Information-centered representation of retrievals with limited degrees of freedom for signal: Application to methane from the Tropospheric Emission Spectrometer

Date: Journal Article

Venue: Journal of Geophysical Research Atmospheres

Citation:

Payne, V. H., S. A. Clough, M. W. Shephard, R. Nassar, and J. A. Logan (2009), Information-centered representation of retrievals with limited degrees of freedom for signal: Application to methane from the Tropospheric Emission Spectrometer, J. Geophys. Res., 114, D10307, doi:10.1029/2008JD010155.

Resource Link: http://www.agu.org/pubs/crossref/2009/2008JD010155.shtml

Remote measurements of trace gas profiles from nadir‐viewing instruments are often retrieved and/or reported on a fine grid containing more levels than the number of independent pieces of information in the measurement. Such profiles contain a priori information, which complicates interpretation. For scientific analyses of these data it is desirable to move to a representation in which measurement information is dominant and the influence of a priori information is minimal. Presented here is a postprocessing approach using a simple algorithm to transform each retrieved profile to an appropriate, geographically varying coarse grid. The representation is chosen such that the averaging kernel is close to unity for regions of the atmosphere where the retrieval has most information. The approach takes advantage of the sensitivity characterization allowed by retrieval on a fine grid, while reducing the influence of the a priori, accounting for spatial and temporal variations in the sensitivity of the measurement to the true atmosphere, and preserving obvious physical meaning in the end product. The example used to demonstrate the approach is the methane product from the Tropospheric Emission Spectrometer (TES), which contains 0.5–2.0 degrees of freedom for signal, depending on season and location. The TES methane has been postprocessed, and the end product has been compared with results from GEOS‐Chem, a global chemical model. Results show realistic latitudinal gradients from the TES data. Model/measurement differences also show large‐scale features over Indonesia that we attribute to tropical biomass burning in the summer/fall.