Two Way Transmission Calculations

In the thermal-infrared, contribution functions (identical to the Jacobians, or functional derivatives, for temperature) are calculated analytically at every step of a NEMESIS inversion, and output in the nemesis.cov or kk.out files. However, when we wish to calculate the two-way transmission in reflected sunlight, a more convoluted use of individual subroutines is required.

Editing the PATH File:

You’ll need to write your own .pat file and run CIRSdrv_wave. In the .pat file you’ll split up the atmosphere into N layers as usual and then list N reflatm calculations that calculate the two-way transmission (at your choice of incident and reflected angles) to the bottom of whichever layer (defined by layer) you want.

Defining a reflected atmospheric path requires editing the .pat file to include N blocks of the following (one for each layer in your model):

• reflatm (this is the keyword)
• angles followed by the incident and reflected angles and the number of the layer from which the reflection occurs (reflection happens at base of layer)
• (no)cg Sets Curtis Godson flag as required
• (no)absorb Sets absorption flag as required

Note also that the calculated transmission may depend on how you choose to divide up the atmosphere into layers - typically we'd use LAYTYP=1 to subdivide the layers in equal units of log(p).

Running CIRSdrv_wave:

For the reflecting atmosphere paths, CIRSdrv_wave will calculate the transmission at each level as a function of wavenumber/wavelength. The transmission=0.5 level can then be estimated. Alternatively, the natural logarithm of the transmission provides the optical depth, allowing you to plot the unity optical depth level.

CIRSdrv_wave will require the following inputs:

• An operational filename (i.e., the root of your .pat file)
• ISPACE in wavelength or wavenumber.
• Whether to use ktables of line-by-line.
• Which planet you're modelling (to estimate solar distance).
• An option to use fast post-calculation or slow regridding (choose the former, which uses the wavelengths in your ktable).
• Which scattering model to use: scloudwave11 is the one used in regular NEMESIS runs.
• A surface temperature (not applicable to giant planets).

Notes

• The simple two-way transmission calculation ignores the multiple-scatteriing processes in which photons can bounce around multiple times before being absorbed (dependent on the single-scattering albedo). Incoming photons would be attenuated with some of these photons providing a source function offsetting those lost. The peak of this source function would be near tau_inbound = 1 and be distributed around it. Outbound photons would originate from roughly tau_outbound = 1, modified by the strength of the source function. The rigorous way to determine this would be to determine the contribution function, which is not trivial to calculate with scattered radiation. Thus, consider the two-way transmission method to be near the “top” of region from where UV photons originate.
• Note that Raman scattering will allow photons to scatter to longer wavelengths where photons may be able to make it deeper.
• When using the REFLATM block described above, note that 'scatter' is not a recognised keyword, so although you specify the use of a cloud scattering subroutine (scloud11wave.f), it's not actually being used because we don't have 'scatter' in one of the 'atm' blocks! So all with have is the extinction optical depth due to Rayleigh scattering, and no multiple scattering.
• The ‘scatter’ command only works in ‘atm’ blocks not ‘reflatm’ blocks, which purely calculate 2-way transmissions. If you do ‘atm’ that’s a calculation for the whole atmosphere (not individual layers) and CIRSdrv_wave will compute the scattered and thermally emitted radiation at the angles defined in the .sca file. There’s no simple way of computing contribution functions for scattering other than doing a load of scattering computations where you perturb the haze/cloud opacity in successive layers individually and then subtract from the calculation for the default unperturbed atmosphere. That's why you tend to see 2-way transmission calculations in papers.