Ensemble Kalman smoother for soil moisture data assimilation

To study the influence of soil texture and organic matter on soil moisture assimilation when observation data are rare, we develop a soil moisture assimilation scheme based on ensemble Kalman filter (EnKF) and ensemble Kalman smoother (EnKS). Surface soil moisture observations obtained from Arou station in the upper reaches of Heihe river basin in 2008 are assimilated into simple biosphere model 2 (SiB2) by using EnKF and EnKS respectively to analyze the influences of soil texture and organic matter and observation frequency on estimation. Results showed that influences of soil texture and organic matter on soil moisture estimation are large at surface layer but little at deep layer. Both EnKF and EnKS can significantly improve the accuracy of soil moisture estimation at surface and root zone by assimilating surface soil moisture observations, and EnKS is little better than EnKF. High accuracy of soil moisture estimation can be obtained by EnKS when observation data are scarce.