Elsevier

Journal of Hydrology

Volume 392, Issues 3–4, 15 October 2010, Pages 150-163
Journal of Hydrology

Freshwater recharge into a shallow saline groundwater system, Cooper Creek floodplain, Queensland, Australia

https://doi.org/10.1016/j.jhydrol.2010.08.003Get rights and content

Summary

Freshwater lenses have been identified as having penetrated the shallow regional saline groundwater beneath the Cooper Creek floodplain near Ballera (south-west Queensland). Piezometers were installed to evaluate the major-element chemistry along a floodplain transect from a major waterhole (Goonbabinna) to a smaller waterhole (Chookoo) associated with a sand dune complex. The floodplain consists of 2–7 m of impermeable mud underlain by unconsolidated fluvial sands with a saline watertable. Waterholes have in places scoured into the floodplain. The transect reveals that groundwater recharge takes place through the base of the waterholes at times of flood scour, but not through the floodplain mud. Total dissolved solids rise with distance from the waterhole and independently of the presence of sand dunes. Stable water isotopes2H and δ18O) confirm that recharge is consistent with, and dependant on, monsoonal flooding events. Following floods, the waterholes self-seal and retain water for extended periods, with sulfate-δ34S and δ18O isotopes suggesting bacterial reduction processes within the hyporheic zone, and limited interaction between the surface water and groundwater during no-flow conditions.

The area occupied by the freshwater lenses (TDS < 5000 mg/L) is locally asymmetrical with respect to the channel flow direction, extending down gradient along distances of ∼300 m.

Introduction

Cooper Creek (Fig. 1), with a total catchment area of 306,000 km2 is the longest and probably most ecologically important dryland river in Australia (Kingsford et al., 1999). It starts at the confluence of the Barcoo and Thomson Rivers (Fig. 1) and is fed from the north-western slopes of the Great Dividing Range. Its middle to lower course is characterised by a multiple-channel floodplain where interacting aeolian and fluvial sedimentation and gradually reducing river discharges over the Late Quaternary have resulted in a unique stratigraphic and geomorphic setting (Nanson et al., 1988, Nanson et al., 2008, Knighton and Nanson, 2000, Maroulis et al., 2007). Local geomorphic features such as remnant dunes, mud-capped waterholes and floodplain, combined with large hydrological variability, result in huge uncertainties in recharge and groundwater evolution processes. Research presented here has identified shallow groundwater of low-salinity forming freshwater lenses overlying the saltier regional groundwater widespread beneath the floodplain.

Low-salinity groundwater surrounded by higher salinity water has been described from islands (Scheneider and Kruse, 2003) and near-shore marine areas, where freshwater lenses totally surrounded by seawater are a legacy of lower sea-level during the Late Pleistocene (Kooi and Groen, 2001). In fully continental settings, low-salinity groundwater lenses have been detected in the depressions of hyper-arid regions such as Central Oman lying above regional saline groundwater (Young et al., 2004). Within the Australian context, freshwater lenses have been identified and discussed by Thorne et al. (1990) in the lower course of the Murray River and recently reassessed by Cartwright et al. (2010), however, these are generally located within gaining streams, where lenses provide base flow to rivers during low flow conditions. In contrast, the freshwater lenses identified within the Cooper Creek floodplain are fed by episodic flooding events in fully loosing streams, with water tables well below maximum depths in the waterholes along the studied transect. The hydrogeochemical transect across several waterholes shows a link between the size of the waterhole and that of the associated freshwater lens. These shallow groundwater features appear to be vital in supporting the vegetation along and near the channels and, therefore, in maintaining a delicate ecological balance along the waterholes and on the floodplain generally. Furthermore, they provide a non-disturbed analogue to understand how other presently wetter regions, particularly those along the Murray-Darling basin, could evolve in the future if their climate becomes even more variable. This would also apply to totally regulated river systems with small ecological allocations that deprive shallow groundwater systems from significant recharge along their channels.

To better understand the role that the local geomorphology has on the dynamics of shallow groundwater, a field study was undertaken within a channel, dune and floodplain system (Goonbabinna–Chookoo) in order to determine the likely influence of these landforms on recharge processes and hydrogeochemistry. Groundwater and surface water samples were obtained along a transect across these landscape features and analysed for major and minor element chemistry, water stable isotopes (δ18O and δ2H) and dissolved sulfate isotopes (sulfate-δ34S and δ18OSO4).

