Development of a Deep Learning Surrogate Model to Simulate MetOp ASCAT Observations with Land Surface Parameters

In this study, deep neural networks are employed to act as a surrogate model between the Meteo France land surface model and Advanced SCATterometer (ASCAT) satellite observations. This provides a measurement operator for the assimilation of ASCAT satellite derivations into this model. Currently, TU Wien uses ASCAT measurements to retrieve soil...