with thinner soils therecharge may be fresher and less depleted; mixing of thisrecharge with more saline diffuse recharge that hasundergone evapotranspiration and ion depletion in thickersoils causes a concomitant downflow increase in ground-water salinity and decrease in ion/Cl– ratios up to salinitesof ~2,000 mg/L TDS (the value at which seawater ion/Cl–
ratios become established in the catchment). There is afurther increase in groundwater salinity without anysignificant change in ion/Cl– ratios due to the continuingaddition down the groundwater flow path of salinerecharge with seawater ion/Cl– ratios.
This explanation of groundwater chemical trends doesnot require that cation exchange and/or mineral precipita-tion occur within the saturated zone of the aquifer, anddemonstrates that much of the chemical evolution ofgroundwaters normally attributed to groundwater-aquiferinteractions in fact occurs prior to recharge.
The higher salinity groundwaters in the study area withion/Cl– ratios typical of seawater are thousands of yearsold, implying that the cation-removal processes in theunsaturated zone have been operating for this time period.It is therefore possible that a steady-state scenario hasbecome established at the study site; the soil clays areprobably saturated with respect to cation exchange, andthe cations supplied in rainfall are partly exported inorganic matter and partly incorporated into rechargeinfiltrating downwards into the groundwater.
The soil processes in the Mount William Creekcatchment have been identified elsewhere, and there areclear analogies between groundwater evolution in the
1372
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
study area and that in the Murray-Darling Basin and otherpredominantly unconfined alluvial aquifer systemsthroughout the world, suggesting that the evolutionarymodel presented here may have broad application.
Acknowledgements This research was supported by the WimmeraCatchment Management Authority and an Australian postgraduateaward from the Australian Government Department of Education,Science and Training. Groundwater dating was supported by AINSEGrant 05/172. The authors also acknowledge the contribution of threeanonymous reviewers whose comments substantially improved the finalmanuscript.
References
Acworth I, Jankowski J (2001) Salt source for dryland salinity:evidence from an upland catchment on the Southern Tablelandsof New South Wales. Aust J Soil Res 39:39–59
Arad A, Evans R (1987) The hydrogeology, hydrogeochemistry andenvironmental isotopes of the Campaspe River aquifer system,north-central Victoria, Australia. J Hydrol 95:63–86
Artaxo P, Orsini C (1987) Pixe and receptor models applied toremote aerosol source apportionment in Brazil. Nucl InstrumMethods Phys Res B22:259–263
Benedetti MF, Dia A, Riotte J, Chabaux F, Gerard M, Boulegue J,Fritz B, Chauvel C, Bulourde M, Deruelle B, Ildefonse P (2003)Chemical weathering of basaltic lava flows undergoing extremeclimatic conditions: the water geochemistry record. Chem Geol201:1–17
Bennetts DA, Webb JA, Stone DJM, Hill DM (2006) Understandingthe salinisation process for groundwater in an area of south-eastern Australia, using hydrochemical and isotopic evidence. JHydrol 323:178–192
Bennetts DA, Webb JA, McCaskill M, Zollinger R (2007) Drylandsalinity processes within the discharge zone of a local groundwatersystem, Southeastern Australia. Hydrogeol J 15:1197–1210
Blackburn G, McLeod S (1983) Salinity in atmospheric precipita-tion in the Murray-Darling drainage basin, Australia. Aust J SoilRes 21:411–434
Blake R (1989) The origin of high sodium bicarbonate waters in theOtway Basin, Victoria, Australia. In: Miles (ed) Water-rockinteraction. Balkema, Rotterdam, The Netherlands
Bormann ME (2004) Temporal and spatial trends in rainwaterchemistry across central and western Victoria. Honours Thesis,La Trobe University, Melbourne, Australia, 86 pp
Bureau of Meteorology (2003) Climate data for stations 079105 and079034. Climate and Consultancy Section, Victorian RegionalOffice, Bureau of Meteorology, Melbourne
