The MIPAS instrument, operating on board of the Envisat satellite, with its broad band and high resolution measurements has posed a major challenge to operational retrieval techniques. The code developed for the operational analysis of MIPAS measurements has proved that near-realtime operation is possible and that a three-dimensional picture of the atmospheric composition can be retrieved from satellite observations. The features and performances of MIPAS retrieval code are briefly recalled.
The breakthrough in retrieval techniques obtained with MIPAS can be the basis for further significant improvements. The correlation that exists among the observations and among the target parameters are more rigorously and more efficiently accounted for if the retrieval is made handling simultaneously correlated observations and correlated target parameters. This is the case of two-dimensional retrieval and multi-target retrieval. Another direction for further improvements is that of the effect of model errors. These can be better accounted either with the use of the variance covariance matrix of these errors or with the simultaneous retrieval of the main model errors. These retrieval approaches can be very demanding in terms of computing resources, but they change the ultimate accuracy possible with remote sensing techniques and must be taken into account in the planning of future instruments for environment and climate.