This paper presents the reconstruction of a wind field from three-beam scat
terometer measurements under the framework of a neural network. A neural ne
twork is adopted to implement the inversion of a geophysical model function
(GMF) that relates the scatterometer measurements of normalized radar cros
s section to surface wind speed and direction. To illustrate the functional
ity and applicability of the neural network, a set of wind fields generated
by means of the Monte Carlo simulation are used. At each sample point of t
he wind field, the speed and direction are simulated, Then, a GMF CMOD4 is
used to synthesize the normalized radar cross section at three pointing ant
ennas according to the ERS-1 configuration. In such a case, the neural netw
ork is constructed to model the inverse transfer function. For inputs, a pi
xel-based and area-based scheme are considered. The network training is acc
omplished by mapping input-output pairs that are randomly selected from the
database of simulated wind fields. The effectiveness of the neural network
as an inverse transfer function is validated. Four data sets of ERS-1 scat
terometer data over the western Pacific were selected for case study. Inter
comparison with other method concludes that the use of neural network has i
ts indispensable advantages and better retrieval accuracy can be obtained.