| Summary: | Soil moisture (SM) is important in understanding Earth's hydrologic cycles. However, an overall performance of multisource SM products is still unclear due to a lack of comprehensive validation. Using SM simulated by hydrologic models as a reference to perform validation is a promising alternative since SM simulation is not restricted by its coverage or scale. In this study, an integrated hydrologic model (ParFlow-CLM) forced by hydrometeorological and agricultural water use reanalysis data is built in the middle reaches of the Heihe River Basin (HRB), a typical managed agricultural region in Northwestern China. Using the ParFlow-CLM simulated SM data as the validating reference, ten SM products, including four single-source RS SM, three merged SM, and three assimilated SM products are systematically assessed by a comprehensive evaluation framework composed of fifteen statistical performance indicators. For validation, the results show that ParFlow-CLM SM simulations agree with the time domain reflectometry probe and cosmic ray neutron probes observations. The merged SM and assimilated SM outperform the single-source RS SM products if compared with ParFlow-CLM simulated data. Results from the intercomparison demonstrate that hydraulic conductivity and leaf area index are the dominant factors in SM spatial variations based on the generalized additive model. The statistical linkage indicates that mean absolute deviation, uncertainty at 95% (U95), Nash–Sutcliffe's efficiency, and combined performance index serve as substitutes for quantifying the relative uncertainty of SM. This study paves a way for model-data intercomparison of SM products in the HRB as well as in other arid and semiarid basins.
|
|---|