RFM Spectroscopic Data

10JAN21

Line Parameters - HITRAN .par files

List of HITRAN parameters used by the RFM
SymbolParameter Units
ν0 Transition Wavenumber cm-1
S Line Intensity cm-1/(molec cm-2)
γair Air-broadened width cm-1 atm-1
γself Self-broadened width cm-1 atm-1
EL Lower-state energy cm-1
n Temperature dependence of γair [n/a]
δ Pressure shift cm-1 atm-1
V Vibrational quanta (for non-LTE) [n/a]
Q Rotational quanta (for line-mixing) [n/a]
HITRAN [
Web-site] is the most widely-used compilation of spectroscopic line parameters. The basic form consists of text records of 160-characters (.par file), one record for each transition, ordered by increasing wavenumber. The 2016 edition contained over 5 million records spanning 0–36000 cm-1. These are largely selected on the basis of their relevance for modelling the earth's atmosphere.

The parameters used by the RFM are listed in the table on the right.

While the RFM v5 can use this .par file directly, it can be quite slow, and the recommendation is to use the hitbin program to convert it to a binary format.

There are also some local FORTRAN programs for manipulating the HITRAN parameter files:

subhit
Subset HITRAN .par file. Can be used to reduce original .par file to containing only molecules/spectral range of interest.
mrghit
Merge HITRAN .par files. Can be used to merge different .par files, eg to replace molecule within the original file with a different set.

Line Parameters - HITRAN database

List of selected HITRAN database parameters in the RFM-basic format
SymbolParameter Units Format
ν Transition wavenumber cm-1 F12.6
- Molecule ID I2
- Isotopologue ID I2
S Line Intensity cm-1/(molec cm-2) E10.3
E" Lower-state energy cm-1 F10.4
γair Air-broadened halfwidth cm-1 atm-1 F6.4
γself Self-broadened halfwidth cm-1 atm-1 F5.3
nair Temperature exponent of γair F7.4
nself Temperature exponent of γself F7.4
δair Pressure shift (air) cm-1 atm-1 F9.6
δself Pressure shift (self) cm-1 atm-1 F9.6
Yair Line Coupling (air) cm-1 E10.3
Yself Line Coupling (self) cm-1 E10.3
- Transition ID I12
A limitation of the standard HITRAN .par format is that it does not allow for any additional line parameters beyond the basic set required for the standard Voigt lineshape.

Nowadays HITRAN is actually a proper database, and the 160-character .par line list is just one way of accessing the information. In fact the user has complete flexibility in extracting any subset of parameters in any order in different formats. Unfortunately this flexibility isn't conducive to getting data into the RFM which expects a fixed record structure.

The immediate solution (for RFM v5.10 onwards) is to define an 'RFM-basic' structure containing the variables shown by the table on the right - these are all parameters selectable via the HITRAN database menu system once you get to 'Select Output Options'. The other things you'll need are

  • Field separator: [space]
  • Line endings: Linux/Unix/Mac (LF)
  • Tick box for Fixed Width Format
  • Tick box for Output header line
Assuming that all works the resulting file should like something like this: [Sample]. Note that the field names in the header record aren't aligned with the corresponding positions in the data record. I realise it would be simpler all round if other HITRAN users could download data with this pre-defined format that I've set up for myself - I'll have to ask the HITRAN people about that.

This structure represents the mininum set of parameters the RFM needs for the standard Voigt line shape plus parameters required for first-order line-coupling. And since the database also allows for 'self' (rather than just 'air') values of temperature exponent and pressure shift, I've added those as well to improve the modelling of non-terrestrial atmospheres (eg CO2 on Mars). The Transition ID is just a unique index assigned by HITRAN for each line - the RFM doesn't need it but I thought it might be useful to extract it.

I've had to omit vibrational quanta from this structure since trying to decode the way these are returned by the database is just too horrible to contemplate (it's not just the 15-character strings used in the .par file). So you can't use the RFM-basic structure for non-LTE or band selection. If it's an issue, let me know and I'll come up with something.

But ... before you get too excited: just because the HITRAN database allows for all these extra parameters doesn't actually mean that values have been provided; in most cases you'll find these fields filled with '#####' characters, indicating missing data (as you'll see from the [Sample]). In these cases the RFM defaults to using the air values (or if Yair is also missing, 0.0).

For now, more realistically, if you have access to such information as line-coupling parameters, you can use this as a template to get your own data into the RFM, which is what I've had to do to test it.

Once I'm satisfied that this works, the next step will be to define an RFM-HTP structure to include extra parameters required for modelling the Hartmann-Tran lineshape (which also requires different line-coupling values).

