Estimation of atmospheric temperature and humidity profiles from MODIS data and radiosond data using artificial neural network

Abstract

The aim of this study is to test the quality of the neural network for retrieving the temperature and humidity by comparison with the radiosond values and a linear regression method. Remote sensed images give useful information about the atmosphere. In this article, MODIS data is used to retrieve temperature and humidity profiles of the atmosphere. Two methods of linear regression and artificial neural network are used to retrieve the temperature and humidity profiles. A multilayer feed-forward neural network is tested to estimate the desired geophysical profiles. Retrievals are validated by comparison with coincident radiosond profiles.