The Chookoo dune-floodplain complex is located within the middle reaches of Cooper Creek in the Channel Country of Queensland in the north-eastern portion of the Lake Eyre basin (Fig. 1A and B). At Windorah, Cooper Creek is formed at the confluence of the Thomson and Barcoo Rivers, whose headwaters lie partly in the Great Dividing Range and are fed largely by the Australian summer monsoon. Local rainfall events down the system can generate additional runoff typically during La Niña events (Evans and Allan, 1992). Synoptic processes in northern Australia and the Lake Eyre basin are strongly dependant on El Niño-Southern Oscillation (ENSO) phenomenon. In general, higher rainfall, river discharges and lake levels are experienced during La Niña conditions, while the reverse occurs during El Niño phases (Kotwicki and Allan, 1998). This is due to the enhanced ability of the Australian summer monsoon to penetrate into the interior of the continent during La Niña phases (Allan, 1988). The meteorological station closest to the Chookoo complex, Ballera, is located ∼25 km to the north-west (Fig. 1B). The average January and July maximum temperatures are 40.6 °C and 20.9 °C, respectively, while the average precipitation for the same months is 24 mm and 2.4 mm, respectively (Fig. 2). The mean annual rainfall is 127 mm/a, whereas evaporation at Windorah is 2900 mm/a, increasing steadily to more than 3600 mm/a near Lake Eyre. During the sampling periods of September 2003 and July 2004, the local rainfall was 9.2 mm and 0 mm respectively. Whilst local rainfall is important in maintaining standing waterhole levels, its contribution to the overall hydrology of the Cooper Creek floodplain is minor. It is the high magnitude summer rains associated with synoptic incursions of the Australian monsoon into the headwaters of Cooper Creek that generate the major flood events. Rainfall in the headwaters of the catchment (e.g. Barcaldine) are more likely to generate major flooding downstream than local rainfall events (Fig. 2).

The major rivers draining the Lake Eyre basin experience some of the largest discharge variability in the world (Fynlayson and McMahon, 1988, Knighton and Nanson, 1994). As flow stage rises, Cooper Creek flows through an anabranching multiple-channel system, with heavy transmission losses downstream (Knighton and Nanson, 1994, Knighton and Nanson, 2001). During periods of low flow, the Channel Country rivers are divided into a series of waterholes, which are enlarged segments of channel scoured by flow concentration during floods (Knighton and Nanson, 1994, Knighton and Nanson, 2001). These waterholes can often retain a standing body of water over the year, some for several years, before flow re-commences, and thus are vital for the sustenance of the local ecosystems which are well adapted to ‘boom and bust’ population events (Kingsford et al., 1999).

In contrast to the hydrogeology of the Great Artesian Basin that lies beneath this region of the Lake Eyre basin at about 1500 m depth (Radke et al., 2000), the shallow groundwater zone is poorly understood. A limited number of bores have been drilled with the express purpose of interpreting the shallow aquifers, however, they are scattered and the data from these are limited in detail and, as such, regional processes in the Lake Eyre basin remain elusive (McMahon et al., 2005). Costelloe et al. (2005a) endeavoured to model recharge into these Quaternary sediments from the major rivers draining the Lake Eyre basin, and while partial recharge was observed, overall recharge was found to be only a fraction of the peak discharge (>1%), and recharge through clay- versus sand-dominated streams was an order of magnitude less. Similar inferences have been made by Knighton and Nanson (2001). Costelloe et al. (2005b) made a similar argument based on salt store and release in floodplains of the Neales River in the western Lake Eyre basin.

Maroulis et al. (2007) provided detailed grain size and stratigraphic information for the current study site, and from this we estimate the hydraulic conductivity (K) to be in the range of ∼1.2–19.2 cm s−1 (with D10 values 1–4 mm for the aquifer sands), using the method of Shepherd (1989). These high K values reflect the coarse-grained and well sorted nature of the fluvial sands which comprise the aquifer, and are consistent with expected values from similar environments (Shepherd, 1989). This study provides cross-sectional stratigraphic and hydrological information, which reveal minor variations in hydraulic head (Fig. 3), most likely associated with local recharge mounds close to the channels. Although the regional hydrogeological trends are not captured in this study, we suggest that the hydraulic gradient must be fairly consistent with the very low regional valley slope of 0.0003. The scarcity of hydrogeological knowledge regarding the shallow groundwater systems in this region of Australia is a reflection of its remoteness, and also the difficulty in obtaining detailed and continuous hydrogeological information. Given these limitations, this study utilises geochemical methods to obtain a snapshot of the shallow groundwater systems in this under studied area.

