Joint retrievals

Consider a region of the thermal-infrared where the weighting functions peak in the upper troposphere and where, say, ammonia is strongly absorbing. Suppose that the simulated spectrum is less bright than the measured spectrum. This would suggest that the abundance of ammonia in our first simulated spectrum is too high near the peak of the weighting function and should thus be reduced. However, it could also mean that the assumed atmospheric temperatures are too low and must be increased. Or it could mean that the abundance of some other constituent such as aerosols needs to be modified. It could also conceivably mean that all three need to be modified in some way!

In such cases, a spectral region could first be selected that is sensitive only to well-mixed absorbers and a linear temperature retrieval conducted to give the "true" temperature profile. This temperature profile could then be used in an ammonia composition nonlinear retrieval, using another spectral region dependent on both temperature and ammonia. However, in some cases it is found to be more effective to consider both spectral regions together and retrieve both ammonia and temperature simultaneously in a nonlinear joint retrieval. Such an approach is useful in cases where the number of spectral points is limited or where there is significant noise.

Was this article helpful?

0 0

Post a comment