Meterological forcing for CLM should have taxmode of cycle or limit, but NOT extend.

[escomp]

Meterological forcing for CLM should have taxmode of cycle or limit, but NOT extend.

[mvertens]

@ekluzek - can you please point out which taxmodes need to be changed from extend and what the new values should be.
If you just want to go ahead and change them in the stream_definition.xml and issue a PR - that would be great and very helpful.

[ekluzek]

Yes, I have a set of changes ready. I just need to make a PR with it. And a little more testing to make sure I'm not messing anything up.

[mvertens]

Great. Thank you.

[ekluzek]

In my testing I found out that this can't be done for the main global forcing, since we use 1901 forcing for 1850-1900. So I'm only applying this for the single point forcing and NLDAS2 options.

[ekluzek]

OK, a new issue I found is that the limit shows up in the simple cdep tests -- because they aren't careful to start and end on the time extents of the 1PT forcing datasets. My tests in CTSM are careful to do that. This is actually doing what I think it should though -- because a user can't actually do meaningful runs outside the period of data. So it's catching what I think it should. But, I'll have to modify the tests in the aux_cdeps test list so that they actually work.

[ekluzek]

In looking at this a little bit I see that MCT is setup this way as well. That's likely why it's setup this way as well. I'd say it's a mistake there as well though. This sets up the model in a way that a user can get nonsensical results -- but not realize what the problem is. It's easier for us as code maintainers possibly, but we really don't want to allow a situation where users can run a case that gives nonsensical results. It can be very difficult for them to diagnose what's wrong.