Combined effects of multiple factors on spatiotemporally varied soil moisture in China’s Loess Plateau

Identifying the factors controlling spatiotemporally varied soil moisture (SM) is fundamental for water resources management, but the combined effects of multiple factors across multiple timescales have been rarely considered. The objective of this study is to identify the factor(s) individually or simultaneously controlling SM in China’s Loess Plateau, a region with arid climate experiencing substantial vegetation change because of a revegetation project implemented in 1999. With SM product derived from Global Land Evaporation Amsterdam Model Version 3.0a, we investigated large-scale SM dynamics for surface soil (0–10 cm) and root zone (10–250 cm) for 1982–2015, and considered environmental factors at the regional scale (e.g. precipitation PCP, mean temperature TMP, potential evapotranspiration PET, normalized difference vegetation index NDVI, land use types, elevation, slope gradient, clay fraction, and reference bulk density) and at the global scale (e.g. El Niño/Southern Oscillation ENSO, Pacific Decadal Oscillation PDO, North Atlantic Oscillation NAO, and Arctic Oscillation AO). Spatially, the mean SM in both surface and deep layers decreases with a gradient from the southeast to the northwest. Temporally, SM tends to decrease in the southeast while increases in the northwest, implying a pattern of “wet gets drier, dry gets wetter”. Individually, PET and ENSO are the most important regional and global factors controlling SM, respectively. Simultaneously, PET, PCP, and NDVI have combined effects, while the global factors exhibit no combined effects. The regional and global factors both influence SM at a scale of 8–16 months across the whole study period. In addition, the effects of regional factors exhibit at the scale of 32–64 months but tend to disappear after 1999. The global factors have smaller impacts than regional factors, but they tend to indirectly influence SM through perturbing regional climate. The identified factors highlight the complicated changes and controlling factors of SM, which provides important information for land use management and water resources management. The employed methods are effective in identifying the combined effects of multiple factors and can be extrapolated to other regions.