NetCDF start date when restarting smrf_ipysnobal crashed run

[micah-prime]

The way this is written write now, if a smrf_ipysnobal run (when forcing data is passed in the queue) is restarted, the smrf start date must also change. This should be fine as long as the forcing data is not output as well (currently not an option for this type of run). This issue this could cause is that the re-initialized smrf run will try to set a new start date for indexing in the netCDF, and this could cause issues. Not very pressing right now.