The middle reaches of Cooper Creek are located on a complex sequence of nested, gently warped basins and associated near-horizontal sedimentary sequences that are Early Palaeozoic to Cenozoic in age (Nanson et al., 2008, and references therein). Structural deformation during the Cenozoic has produced a land surface of broad anticlines and synclines, establishing the spatial framework for the development of the present land surface and upon the direction and orientation of the contemporary Cooper Creek drainage system (Senior, 1968, Senior et al., 1978, Maroulis et al., 2007).

In the absence of glaciation, changes in precipitation and wind strength over the Late Quaternary have resulted in alternating fluvial, aeolian and lacustrine sedimentation over much of continental Australia. In the currently arid anabranching floodplain-channel system of Cooper Creek, this is manifest as extensive late Quaternary fluvial (MIS 8–5) and aeolian (MIS 4–2) sand bodies overlain by thick (2–7 m) floodplain and channel mud deposits (MIS 2–1) (Rust and Nanson, 1986, Nanson et al., 1988, Maroulis et al., 2007). This dense impermeable alluvial ‘mud capping’ is the result of the much reduced late Pleistocene and Holocene transport capacity of the Cooper Creek system, and is in places punctuated at the surface by remnant aeolian sand dunes stratigraphically connected to the underlying fluvial sand bodies. The focus of our study, the Chookoo dune-floodplain complex (Fig. 1B), is such a system, where the Quaternary sand bodies constitute the main shallow semi-confined aquifers of the region, and the local water table is on average ∼10 m below the floodplain surface (Fig. 3).

An additional characteristic of the anabranching channel network is the presence of enlarged channel segments called waterholes (or billabongs). These waterholes form under a variety of geomorphic conditions, and are the only sections of the channel system to reach significant depths (Knighton and Nanson, 2000), and thus are also where recharge from the channel to the shallow groundwater is most likely to occur.

Section snippets

Sample preparation and collection

Groundwater samples were initially collected during exploratory work in September 2003, followed by a full groundwater sampling in July 2004. The groundwater samples were complemented with spot surface samples collected in the upper catchment in July 2006 following a minor flooding event. A set of 12 piezometers were drilled in September 2003, and a further 22 in July 2004, along a transect from the Goonbabinna waterhole, south and across the Chookoo waterhole and nearby sand dunes (Fig. 1B,

Salinity and major dissolved ions

Total dissolved solids (TDS) and major ion ratio differences between surface and shallow groundwater are considerable (Table 1, Fig. 4). Surface water TDS generally relates to its position within the catchment (upstream–downstream), type of waterhole (major or secondary), and the time lapsed since waterhole isolation (surface flow stops). Groundwater on the other hand shows variations directly related to its distance from the waterhole.

Freshwater lenses and ecological sustainability

This investigation of freshwater lenses on the Cooper Creek floodplain has important implications for rangeland management and regional ecology. The retention of surface waters in Channel Country rivers such as Cooper Creek, on which many aquatic, avian, mammalian and plant species depend, is directly related to waterhole morphology and corresponding ground and surface water interaction. Costelloe et al. (2008) found that Eucalyptus coolabah (the dominant riparian tree species in arid zone

Conclusions

  • (1)

    Shallow groundwater, in proximity to waterholes along the transect in this study of Cooper Creek forms freshwater lenses which mix progressively with saline regional groundwater with increasing distance from waterholes. The lenses are asymmetrical and spread downstream with TDS < 5000 mg/L for distances of up to ∼300 m from the waterholes, increasing to TDS between 5000 and 15,000 mg/L at 1000 m. Higher TDS concentrations in groundwater are found farther south in central areas of the floodplain.

