Validation of the QSCAT NRCS on the advanced neural network NSCAT GMF and estimation of neural network QSCAT GMF

The authors present an estimation of the geophysical model function (GMF) of the QSCAT scatterometer done by using neural network methodology. This GMF which is denoted QSCAT-NN was calibrated with collocated ECMWF wind vectors and QSCAT σ₀ measurements. Several elementary tests show the good quality of QSCAT-NN. Since the frequency of QSCAT is the same as this of NSCAT, they also tested the validity of a Neural Network NSCAT GMF (NSCAT-NN-2) for representing the QSCAT GMF It is found that NSCAT-NN-2 also is a good estimate of the GMF of QSCAT. Besides the authors have estimated a specific neural network for determining the conditional variance of QSCAT measurements following the previous works on NSCAT measurements