Chloride imbalance in a catchment undergoing hydrological change: Upper Barwon River, southeast Australia

Chloride imbalance in a catchment undergoing hydrological change: Upper Barwon River, southeast Australia

Applied Geochemistry 31 (2013) 187–198 Contents lists available at SciVerse ScienceDirect Applied Geochemistry journal homepage: www.elsevier.com/lo...

1MB Sizes 0 Downloads 8 Views

Applied Geochemistry 31 (2013) 187–198

Contents lists available at SciVerse ScienceDirect

Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem

Chloride imbalance in a catchment undergoing hydrological change: Upper Barwon River, southeast Australia Ian Cartwright ⇑, Benjamin Gilfedder, Harald Hofmann School of Geosciences, Monash University, Clayton, Vic. 3800, Australia National Centre for Groundwater Research and Training, GPO Box 2100, Flinders University, Adelaide, SA 5001, Australia

a r t i c l e

i n f o

Article history: Received 9 August 2012 Accepted 4 January 2013 Available online 12 January 2013 Editorial handling by R. Fuge

a b s t r a c t Documenting whether surface water catchments are in net chemical mass balance is important to understanding hydrological systems. Catchments that export significantly greater volumes of solutes than are delivered via rainfall are not in hydrologic equilibrium and indicate a changing hydrological system. Here an assessment is made of whether a saline catchment in southeast Australia is in chemical mass balance based on Cl. The upper reaches of the Barwon River, southeast Australia, has total dissolved solids, TDS, concentrations of up to 5860 mg/L and Cl concentrations of up to 3370 mg/L. The high river TDS concentrations are due to the influxes of groundwater with TDS concentrations of up to 68,000 mg/L. Between 1989 and 2011, the median annual Cl flux from the upper Barwon catchment was 17.8  106 kg (140 kg/a/ha). This represents 340–2230% of the annual Cl input by rainfall to the catchment. Major ion and stable isotope geochemistry indicate that the dominant source of solutes in the catchment is evapotranspiration of rainfall, precluding mineral dissolution as a source of excess Cl. The upper Barwon catchment is not in chemical mass balance and is a net exporter of solutes. The chemical imbalance may reflect the transition within the last 100 ka from an endorheic lake system where solutes were recycled producing shallow groundwater with high TDS concentrations to a better drained catchment. Alternatively, a rise in the regional water table following land clearing may have increased the input of groundwater with high TDS concentrations to the river system. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction High concentrations of total dissolved solids (TDS) in surface water, groundwater, or soils, is a common feature in Australia and other semi-arid regions such as Argentina, China, western USA, the Middle East and India (Ghassemi et al., 1995). There may be multiple causes of high TDS concentrations in surface water or groundwater. For example, much of the saline groundwater (TDS locally in excess of 100,000 mg/L) in the Murray Basin in southeast Australia results from a combination of evaporation in a semi-arid climate and high transpiration rates of the native vegetation (Allison et al., 1990; Herczeg et al., 2001; Cartwright et al., 2004, 2008). Input of this high TDS groundwater into rivers and lakes as baseflow results in surface water systems with locally high TDS concentrations. Wetlands and lakes developed on the basalt plains of the Pliocene to Pleistocene Newer Volcanics Province in southeast Australia are commonly brackish to hypersaline with TDS concentrations up to 300,000 mg/L (Joyce, 1988; Bennetts et al., 2006, 2007; Dahlhaus et al., 2008; Cartwright et al., 2009; Tweed et al., 2011). The high TDS of these surface water bodies is ⇑ Corresponding author at: School of Geosciences, Monash University, Clayton, Vic. 3800, Australia. Tel.: +61 399054887. E-mail address: [email protected] (I. Cartwright). 0883-2927/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apgeochem.2013.01.003

due to poor drainage that allows evaporation to occur combined with the fact that many are throughflow systems or the sites of local groundwater discharge. Recharge from these lakes results in shallow groundwater with high TDS concentrations. Many of the groundwater and surface water systems in southeast Australia have probably been saline for many thousands of years (Dahlhaus et al., 2000, 2008), which is generally referred to as primary salinity. Secondary salinity is where anthropogenic modification of hydrological systems increases TDS concentrations. In Australia, the removal of native vegetation combined with irrigation increased recharge rates and caused the rise of regional water tables. Groundwater TDS concentrations subsequently increased due to the mobilisation of solutes stored in the unsaturated zone, upward migration of high TDS water from underlying aquifers, and/or increased evapotranspiration when the water table is within a few metres of the surface (Allison et al., 1990; Simpson and Herczeg, 1994; Bradd et al., 1997; Jankowski and Acworth, 1997; Acworth, 1999; Acworth and Jankowski, 2001; Clarke et al., 2002; Walker et al., 2002; Taylor and Hoxley, 2003; Cartwright et al., 2004). A rise in the water table also increases baseflow to surface water bodies that may result in an increase in river and lake TDS concentrations. While there has been much research on the distribution of salinity and the origins of the solutes, there has been less emphasis

188

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

on determining whether catchments are in hydrological and chemical mass balance. Simpson and Herczeg (1994) determined that the River Murray exported approximately twice the annual input of Cl from the Murray-Darling Basin, which they attributed to rising water tables and increased export of groundwater-derived solutes via the river system. However, most studies of chemical imbalances are from small forested upland catchments where imbalances in Cl fluxes may occur as a result of changes to the amount of Cl stored by organic matter and soils (e.g., Bastviken et al., 2006, 2007; Svensson et al., 2012). Areas of primary salinity may be in steady state such that the water and solute fluxes within the catchment are not changing over time. By contrast, because secondary salinity results from perturbation of the hydrological balance resulting in changes to water and solute fluxes, these catchments may not be in chemical equilibrium. 1.1. Solute and water balances in catchments If the boundary of the surface water system is used to define the catchment, solutes in surface water and groundwater in that catchment are derived from precipitation (rainfall, snowfall, and dry deposition) and the dissolution of minerals (Fig. 1). If the boundaries of the surface water and groundwater systems do not coincide, there also may be input of solutes from adjacent catchments via groundwater inflows. Solutes are exported from catchments via the river and groundwater systems, and solutes may also be consumed by mineral precipitation. The components of the water balance are similar but also include evapotranspiration as an export term. The water and solute balances are thus:

P þ GWi ¼ SWo þ ET þ GWo þ DST

ð1Þ

P  X P þ GWi  X GWi þ RRX ¼ SWo  X SWo þ GWo  X GWo þ DST  X ST

ð2Þ

where P is precipitation, SWo is surface water outflow, ET is evapotranspiration, GWi is groundwater inflow from adjacent catchments, and GWo is groundwater outflow from the catchment (all terms have units of m3/a). For catchments where the groundwater and surface water boundaries coincide Gwi = 0 m3/a. DST accounts for changes to groundwater held in storage in the catchment due to changes in soil moisture, surface storage, or due changes in the water table (in m3/a). DST > 0 implies an increase in water held in storage due to, for example, a rise in the water table. X is the concentration of solute X (in mg/m3) in the various components of the system. RRX is a reaction rate term for the whole catchment (in mg/a) for all mineral reactions that produce or consume solute X; RRX = 0 mg/a for conservative solutes. Other terms could be included in the mass balance in specific catchments (for example to take into account groundwater or surface water extraction or the diversion of surface water from adjacent catchments). Additionally it is possible to subdivide terms (such as distinguishing between direct groundwater outflows and groundwater that is exported from the catchment as baseflow to the rivers). There is a growing database of climate and hydrology data. Long term (decades to centuries) rainfall, evaporation and, transpiration records are available in many parts of the world. Databases such as the United States Geological Survey Water Data for the Nation program in the USA (www.waterdata.usgs.gov) and the National River Flow Archive in the UK (www.ceh.ac.uk/data/nrfa) contain river discharge records that span several decades or longer; groundwater levels and river and groundwater chemistry data are also available from the United States Geological Survey and from the British Geological Survey (www.bgs.ac.uk/research/groundwater). Combined with estimates of rainfall chemistry (e.g., Commonwealth Scientific and Industrial Research Organisation, 2012), these data allow many of the solute fluxes in Eq. (2) to be constrained. Groundwater fluxes, however, are difficult to calculate (although

Surface water catchment boundary

Possible groundwater flow from adjacent catchment (GWi, GW*XGwi)

Gauge Basement

Surface water outflow (SWo, SW*XSwo)

Evapotranspiration (E) Precipitation (P, P*XP)

Basement

Surface water outflow (SWo, SW*XSwo) Groundwater outflow (GWo, GW*XGwo)

Sediments

Groundwater flow Water Table Surface water flows

Water table fluctuation (ΔST, ΔST*XST) Mineral dissolution / precipitation (ΣRx)

Fig. 1. Schematic depiction of solute and water fluxes in a catchment; terms are defined in the text.

