This is a bug report from 14 Sep 2004. It describes a problem with RT where energy is not conserved. The original description is as follows (the "other problems" referred to below are probably solved by now, I think they were convergence problems in the molecular solver):
I have dug up the test case that shows the RT problem. You need to change GrnVryDepth to run it. The model runs for a long time, and it runs into lots of other problems which muddles the issue. The problem is that if you integrate over the transmitted spectrum, it has far less energy than the incident spectrum. Since this is a closed geometry, the energy in the spectrum should be conserved.
I have not modified anything from the original input file / code. I also did not test whether it still produces the same behavior with the current code.
