The Analysis of Capabilities of Neural Networks in CO2 Sounding with Spaceborne IPDA-Lidar with the Use of Different A Priori Data

Abstract

The possibility of retrieving, using a neural network, the columnar carbon dioxide concentration profile when sounding from a space orbit of 450 km and from a balloon at altitudes of 23 and 10 km are analyzed. The use of a priori data on temperature, pressure, and reflected and scattered signals is considered. The errors of retrieval of the columnar CO₂ are 0.15% and 0.5% at altitudes lower than 2 km for lidar with a telescope diameter of 1 m and laser pulse energy of 50 μJ at a resolution of 60 km.