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

they may be broadly estimated from hydraulic conductivities and hydraulic heads). Additionally, there are very few constrains on catchment-wide rates of mineral dissolution or precipitation. While it is difficult to perform accurate water or solute balances without a detailed knowledge of all of the terms in Eqs. (1) and (2), some constraints on the overall chemical mass balance within catchments may be made. Specifically, if the flux of a conservative tracer out of a catchment via the river system is larger than the input from rainfall (i.e., where SWo  XSWo > P  XP) the catchment is a net exporter of solutes. Several scenarios may explain this type of chemical imbalance, such as: (1) rising water tables increasing baseflow to rivers, which results in the increased export of solutes from groundwater via the river systems; (2) draining of saline marshes that causes the export of solutes that have accumulated over long periods in the soil zone and shallow groundwater; (3) degradation of soils that contain organically-bound solutes and/ or; (4) rivers in coastal areas exporting solutes that accumulated in groundwater from ocean water during previous higher sea levels. All of these scenarios result from landscapes undergoing hydrological change and the chemical imbalance is an indicator that changes are occurring. In this study the causes of high TDS surface water in the upper Barwon catchment in southeast Australia are examined to determine whether the catchment is a net exporter of solutes. By utilising continuous (up to 22 year) records of electrical conductivity (EC) to construct solute fluxes, issues with representatively of periodic geochemical sampling can be overcome. The Barwon catchment is one of a number of catchments in southeast Australia that have been identified as high salinity (Coram et al., 2000; Dahlhaus et al., 2008). Because this study addresses whether this catchment is in chemical mass balance, the results will aid in assessing whether the salinity is primary and the timescales over which it may persist if the catchment is undergoing hydrologic change. Understanding the chemical mass balance is also important in predicting possible future changes to the TDS concentration of groundwater and surface water. While the focus is on a single catchment, the methodology outlined in this study is applicable to catchments globally. 1.2. Local geology and hydrogeology The Barwon catchment occupies 2700 km2 of southern Victoria, Australia (Fig. 2) and includes three major river systems- the Barwon, Leigh and Moorabool Rivers (Corangamite Catchment Management Authority, 2005). Due to the distribution of groundwater and surface water monitoring water infrastructure, this study focuses on the upper catchment of the Barwon River, defined herein as the part of the catchment that drains through the Inverleigh gauging station upstream of the confluence with the Leigh River (Fig. 2a). The headwaters of the Barwon River drain the northern slopes of the Otway Ranges where the surface geology comprises the Mesozoic–Cainozoic sediments of the Gelibrand Marl, Clifton Hill Formation, and the Eumeralla Formation (Corangamite Catchment Management Authority, 2005; Dahlhaus et al., 2008; Department of Primary Industries, 2012). Most of the upper catchment comprises basaltic flows and pyroclastic deposits of the Piocene–Pleistocene Newer Volcanics Province that are interbedded with Tertiary marine and freshwater sediments and which overlie the Cainozoic sediments. Holocene alluvial deposits are developed along the river courses. The landscape of the upper Barwon catchment is relatively young. The river cuts through a steep-sided basalt plateau that is 50–75 m above the river valley (Hills, 1975; Joyce, 1988). The lava flows are younger than 4.5 Ma with some scoria deposits and lava shields as young as 100 ka (Price et al., 2003). The eruption of these basalts buried the pre-existing landscape, blocking drainage courses

189

and forming lakes in depressions in the lava surface. From the analysis of landforms, notably lake terraces, a range of dating techniques (including 14C and thermoluminescence), and the distribution of lake sediments, the hydrological changes in the upper Barwon catchment and the basalt plains in general may be reconstructed (Dahlhaus et al., 2008). Lavas blocked the Leigh and Barwon Rivers producing a lake of 1800 km2 that extended SW from Inverleigh into the upper Barwon catchment and part of the adjacent Corangamite catchment (Hills, 1975; Jenkin, 1988: Fig. 2a). This lake may have been endorheic except at high water levels (as are many present day lakes within the Newer Volcanics Province: Dahlhaus et al., 2008; Tweed et al., 2011). Eroding of the lake outflow near Inverleigh gradually drained the lake leaving a number of remnant lakes and wetlands, including the 1550 ha Lake Murdeduke (Hills, 1975; Jenkin, 1988). Many of these lakes and wetlands are the sites of local groundwater discharge or throughflow systems and are brackish to hypersaline (Dahlhaus et al., 2000, 2008; Bennetts et al., 2006; Cartwright et al., 2009; Tweed et al., 2011). The Barwon region is 30 km from the coast and has a temperate climate with annual rainfall varying from 1030 mm at Forrest to 630 mm at Winchelsea. The majority of rain occurs in the austral winter (July–September) and potential evaporation over the summer months exceeds rainfall (Bureau of Meteorology, 2012). Total annual flows of the Barwon River at Inverleigh between 1974 and 2011 were 2200–330,000 ML. River discharge has a strong seasonality with 60% of the annual flow occurring between July and September. The headwaters of the Barwon catchment largely comprises native eucalypt forest and plantation forestry; however much of the upper catchment is cleared and utilised for grazing. Regional groundwater in the deeper Mesozoic–Cainozoic aquifers flows from the recharge area in the Otway Ranges to the NE, approximately parallel to the Barwon River Valley (Witebsky et al., 1995; Petrides and Cartwright, 2006). Flow in the shallower basaltic and alluvial aquifers is broadly in the same direction; however, there are variations due to local topography and there are numerous local groundwater discharge points in wetlands and lakes (Corangamite Catchment Management Authority, 2005). Due to the presence of the Gellibrand Marl it is unlikely that the deeper and shallower aquifers are hydraulically connected and hydraulic connection between the Barwon River and the Mesozoic–Cainozoic aquifers only occurs where those units crop out (Witebsky et al., 1995). Groundwater in the Mesozoic–Cainozoic aquifers has TDS concentrations of 100–4500 mg/L (Petrides and Cartwright, 2006). The TDS concentration of groundwater in the basaltic and alluvial aquifers generally increases down catchment and much of the shallow groundwater in the upper Barwon catchment has TDS concentrations of 3500–13,000 mg/L (Water Resources Data Warehouse, 2012: Fig. 2b). The TDS distribution shown in Fig. 2a is generalised and numerous local variations exist. In particular, there are several instances of shallow (<10 m) groundwater and soil water with much higher TDS concentrations in and around low-lying marshes and wetlands on the river floodplain (Fig. 2b). The head gradients in these regions are generally upwards and they are sites of local groundwater discharge (Dahlhaus et al., 2008). 1.3. Water sampling and analytical techniques Surface water was sampled from the Barwon River in seven sampling campaigns between April 2011 and May 2012 at a variety of flow conditions. The sampling round on 23/7/2011 represents flood conditions where the river had breached its banks in several places while those on the 8/2/2012 and 10/5/2012 represent low flow conditions. Water was collected from 0.5 to 1 m below the surface of flowing sections of the river using a sampler attached to a pole; sampling sites are shown in Fig. 2b and in Appendix 1. Groundwater was extracted from monitoring bores with screens of 1–3 m using

