Stochastic soil moisture dynamic modelling: a case study in the Loess Plateau, China

Abstract

ABSTRACTSoil moisture is a key factor in the ecohydrological cycle in water-limited ecosystems, and it integrates the effects of climate, soil, and vegetation. The water balance and the hydrological cycle are significantly important for vegetation restoration in water-limited regions, and these dynamics are still poorly understood. In this study, the soil moisture and water balance were modelled with the stochastic soil water balance model in the Loess Plateau, China. This model was verified by monitoring soil moisture data of black locust plantations in the Yangjuangou catchment in the Loess Plateau. The influences of a rainfall regime change on soil moisture and water balance were also explored. Three meteorological stations were selected (Yulin, Yan'an, and Luochuan) along the precipitation gradient to detect the effects of rainfall spatial variability on the soil moisture and water balance. The results showed that soil moisture tended to be more frequent at low levels with decreasing precipitation, and the ratio of evapotranspiration under stress in response to rainfall also changed from 74.0% in Yulin to 52.3% in Luochuan. In addition, the effects of a temporal change in rainfall regime on soil moisture and water balance were explored at Yan'an. The soil moisture probability density function moved to high soil moisture in the wet period compared to the dry period of Yan'an, and the evapotranspiration under stress increased from 59.5% to 72% from the wet period to the dry period. The results of this study prove the applicability of the stochastic model in the Loess Plateau and reveal its potential for guiding the vegetation restoration in the next stage.