Retrieval of carbon dioxide vertical concentration profiles from satellite data using artificial neural networks

Abstract

Abstract. In this paper, we derive vertical distributions of carbon dioxide atmospheric concentration from satellite data using a retrieval algorithm based on an artificial neural network (ANN) technique. Sensitivity studies were made to selec tthe most appropriate sensor channels. A MultiLayer Perceptron (MLP) ANN was implemented and tested for a large and diversified dataset. Here we focused on the retrieval of vertical Carbon Dioxide concentration profiles using SCIAMACHY channel 6 (1000-1700 nm) nadir measurements. The results show we can accurately and efficiently obtain carbon dioxide profiles by using this approach.