190

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

Zone 54

5900000

70000

75000

Zone 55

25000

(a)

(b) 20 km Ba Lake Murdeduke

o rabo

RM(523)

l Riv

River

5800000

WI(686) BC(13600) DM KR(587) (639)

Moo

Leigh

BI(482) DC(280)

er

Cor

Wi

Col 5700000

In IG

Mdk

IG (1200)

Ge

BK (318)

Bar

won

GE (167)

Co

PC(886) MC (376) MU (195)

FO(128)

Saline Discharge Gauging station Other sampling point Groundwater samples Extent of basalts

Fo

Lakes Rivers Settlement Gauge

Barwon River

Catchment boundary

Bass Straight

(c) Groundwater TDS (mg/L) <1000 1001-3500 3501-13,000

Groundwater Flow

Upper surface water catchment “Additional” groundwater catchment

Fig. 2. (a) Map of the Barwon catchment showing major rivers, general groundwater flow paths and the location of the upper Barwon surface and groundwater catchments. Settlements: Co = Colac; Fo = Forrest; Ge = Geelong; In = Inverleigh; Wi = Winchelsea. Lakes: Col = Colac; Cor = Corangamite; Mdk = Murdeduke. IG = Inverleigh Gauge that is used to define the extent of the catchment. Coordinates are from the Australian Map Grid (zones 54 and 55). (b) Groundwater salinities in the upper Barwon Catchment and the location of sampling sites (BC = Birregurra Ck; BI = Birregurra; BK = Boundary Creek; DC = Deans Marsh-Colac Rd.; DM = Deans Marsh; FO = Forrest; GE = Gerangamete; IG = Inverleigh Gauging Station; MC = Matthews Ck; MU = Muroon; PC = Pennyroyale Ck; RM = Ricketts Marsh; WI = Winchelsea). Numbers in brackets are EC values of surface water on 11/11/2011 (Appendix 1). Dashed line shows the southward extent of the basalts. (c) Location of Barwon Catchment. Data from Corangamite Catchment Management Authority (2005), Dahlhaus et al. (2008), Water Resources Data Warehouse (2012).

an impellor pump. In excess of three bore volumes were extracted prior to sampling or the bore was pumped dry and allowed to recover. Samples of ‘‘soil water’’ were collected from marshes and wetlands on the floodplain. The pH, EC and temperature were measured in the field using a TPS meter and probes. Alkalinity was determined on the day of sample collection using Hach digital titrators and reagents. Samples for major ion analysis were collected in HDPE bottles filled to overflowing and stored at 4 °C until analysis. Cations were determined using a ThermoFinnigan ICP-MS at Monash University on samples that had been filtered through 0.45 lm cellulose nitrate filters and acidified to pH 2 using 16 N ultrapure HNO3. Anions were determined in unacidified samples using a Metrohm ion chromatograph at Monash University. TDS is calculated as the sum of the dissolved anions and cations, and charge balances calculated using PHREEQC (Parkhurst and Appelo, 1999) are within 5%. Stable isotope ratios were measured at Monash University using Finnigan MAT 252 and ThermoFinnigan DeltaPlus Advantage mass spectrometers. The d18O values were measured via equilibration with He–CO2 at 32 °C for 24–48 h in a ThermoFinnigan Gas Bench. The d2H values were measured via reaction with Cr at 850 °C using an automated Finnigan H/Device. The d18O and d2H values were measured relative to internal standards that were calibrated using IAEA SMOW, GISP, and SLAP standards. Data were normalised following Coplen (1988) and are expressed relative to V-SMOW where d18O and d2H values of SLAP are 55.5‰ and 428‰, respectively. Precision (1r) is

d18O = ±0.15‰ and d2H = ±1‰. Major ion and stable isotope geochemistry are presented in Appendix 1. Additional groundwater geochemistry, river discharge, and river EC values were obtained from the Water Resources Data Warehouse (2012) that is maintained by the Department of Sustainability and Environment, Victoria. River EC and discharge is monitored continuously at several sites in the Barwon catchment (Fig. 2b). Data is recorded as an instantaneous in-river measurement on a sub daily basis (typically 30–60 min) and EC records extend from 1989 to the present day. The records have <5% missing data and the gaps are distributed throughout the record (i.e., there is no bias for extreme events not being recorded). For this study average daily EC values were produced from the sub-daily measurements; inspection of the data indicates that EC varies by <10%, and generally by <5% over 24 h. Unpublished EC and head data are also available for Department of Primary Industry salinity monitoring bores that sample mainly shallow (<20 m) groundwater and which are clustered in many of the areas designated as having salinity issues in Fig. 2b. 2. Results 2.1. Spatial distribution of river, soil water, and groundwater TDS concentrations EC is a qualitative measure of groundwater or surface water salinity and an assessment of chemical mass balances ideally

191

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

L near Forrest to as high as 1200 mg/L at Inverleigh. An increase in TDS concentrations down catchment was observed in the other sampling rounds (Appendix 1) and is also evident in the continuous EC data that increase from a median value of 850 lS/cm at Rickets Marsh to 1830 lS/cm at Inverleigh (Table 1), which corresponds to an increase in TDS concentrations from 600 mg/L to 1300 mg/L. Groundwater from bores completed in the basalt and Tertiary aquifers have TDS concentrations that range from 300 to 20,000 mg/L (Water Resources Data Warehouse, 2012). EC values from shallow (<20 m) bores in the low-lying poorly-drained regions on the floodplain that are identified as having salinity issues are 1000–76,300 lS/cm (Appendix 1, Department of Primary Industry, unpublished data), which corresponds to TDS concentrations of 760–68,000 mg/L (Fig. 3). Soil water from the marshes in the upper Barwon catchment has TDS concentrations of 115– 8820 mg/L (Appendix 1).

25000

TDS Cl Groundwater River

TDS/Cl (mg/L)

20000

r2 = 0.989 15000

10000

r2 = 0.990 5000

2.2. River, groundwater, and soil water geochemistry 0 0

5000

10000

15000

20000

25000

30000

Figs. 4 and 5 summarise the major ion geochemistry of river water, soil water and groundwater from the Barwon catchment. With increasing TDS, Cl concentrations in the river water increase from 46% to 97% relative molar abundance of total anions and HCO3 decreases from 53% to 1%. Sulphate comprises up to 24% of total anions but is not correlated with TDS; surface water with the higher SO4 concentrations is from Boundary Creek (Fig. 4). Sodium is the most abundant cation (61–86% of total cations on a molar basis) and the relative abundance of Na increases with increasing TDS (Fig. 4). Calcium, Mg and K constitute 2–16%, 11– 20%, and 0–5% of total cations, respectively. Calcium and Mg are present in higher relative abundances in the waters with the lowest TDS. Groundwater and soil water from the marshes on the floodplain have similar major ion chemistry to that of the river water. Molar Cl/Br ratios of the Barwon River are 520–830 and

EC (uS/cm) Fig. 3. Correlation of Cl and EC and TDS and EC in surface water and groundwater from the upper Barwon. The calculated r2 values and regression lines are for the dataset as a whole. Data from Appendix 1 and Water Resources Data Warehouse (2012).

requires concentrations of TDS or solutes such as Cl. There are good correlations between EC and TDS (r2 = 0.989) and EC and Cl (r2 = 0.990) for groundwater and surface water samples in the upper Barwon catchment (Fig. 3). This allows TDS and Cl concentrations to be robustly estimated from EC values. The spatial changes in river TDS concentrations are illustrated using data from 11th November 2011 (Fig. 2b). The TDS concentration of the Barwon River increased down catchment from <200 mg/

60

20

20

g +M 40

Barwon Tributary Groundwater Soil water

Ca 60

C 40 l+SO

4

80

80

100

0

SO

0

20 40

80

20

60

60 +K Na 40

Mg 60

80

40

20

6 HC0 O 40 3

10

0

10

20

0

0

0

20

0

60

80 0

10

40 20 0

40

0

Ca

20

0 10

60

40

80

80

80

4

80

60

100

100

Boundary Creek

0

20

40

Cl

60

80

100

Fig. 4. Piper plot of river water, soil water and groundwater from the Upper Barwon catchment (data from Appendix 1 and Water Resources Data Warehouse, 2012). Arrows show trends with increasing TDS, the samples from Boundary Creek are outlined.

