I have raised some questions regarding the usage of MODTRAN by Ed Caryl on NoTricksZone. Ed Caryl has applied MODTRAN here, here and here. In the last of those blog posts Ed Caryl has made a validation run and compared a MODTRAN run with observations of downwelling radiation (also known as back radiation) at the South Pole. However, he also states that the version of MODTRAN used by him is different than the version connected to the web interface offered by the University of Chicago used by me.
For the purpose of comparison I have carried out the following two MODTRAN runs at South Pole conditions. The pictures below show the input data, the diagrams of the back radiation and part of the atmospheric profiles for the two runs.
I have got similar results as Ed Caryl for the South Pole (his diagram is to the right here) with a very low temperature at the ground of 228 K (-45 C) which may well be representative of the South Pole. I have also assumed that the height of the South Pole is 3 km. The legend in the upper right corner of the diagrams confirms that we are not using the same MODTRAN version.
My first run is with the original profile of the relative humidity for the 1976 US Standard Atmosphere. This is obtained by using the parameter Water Vapor Scale equal to one. The output file shows that in that case the relative humidity is a little less than 50% in most of the troposphere. However, although with similarities the resulting diagram from this first run is much less similar to the diagram by Caryl than the one from the second run.
In the second run I have used 50 as the value of the Water Vapor Scale parameter. The output file shows that the MODTRAN program does not allow greater values than 100% for the relative humidity. Consequently with this value the relative humidity becomes 100% in the whole troposphere.
From this behaviour of the MODTRAN program follows that you get the same relative humidity profile and the same simulation results with the Water Vapor Scale parameter equal to 50 and equal to 90. This agrees with what Ed Caryl found in his MODTRAN runs where a parameter, interpreted by him as relative humidity, was varied from 50 to 90.
However, since we are using different versions of MODTRAN it is possible that the version used by Ed Caryl has another way of handling the relative humidity profiles than the version used by me.
Right click for bigger pictures.