Testing of Soil Moisture Prediction Model for Army Land Managers

Abstract

The ARC/INFO geographic information system (GIS) has been combined with the ARX spatial expert system and the SMSP soil moisture‐soil strength model to give the managers of a major Australian army training base the ability to predict damage from exercises. Soil types on the base, their wetness indices, and moisture status are biophysical inputs to the damage prediction model. ARX is used to predict the soil types and wetness indices. These results and antecedent rainfall are then used by SMSP to predict soil moisture and soil strength. ARC/INFO is used to store and map the results. In this paper, we test the accuracy of the soil type and soil moisture predictions. Field trials demonstrated that the soil type predictions are acceptable, but that SMSP is unable to accurately predict soil moistures at individual sites, although predictions over transects of 1.5 km are acceptable.