192

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

1800

(a)

1600 1400 1200

10

1000 800

Rainfall

600 400 200 0 1.4

(b)

1.2

GMWL

-10

Evaporation

-20

-30

Average Melbourne rainfall

1.0

Na/Cl

MMWL

Barwon Tributary Groundwater Soil water

0

δ2H (‰ V-SMOW)

Cl/Br

(a)

Barwon Tributary Groundwater Soil water

-40

0.8

23/7/11

0.6 -50

0.4

-8

-6

-4

0

-2

2

4

δ18O (‰ V-SMOW)

0.2 0.0 0.3

(b)

-3.0

(c) δ18O (‰ V-SMOW)

-3.4

Ca/Cl

0.2

0.1

-3.8

-4.2

-4.6

0 0.4

(d)

-5.0 0

0.3

20

40

60

80

100

Mg/Cl

Distance (km from Forrest) Fig. 6. (a) Stable isotope ratios of river water, soil water and groundwater from the Upper Barwon catchment (data from Appendix 1). GMWL is the Global Meteoric Water Line and MMWL is the local Melbourne Meteoric Water Line (Cartwright et al., 2008). The arrowed line shows the general trend (slope of 4.5) expected for evaporation of water with an initial isotopic composition similar to that of local rainfall (c.f., Herczeg et al., 2001; Cartwright et al., 2008). Samples collected on 23/ 7/2011 are highlighted. (b) Variation in d18O values with distance along the river from Forrest (Fig. 2) for the 11/11/2011 sampling round.

0.2

0.1

0 0.4

(e)

K/Cl

0.3

0.2

0.1

0 0

4000

8000

12000

16000

20000

TDS (mg/L) Fig. 5. Variation of molar Cl/Br (a), Na/Cl (b), Ca/Cl (c), Mg/Cl (d) and K/Cl (e) ratios with TDS in river water, soil water and groundwater from the Upper Barwon catchment (data from Appendix 1 and Water Resources Date Warehouse, 2012). Rainfall ratios are from Blackburn and McLeod (1983).

do not vary with TDS; Cl/Br ratios of the soil water are 410–830 and those of groundwater are 540–730 (Fig. 5a). These Cl/Br ratios are similar to those of the oceans (650) and coastal rainfall (Davis et al., 1998). Molar Na/Cl ratios vary between 1.3 and 0.7 (Fig. 5b),

Ca/Cl ratios are 0.001–0.25 (Fig. 5c), Mg/Cl ratios are 0.05–0.35 (Fig. 5d), and K/Cl ratios are 0.001–0.37 (Fig. 5e). Cation/Cl ratios are generally higher in waters with TDS < 500 mg/L. The Na/Cl ratios of most groundwater and surface water are close to those of local rainfall whereas the Ca/Cl, Mg/Cl, and K/Cl ratios are substantially lower than those of local rainfall (Blackburn and McLeod, 1983). Groundwater and surface water from elsewhere in the Newer Volcanic Province basalt plains of southeast Australia has similar geochemistry to that in the Barwon catchment (Bennetts et al., 2006, 2007; Raiber and Webb, 2006; Cartwright et al., 2009). Most river and soil water have d18O and d2H values that lie close to the global and Melbourne meteoric water lines and which cluster around the average values of rainfall in Melbourne (90 km to the NE: Fig. 6a). Some river and soil waters, notably from Birregurra Creek, define a trend with a slope of 4–5 below the meteoric water line. There are some systematic differences between the d18O and d2H values from the different sampling rounds. The main Barwon River during the flood event on 23/7/2011 had d18O and d2H values of 6.1‰ to 5.4‰ and 41‰ to 37‰, respectively, which were the lowest recorded during the sampling period. The Barwon River in the two preceding rounds had d18O

193

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

and d2H values of 5.2‰ to 4.5‰ and 32‰ to 27‰, respectively, while d18O and d2H values in those in the round following the flood event (11/11/2011) were 4.6‰ to 3.2‰ and 26‰ to 20‰, respectively. In some sampling rounds there is a systematic increase in d18O values of 1–2‰ downstream between Forrest and Inverleigh (Fig. 6b); although such regular trends were not always observed. Except for Birregurra Creek, the variation in d18O and d2H values in the tributaries over time is similar to that of the Barwon River. 2.3. Temporal variation in river geochemistry and Cl fluxes The continuous EC and discharge records in the upper Barwon catchment allow the temporal variation of river geochemistry to be assessed. Firstly, daily averages of the EC were calculated from the sub-daily records. Subsequently, average daily Cl concentrations were estimated using the correlation between EC values and Cl concentrations of the river water (Fig. 3, Appendix 1). Finally, the daily discharge data were combined with the Cl concentrations to calculate average daily Cl fluxes (Fig. 7a). Calculated Cl concentrations at the lowermost gauging station in the study area Inverleigh range from 3 to 3370 mg/L. TDS concentrations at Inverleigh estimated from the correlation between EC values and TDS (Fig. 3) are between 5 and 5860 mg/L. Average daily discharge at Inverleigh ranges from 0 to 48,850 ML (zero discharge was recorded for several days in the summers of the drought years of 2007 and 2009 at many gauges in this and adjacent catchments and reflects actual conditions rather than a gauge malfunction). There is a broad inverse correlation between Cl concentrations and discharge (Fig. 7a) and a broad positive correlation between Cl concentrations and the net flux of Cl (Fig. 7b). The latter is due to the relative variation in discharge being higher than that in solute concentration. For 1989–2011 the median annual Cl flux at Inverleigh is 17.8  106 kg/a (Table 1), with Cl fluxes in individual years varying from 4.2  106 to 64  106 kg/a (Table 2, Fig. 8). Annual solute fluxes correlate with annual discharge (Table 2, Fig. 8), again reflecting the relative variability of discharge and Cl concentrations. Median Cl fluxes at the other gauging stations calculated in a similar manner are 14.5  106 kg/a at Winchelsea, 13.8  106 kg/a at Kildean Rd., and 7.4  106 kg/a at Ricketts Marsh; the Birregurra Creek tributary has a median annual Cl flux of 1.8  106 kg/a (Table 1). The area averaged Cl fluxes for the five gauges for the whole monitoring period are 125–186 kg/a/ha (Table 1). Area averaged fluxes for the Inverleigh gauge for individual years between 1989 and 2011 are 33–503 kg/a/ha (Table 2). 3. Discussion The combination of geochemical, groundwater head, and river discharge data allow a comprehensive understanding of processes in the upper Barwon catchment.

