Design of Neural Network algorithms
for the retrieval of tropospheric ozone from satellite data

F. Del Frate, P. Sellitto, D. Solimini

Abstract

This paper reports on the development of Neural Networks algorithms for tropospheric ozone retrieval from ESA-Envisat SCIAMACHY measurements. We present and discuss some retrieval experiments from UV/VIS simulated data, based upon a combined sensitivity analysis performed with the aid of the UVSPEC radiative transfer model and a neural Extended Pruning procedure. In particular, the role of UV and VIS information budget is here exploited and critically discussed.

Key words: tropospheric ozone, neural networks, air pollution

Proc. ‘Envisat Symposium 2007’, Montreux, Switzerland, 23–27 April 2007 (ESA SP-636, July 2007)

© 2007 ESA.

.