Convergence Space Experiment: Retrieving the Water-Vapor Profile of the Atmosphere by Means of Artificial Neural Networks

In this work, the possibility of retrieving the profile of absolute humidity of the atmosphere by means of an artificial neural network is investigated based on modeling radiometric data of the MIRS passive microwave complex, which is a part of the scientific equipment of the Convergence space experiment. The process of modeling MIRS radiometric data is described. Optimum characteristics of the neural network are selected. The necessity of information about the atmospheric temperature profile for the best accuracy in solving the inverse problem is shown. The advantages of using “differential” channels in the 22-GHz absorption band for retrieving the humidity profile are demonstrated. The expected errors in retrieving the atmospheric humidity profile in the course of the Convergence space experiment at heights from 0 to 10 km are presented.