Neural Network Estimation of the GMFs of the ERS1/2 Scatterometers. Comparison with other GMFs.

   C. Mejia, S. Thiria, F Badran and M. Crepon (LODYC, UPMC)

  A new Geophysical Model Function (GMF) for the ERS-1
  scatterometer was computed by using neural networks (NN).
  The NN-GMF was calibrated with ERS-1 scatterometer sigma0
  collocated with ECMWF analyzed wind vectors. Systematic
  comparisons with the ESA CMOD4-GMF and the IFREMER
  CMOD2-I3-GMF were done. The NN-GMF RMS is better than the
  CMOD4 and CMOD2-I3 RMS. The dynamical range of the NN-
  GMF is smaller than the CMOD4-GMF and the CMOD2-I3-GMF.
  The NN-GMF gives smaller sigma0 values at high wind speed
  than the CMOD4 and CMOD2-I3 and larger values at small wind
  speed. A technic to compute error bars on the NN-GMF is
  derived.
  We also present a NN-GMF for ERS2. It is found that the two GMF
  are quite different

Keywords: ESA European Space Agency - Agence spatiale europeenne, observation de la terre, earth observation, satellite remote sensing, teledetection, geophysique, altimetrie, radar, chimique atmospherique, geophysics, altimetry, radar, atmospheric chemistry