Determination of the Best Pixel Resolution and Window Size of DEM for Digital Mapping of Soil Clay Content

Although a better understanding and quantitative knowledge of digital elevation model scale will help to improve soil predictions, the influence of pixel size has not been investigated in detail. The aim of this study was to investigate the role of spatial scale on soil clay content prediction by empirically testing the interaction between pixel resolution and window size with regression tree model. In two different areas in terms of their geomorphology and soil (area 1, Maybod located in Yazd province covered 400 km2; area 2, Iasokand located in Kurdistan province covered 400km2), 120 surface soil samples (0-30 cm) were taken and their clay contents were measured. From 121 digital elevation models representing different scales, 22 attribute were extracted and used for soil clay content prediction. Results showed that Maybod area had the minimum RMSE (9.0%) and maximum R2 (0.47) and dependence of tree model on pixel size was significant for clay prediction[H1] ; however, in Iasokand area, the minimum RMSE (5.65%) and maximum R2 (0.77) were obtained and window size was significant for clay prediction.