Absorption Cross-sections

Certain molecules, usually the more complex type, have lines which are too closely spaced to be resolved. For these, the absorption coefficient k(ν) [cm2/molec] is tabulated directly from lab measurements under particular pressure and temperature conditions. These are the HITRAN Absorption Cross-Section data files.

In HITRAN 2016 the data for each tabulation of k(ν) at a particular (p,T), (p in this case is measured in Torr, 760 Torr = 1 atmosphere) is provided as a separate file. Each tabulation is on a regular spectral grid, although not necessarily the same for all (p,T) — at lower pressures the spectral features start to resolve and higher spectral sampling is necessary.

The RFM expects cross-section data as a single file for each molecule, which can be constructed by simply concatenating the HITRAN files in any order of (p,T). Some caution is required here: the HITRAN data are often a collection from different sources and simply combining all the available data can lead to systematic errors when interpolating in the (p,T) domain (particularly noticeable when evaluating Jacobians). You are advised to only concatenate data from a single source.

Eg using the linux cat command to create a combined file sf6.xsc for data in the spectral range 925–955 cm-1

cat SF6*925*xsc > sf6.xsc

However, where the cross-section data covers different bands (eg 810–880 cm-1 and 1050–1120 cm-1 for CFC-11) all the data for each band should be concatenated first, and then data for the individual bands concatenated. [Note: for RFM v5.10 and later versions, the files can be concatenated in any order] For example, using the linux cat command with the CFC-11 (CCl3F) data to create a combined file f11.xsc

cat CCl3F*810*xsc CCl3F*1050*xsc > f11.xsc

HITRAN Collision-Induced Absorption

Collision Induced Absorption data are a relatively new (since 2012) third type of HITRAN spectroscopic data. As with
absorption cross-sections these are tabulations of absorption coefficient as a function of wavenumber c(ν) but representing the absorption due to collisions between a specific pair of molecules rather than generic pressure broadening. As a consequence the absorption coefficient c is in units of cm5/molec2 (so that, mulitplied by the product of the densities of the two molecules, the result cρ1ρ2 is cm-1, as with conventional absorption kρ). The tabulations are only functions of temperature.

The data are available as single files for each pair of molecules, with all temperatures and spectral ranges contained in the same file. The ordering is band-by-band, and within each band, tabulations for different temperatures.

The RFM can use these data files directly.

  • *CIA section of driver file

HITEMP Data

HITEMP data is a line parameter compilation including many more transitions for the major molecules which are only apparent at higher temperatures.

Since it is in the same 160-character format as the standard HITRAN line list the RFM can use these files either directly (RFM v5) or converted to binary format using hitbin. However since HITEMP contains only a limited set of molecules you might first want to merge this with the standard HITRAN list using mrghit.

RFM v5 uses the 2016 Gamache TIPS functions which are valid up to 9000K.

UV Data

Radiative transfer in the ultra-violet is dominated by molecular scattering so the RFM cannot be used to calculate meaningful radiances. However it is still useful for calculating the molecular absorption component.

The HITRAN 2016 line database extends into the UV (>25 000 cm-1, <0.4 µm) but only for a limited set of molecules with well-resolved lines: H2O (<25 711 cm-1), OH (<35 875 cm-1), HF (<32 352 cm-1) and H2 (<36 406 cm-1)

For other gases, absorptions are available as cross-section tabulations, in the same format as for the infrared heavy molecules, and so can be handled by the RFM in the same way. Note that for molecules which are normally represented in the infrared by HITRAN line data (eg O3) to specify cross-section data you have to append the letter 'x' (eg 'O3x') in the RFM driver table *GAS section — see also Recognised Molecules below, molecule numbers 171–184.

GEISA Data

GEISA [Web-site] is an alternative source of spectroscopic line parameters and cross-section parameters.

GEISA line data [format] contains the same set of essential parameters required by the RFM, but in a different format. Also the list of molecules and isotopologues is slightly different to HITRAN. The RFM cannot read GEISA data records directly, but a Fortran program is provided to convert GEISA to HITRAN .par format, which can then be read directly by the RFM, or converted to binary.

geihit
Convert GEISA 2015 data to HITRAN 2016 .par format

GEISA absorption cross-section data [format] are also in a different format to HITRAN (although often from the same orginal source). As with the line parameter data, there is a Fortran convversion program:

geixsc
Convert GEISA 2015 absorption cross-section data to HITRAN .xsc format

Recognised Molecules

The following is a list of all the molecules (Names in Red) recognised by RFM. If the last letter is 'q' this is a line molecule which is normally expected to be represented by tabulated cross-section data, and vice-versa for 'x'.

