Quantifying the effect of surface heterogeneity on soil moisture across regions and surface characteristic

Abstract

Modeling soil moisture (SM) spatial distribution is important for hydrological forecasting, runoff estimation and catchment management and other applications. The objective of this study was to quantify the effect of surface heterogeneity including vegetation, soil, topography etc. on SM at the regional and surface characteristics classes (SCC) scales. Landsat 8 images, ASTER digital elevation model, land cover map, climatic data, and SM measured at 148 locations in the Balikhli-Chay catchment of Iran were used. Greenness, brightness, wetness, and land surface temperature are the surface biophysical characteristics and the topographical characteristics including elevation, slope, aspect, Topographic Wetness Index (TWI) and solar local incident angle calculated using Tasseled Cap Transformation, Single Channel algorithm used in this study. The Triangular method and the bootstrapping model were used to model SM and mitigate prediction uncertainty. Correlation coefficient (r) and Root Mean Square Error (RMSE) between the modeled and measured SM values were used to evaluate the modeled SM. The mean r and RMSE between the modeled and measured SM for different months were 0.83 and 1.84 (volumetric percentage), respectively. The degree of heterogeneity of the spatial distribution of SM in different classes of surface biophysical and topographical characteristics was different. The impact of surface heterogeneity on SM varied across regional and SCC scales. The mean r values between SM and surface biophysical and topographical characteristics in the regional and SCC strategies for different months were 0.41 and 0.54, respectively. Using the SCC scale instead of the regional scale heterogeneity information can increase the accuracy of SM modeling.