A Variational Method for Estimating Near-Surface Soil Moisture and Surface Heat Fluxes

A variational data assimilation method is proposed to estimate the near-surface soil moisture and surface sensible and latent heat fluxes. The method merges the five parts into a cost function, i.e., the differences of wind, potential temperature, and specific humidity gradient between observations and those computed by the profile method, the difference of latent heat fluxes calculated using the ECMWF land surface evaporation scheme and the profile method, and a weak constraint for surface energy balance. By using an optimal algorithm, the best solutions are found. The method is tested with the data collected at Feixi Station (31.41 degrees N, 117.08 degrees E) supported by the China Heavy Rain Experiment and Study (HeRES) during 7-30 June 2001. The results show that estimated near-surface soil moistures can quickly respond to rainfall, and their temporal variation is consistent with that of measurements of average soil moisture over 15-cm top depth with a maximum error less than 0.03 m(3) m(-3). The surface heat fluxes calculated by this method are consistent with those by the Bowen ratio method, but at the same time it can overcome the instability problem occurring in the Bowen ratio method when the latter is about -1. Meanwhile, the variational method is more accurate than the profile method in terms of satisfying the surface energy balance. The sensitivity tests also show that the variational method is the most stable one among the three methods.