3.1. Geochemical processes The major ion and stable isotope geochemistry of the groundwater and surface water in the upper Barwon catchment constrains the major geochemical processes. The d18O and d2H values of groundwater are similar to those of local rainfall, implying that despite the high TDS concentrations of the groundwater there is no connate water in the catchment. The observation that molar Cl/Br ratios in both surface and groundwater cluster around those of coastal rainfall and do not increase with increasing TDS is most consistent with evapotranspiration being the dominant process controlling water salinity (c.f., Davis et al., 1998; Herczeg and Edmunds, 2000; Cartwright et al., 2004, 2006). Halite dissolution is the main alternative mechanism that can increase groundwater TDS concentrations; however, halite has molar Cl/Br ratios of 104–105 (McCaffrey et al., 1987; Herczeg and Edmunds, 2000; Cartwright et al., 2004) and significant halite dissolution results in high TDS waters with Cl/Br ratios that are one to two orders of magnitude higher than those observed in the groundwater or surface water from the upper Barwon catchment. In addition, there is no reported halite in the sediments or soils of this region. Molar Na/Cl ratios in both groundwater and surface water are close to those in local rainfall. In the absence of halite dissolution, the main source of Na aside from rainfall is from the weathering of silicate minerals such as plagioclase feldspar or evaporite minerals (e.g., trona or mirabilte: Herczeg and Edmunds, 2000). Significant weathering of these minerals results in high Na/Cl ratios most notably in low TDS waters where only a small degree of weathering is required to produce a noticeable change in geochemistry. The Na/Cl ratios as high as 1.3 in some surface water with TDS < 500 mg/L are most probably due to mineral weathering; however, the observation that Na/Cl ratios in the majority of groundwater and surface water are similar to those of local rainfall again implies that evapotranspiration is the dominant process controlling water geochemistry. The observations that Ca/Cl and Mg/Cl ratios are higher than those in rainfall in some surface water with TDS < 500 mg/L is also explained by minor weathering of silicates (such as plagioclase, pyroxene, or hornblende), carbonates, and/ or gypsum (Herczeg and Edmunds, 2000). Carbonate and gypsum precipitation is also likely to occur during evapotranspiration (Bennetts et al., 2006; Cartwright et al., 2009). The precipitation of these minerals explains the low Ca/Cl and Mg/Cl ratios in much of the groundwater and surface water and the decrease of Ca and Mg relative to Na and HCO3 relative to Cl as TDS increases. 3.2. Controls on river TDS concentrations The tributaries (especially Birregurra Creek) generally have higher TDS concentrations than the main river (Appendix 1, Fig. 2b); however, the downstream increase in TDS in the Barwon River is not the result of tributary inflows. For example, the average annual discharge of Birregurra Creek between 1994 and 2012 was

Table 1 Summary of Cl inputs and exports for the gauges in the Upper Barwon catchment.

a b c d

Gauge

Area (km2)

Median EC (lS/cm)a

Cl input (106 kg/a)b

Median Cl flux (106 kg/a)

Median Cl flux (kg/a/ha)

Excess Cl (%)b

Average Discharge (103 ML/year)

Discharge (%)b

Ricketts Marsh Kildean Lane Winchelsea Inverleighc Birregurra Ckd

590 865 1050 1270 102

850 1082 1250 1830 6440

0.37–2.4 0.54–3.6 0.63–4.3 0.80–5.2 0.064–0.42

7.38 13.8 14.5 17.8 1.84

125 160 140 140 186

310–1990 380–2560 340–2220 340–2230 450–2940

49.7 58.3 49.2 80.0 2.0

8–13 7–10 4–7 6–10 2–3

Data from Victorian Water Resources Data Warehouse. Range estimated using extremes of rainfall and Cl concentrations in rainfall as discussed in text. Lowermost gauge representing export from upper catchment as a whole. Tributary.

194

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198 Table 2 Ranges of Cl fluxes and excess Cl at the Inverleigh gauge.

104

(a)

Year

Cl flux (106 kg/a)

Cl flux (kg/a/ha)

Average discharge (103 ML/day)

Min Cl excessa (%)

Max Cl excessa (%)

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

47.7 36.1 25.0 63.9 36.1 28.4 16.4 33.9 10.8 9.9 6.6 13.0 30.1 17.8 20.2 16.0 12.0 4.2 11.7 5.0 9.7 19.6 22.0

376 285 197 503 284 224 129 267 85 78 52 102 237 140 159 126 94 33 92 40 77 154 173

297.5 112.3 173.7 224.2 127.1 55.1 59.1 155.3 26.2 31.3 10.1 38.5 128.1 40.5 60.1 56.1 33.7 8.9 43.2 9.1 25.9 70.9 60.8

919 696 481 1230 695 547 316 653 208 192 128 250 580 342 390 308 231 81 226 97 187 377 423

5969 4520 3126 7988 4515 3552 2049 4240 1349 1245 829 1622 3768 2223 2530 2001 1501 525 1466 628 1216 2452 2749

Max Min

63.9 4.2

503 33

297.5 8.9

1230 81

7988 525

Cl (mg/L)

103

10

2

101

100

10-1 107

(b)

Cl Flux (kg/day)

106 105 104 103 102

a Maximum and minimum Cl excesses are calculated from the range of rainfall and Cl concentrations of rainfall as outlined in the text.

101 100 10-3

10-1

101

103

105

Discharge Fig. 7. Variation in Cl concentrations (a) and Cl fluxes (b) of the Barwon River at the Inverleigh Gauging Station with discharge (data from Water Resources Date Warehouse, 2012).

2000 ML compared with 80,000 ML/day of the Barwon River at Inverleigh (Water Resources Data Warehouse, 2012). The median annual Cl flux from Birregurra Creek is 1.8  106 kg which is <10% of the median annual Cl flux of 17.8  106 kg at Inverleigh (Table 1). Matthews Creek and Pennyroyale Creek are ephemeral and contribute <10% of the flow to the upper Barwon, while Boundary Creek accounts for 5–15% of the flow. The TDS concentrations of these tributaries are only marginally higher than the main Barwon River (Appendix 1) and thus they are not major contributors of solutes. Furthermore, the largest increase in surface water TDS concentrations occurs in the stretch of river between Kildean Lane and Inverleigh (Fig. 2b) that is devoid of major tributaries. Most river and soil water from the floodplain marshes have d18O and d2H values that lie close to the meteoric water line (Fig. 6a). The surface waters with d18O and d2H values that define a trend beneath the meteoric water line have undergone evaporation and inriver evaporation is implied by the downstream increase in d18O values (Fig. 6b). However, the shift in d18O values is <4‰ in the main river, which may be achieved by <20% evaporation, and <8‰ in the tributaries, implying <40% evaporation (Gonfiantini, 1986). Far greater degrees of evaporation would be required to increase the TDS concentrations from those of rainfall (10–15 mg/L) to those in Appendix 1. Overall, these data imply that the high TDS concentrations in the Barwon River are not the result of in-river evaporation. In common with rivers with high TDS concentrations elsewhere (Allison et al., 1990; Farber et al., 2005), it is concluded that the high TDS concentrations of the Barwon River reflects the inflows of shallow groundwater and/or water from the unsaturated zone rather than in-river processes. While there are no near-river bores to confirm the relationship between river and groundwater

levels, the conclusion that the Barwon River receives groundwater inflows is supported by the following observations: (1) the major ion geochemistry of the Barwon river is similar to that of the groundwater and the soil water (Figs. 4 and 5); (2) the upper Barwon River rarely has zero flow, whereas losing streams in southeast Australia have extended periods of no flow over summer (Allison et al., 1990); and (3) head gradients in the shallow groundwater on floodplains of the upper Barwon catchment are dominantly upwards (Fig. 9). The saline wetlands and marshes that are abundant in the upper Barwon catchment probably play an important role in producing high TDS concentrations of groundwater and surface water. The head gradients in these regions are normally upwards (Fig. 9) implying that they are local groundwater discharge points. Some of the marshes and lakes may be throughflow systems as has been described in adjacent catchments (Bennetts et al., 2006; Cartwright et al., 2009; Tweed et al., 2011). Following periods of high rainfall the water table rises and head gradients in these areas temporarily reverse (Fig. 9). The reversal of head gradients causes these local discharge sites to revert to sites of groundwater recharge, which results in high TDS water being input into the shallow aquifers (c.f., Cartwright et al., 2004). This saline shallow groundwater provides a source of high TDS baseflow to the Barwon River. Lake Murdeduke (Fig. 2) is a throughflow lake with high TDS groundwater at its southern margin (Tweed et al., 2011); this groundwater forms baseflow to the Barwon River between Winchelsea and Inverleigh and probably accounts for much of the increase in TDS concentrations of the river along this reach. The Cl vs. discharge trend in the Barwon River is irregular (Fig. 7b). Some of this irregularity is due to individual high flow events being characterised by initial rises in Cl as discharge increases. This may represent the flushing of saline water from the marshes and soils during the first stages of overland flow (c.f., Obermann et al., 2009), which represents an additional pathway of saline surface water into the Barwon River. Overall, the geochemistry of groundwater and surface water in the Barwon catchment implies: (1) Cl is dominantly derived from rainfall; (2) evapotranspiration is the dominant process controlling

195

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

350 Discharge Cl Flux

60

300

2011

2010

2009

2008

2007

2005

2006

2004

2002

2003

2001

2000

0

1999

0

1998

50

1997

10

1996

100

1995

20

1994

150

1993

30

1992

200

1991

40

1989

250

1990

50

Average discharge (103 ML/day)

Average Cl Flux (106 kg/year)

70

Year Fig. 8. Annual variation in river discharge and Cl flux of the Barwon River at the Inverleigh Gauging Station between 1989 and 2011 (data from Table 2).

