Speeding up the GOES-R Legacy Atmospheric Sounding Algorithm: a Blend of the CRTM Forward Model and the Fast Analytical Jacobian Scheme

The computing time is always a concern when it comes to the 1D-VAR atmospheric profile sounding retrieval. The GOES-R Algorithm Working Group have selected the community radiative transfer model (CRTM) as the operational RTM which is associated with a very time-consuming K-Matrix method for Jacobian calculation. Here we present a method to mix the CRTM forward model with a fast analytical Jacobian method in the legacy profile sounding algorithm. The test with EUMETSAT's SEVIRI measurements shows that our efforts significantly speed up the calculation with a comparable precision.