Cardenal J, Benavente J, Cruz-Sanjulian JJ (1994) Chemicalevolution of groundwater in Triassic gypsum-bearing carbonateaquifers (Las Alpujarras, southern Spain). J Hydrol 161:3–30
Cartwright I, Weaver TR, Fulton S, Nichol C, Reid M, Cheng X(2004) Hydrogeochemical and isotopic constraints on theorigins of dryland salinity, Murray Basin, Victoria, Australia.Appl Geochem 19(8):1233–1254
Chorover J, Kretzschmar R, Garcia-Pichel F, Sparks D (2007) Soilbiogeochemical processes within the Critical Zone. Elements3:321–326
Cayley RA, Taylor DH (2001) Ararat: 1:100,000 map areageological report. Geological Survey of Victoria Report 115,Geological Survey of Victoria, Melbourne, 324 pp
de Mello WZ (2001) Precipitation chemistry in the coast of themetropolitan region of Rio de Janeiro, Brazil. Environ Pollut114:235–242
Dogramaci SS, Herczeg AL (2002) Strontium and carbon isotopeconstraints on carbonate-solution interactions and inter-aquifermixing in groundwaters of the semi-arid Murray Basin,Australia. J Hydrol 262:50–67
Drever JI, Smith CL (1978) Cyclic wetting and drying of the soilzone as an influence on the chemistry of groundwater in aridterrains. Am J Sci 278:1448–1454
Dyson PR (1983) Dryland salting and groundwater discharge in theVictorian Uplands. Proc R Soc Vic 95(3):113–116
Dyson PR, Jenkin JJ (1981) Hydrological characteristics of soilsrelevant to dryland salting in central Victora. Soil ConservationAuthority of Victoria, Melbourne
Edwards MD (2006) A hydrological, hydrogeological and hydro-geochemical study of processes leading to land and watersalinisation in the Mount William Creek Catchment, southeasternAustralia. PhD Thesis, LaTrobe University, Melbourne, 263 pp
Edwards MD, Webb JA (2006) The effects of lithology, soil andvegetation on recharge estimates in an upland catchmentaffected by dryland salinity: Mt William Creek, westernVictoria. 10th Murray Darling Basin Groundwater Workshop,Canberra, September 2006
Elliot T, Andrews JN, Edmunds WM (1999) Hydrochemical trends,palaeorecharge and groundwater ages in the fissured Chalkaquifer of the London and Berkshire Basins, UK. ApplGeochem 14:333–363
Fink D, Hotchkis M, Hua Q, Jacobsen G, Smith AM, Zoppi U,Child D, Mifsud C, van der Gaast H, Williams A, Williams M(2004) The ANTARES AMS facility at ANSTO. NIM B
Fryar AE, Mullican WF, Macko SA (2001) Groundwater rechargeand chemical evolution in the southern high plains of Texas,USA. Hydrogeol J 9:522–542
Garcia-Pichel F, Johnston SL, Youngkin D, Belnap J (2003) Smallscale vertical distribution of bacterial biomass and diversity inbiological soil crusts from arid lands in the Colorado Plateau.Microb Ecol 46:312–321
Garrels RM, Mackenzie FT (1967) Origin of the chemicalcompositions of some springs and lakes. In: Gould RF (ed)Equilibrium concepts in natural water systems. AmericanChemical Society, Washington, DC, pp 222–242
Guler C, Thyne GD (2004) Hydrologic and geologic factorscontrolling surface and groundwater chemistry in IndianWells-Owens Valley area, southeastern California, USA. JHydrol 285:177–198
Harrison A (1993) Hydrogeological assessment of salinity process-es: Mount William Creek Catchment. Report 1993/21, RoyalWater Commission, Melbourne
Heathcote JA (1985) Carbonate chemistry of recent chalk ground-water in a part of East Anglia, UK. J Hydrol 78:215–227
Herczeg AL, Dogramaci SS, Leaney FWJ (2001) Origin ofdissolved salts in a large, semi-arid groundwater system:Murray Basin, Australia. Mar Freshw Res 52:41–52
Hopmans P, Flinn DW, Farrell PW (1987) Nutrient dynamics offorested catchments in southeastern Australia and changes inwater quality and nutrient exports following clearing. For EcolManage 20:209–231
Hudson RO, Golding DL (1997) Controls on groundwater chem-istry in subalpine catchments in the southern interior of BritishColumbia. J Hydrol 201:1–20
Hutton JT, Leslie TI (1958) Accession of non-nitrogenous ionsdissolved in rainwater to soils in Victoria. Aust J Agric Res 9:59–84
Jankowski J, Acworth I (1993) The hydrogeochemistry of groundwaterin fractured bedrock aquifers beneath dryland salinity occurrencesat Yass, NSW. AGSO J Aust Geol Geophys 14:279–285
