Document Type: research
Soil loss erosion is one of the most serious environmental problems (widespread globally), which is a menace to sustainable ecosystems and agriculture. As the previous studies show, the world’s highest soil loss rates due to erosion are in three continents, i.e. Asia, Africa, and South America. A new method was proposed to statistically evaluate the most appropriate cell size for LS factor input and to study the effects of using the appropriate cell size in calculating the erosion’s total soil loss. Different models have been used. Among others, Revised Universal Soil Loss Equation (RUSLE) is used in this study. This model needs five parameters such as slope length and steepness (LS), crop cover (C), rainfall erosivity (R), soil erodability (K), and prevention practice (P) with the help of Geographical Information System (GIS) using raster technique whereby all the parameters are mapped in the form of grid layers of a specific cell size. The proposed methodology shows a way to comprehend how to implement and estimate annual soil loss erosion. Different celsize are selected and applied to data of Nibong Tebal Penang as a sample test. LS factors have been comprised where semivariogram models are fitted to the height information based on the 20-m contourline topographic map. The results show that with increasing cell size up to 50m, the nugget effect decreases and spatial dependency increases. The best spatial dependency and high variances and diversity of 50 m cell spacing, the Digital Elevation Model (DEM) of this cell size is found the most appropriate for such dataset. According to the results, by using geostatistical techniques which can identify the best DEM cell size in order to make the suitable raster analysis for decision in the case of DEM spacing, could be applied to select a suitable cell spacing in DEM to predict topographical factor in soil erosion modeling. Basically, these techniques lead to find DEMs 50m from topographic map with 20m interval contour lines. In general, the results of this study have confirmed that the geostatistical analysis and statistical approaches together can be applied to select an adequate cell spacing in DEM and also to predict topographical factor in RUSLE model.