150 Head reversals during recharge events

Head (m AHD)

149

148

147 4.2m 17.0m 146 Aug/87

Aug/89

Aug/91

Aug/93

Aug/95

Aug/97

Aug/99

Date Fig. 9. Typical variation in groundwater heads in shallow bores from discharge areas in the upper Barwon catchment (depths indicate bore depth below the surface). Hydraulic gradients are downwards during recharge events and upwards at other times. Data from unpublished Department of Primary Industries database.

TDS concentrations; (3) mineral weathering and precipitation are relatively minor processes; and (4) the river is exporting solutes from shallow groundwater that enters the river as baseflow and from surface runoff from the saline marshes and wetlands. These conclusions are similar to those of other studies of groundwater and surface water chemistry in adjacent regions (e.g., Bennetts et al., 2006, 2007; Raiber and Webb, 2006; Cartwright et al., 2009). 3.3. Assessment of geochemical mass balance The calculated Cl fluxes from the upper Barwon catchment allow an assessment of whether the system is in chemical mass balance. These calculations again use the data from the Inverleigh gauge. There is a strong rainfall gradient across the area from 1030 mm in the upper catchment to 630 m at Winchelsea (Bureau of Meteorology, 2012). For the calculations these two values are used as the upper and lower bounds of net precipitation. The actual area weighted average rainfall will be between these two values (Bureau of Meteorology, 2012); however, using these extreme values provides a sensitivity test for the calculations. The

catchment area at the Inverleigh gauge is 1270 km2 (1.27  109 m2) (Water Resources Data Warehouse, 2012) and thus the net annual precipitation in the catchment is 7.99  108 to 1.31  109 m3. The Cl concentrations in rainfall and dry deposition in southeast Australia for areas away from the coast are 1–4 mg/L (Blackburn and McLeod, 1983; Commonwealth Scientific and Industrial Research Organisation, 2012). Using these values, the average annual delivery of Cl to the upper Barwon catchment is 0.8  106 to 5.2  106 kg (Table 1); the major ion geochemistry implies that this is the dominant source of Cl in groundwater and surface water in this catchment. The median annual export of Cl during the monitoring period was 17.8  106 kg/a, which represents 340–2230% of the calculated input. Since groundwater also exports Cl out of the catchment, these values are minima. Using the same range of rainfall totals and Cl concentrations the median annual export of Cl at the other gauging stations is: 14.5  106 kg (340–2220%) at Winchelsea; 7.38  106 kg (310–1990%) at Rickets Marsh; 13.8  106 kg (380–2560%) at Kildean Lane; and 1.84  106 kg (450–2940%) at Birregurra Creek (Table 1). Annual water fluxes between 1989 and 2011 calculated from the discharge data and rainfall estimates are 6–10% of rainfall at Inverleigh, 4–7% at Winchelsea, 7–10% at Kildean Rd., 8–13% at Ricketts Marsh, and 2–3% at Birregurra Creek (Table 1). There are numerous assumptions and uncertainties in these calculations that need to be assessed to determine whether they can account for the mismatch between Cl input and output. (1) Input of Cl: Variation in the rainfall and consequently variations in the delivery of Cl must be taken into account. The calculations used rainfall totals that span those of the headwaters (Forrest) and the outflow from the upper catchment (Winchelsea). The average rainfall within the catchment is intermediate between that at Forrest and Winchelsea (Bureau of Meteorology, 2012), and the range in rainfall of ±200 mm used in these calculations is greater than the annual variability of rainfall at either of those stations between 1900 and 2011. It is highly unlikely that long-term area-averaged rainfall in the catchment would be outside these bounds. While the Cl concentration of rainfall is subject to uncertainty, the range of 1–4 mg/L covers all stations in SE Australia except for those within a few km of the coast (Blackburn and McLeod, 1983; Commonwealth Scientific and Industrial Research Organisation, 2012) and is within the range of rainfall globally (Drever, 1997). Only if Cl con-

196

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

centrations were far in excess of those assumed in this study would the mismatch between calculated inputs and outputs be significantly reduced. (2) Stream flow: The variation in Cl fluxes is largely a function of stream discharge (Fig. 7b), which in turn varies with annual precipitation. Thus if the stream flow record included mainly high precipitation years it may overestimate the rate at which solutes are exported. Aside from 2 years (2006 and 2008) the calculated Cl fluxes of 4.2  106 to 64  106 kg/a between 1989 and 2011 are all greater than estimated annual Cl influx (0.8  106 to 5.2  106 kg). For 2006 and 2008 the Cl exports are only lower than the estimated influxes if the lower estimates of rainfall and Cl concentrations in precipitation are used (and even then they are within 81% and 97% of the estimated input). The monitoring period includes the significant drought period of 2000–2009 when rainfalls were lower than average as well as higher rainfall periods such as 1989–1993 and 2010–2011. The average annual river discharge at Inverleigh between 1989 and 2011 is within 10% of the average annual discharge between 1962 and 2011, indicating that the monitoring period is representative. (3) Other sources of Cl: The calculations assume that groundwater only exports Cl from the system. Significant inflows of groundwater with high TDS concentrations from outside the surface water catchment would represent a possible additional source of Cl. The boundaries of the surface water and groundwater systems coincide in most of the Barwon catchment; however, groundwater may infiltrate from the adjacent Corangamite catchment to the Barwon catchment in the NW of the study area (Fig. 2), which represents a groundwater inflow flux (Eqs. (1) and (2)). At steady state, the additional Cl from groundwater inflows from this region is that which is delivered via rainfall. The area from which these groundwater inflows are derived is 20% of surface water catchment area of the Inverleigh gauge. Even if all of the Cl derived from rainfall from this additional area were transferred via groundwater inflows to the Barwon River system, it could not account for the large mismatch between Cl input and output. Additionally gauging stations up catchment of where these groundwater inflows may occur, such as Ricketts Marsh, show similar large excesses of Cl to the catchment as a whole (Table 1). (4) Other uncertainties: Other uncertainties include those in the measured surface area of the catchment upstream of the various gauges. Re-measurement of areas using digital topographic maps produced area estimates that were within a few percent of those in Table 1. There will also be minor uncertainties in the EC measurements (most commercial EC loggers are precise to within 0.5%) and in the conversion of EC values to Cl concentrations; however, these are small relative to the magnitude of the Cl imbalance. There is a small reservoir in the headwaters of the Upper Barwon River upstream of Forrest. The capacity of this reservoir is only 1.9  107 m3 or <3% of the rainfall in the catchment and it has minimal impact on the water balance. There are also possible errors in the ratings curves from the gauges used to estimate discharge. A recent assessment of the ratings curves suggests that errors may be as high as 30% at low flow but are <5% at medium or high flow conditions (Scanlon et al., 2008). Since the majority of Cl is exported at high flow (Fig. 7b), this has a minor impact on the calculated values.

