normal FATES crash Sierra CZ2 site

[JunyanDing]

I ran the normal FATES for pine and oak individually with different combinations of root parameters and vcmax and bb_ slope on Sierra CZ2 site. Oak was set to deciduous. The oak case crashed on 3rd year in Sep. The pine runs ok. The log file is attached. cesm.log.txt

[ckoven]

Hi @JunyanDing looks like the line numbers don't line up exactly with the code in master, but are you running into this scenario? if so, perhaps reduce slamax? https://github.com/NGEET/fates/blob/master/biogeochem/FatesAllometryMod.F90#L702

[rgknox]

@JunyanDing , could you provide the lnd.log file as well? Also, if you could provide the specific parameter file, or run script, that would help reproduce what is happening.

Also, can you provide the branch, and/or hash of the fates code you are using?

[JunyanDing]

The FATES hash is Faddc742 I ran multiple instance cases. There are 162 lnd.log file. I put 10 of them with the run script in a zip file. lnd_0009.log.16895540.zip. Thanks!

Thanks!

[rgknox]

hash : addc742

Thanks @JunyanDing , I see that this hash is part of 1.22.0_api.7.0.0

This hash was from a pretty unstable period of time from our model (it was crashing a lot then). I would recommend avoiding anything: after sci.1.21.0 and before sci.1.26.2.

master is also, provisionally, stable

https://github.com/NGEET/fates/releases

[JunyanDing]

Thanks Ryan, I will try another branch. And go through the log files. I found the code that caused the problem is is FateAllometryMod.F90 (line 712 in that version)

if( (treelai + tree_sai) > (nlevleafdinc_ed) )then write(fates_log(),) 'The leaf and stem are predicted for a cohort, maxed out the array size' write(fates_log(),) 'lai: ',treelai write(fates_log(),) 'sai: ',tree_sai write(fates_log(),) 'lai+sai: ',treelai+tree_sai write(fates_log(),) 'nlevleaf,dinc_ed,nlevleafdinc_ed :',nlevleaf,dinc_ed,nlevleafdinc_ed call endrun(msg=errMsg(sourcefile, LINE)) end if

[ckoven]

Hi @JunyanDing this is a known issue, see here: https://github.com/NGEET/fates/issues/420#issue-364630134

[rgknox]

Hopefully that problem will not be an issue with newer code, which may have been fixed via: #501

[JunyanDing]

Need to update the codes more often :)

[rgknox]

@JunyanDing , it looks like this was resolved, will close, but re-open if needed