CRAZY!! PROBLEM in WRF + SWAN in namelist.input

It took me two days to narrow down this issue, but I think it is quite crazy and worth mentioning it here. Maybe someone else has already experienced it or it is a well known issue, I am quite new to COAWST so I don't know.

I discovered it because I didn't start from the Sandy case to build my WRF case, but using my own files instead.

**If for any reason the keyword time_step appears AFTER the keyword max_dom in the WRF namelist.input file, the simulation will crash with a segmentation fault error or hang, depending how far these two parameters are in the wrong order.**

My hypothesis is that max_dom is read by COAWST and then some calculations are done with it, in which the time_step is required. If the latter has not been read yet, there is your segmentation fault.

I was able to reproduce this with both my case and with the Sandy (in WRF + SWAN mode).

Hope this helps somebody. It certainly would have helped me.

SS

Reproduce the failure in the WRF + SWAN Sandy case using namelist.input, placing time_step after max_dom and comparing the resulting crash or hang with the correct ordering. Trace how these two keywords are read and verify that the simulation completes regardless of their order.

Written by the indexing model from the issue text.