Mapping Salinity Using Decision Trees and Conditional Probabilistic Networks

− This paper examines the use of different classifiers for integrating multi-temporal remotely sensed data with landform data derived from digital elevation models to produce maps showing areas affected by salinity in the south west agricultural region of Western Australia. Decision trees are used to map saline areas in the Ryan’s Brook catchment, located approximately 50 kilometres southwest of Kojonup, WA. The results are compared with maximum likelihood classification techniques using single-date Landsat TM imagery. The non-parametric decision tree classifiers combine multi-temporal Landsat TM data with landform data derived from digital elevation models to produce more accurate salinity maps. However, the maps exhibited large amounts of noise and showed errors which might be improved by incorporating prior knowledge about the relationships between input attributes and their relationship with salinity. A conditional probabilistic network is used to impose a known relationship between input attributes and salinity status. In this way, changes in salinity through time can be modelled using all of the available Landsat TM data. The results show a large improvement on the maximum likelihood and decision tree classifiers. The network is used to produce a time-series of landcover and salinity maps for the upper Blackwood and Frankland-Gordon catchments.