Gome ozone profiles retrieved by neural network techniques: A global validation with lidar measurements
Résumé
Ozone profiles retrieved from Global Ozone Monitoring Experiment (GOME, flying on ERS-2 satellite) spectra from July 1995 to June 2003 by means of 2 independent neural network (NN) schemes have been validated with ozone lidar measurements performed at different stations belonging to the network for the detection of atmospheric composition changes (NDACC). The retrieval and the whole validation have been carried out by using the performances and resources of the European project Enabling Grid for E-sciencE (EGEE) and of a local Grid at the European Space Research Institute of the European Space Agency (ESRIN/ESA). Roughly 1800 collocated profiles have been found, in tropical, mid-latitude and high-latitude regions; for each lidar station the differences between GOME and lidar profiles have been evaluated and the global performance of the proposed NN approaches has been critically discussed. The results indicate the potentialities for obtaining reliable ozone field analysis on global scale, including detailed altitude resolved trend analysis.