Subject: 	RE: RTTOV Query Ref: 10987/10988
Date: 	Mon, 8 Dec 2014 17:18:15 +0000
From: 	Hocking, James <james.hocking@metoffice.gov.uk>
To: 	'patrick.eriksson@chalmers.se' <patrick.eriksson@chalmers.se>



RTTOV query

Hi Patrick,

FASTEM is not available as a separate module so you will have to extract
it from the RTTOV code unfortunately.  It is also unfortunately the case
that the FASTEM code as it stands is not well documented: there are now
6 versions of the model which have involved incremental changes and
additions to the code.

I recommend looking at Table 14  (p42) in the RTTOV v11 user guide which
lists the 6 versions and gives references which provide information on
the science behind the code and should help decipher what is going on.
FASTEM-6 is now the recommended version because it fixes a problem with
the azimuthal dependence which has existed since at least FASTEM-4.
There are some caveats regarding the use of FASTEM-6 in Table 14.

The coefficients used by FASTEM 1-3 are in
src/main/mod_rttov_fastem3_coef.F90.

The coefficients used by FASTEM 4-6 are in
src/main/mod_rttov_fastem5_coef.F90.

FASTEM 4-6 are implemented in rttov_fastem5.F90. This subroutine is
called by rttov_calcemis_mw.F90 and this latter subroutine also
implements FASTEM 1-3.

There shouldn’t be any problems if you want to implement FASTEM in ARTS
in respect of licenses. However we would like to ensure that if FASTEM-X
is implemented in ARTS then it is the same FASTEM-X as is implemented in
RTTOV. FASTEM has also been included in the CRTM (a fast model similar
to RTTOV) for a long time, but it recently became apparent that there
are significant differences in the implementations of FASTEM-5 between
RTTOV and CRTM. This is a situation we would like to avoid in the future.

Our plans for RTTOV v12 (due for release end of 2016) include
consolidating the FASTEM versions into a single version. We hope to also
make this available as a stand-alone module/subroutine which other RTM
developers can use. This will be based on FASTEM-6.

Regarding your specific questions:

-Frequency is in GHz

-You are right: emissivity(1:4) corresponds to eqn 13a-d in the FASTEM-4
paper.

-In RTTOV the surface-reflected downwelling atmospheric emission is
treated according to the assumption of specular reflection when the RT
equation is integrated. FASTEM modifies the surface emissivity in such a
way to account for the anisotropic downwelling radiation.
“Transmittance” is the top-of-atmosphere to surface transmittance. This
is described in http://nwpsaf.eu/deliverables/rtm/evalfastems.pdf. I’d
assume for ARTS that you would not want to include this since ARTS is a
fully 3-D RT model (if I remember rightly!)

Best regards,

James

