Practical modelling for management in data-limited catchments

Abstract

Dryland salinity is an accelerating problem that threatens some of Australia's most productive agricultural land. The problem with using models for developing management options in dryland areas is the general lack of data: there is no shortage of complex physical models, but we have little data with which to calibrate them. Not only is there the problem of having confidence in the equations and parameters used in the models, but it is difficult to follow the information trail that leads to the conclusions. This is emphasized when the outputs of one model are used as inputs to another. In the case of dryland salinity, the catchment is first broken into so-called homogeneous areas to drive a surface water balance model, which drives a groundwater model, which in turn drives an economics model, the results of which influence management decisions. What confidence do we have in the model outputs, at any of these stages, upon which decisions worth millions of dollars are made? This work presents an examination of field data leading from a conceptual model to a simple numerical model of groundwater flows in the Liverpool Plains. We maintain the critical process interactions and derive a simple flow model, whose output can be coupled with production and economic models. The Liverpool Plains is a National Dryland Salinity Program focus catchment. As such, there has been more work done and more data available than for most catchments in Australia. Given a need to link physical and economic models and follow the information trails, we require conceptual and numerical models that include the critical processes and interactions, described by the simplest equations and fewest parameters. With these models in place, we present common management scenarios, and draw conclusions about the current and prospective state of the system, and the modelling exercise.