Internally the RFM assigns each molecule a number (set in subroutine molidx_sub.f90), although for tabulated cross-section data (.xsc files) it is possible to add arbitrary molecules which are, internally, assigned temporary numbers. This is possible for cross-section molecules since it is only necessary to associated the molecule's name with a matching cross-section file and atmospheric profile. However, for molecules described by line parameters, additional information such as molecular weight (for Doppler widths), permitted isotopologues and TIPS data is also required, requiring amendments to the code.
Molecules 1–49 follow the assigments for HITRAN 2016, 50–53 are extra molecules assigned these indices in the associated TIPS data, and 60–63 are assigned to additional molecules defined in the GEISA database.
ID Molecule ID Molecule ID Molecule ID Molecule ID Molecule
1 H2O 2 CO2 3 O3 4 N2O 5 CO
6 CH4 7 O2 8 NO 9 SO2 10 NO2
11 NH3 12 HNO3 13 OH 14 HF 15 HCl
16 HBr 17 HI 18 ClO 19 OCS 20 H2CO
21 HOCl 22 N2 23 HCN 24 CH3Cl 25 H2O2
26 C2H2 27 C2H6 28 PH3 29 COF2 30 SF6q
31 H2S 32 HCOOH 33 HO2 34 O 35 ClONO2q
36 NO+ 37 HOBr 38 C2H4 39 CH3OHq 40 CH3Br
41 CH3CNq 42 CF4q 43 C4H2 44 HC3N 45 H2
46 CS 47 SO3 48 C2N2 49 COCl2 50 SO
51 C3H4 52 CH3 53 CS2
60 GeH4 61 C3H8q 62 HNC 63 C2H6q
Key Standard HITRAN line molecules
HITRAN line data normally/better represented as cross-sections
Additional line molecule indices defined in 2016 TIPS data
Additional line molecule indices defined in the GEISA database
Molecules 100+ are RFM-specific indices for species represented by infrared absorption cross-section data.
ID Molecule ID Molecule ID Molecule ID Molecule ID Molecule
100 Aerosol
General Heavy Molecules
101 ClONO2 102 N2O5 103 SF6 104 CCl4 105 HNO4
106 SF5CF3 107 BrONO2 108 ClOOCl 109 X109 110 X110
CFCs (ChloroFluoroCarbons)
111 F11 (CCl3F) 112 F12 (CCl2F2) 113 F113 (C2Cl3F3) 114 F114 (C2Cl2F4) 115 F115 (C2ClF5)
116 F116 (C2F6) 117 F13 (CClF3) 118 F14 (CF4)
HCFCs (HydroChloroFluoroCarbons)
121 F21 (CHCl2F) 122 F22 (CHClF2) 123 F123 (CHCl2CF3) 124 F124 (CHClFCF3) 125 F141b (CH3CCl2F)
126 F142b (CH3CClF2) 127 F225ca (CHCl2CF2CF3) 128 F225cb (CClF2CF2CHClF)
HFCs (HydroFluoroCarbons)
131 F125 (CHF2CF3) 132 F32 (CH2F2) 133 F134 (CHF2CHF2) 134 F134a (CFH2CF3) 135 F143 (CF3CH3)
136 F143a (CF3CH3) 137 F152a (CH3CHF2) 138 F365mfc
MHCs (Methylated Hydrocarbons)
141 CH3OH (Methanol) 142 CH3CN (AcetoNitrile) 143 CH3CHO (Acetaldehyde) 144 Acetone (CH3COCH3) 145 PAN (CH3C(O)OONO2)
NMHCs (Non-Methylated Hydrocarbons)
151 C2H6x (Ethane) 152 C3H8 (Propane) 153 C6H6 (Benzene) 154 C2H2x (Acetylene) 155 C2H4x (Ethylene)
Halocarbons
161 C4F8 (Octafluorocyclobutane)
UV Cross-sections of simple molecules
171 O3x 172 N2Ox 173 SO2x 174 NO2x 175 H2COx
176 BrO 177 NO3 178 OClO
UV Cross-sections of heavy molecules
181 C7H8 (Toluene) 182 oxylene (o-C8H10) 183 mxylene (m-C8H10) 184 pxylene (p-C8H10)
Key Extinction cross-section [km-1]
Molecular cross-section [cm2/molec]
Dummy names for user-supplied cross-section molecules
Cross-section normally represented as HITRAN line data
Additional GEISA Cross-section molecules
UV Cross-sections
Some molecules are represented by both HITRAN line parameters and as cross-section data.

  • Where the RFM default is to use cross-section data (eg ClONO2, ID=101) the user can select the line data instead by adding 'q' after the molecule name (eg ClONO2q, ID=35).
  • Where the default is to use line data (eg C2H6, ID=27), the cross-section data can be selected by adding 'x' after the molecule name (eg C2H6x, ID=151).