Evaluating controls of soil properties and climatic conditions on the use of an exponential filter for converting near surface to root zone soil moisture contents

摘要

Root zone soil moisture (RZSM) is an important state variable for understanding various land surface and ecohydrological processes. Due to the lack of field measurements, different methods have been proposed to estimate RZSM, including the use of exponential filters to predict RZSM from remotely sensed near surface soil moisture data. However, inconsistent findings about the controls on the optimal characteristic time length Taps, which is used in the exponential filter method, have been reported in the literature. To reconcile these inconsistent findings and further improve the use of the method, the impacts of soil properties and climatic conditions on T-opt were assessed in this study using observed and modelled soil moisture datasets. Daily soil moisture data, daily meteorological records, and soil properties were retrieved from the Automated Weather Data Network (AWDN) and the Soil Climate Analysis Network (SCAN) within the continental United States. Data from the AWDN stations showed that Tut was mostly controlled by soil texture (e.g., a negative correlation with the sand fraction and a positive one with the clay fraction) as compared to climatic conditions. However, at SCAN stations, T-opt was mostly affected by precipitation (P), and no significant correlation was found between T-opt and soil texture. The difference in controlling factors between ADWN and SCAN stations can be largely attributed to the higher spatial variability in P across the SCAN stations, which overrode the impacts of soil properties on T-opt. A 1-D vadose zone model was also utilized to simulate soil moisture at selected SCAN sites using a generated soil hydraulic parameter dataset. The simulation results further demonstrated the negative relationship between Toot and P observed for the SCAN stations. Moreover, the simulation results revealed that T-opt was larger under vegetated conditions than under bare surface conditions. Under the same climatic conditions at each simulated site, which could be deemed as reduced variability in P, significant correlations existed between T-opt and van Genuchten parameters. In particular, T-opt was shown to be smaller for coarser soils, which was consistent with the results observed from AWDN stations. The findings of this study offer additional insights into the use of the exponential filter method for estimating RZSM from near surface soil moisture measurements. (C) 2017 Elsevier B.V. All rights reserved.