A Neural Network Approach for Wind Retrieval from the ERS-1 Scatterometer Data Part 1 - Determination of the Geophysical model function of ERS-1 Scatterometer for OCEANS 94 OSATES C. Mejia, S. Thiria, M. Crépon Laboratoire d'Océanographie Dynamique et de Climatologie, Lodyc, UPMC-CNRS, Paris F. Badran Conservatoire Nationale des Arts et Metiers Cedric, CNAM, Paris Abstract — The objective of the present work is to compute a new Geophysical Model Function (GMF hereinafter) for the ERS-1 scatterometer by the use of neural networks (NN hereinafter). This NN-GMF is calibrated with ERS-1 scatterometer sigma0 collocated with ECMWF analysed wind vectors. In order to check the validity of the NN-GMF systematic comparisons with the ESA's CMOD4-GMF (version 2 of 3-25-93) and the IFREMER's CMOD2-I3-GMF are done. The GMF is used in many algorithms to retrieve the scatterometer wind. Résumé — Nous proposons dans ce papier une méthode pour calculer la fonction du modèle geophysique (GMF) du diffusiometre du satellite ERS-1. Cette méthode est fondée sur des techniques de réseaux de neurones (NN). La fonction anisi obtenue, NN-GMF, est étalonnée grâce à la collocation des sigma0 mesurés par ERS-1 avec les vecteurs de vent donnés par le modèle ECMWF. La validité de la méthode NN-GMF est verifiée en la comparant avec la méthode CMOD4-GMF (version 2 du 25-3-93) de l'ESA et avec la méthode CMOD2-I3-GMF de l'IFREMER. 1. INTRODUCTION The transfer function allowing to compute the wind from the scatterometer signal is very difficult to be determined. It is a non linear function which may have ambiguities on the direction. Several algorithms have been proposed to model the wind retrieval transfer function. Most of them are based on the inversion of the Geophysical Model Function (GMF) which gives the sigma0 with respect to the wind vector. The study of the GMF is then of a fundamental interest. Furthermore the GMF can give useful information on the behaviour of the scatterometer. The present study is devoted on the modelling of the GMF by the use of Neural Networks. Neural Networks have been used with success by the present team to retrieve the wind vector from the ERS-1 scatterometer data. The methodology is described in [2]. In the present study we propose to determine a new GMF the Neural Networks is calibrated onto ECMWF analysed wind vectors collocated with scatterometer sigma0. 2. THE NEURAL NETWORKS GEOPHYSICAL MODEL FUNCTION (NN-GMF) ALGORITHM A. The geophysical Problem Scatterometers are active microwave radar which accurately measure the power of transmitted and back scatter signal radiation in order to calculate the normalised radar cross section (σ 0 ) of the ocean surface. The σ 0 depends on the wind speed, the incidence angle (which is the angle θ between the radar beam and the vertical at the illuminated cell, Fig. 1) and the azimuth angle (which is the horizontal angle χ between the wind and the antenna of the radar). Empirically based relationship between σ 0 and the local wind vector can be established which leads to the determination of a geophysical model function. θ antenna 1 antenna 2 χ Satellite trajectories swath direction wind wind antenna 3 satellite track Fig. 1. Definition of the different geophysical parameters. Recently different GMF have been proposed for the ERS-1 scatterometer. One can mention the IFREMER CMOD2-I3- GMF which is similar to (1) and is denoted now: σ 0 = a + b cos(χ) + c cos(2χ) (1) and the ESA CMOD4-GMF which is of the form: