CSIRO PUBLISHING www.publish.csiro.au/journals/ajb Australian Journal of Botany, 2006, 54, 193–205 Tree water sources over shallow, saline groundwater in the lower River Murray, south-eastern Australia: implications for groundwater recharge mechanisms K. L. Holland A,D , S. D. Tyerman B , L. J. Mensforth C and G. R. Walker A A CSIRO Land and Water, PMB 2, Glen Osmond, SA 5064, Australia. B University of Adelaide, School of Agriculture and Wine, PMB 1, Glen Osmond, SA 5064, Australia. C Department of Water, Land and Biodiversity Conservation, GPO Box 2834, Adelaide, SA 5001, Australia. D Corresponding author. Email: kate.holland@csiro.au Abstract. The decline of riparian vegetation in the lower River Murray, south-eastern Australia, is associated with a reduction in flooding frequency, extent and duration, and increased salt accumulation. The plant water sources of healthy Eucalyptus largiflorens trees growing over highly saline (>40 dS m -1 ) groundwater were investigated during summer when water deficit is greatest. The study found low-salinity soil water overlying highly saline groundwater at most sites. This deep soil water, rather than the saline groundwater, was identified as the plant water source at most sites. Stable isotopes of water and water potential measurements were used to infer how the deep soil water was recharged. The low-salinity, deep soil water was recharged in the following two ways: (1) vertically through the soil profile or via preferential flow paths by rainfall or flood waters or (2) horizontally by bank recharge from surface water on top of the saline groundwater. Vertical infiltration of rainfall and floodwaters through cracking clays was important for trees growing in small depressions, whereas infiltration of rainfall through sandy soils was important for trees growing at the break of slope. Bank recharge was important for trees growing within ∼50 m of permanent and ephemeral water bodies. The study has provided a better understanding of the spatial patterns of recharge at a scale relevant to riparian vegetation. This understanding is important for the management of floodplain vegetation growing in a saline, semi-arid environment. Introduction In regions that experience temporal water deficits, the distribution of terrestrial plants is most often regulated by the availability of water during critical periods (Kozlowski 1968). The decline of riparian vegetation in the lower River Murray, south-eastern Australia has been linked to a reduction in flooding frequency, extent and duration, and increased salt accumulation (Murray–Darling Basin Ministerial Council, MDBMC 1999; Slavich et al. 1999). It is estimated that 40% of the floodplain vegetation is currently affected by salinisation, comprising vegetation communities dominated by unhealthy or dead trees and halophytes (Department for Environment and Heritage, DEH 2004). To better manage these floodplains, we need to understand the processes controlling long-term plant water availability and vegetation health (Eagleson 1982; Nemani and Running 1989; Hatton and Evans 1998). Stable isotope techniques have been used in conjunction with water-potential measurements to determine from where plants were obtaining their water. The plant water sources that have been identified by using these techniques include hydraulic redistribution, where trees transport water from the groundwater into the surface soils (Dawson 1993; Dawson and Pate 1996; Burgess et al. 1998; Caldwell et al. 1998), and from the surface soils, following rainfall into the deeper root zone (Burgess et al. 2001). Isotopic methods have also been used to differentiate plant water sources spatially and temporally, where riparian trees have access to water from creeks, soil and groundwater (White et al. 1985; Dawson and Ehleringer 1991; Flanagan et al. 1992; Thorburn et al. 1993a; Mensforth et al. 1994; Dawson and Pate 1996; Mensforth and Walker 1996; Chimner and Cooper 2004). The floodplains of the lower River Murray are typically vegetated by a mixture of river red gum (Eucalyptus camaldulensis), black box (E. largiflorens) and lignum (Muehlenbeckia florulenta) (O’Malley and Sheldon 1990). River red gums cover ∼20% of the vegetated area (DEH 2004) and tend to grow in less saline, more frequently flooded parts of the floodplain, typically adjacent to creek beds. Whereas black box cover ∼30% of the vegetated area (DEH 2004) and are found at higher elevations away from the creeks, but with access to shallow groundwater © CSIRO 2006 10.1071/BT05019 0067-1924/06/020193