Efficient prediction of profile mean soil water content for hillslope-scale Caragana korshinskii plantation using temporal stability analysis

Spatio-temporal variability in soil water content (SWC) increases the difficulty in soil water resource management and vegetation construction in arid and semi-arid regions. Our objectives were to investigate the spatial variability and temporal stability of SWC and identify representative locations for reliably predicting the profile mean SWC (MSWC) of a 33-year-old Caragana korshinskii plantation on typical sunny and shady hillslopes on the Loess Plateau, China. The profile SWC (0-180 cm) in three sampling belts and five hillslope positions (upper position, UP; upper-middle position, UMP; middle position, MP; lower-middle position, LMP; lower position, LP) were measured 27 times for both the hillslopes in 2018 and 2019. The mean relative difference (MRD), the standard deviation of the MRD (SDRD), and an index of temporal stability (ITS) were used to identify representative locations of the MSWC. The results indicated that the SWC presented different horizontal but similar vertical change trends in the five hillslope positions. The profile MSWC on the sunny hillslope (8.89%) was lower than that on the shady hillslope (9.44%). The mean Spearman correlation coefficients of the temporal patterns ranged from 0.6 to 0.8 for the five soil layers on both hillslopes. The characteristics of temporal stability differed in both temporal persistence and the representative locations for the five soil layers on the sunny and shady hillslopes. The mean ITS values generally decreased with increasing soil depth. Ultimately, the representative locations for estimating the profile MSWC were determined to be the MP-2 location on the sunny hillslope and the MP-3 location on the shady hillslope based on the MRD, SDRD, and ITS values. The representative locations have the lowest ITS values and represent the MSWC accurately, with root mean square error (RMSE) and mean error (ME) values of less than 1%. In addition, the soil organic carbon, clay, silt, and sand contents greatly affect the spatial and temporal distributions of the soil water at the hillslope scale. Our study provides a rapid and accurate method for determining profile MSWC in C. korshinskii plantations on typical sunny and shady hillslopes and could also support valuable guidance for the rational management of hillslope-scale SWC resources for similar dry regions.