Error message about PARPROF_PFT_DIR_Z

[jennykowalcz]

Hi all! I'm looking for help understanding an error message I have not encountered before. My FATES run failed after ~195 years with:

I rebuilt with DEBUG=TRUE and now it reads:

This is a full-complexity FATES run over tropical S. America with two evergreen tropical tree PFTs representing early and late successionals. I'm using two-stream radiation and have nclmax set to 3.

[jennykowalcz]

It looks like the third subscript of this array corresponds to nlevleaf

[ckoven]

It looks like it might plausibly be caused by the underlying problem that #1269 tries to solve? Maybe try pulling that branch or cherry-picking the commit into your branch and see if that fixes it?

[jennykowalcz]

Aha, okay, thanks, I'll try that and report back!

[jennykowalcz]

Darn, sadly that didn't fix it. It still crashes at the same point with the same error after cherry-picking the commit into my branch.

Digging into the code, I see PARPROF_PFT_DIR_Z is allocated as parprof_pft_dir_z(ncan,numpft,nveg) where nveg = maxval(this%ncan(:,:)) . (In the main branch ncan is changed to nleaf with this commit ).

@rgknox does anything jump out at you here?

I'm going to try variations of this case with Norman radiation and also with the main branch..

[jennykowalcz]

If I switch to Norman radiation and continue from the last restart at 190 years, it gets past the crash point and completes the 10 year block successfully. Next I will try with two-stream radiation but with the main branch.

[jennykowalcz]

Moved to issue 1279