Abstract:
To address the gap in bridging global and smaller
modelling scales, downscaling approaches have been reported as
an appropriate solution. Downscaling on its own is not wholly
adequate in the quest to produce local phenomena, and in this
paper we use a physical downscaling method combined with data
assimilation strategies, to obtain physically consistent land surface
condition prediction.
Using data assimilation strategies, it has been demonstrated
that by minimizing a cost function, a solution utilizing imperfect
models and observation data including observation errors is
feasible. We demonstrate that by assimilating lower frequency
passive microwave brightness temperature data using a validated
theoretical radiative transfer model, we can obtain very good
predictions that agree well with observed conditions.