How shallow and how many points of measurements are sufficient to estimate the deep profile mean soil water content of a hillslope in the Loess Plateau?

Soil water content (SWC) measurements for deep soil profiles involve large and expensive expenditures of time and labor. However, understanding of soil water dynamics for deep soil profiles is very important for soil and water resource management and hydrological modeling. This study was aimed at (1) testing whether the mean SWC for the deep profile of a hillslope could be well estimated using the SWC measurements only made at a limited number of time stable locations (TSLs), which were identified from the SWC measurements of the shallow profile (i.e., within 0-100 cm) and (2) subsequently determining the optimal depth of the shallow profile and the optimal number of TSLs using the relative bias to the mean (RBM) SWC and the corrected Alcaike information criterion (AICc). SWC in the 0-300 cm profile was measured at 40 locations along a 243 m long transect on a hillslope where is located in the semiarid region of the Chinese Loess Plateau with a mean annual precipitation (MAP) of 437 mm. A total of 46 SWC measurements made over three growing seasons at each location were divided into the calibration and validation periods. The same method was repeated on the other two hillslopes in the semiarid and semi-humid regions with the MAP of 505 and 580 mm, respectively, for further assessing the effect of soil moisture conditions on the optimal number of TSLs and the optimal depth of the shallow profile. Results indicated that there were significantly correlations between the spatial patterns of SWC in the shallow soil profiles (0-10, 0-20, ... , 0-100 cm) and that in the 0-300 cm deep soil profile (p < 0.01), and the mean SWC in the deep profile on the hillslope could be well estimated using the SWC measurements only at several TSLs identified from the SWC measurements in the shallow soil profile (within 0-100 cm depth). Moreover, the optimal number of TSLs and the optimal depth of the shallow profile varied with soil moisture conditions. In the semiarid region, the optimal depth of the shallow profile was 0-40 cm and the optimal number of TSLs varied from 2 to 3; in the semi-humid region, the optimal depth of the shallow profile was 0-30 cm and the optimal number of TSLs was 3. The RBM values between the predicted and measured mean SWCs in the 0-300 cm deep profile during the calibration and validation periods were in the range of 1.6 to 4.0% on three hillslopes, which are regarded as the acceptable prediction accuracy in the hydrology community. The proposed approach could greatly reduce the expenditures of time and cost in monitoring the mean SWC for the deep profile at various spatial scales and could be applied well to the regions with similar hydrological and climatic conditions.