3.4. Causes of the Cl imbalance The data above imply that the upper Barwon catchment exports 340–2230% of the Cl that is input via rainfall. These values may be

underestimates as they do not take into account solutes exported via the groundwater system. It is proposed that the Cl imbalance in the upper Barwon catchment is due to changes in the hydrological balance and two scenarios are suggested. Firstly, following the eruption of the Newer Volcanic Plains Basalts, the upper Barwon contained a lake system with an outflow near Inverleigh that also extended into the adjacent Corangamite catchment (Hills, 1975; Joyce, 1988). The basalts in this region are younger than 4.5 Ma and some are as young as 100 ka (Price et al., 2003). Lake Corangamite and other lakes that were originally part of the larger lake system are developed on the younger basalt deposits. Erosion of the basalts gradually reduced lake levels creating a series of disconnected lakes and wetlands in the upper Barwon and Corangamite catchments. Dating of lunettes in the adjacent Corangamite catchment indicates that lake levels fell between 30 and 10 ka (Edwards et al., 1996; Dahlhaus et al., 2008), which may reflect the draining of the larger lake system. The numerous saline wetlands and Lake Murdeduke in the upper Barwon catchment are remnants of this older lake (Hills, 1975; Dahlhaus et al., 2008). Lakes across the Newer Volcanic Province in southeast Australia are generally brackish to hypersaline and include examples that lie above the local water table which act as recharge areas, throughflow systems, and discharge lakes that lie in topographic lows. The discharge lakes locally revert to recharge zones when brines become sufficiently dense that hydraulic gradients are reversed or when rainfall and runoff raises the lake level above that of the water table. Groundwater around the saline lakes throughout the Newer Volcanics Province has very highly TDS concentrations due to it being recharged from these lakes and wetlands (Edwards et al., 1996; Bennetts et al., 2006; Dahlhaus et al., 2008; Cartwright et al., 2009; Tweed et al., 2011). Breaching of the upper Barwon lake system would allow the upper basin to drain, and Cl that accumulated when the area was a lake basin may be being exported by the river. An alternative scenario is that land-clearing in the upper Barwon catchment has caused increased recharge that has resulted in a rise of the regional water table (c.f., Allison et al., 1990; Walker et al., 2002). In turn this will increase the baseflow component to the rivers resulting in the export of Cl that has relatively long residence times. While groundwater levels in the upper Barwon catchment have varied little over the last 50–60 a (Water Resources Data Warehouse, 2012), the area was cleared for agriculture over 150 years ago and the rise in the water table may have preceded routine groundwater monitoring. Given the geological and land use history of the region, it is possible that both scenarios apply. Breakdown of soils and the release of Cl bound to organic matter may also account for excess Cl in catchments (Svensson et al., 2012). The magnitude of Cl fluxes for the upper Barwon catchment as a whole (median values of 125–186 kg/a/ha: Table 1) are far higher than the small forested catchments where excess Cl produced by soil degradation has been documented. The long-term magnitude of excess Cl exported from the upper Barwon via the Inverleigh Gauge (340–2230%) exceeds the estimates of up to 130% Cl in those catchments (Svensson et al., 2012). It is unlikely that the store of Cl in the 2–3 m of soil in the catchment is sufficient to explain the very large Cl excesses in this catchment. Regardless of the cause, the upper Barwon catchment is adjusting to a new hydrological balance. Both the removal of lakes and wetlands and land clearing results in a decrease in evapotranspiration that results in both higher recharge rates and a lowering of groundwater TDS concentrations. Eventually, the groundwater system in the upper Barwon catchment will become less saline; this will result in a lowering of the TDS concentrations of the river as much of the solute load in the Barwon River is provided by baseflow. Given the calculated excesses of Cl in the Barwon River, its TDS could be reduced by over an order of magnitude. The likely

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

timescales over which this may happen may be broadly estimated. Assuming an area of 1270 km2 for the upper Barwon catchment, a depth of the shallow aquifers of 20–30 m, a porosity of 25%, and Cl concentrations of groundwater of 5000–25,000 mg/L yields estimated stores of Cl in the groundwater of 3.2  1010– 2.4  1011 kg, which are 3–4 orders of magnitude greater than the average annual Cl export. While these values are subject to uncertainty, they indicate that it may take several thousand years for the catchment to achieve chemical mass balance. In common with most hydrological systems, the change to chemistry in a catchment takes far longer than the change to the water balance. 4. Conclusions The upper Barwon River exports significantly more solutes than are input from rainfall, indicating that the catchment is undergoing hydrological changes. The methodology outlined here identifies catchments that are net exporters of solutes and may be used to identify hydrological systems that are undergoing change due to either anthropogenic or natural causes. It may also help resolve the debate as to whether land clearing is contributing to high surface water TDS concentrations in individual catchments (c.f. Dahlhaus et al., 2000, 2008). The methodology is easiest applied to conservative solutes, such as Cl, that are not produced or consumed by mineral reactions and to basins where there is no groundwater input from beyond the limit of the surface water catchment. In such cases, if solute export via the river system is greater than the solute input via rainfall the catchment must be a net exporter of solutes and the calculated imbalances are minima as there will generally be additional export of solutes via underlying groundwater systems. Acknowledgements Fieldwork and analyses were made with the help of Massimo Raveggi, Rachelle Pierson, Justin Wu, Anna Cartwright and Tamie Weaver. Funding for this project was provided by Monash University and the National Centre for Groundwater Research and Training (Program P3). The National Centre for Groundwater Research and Training is an Australian Government initiative supported by the Australian Research Council and the National Water Commission. The comments of two anonymous referees helped clarify our ideas. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apgeochem. 2013.01.003. References Acworth, R.I., 1999. Investigation of dryland salinity using the electrical image method. Aust. J. Soil Res. 37, 623–636. Acworth, R.I., Jankowski, J., 2001. Salt source for dryland salinity – evidence from an upland catchment on the Southern Tablelands of New South Wales. Aust. J. Soil Res. 39, 39–59. Allison, G.B., Cook, P.G., Barnett, S.R., Walker, G.R., Jolly, I.D., Hughes, M.W., 1990. Land clearance and river salinisation in the western Murray basin, Australia. J. Hydrol. 119, 1–20. Bastviken, D., Sanden, P., Svensson, T., Stahlberg, C., Magounakis, M., Oberg, G., 2006. Chloride retention and release in a boreal forest soil: effects of soil water residence time and nitrogen and chloride loads. Environ. Sci. Technol. 40, 2977– 2982. Bastviken, D., Thomsen, F., Svensson, T., Karlsson, S., Sanden, P., Shaw, G., Matucha, M., Oberg, G., 2007. Chloride retention in forest soil by microbial uptake and by natural chlorination of organic matter. Geochim. Cosmochim. Acta 71, 3182– 3192.

