Townley, L.R., Rainford, H.J., Sander, G.C., and Bligh, K.J. (1987), Field Measurements and Modelling of Rainfall-Runoff and Soil Erosion at Berkshire Valley, W.A., UWA Environmental Dynamics Report ED-87-200, 95pp.

Wheat in Western Australia is grown on extremely shallow soils, and although volumes of eroded soil are small, the fine suspended sediments carried by surface runoff have been shown to carry 3 to 4 times as much phosphorus and nitrogen as are present in undisturbed soil. Experiments have shown that different land treatments, such as cropping without tillage, can reduce surface runoff and suspended sediment volumes by factors of 5 and 8, respectively, relative to three tillage treatments per season. Thus there are good reasons for wanting to be able to quantify the effect of different land treatments on surface runoff volumes and mobilisation of fine sediments.

The purpose of this Project, on rainfall-runoff and soil erosion modelling in Western Australian agricultural catchments, was to develop a quantitative basis for the WA Department of Agriculture's current recommendations on methods of surface water control in the wheatbelt of southwestern WA. The research focussed on the Department's research catchment at Berkshire Valley. Instruments were installed to monitor rainfall, surface runoff within contour bays, suspended sediment concentrations in surface runoff, and shallow groundwater levels in the top 10-60 cm of soil. Good field data were obtained during 17 rainfall events between May and October 1986. These data provide strong evidence for the role of shallow groundwater flow in reducing rates of surface runoff and therefore rates of soil erosion.

Attempts to use computer models to simulate the measured data were not as successful as had been hoped. Two models were implemented on microcomputers and tested extensively. ANSWERS is relatively easy to set up for individual contour bays, and is capable of simulating a two-dimensional network which covers the plan area of a small catchment. Its algorithms are simple and it is clearly capable of simulating complex events with good accuracy. HILLS, on the other hand, offers the advantage of being able to predict shallow groundwater levels as well as surface flows. However the latter model is much harder to calibrate and appears to be less robust, in the sense of its ability to converge over a wide range of parameter values.

In general, the models need to be much more user-friendly before they can be used regularly in practice. Both are event-based models and are particularly sensitive to antecedent moisture conditions. They probably need to be run in parallel with a long-term soil moisture accounting model. The Project did not achieve its goal of providing tools to support the Department of Agriculture's recommendations on methods of surface water control, but has provided data for future analysis, and understanding about the difficulties of practical rainfall-runoff and erosion modelling in Western Australian conditions.


Copyright © 2005 by Lloyd Townley
Last revised: 6 May 2005