GLOBAL TROPOSPHERIC OZONE RETRIEVALS FROM OMI DATA BY MEANS OF NEURAL NETWORKS

摘要

In this paper, the design and validation of a neural networkrnalgorithm for tropospheric ozone column retrievalsrnfrom Ozone Monitoring Instrument (OMI) are discussed.rnThe algorithm estimates the ozone column between thernsurface and the thermal tropopause by combining OMIrnultraviolet reflectances with temperature and tropopauserninformation, as well as with a first guess for the troposphericrnozone column provided by a satellite climatology.rnThe algorithm was extensively validated againstrnozonesonde measurements, and first comparisons werernmade with Chemistry/Transport Model (CTM) simulations.rnThe results show encouraging retrieval capabilities.