GROUNDWATER-FLOW MODELING TO ASSIST DRYLAND SALINITY MANAGEMENT OF A COASTAL-PLAIN OF SOUTHERN AUSTRALIA

Groundwater flow modelling has been undertaken for an area of 10 500 h a within the regional unconfined aquifer system of a coastal plain of southern Australia, in the vicinity of the town of Cooke Plains, to pr edict the impact of various land management options (including recharg e reduction and discharge enhancement) on the extent of land salinisat ion caused by shallow saline watertables. The model was calibrated aga inst field data collected over 6 years. Sensitivity analysis was perfo rmed to assess the influence of mesh size, boundary conditions, and aq uifer parameters, and particularly rates of recharge and evaporative d ischarge, on groundwater levels. These were varied until the model was shown to be capable of simulating seasonal trends and regional and lo cal flow patterns. The model was then used to predict the impact of th e management options on groundwater levels. The results showed that co ntinuing current annual crop-pasture rotations will result in watertab le rises of approximately 0.2 m in 20 years (significant in this setti ng), with a further 50 ha of land salinised. A reduction in the rates of groundwater recharge through the establishment of high water-use pe rennial pastures (e.g. lucerne) showed the most promise for controllin g groundwater levels. For example, a reduction in recharge by 90% woul d result in watertable declines of 0.6-1.0 m within 5-10 years, with t he return to productivity of 180 ha of saline land. Small-scale (say < 100 ha) efforts to reduce recharge were found to have no significant i mpact on groundwater levels. Enhanced groundwater discharge such as pu mping from a windmill was found to be non-viable due to the relatively high aquifer transmissivity and specific yield. The modelling approac h has enabled a relatively small area within a regional aquifer system to be modelled for a finite time (20 years) and has shown that extens ion of the boundaries of the model would not have altered the predicte d outcomes. Furthermore, the analysis of sensitivity to cell size in a n undulating landscape where net recharge areas can become net dischar ge areas with only small increases in groundwater level is novel, and has helped to build confidence in the model. Modelling has demonstrate d that dryland salinisation can be controlled by reducing groundwater recharge over substantial tracts of land, and is not dependent on rech arge reduction over an extensive area upgradient, at least over the ne xt 20 years.