Estimation of Soil Physical Properties Using Remote Sensing and Artificial Neural Network

Passive microwave remote sensing techniques have been successfully used to obtain spatial and multitemporal surface soil moisture data over large areas. Measurement of fine-resolution soil physical properties over large areas are, however, rarely available. In this study, we explore the possibility of inferring soil physical properties from a multitemporal remotely sensed brightness temperature and soil moisture maps. We construct two Artificial Neural Network models based on physical linkages among space-time distribution of brightness temperature, soil moisture, and soil media properties. Using a sequence of remotely data from Washita ʹ92 experiment, we show that it is possible to infer soil texture from multitemporal brightness temperature and soil moisture data.