Abstract:
Satellite remote sensing observations of soil moisture
and land surface temperature (LST) may be assimilated into a land
surface model (that is either driven by observed meteorological
forcing data or coupled to an atmospheric model).
The hope is, of course, that the assimilation provides
superior estimates of land surface conditions that can
subsequently be used in the initialization of weather or seasonal
climate forecasts.
There are, however, serious obstacles to success in land data assimilation. Assimilating satellite data into the land model without adequate treatment of the sometimes severe biases between the land model fields and the satellite data creates serious imbalances in the model-generated mass and energy fluxes.
In this paper, we investigate the biases between the land model fields and satellite retrievals from different platforms for soil moisture and LST. We also demonstrate how such biases can be addressed through a scaling approach. In particular, the retrievals from each sensor are scaled to the land model's climatology before they are assimilated into the land model. After assimilation, the merged land surface product may be scaled back into the climatology of the satellite retrievals if the application calls for it.