Thanks a bunch for maintaining this code. I'm using this to compute radiative transfer through some Earth-like GCM output, and after switching from the older version of SBDART that was available online as a compressed archive (I forget where) to the version here, I find that more often than not SBDART fails to converge and I get a bunch of nans, where previously the code ran fine (though I suspect the output with IOUT=7 may not have been correct in the older version--thus moving to this version). This is for a static (no advection) cloud-free atmosphere. In addition to the usual too-large temperature step warning, I get "matrix near singular" errors from UPISOT, UPBEAM, and SOLVE0, and "convergence problems" from ASYMTX. I've attached my INPUT, atms.dat, usrcld.dat, and output files. Changing the output mode does not seem to lead to a change in behavior; I get nans no matter what.
Hi Paul,
Thanks a bunch for maintaining this code. I'm using this to compute radiative transfer through some Earth-like GCM output, and after switching from the older version of SBDART that was available online as a compressed archive (I forget where) to the version here, I find that more often than not SBDART fails to converge and I get a bunch of nans, where previously the code ran fine (though I suspect the output with IOUT=7 may not have been correct in the older version--thus moving to this version). This is for a static (no advection) cloud-free atmosphere. In addition to the usual too-large temperature step warning, I get "matrix near singular" errors from UPISOT, UPBEAM, and SOLVE0, and "convergence problems" from ASYMTX. I've attached my INPUT, atms.dat, usrcld.dat, and output files. Changing the output mode does not seem to lead to a change in behavior; I get nans no matter what.
Any thoughts?
All the best, Adiv Paradise University of Toronto
sbdart_in_out.zip