glacier (glc) output in ./cat ?

[gmao-rreichle]

Diagnostic (MAPL HISTORY) output from landice ("glc") ends up in the ./cat directory, where "cat" is short for catchment, i.e., non-glaciated land. This is somewhat inconsistent. I think we have two options to fix this:

1. rename ./cat to ./diag (as in "model diagnostics").
2. add another output dir ./glc (next to ./cat)

I think option 1) would match what the GCM and DAS are doing. It's a bit more invasive because we change a long-standing output dir name. To mitigate, we could add a link name ./cat that points to the new ./diag dir.

Option 2) is less invasive but it may require hardcoding collection names (so the post-proc script knows what to put into ./cat and what to put into ./glc). That'd be somewhere between hard-to-maintain and impossible (Collection names can be made up by users as they please).

A third option may be to do nothing and just accept the inconsistent naming. But looking ahead to integrating lakes into GEOSldas, it'd only get worse, so it may be best to fix it now.

Thoughts?

@weiyuan-jiang @biljanaorescanin @lcandre2