Retrieval of Height Profiles of Trace Gases by Optimal Parameter Estimation

Abstract

While most experiments on the subject of atmospheric physics emphasize on the detection of ozone exclusively, the GOME project was designed to monitor about a dozen additional trace gases in order to investigate the process of ozone depletion. GOME is part of the ERS2 misson and was launched in April 1995.

By the means of GOME it is possible to obtain total coverage of the earth’s atmosphere for the first time within only three days in a spectral range from 240 nm to 790 nm.

Effective processing of this tremendous amount of data is only possible by developing fast retrieval algorithms. While there are several approaches for a forward model, a fast solution for the invers problem is not found yet, nevertheless important.

The classical way to solve this problem is a Gauss-Newton iteration, where the forward problem is solved for each set of parameters, while the parameters representing the height profiles of trace gas concentrations are altered for optimized output. This implies repeated calculations of the forward model and of weighting functions making this process time consuming.

This workgroup investigates a different approach. Since the forward model is based on an integro-differential equation that is solved by means of finite elements, it is possible to combine the outer parameters with the ones describing the lattice. By doing so, it is possible to solve the invers problem while iterating a slightly modified forward problem thus avoiding to solve the time consuming forward problem many times. This approach is expected to be at least two orders of magnitude less time consuming than the former.