SUMMARY OF CHANGES OF iSHAI VERSION 2018.1 WITH RESPECT TO PREVIOUS VERSION iSHAI VERSION 2016

GEO-iSHAI algorithm remains based in Jun Li’s algorithm for GOES Sounder physical retrieval algorithm (physical iterative approach with non-linear regression as first guess) but adapted for RTTOV and other GEO imager instruments. Jun Li (CIMSS-Wisconsin) algorithm is also the one proposed as Day-1 algorithm for the GOES-R processing by NOAA (Li, 2010).

GEO-iSHAI release 2018.1 has some major updates due to the introduction of the support of other GEO satellites than SEVIRI. In version 2018.1 Himawari and GOES-R class satellites are also supported. 

The use of RTTOV-11.2 in version 2018.1 implies that the number of RTTOV pressure levels remains in 54 RTTOV pressure levels. For this reason all the coefficients has been calculated using the GEO-iSHAI validation dataset for 2017 year with vertical dimensions adjusted to the used RTTOV pressure levels.

In previous versions there were separated EOFs coefficients files for temperature and humidity that after reading where combined to build the full EOF matrix. In version 2018.1, the full EOFs matrix is read from a single coefficient file.

The GEO-iSHAI code version 2018.1 has been also adapted to ABI instrument on board GOES-R satellites with minor changes in the code and the generation of a set of iSHAI coefficients for GOES-R. In version 2018 Himawari and MSG satellites were supported.

GEO-iSHAI Total Ozone (TOZ) output was introduced as a new output in release 2016. TOZ is calculated from the ozone profile after applying only the non-linear regression step. In version 2016 in the ozone non-linear regressions, the collocated real bias corrected satellite BTs and the temperature profile, the logarithm of ozone profile and the skin temperature from background NWP profile were used as inputs. Thus, neither the physical retrieval step nor the iSHAI retrieved profile (result of FG or physical retrieval step depending on BT_RMS_THREHOLD keyword) were not used in the ozone profile estimation. In version 2018.1; this has been changed and in the ozone non-linear regressions, the collocated real bias corrected satellite BTs and the temperature profile, the logarithm of specific humidity profile and the skin temperature from the iSHAI retrieved profile and the logarithm of ozone profile from the background NWP profile are used as inputs.
 
A summary of the changes of iSHAI version 2016 with respect to previous version former PGE13 SPhR version 2013 is available here.