Acknowledgements

This research was supported through funding provided by GeoQuEST and an ARC Discovery Grant DP0559953 (2006–2008) to Nanson and Jones. We also would like to thank Prof. Allan Chivas for access to IC and ICP-MS facilities at the University of Wollongong. Our sincere thanks to Geoff Black, who operated the truck-mounted power auger under difficult conditions over several field seasons. Special thanks to Dr. S. Hollins (ANSTO) for providing GNIP data and Dr. C. Hughes (ANSTO) for numerous fruitful

References (63)

  • T.A. McMahon et al.

    Understanding the surface hydrology of the Lake Eyre Basin: part 1 – rainfall

    Journal of Arid Environments

    (2008)
  • G.C. Nanson et al.

    Alluvial evidence for major climate and flow regime changes during the middle and late Quaternary in eastern central Australia

    Geomorphology

    (2008)
  • R. Nativ et al.

    Water salinization in arid regions – observations from the Negev desert, Israel

    Journal of Hydrology

    (1997)
  • C. Pierre

    Isotopic Evidence for the dynamic Redox cycle of dissolved sulphur compounds between free and interstitial solutions in marine salt pans

    Chemical Geology

    (1985)
  • E. Playà et al.

    Non-marine evaporites with both inherited marine and continental signatures: the Gulf of Carpentaria, Australia, at 70 Ka

    Sedimentary Geology

    (2007)
  • S. Rose et al.

    Strontium isotope systematics of base flow in Piedmont Province watersheds, Georgia (USA)

    Applied Geochemistry

    (2005)
  • J. Tang et al.

    Sr2+/Ca2+ and 44Ca/40Ca fractionation during inorganic calcite formation: I. Sr incorporation

    Geochimica et Cosmochimica Acta

    (2008)
  • M.E. Young et al.

    Electromagnetic detection of deep freshwater lenses in a hyper-arid limestone terrain

    Journal of Applied Geophysics

    (2004)
  • R.J. Allan

    El Niño southern oscillation influences in the Australasian region

    Progress in Physical Geography

    (1988)
  • Allison, G.B., Barnes, C.J., Hughes, M.W., Leaney, F.W.J., 1983. Effect of climate and vegetation on oxygen-18 and...
  • ANZECC & ARMCANZ, 2000. Australian and New Zealand Guidelines for Fresh and Marine Water Quality....
  • Aranyossy, J.-F., Filly, A., Tandia, A.A., Louvat, D., Ousmane, B., Joseph, A., Fontes, J.C., 1991. Estimation des flux...
  • S.E. Bunn et al.

    Flow variability in dryland rivers: boom, bust and the bits in between

    River Research and Applications

    (2006)
  • I. Cartwright et al.

    Impact of changing climate and land use on the hydrogeology of southeast Australia

    Australian Journal of Earth Sciences

    (2008)
  • J. Costelloe et al.

    Recharge rates to shallow groundwater from streamflow in arid zone catchments

  • J.F. Costelloe et al.

    Spatial and temporal variability of water salinity in an ephemeral, arid-zone river, central Australia

    Hydrological Processes

    (2005)
  • J.F. Costelloe et al.

    Determining loss characteristics of arid zone river waterbodies

    River Research and Applications

    (2007)
  • J.F. Costelloe et al.

    Water sources accessed by arid zone riparian trees in highly saline environments, Australia

    Oecologia

    (2008)
  • D.J. DePaolo et al.

    Evaporation effects on oxygen and hydrogen isotopes in deep vadose zone pore fluids at Hanford, Washington

    Vadose Zone Journal

    (2004)
  • DNRMW, 2008. Surface Water Data. The State of Queensland. Department of Natural Resources Mines and Water....
  • S. Epstein et al.

    Variations of the 18O/16O ratio in natural waters

    Geochimica et Cosmochimica Acta

    (1953)
  • Cited by (63)

    • A review of the use of radiocarbon to estimate groundwater residence times in semi-arid and arid areas

      2020, Journal of Hydrology
      Citation Excerpt :

      Recharge of aquifers in high-rainfall areas is dominantly by the direct infiltration of rainfall (Scanlon et al., 2002; Healy, 2010). By contrast, especially in low-relief areas, substantial recharge in arid and semi-arid regions may occur through flooding of ephemeral rivers (Allison et al., 1985; Scanlon et al., 2002; Scanlon et al., 2006; Morin et al., 2009; Cendón et al., 2010; Meredith et al., 2016). The DIC concentrations in river water are generally substantially higher than those in rainfall.

    View all citing articles on Scopus
    View full text