197

Bennetts, D.A., Webb, J.A., McCaskill, M., Zollinger, R., 2007. Dryland salinity processes within the discharge zone of a local groundwater system, southeastern Australia. Hydrogeol. J. 15, 1197–1210. Bennetts, D.A., Webb, J.A., Stone, D.J.M., Hill, D.M., 2006. Understanding the salinisation process for groundwater in an area of south-eastern Australia, using hydrochemical and isotopic evidence. J. Hydrol. 323, 178–192. Blackburn, G., McLeod, S., 1983. Salinity of atmospheric precipitation in the Murray Darling Drainage Division, Australia. Aust. J. Soil Res. 21, 400–434. Bradd, J.M., Milne-Home, W.A., Gates, G., 1997. Overview of factors leading to dryland salinity and its potential hazard in New South Wales, Australia. Hydrogeol. J. 5, 51–67. Bureau of Meteorology, 2012. Commonwealth of Australia Bureau of Meteorology. . Cartwright, I., Hall, S., Tweed, S., Leblanc, M., 2009. Geochemical and isotopic constraints on the interaction between saline lakes and groundwater in southeast Australia. Hydrogeol. J. 17, 1991–2004. Cartwright, I., Weaver, T.R., Fifield, L.K., 2006. Cl/Br ratios and environmental isotopes as indicators of recharge variability and groundwater flow: an example from the southeast Murray basin, Australia. Chem. Geol. 231, 38–56. Cartwright, I., Weaver, T.R., Fulton, S., Nichol, C., Reid, M., Cheng, X., 2004. Hydrogeochemical and isotopic constraints on the origins of dryland salinity, Murray basin, Victoria, Australia. Appl. Geochem. 19, 1233–1254. Cartwright, I., Weaver, T.R., Tweed, S.O., 2008. Integrating physical hydrogeology, hydrochemistry, and environmental isotopes to constrain regional groundwater flow: southern Riverine province, Murray basin, Australia. In: Carrillo, R.J.J., Ortega, G.M.A. (Eds.), Groundwater Flow Understanding from Local to Regional Scale: International Association of Hydrogeologists Special Publication 11. Taylor and Francis, London, pp. 105–134. Clarke, C.J., George, R.J., Bell, R.W., Hatton, T.J., 2002. Dryland salinity in southwestern Australia: its origins, remedies, and future research directions. Aust. J. Soil Res. 40, 93–113. Commonwealth Scientific and Industrial Research Organisation, 2012. RechargeDischarge Estimation Suite: A Nationally Consistent Approach to Recharge and Discharge Estimation in Data Poor Areas. . Coplen, T.B., 1988. Normalization of oxygen and hydrogen isotope data. Chem. Geol. 72, 293–297. Coram, J.E., Dyson, P.R., Houlder, P.A., Evans, W.R., 2000. Australian Groundwater Flow Systems Contributing to Dryland Salinity. Bureau of Rural Sciences, Canberra. Corangamite Catchment Management Authority, 2005. The Environmental Flow Determination for the Barwon River. Dahlhaus, P., Cox, J., Simmons, C., Smitt, C., 2008. Beyond hydrogeologic evidence. Challenging the current assumptions about salinity processes in the Corangamite region, Australia. Hydrogeol. J. 16, 1283–1298. Dahlhaus, P.G., MacEwan, R.J., Nathan, E.L., Morand, V., 2000. Salinity on the southeastern Dundas Tableland, Victoria. Aust. J. Earth Sci. 47, 3–11. Davis, S.N., Whittemore, D.O., Fabryka-Martin, J., 1998. Uses of chloride/bromide ratios in studies of potable water. Ground Water 36, 338–351. Department of Primary Industries, 2012. Victorian Government Department of Primary Industries earth resources (GeoVic). . Drever, J.I., 1997. The Geochemistry of Natural Waters, 3rd ed. Prentice Hall, New Jersey. Edwards, J., Leonard, J.G., Pettifer, G.R., McDonald, P.A., 1996. Colac 1:250,000 Map Geological Report. Geological Survey Report 98. Department of Natural Resources and Environment, Melbourne. Farber, E., Vengosh, A., Gavrieli, I., Marie, A., Bullen, T.D., Mayer, B., Holtzman, R., Segal, M., Shavit, U., 2005. Management scenarios for the Jordan River salinity crisis. Appl. Geochem. 20, 2138–2153. Ghassemi, F., Jakeman, A.J., Nix, H.A., 1995. Salinisation of Land and Water Resources: Human Causes, Extent, Management and Case Studies. University of New South Wales Press, Sydney. Gonfiantini, R., 1986. Environmental isotopes in lake studies. In: Fritz, P., Fontes, J.C. (Eds.), Handbook of Environmental Isotope Geochemistry, The Terrestrial Environment, vol. 2. Elseveir, Amsterdam, pp. 113–186. Herczeg, A.L., Edmunds, W.M., 2000. Inorganic ions as tracers. In: Cook, P., Herczeg, A. (Eds.), Environmental Tracers in Subsurface Hydrology. Kluwer Academic Publishers, Boston, pp. 31–78. Herczeg, A.L., Dogramaci, S.S., Leaney, F.W., 2001. Origin of dissolved salts in a large, semi-arid groundwater system: Murray basin, Australia. Mar. Freshwater Res. 52, 41–52. Hills, E.S., 1975. Physiography of Victoria. Whitcombe and Tombs, Melbourne. Jankowski, J., Acworth, R.I., 1997. Impact of debris-flow deposits on hydrogeochemical processes and the development of dryland salinity in the Yass river catchment, New South Wales, Australia. Hydrogeol. J. 5, 71–88. Jenkin, J.J., 1988. Geomorphology. In: Douglas, J.G., Ferguson, J.A. (Eds.), Geology of Victoria. Geological Society of Australia, Melbourne, pp. 403–419. Joyce, E.B., 1988. Newer volcanic landforms. In: Douglas, J.G., Ferguson, J.A. (Eds.), Geology of Victoria. Geological Society of Australia, Melbourne, pp. 419–426. McCaffrey, M.A., Lazar, B., Holland, H.D., 1987. The evaporation path of seawater and the coprecipitaion of Br and K+ with halite. J. Sediment. Petrol. 57, 928– 937.

198

I. Cartwright et al. / Applied Geochemistry 31 (2013) 187–198

Obermann, M., Rosenwinkel, K.H., Tournoud, M.G., 2009. Investigation of first flushes in a medium-sized Mediterranean catchment. J. Hydrol. 373, 405–415. Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide to PHREEQC (v.2) – A Computer Program for Speciation, Batch-Reaction, One-dimensional Transport, and Inverse Geochemical Calculations. United States Water-Resources Investigations Report 99-4259, Washington, DC, 312p. Petrides, B., Cartwright, I., 2006. The hydrogeology and hydrogeochemistry of the Barwon Downs Graben aquifer, southwestern Victoria, Australia. Hydrogeol. J. 14, 809–826. Price, R.C., Nicholls, I.A., Gray, C.M., 2003. Cainozoic igneous activity. In: Birch, W.D. (Ed.), Geology of Victoria, Geological Society of Australia Special Publication 23. Geological Society of Australia, Victoria, Melbourne, pp. 361–375. Raiber, M., Webb, J.A., 2006. Application of environmental isotopes to understand recharge mechanisms to a sub-basaltic deep lead system in western Victoria, Australia. Geochim. Cosmochim. Acta 70, A515. Scanlon, P., Western, A., Ozbey, N., 2008. Estimating the Uncertainty in Flow Estimates at Flow Monitoring Sites in Victoria using Australian Standard 3778.2.3. Final Report to Victorian Government, Department of Sustainability and Environment.

Simpson, H.J., Herczeg, A.L., 1994. Delivery of marine chloride in precipitation and removal by rivers in the Murray-Darling basin, Australia. J. Hydrol. 154, 323– 350. Svensson, T., Lovett, G.M., Likens, G.E., 2012. Is chloride a conservative ion in forest ecosystems? Biogeochemistry 107, 125–134. Taylor, R.J., Hoxley, G., 2003. Dryland salinity in western Australia: managing a changing water cycle. Water Sci. Technol., 201–207. Tweed, S., Grace, M., Leblanc, M., Cartwright, I., Smithyman, D., 2011. The individual response of saline lakes to a severe drought. Sci. Total Environ. 409, 3919–3933. Walker, G.R., Zhang, L., Ellis, T.W., Hatton, T.J., Petheram, C., 2002. Estimating impacts of changed land use on recharge: review of modelling and other approaches appropriate for management of dryland salinity. Hydrogeol. J. 10, 68–90. Water Resources Data Warehouse, 2012. Victoria Department of Sustainability and Environment Water Resources Data Warehouse . Witebsky, S., Jayatilaka, C., Shugg, A., 1995. Groundwater development options and environmental impacts Barwon Downs Graben, southwestern Victoria. Victoria Department of Natural Resources and Environment Report.