Using Surface Observations to Increment the Stomatal Resistance Together with the Assimilation of SMOS / SMAP in a Land Data Assimilation System

Great progress has been achieved in the last few years for the initialization of soil moisture in numerical environmental prediction systems with the assimilation of L-band brightness temperatures from the SMOS and SMAP satellites. Although the improved soil moisture has directly benefited hydrological modeling and forecast, it did not necessarily translate into better numerical weather prediction (NWP). Problems with other aspects of land surface modeling and of NWP in general are likely responsible for that. It is well known for instance that land surface schemes or surface-atmosphere turbulent exchanges have several weaknesses in NWP models. Initial conditions of other variables and parameters related to the soil or vegetation could also be blamed. The impact of using surface observations such as screen-level air temperature and humidity or surface skin temperature for the initialization of the surface stomatal resistance is investigated in this study. The assimilation tests are conducted with the ensemble Kalman filter part of the Canadian Land Data Assimilation System (CaLDAS). Results already show that assimilating screen-level observations (in addition to SMOS and SMAP) to initialize surface temperature has a substantial impact on NWP forecast near the surface. Whether or not the same effect can be found with initialization of the stomatal resistance will be discussed at the conference.