Landscape-scale temporal stability of soil water storage within profiles on the semiarid Loess Plateau of China

Abstract

Knowledge of the temporal stability of soil water storage (SWS) at landscape scale is scarce. The recognition of landscape-scale temporal evolution of soil water profiles is critical for soil water management and vegetational restoration in semiarid watersheds. Soil moisture was measured with neutron probes to a depth of 3.0 m on 18 sampling dates at 135 locations along a landscape transect from August 2012 to October 2013. Temporal stability of SWS at a landscape scale and a point scale was examined using Spearman's rank correlation analysis and indices of standard deviation of relative difference and mean absolute bias error, respectively. The mean spatial SWS in the shallow soil layer (0-1.0 m) was relatively more variable temporally than in the deeper soil layers (1.0-3.0 m), and the mean SWS in the deep soil layer (2.0-3.0 m) was more variable spatially. The mean Spearman's rank correlation coefficient increased with increasing soil depth and decreased with increasing time lags between measurements for the deeper soil layers (1.0-3.0 m). The number of temporally stable locations and the accuracy of prediction for predicting the mean SWS increased with increasing soil depth. The temporal stability of the SWS patterns was controlled by soil texture, organic carbon content, bulk density, and saturated soil hydraulic conductivity. Aboveground biomass and site elevation (except for the 2.0-3.0-m layer), however, affected the temporal persistence of SWS relatively weakly. This study provides useful information for estimating mean SWS at the landscape scale and may improve the management of soil water on the semiarid Loess Plateau of China.