Soil moisture is a key factor influencing forest productivity, biodiversity, and ecosystem resilience. Variations in soil moisture, driven by climate change (e.g., droughts, altered precipitation patterns) or deforestation, can disrupt critical ecosystem processes and functions. Consequently, monitoring soil moisture is essential for sustainable forest management. Although geostatistical techniques and satellite-based modeling approaches have been widely employed for mapping soil moisture, their performance under varying topographic conditions (e.g., slope aspect) has received limited attention. This study aimed to evaluate the applicability of these technologies in modeling soil moisture under different topographic conditions (north-facing vs. south-facing slopes) in the semi-arid oak forests of the Zagros Mountains in western Iran. Soil sampling was conducted across three elevation classes, two canopy cover densities (closed and open), and two slope aspects (north and south), with five replications (totaling 60 samples). After laboratory measurement of soil moisture, soil moisture maps were generated using various interpolation techniques based on variogram models. Additionally, Sentinel-2 and Landsat-9 satellite images were used to compute soil moisture indices (NDWI and LSWI). Results revealed that soil moisture varied significantly as a function of slope aspect, with consistently higher values on north-facing slopes across all elevation classes (P-value < 0.05). The highest soil moisture content (12.46%) was recorded in mid- and upper-elevation zones with closed canopy on northfacing slopes, whereas the lowest value (3.92%) was observed on upper south-facing slopes with open canopy. Among the evaluated interpolation methods, Simple Kriging (RMSE = 0.79) performed best for north-facing slopes, while Ordinary Kriging (RMSE = 0.44) was superior for south-facing slopes. Regarding satellite-based indices, LSWI demonstrated higher accuracy in estimating soil moisture compared to NDWI. Ultimately, Landsat-derived soil moisture maps showed greater accuracy than those based on Sentinel imagery. These findings underscore the strong influence of topographic characteristics on soil moisture modeling using both geostatistical and remote sensing approaches.
