## *FIN Section | ||

## RFM Driver Table Sections: Internal spectral grid spacing | ## 26FEB24 |

**Type**- Secondary Section,
activated by
FIN Flag.
**Description**- User-specified Resolution for Fine Mesh Calculation when used in
conjunction with ILS convolution
(to replace default value of 0.0005 cm
^{-1}). **Format**- Single field
Field Type Description Units Range RESFIN D Wavenumber cm ^{-1}0< : ≤1 Or Frequency (if GHZ Flag set) GHz 0 < : I N.Pts/cm ^{-1}cm 1 ≤ : **Type**: I=Integer; D=Double Precision; **Notes**- In the absence of any spectral convolution
the RFM generates spectrally
*sampled*output at the resolution RESSPC specified in the *SPC section of the driver table, so this is also fine mesh resolution for calculations. However, if any spectral averaging is required, either by an explicit instrument line shape convolution (ILS Flag) or using the AVG Flag, the fine mesh resolution for the calculations defaults to 0.0005 cm^{-1}(set by 1/`DEFFIN`in`rfmcon_dat.f90`), irrespective of the output resolution, unless modified by the value supplied here. - The RFM default value of 0.0005 cm
^{-1}is appropriate for mid-infrared limb-viewing calculations, and is set by the requirement to capture the area of Doppler-broadened lines (typical half-width 0.001 cm^{-1}). For nadir-viewing infrared, 0.001 cm^{-1}is considered adequate. The Doppler width narrows in proportion to wavenumber, so microwave limb-viewing is the most extreme case where a value of 30 000 is required to fully capture the Doppler core of a line. - If grid spacings larger than 1 GHz are required, these have to be expressed
in points per wavenumber (1 GHz being approximately 30 pts/cm
^{-1}). [this seems unnecessarily obtuse - I should probably fix this!]

- In the absence of any spectral convolution
the RFM generates spectrally
**Example***FIN 1000 ! Set internal grid to 1000pts/cm-1 (0.001 cm-1 spacing)**Bugs**- [none recent]