Jobbagy EG, Jackson RB (2004) The uplift of soil nutrients byplants: biogeochemical consequences across scales. Ecology85:2380–2389
Kimblin RT (1995) The chemistry and origin of groundwater inTriassic sandstone and Quaternary deposits, northwest Englandand some UK comparisons. J Hydrol 172:293–311
Lawrence CR (1975) Geology, hydrodynamics and hydrochemistryof the southern Murray Basin. Memoirs 30, Geological Surveyof Victoria, Melbourne
Love AJ, Herczeg AL, Leaney FW, Stadter MF, Dighton JC,Armstrong D (1994) Groundwater residence time and palae-ohydrology in the Otway Basin, South Australia: 2H, 18O and14C data. J Hydrol 153:157–187
1373
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8
Ma C, Eggleton RA (1999) Cation exchange capacity of kaolinite.Clay Clay Miner 47(2):174–180
Macumber PG (1991) Interaction between groundwater and surfacesystems in northern Victoria. Department of Conservation andEnvironment, Melbourne
Moss PD, Edmunds WM (1992) Processes controlling acidattenuation in the unsaturated zone of a Triassic sandstoneaquifer (U.K.), in the absence of carbonate minerals. ApplGeochem 7:573–583
Moulton KL, West J, Berner RA (2000) Solute flux and mineralmass balance approaches to the quantification of plant effects onsilicate weathering. Am J Sci 300:539–570
Parkhurst DL, Appelo CAJ (1999) User’s Guide to PHREEQC(Version 2): a computer program for speciation, batch-reaction,one-dimensional transport, and inverse geochemical calcula-tions. US Geol Surv Water Resour Invest Rep 99–4259, 310 pp
Rademacher LK, Clarke JF, Bryant Hudson G, Erman DC, ErmanNA (2001) Chemical evolution of shallow groundwater asrecorded by springs, Sagehen basin: Nevada County, California.Chem Geol 179:37–51
Rayment GE, Higginson FR (1992) Australian laboratory handbookof soil and water chemical methods. Inkata, Melbourne
Rosen M, Jones S (1998) Controls on the chemical composition ofgroundwater from alluvial aquifers in the Wanaka and Wakatipubasins, Central Otago, New Zealand. Hydrogeol J 6:264–281
Salama RB, Wells ASM, Farrington P, Bartle GA (1993) Thechemical evolution of groundwater in the aquifer systems of theYilgarn Craton of Western Australia, CSIRO Division of WaterResources, Perth
Simpson JH, Herczeg AL (1994) Delivery of marine chloride inprecipitation and removal by rivers in the Murray-DarlingBasin, Australia. J Hydrol 154:323–350
Sparks DL (2005) Metal and oxyanion sorption on naturallyoccurring oxide and clay mineral surfaces. In: Grassian VH(ed) Environmental catalysis, Taylor and Francis, London, pp3–36
Spears DA, Reeves MJ (1975) The influence of superficial depositson groundwater quality in the Vale of York. Q J Eng Geol8:255–269
Stewart HTL, Flinn DW (1985) Nutrient losses from broadcastburning of Eucalyptus debris in north-east Victoria. Aust ForRes 15:321–332
Stuyfzand PJ (1999) Patterns in groundwater chemistry resultingfrom groundwater flow. Hydrogeol J 7:15–27
Sutcliffe JF (1962) Mineral salts absorption in plants. Pergamon,London, 194 pp
Taylor JC, Hinczak I (2001) Rietveld made easy: a practical guideto the understanding of the method and successful phasequantifications. Sietronics, Canberra, Australia
Tickell SJ, Humphrys WG (1987) Groundwater resources andassociated salinity problems of the Victorian part of the RiverinePlain. Department of Industry, Technology and Resources
Toth J (1999) Groundwater as a geologic agent: an overview of thecauses, processes, and manifestations. Hydrogeol J 7:1–14
White AF, Blum AE, Schulz MS, Huntington TG, Peters NE,Stonestrom DA (2002) Chemical weathering of the Panolagranite: solute and regolith elemental fluxes and the weatheringrate of biotite. In: Hellmann R, Wood SA (eds) Water-rockinteractions, ore deposits and environmental geochemistry: atribute to David A Crerar. Geol Soc Spec Publ 7:37–60
White AF, Schulz MS, Vivit DV, Blum AE, Stonestrom DA (2006)Controls on soil pore water solutes: an approach for distinguish-ing between biogenic and lithogenic processes. J GeochemExplor 88:363–366
1374
Hydrogeology Journal (2009) 17: 1359–1374 DOI 10.1007/s10040-009-0449-8