Least Dependent component Analysis for trace gases retrieval from satellite data

A new approach is proposed for the retrieval of trace gases, vertical column concentrations, from hyperspectral satellite observations. This technique consists of a semi-blind recovery of concentrations and pure spectra from their linear mixtures. This decomposition is based on the Least Dependent component Analysis (LDA) technique, a method that generalizes the well known Independent Component Analysis, to find least dependent components, after a proper preprocessing for reducing some residual dependencies. Some results are shown for the retrieval of sulphur dioxide SO₂ volcano emission using data from the NASA Ozone Monitoring Instrument (OMI).