The variability in soil water storage on the loess hillslopes in China and its estimation

Soil water storage (SWS) is a critical water resource for vegetation growth in the Loess Plateau. Changes in SWS during the growing season can reflect soil water depletion or recharge conditions in different hydrological years. This study investigated the soil water in three hillslopes covered with artificial forest, natural forest and natural grass in May and October of 2015-2017. Both path analysis and stepwise multiple linear regression were used to determine the direct and indirect effects of most important variables and develop a model for predicting changes in SWS. The results showed that compared with natural forestland and grassland, artificial forestland had low soil water, but deep SWS maintained a balance of inputs and outputs. In all soil layers, during the growing season, total precipitation (X4) was the most important variable affecting changes in SWS. The changes in SWS in surface soil layers were affected by more environmental factors than in deeper soil layers. The natural grassland (X3) and slope position (X12) have different degree of direct and indirect effects by means of antecedent soil water content (X5), even though their correlations were not significant. In addition, changes in SWS at the 0-100 cm, 100-200 cm and 0-400 cm soil layers were adequately predicted using the multiple regression equation, in which R2 reached values from 73.4-93.0%. The model simulating the changes in SWS during the growing season can improve understanding on whether rainfall infiltration can replenish soil water depletion.