REFERENCE FORWARD MODEL
The RFM is a GENLN2-based line-by-line radiative transfer model originally
under an ESA contract to provide
reference spectral calculations for the
It subsequently developed into a general purpose code suitable for a
variety of different
The main algorithm description is in the 2017 paper listed below under
The original code was written in FORTRAN77 and this will continue to be
maintained for the time being with various bug-fixes but no new features.
All future development is in RFM v5,
which is a complete rewrite in FORTRAN90 and compatible with HITRAN 2016
and GEISA 2015.
RFM maintenance and development is currently supported by the
UK National Centre for Earth Observation
The main design features are
I'll consider adding other features that people would find useful, including
combinations of features which are currently not allowed. Just let me know,
along with any bug-reports.
- Easy to use - probably its main advantage
- Robustness - you shouldn't be able to crash the program
- Portability - written in (almost) standard FORTRAN90.
- Self-documented code - just in case you need to make you own modifications
- On-line documentation - these web pages
- Output spectra independent of spectral range
(eg value at 901.3 cm-1 will be the same whether you produce a spectrum from
901.2-901.4 cm-1 or from 500-1000 cm-1)
To get a copy, email me with a brief statement on what you're planning on
using it for (which helps me to anticipate future user requirements)
and I'll email you the source code.
( email@example.com )
The RFM is designed to use the three main types of HITRAN spectroscopic data
There are additional HITRAN data sets such as HITEMP (line data) and UV
(cross-section data) which can be used, with some care.
- Line-by-Line data (160-character .par files), providing the basic
parameters for each transition (wavenumber, intensity, Lorentz half-width
- Absorption Cross sections (.xsc files) representing
tabulations of k(ν,p,T)
where k is in cm2/molecule
- Collision Induced Absorption (.cia files) representing tabuluations of
c(ν,p,T) where c is in
GEISA is an alternative to HITRAN, providing similar line and cross-section
data but in a different format.
While the RFM cannot use GEISA data directly, local
software is available to convert these to HITRAN format.
Things the RFM can handle...
(Not all of these are compatible -
see User's Manual for the full
list of 'flags').
- Spherical or plane-parallel atmospheres, or homogeneous paths
& Instrument Line Shape
- CO2 Line mixing
- Continua for H2O, O2, N2 and CO2
- Non-LTE (via user-supplied vibrational temperatures)
- Select lines of particular isotopes and/or bands
- Jacobians (Weighting functions) for p, T, VMR,
line-of-sight pointing and
surface temperature and emissivity
- Satellite/Balloon & Aircraft/Ground-based viewing geometries
- Surface reflections
- Output spectra of radiance
- Output diagnostics from ray-tracing (including Curtis-Godson integrations)
- Creation/Use of Look-Up Tables of
absorption coefficient k(ν,p,T)
- Horizontal structure in the atmosphere
- Flux calculations
- Different isotopic mixing ratio profiles
Things the RFM cannot handle...
- Sunlight (there is no 'sun' modelled in the RFM)
- Scattering (too difficult)
- Line-mixing other than CO2 (working on it...)
- super-Voigt line parameters (working on it...)
If you publish anything involving the RFM in the scientific literature
I do not expect to be a co-author, or even consulted. However,
I would appreciate it if you sent me a reference for the final publication since
it helps me obtain funding for RFM support.
If you want to cite the RFM, the following is recommended:
- Dudhia, A
Reference Forward Model (RFM)
- Journal of Quantitative Spectroscopy and Radiative Transfer
- 186 (2017) 243–253
Atmospheric, Oceanic and Planetary Physics