search

ESCOMP / SimpleLand

Simple Land Model for CESM --- *** IN DEVELOPMENT *** --- please contact for more info. See supplemental information of https://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-18-0812.1 for a description of SLIM physics. Implementation of SLIM into the main CESM trunk is ongoing. SLIM currently works with the CESM2.1 release, but must be downloaded from this repository until we finish implementing it properly into the main CESM code.
Other
14 stars 7 forks source link

Trouble with branch simualtions #19

Closed ekluzek closed 1 year ago

ekluzek commented 3 years ago

I'm having trouble getting exact restart tests that test: restart, branch, and hybrid startups. It seems to be failing doing the branch type. I've tried various lengths up to 10months or 120 days.

Here's a list of tests that fail:

ERI_D_Ld120.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850
ERI_D_Ld60.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850
ERI_D_Ld90.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850 ERI_D_Lm10.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850
ERI_D_Lm3.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850
ERI_D_Lm6.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850

some of the above fail because of the run length.

ekluzek commented 3 years ago

The ERI test does the full length, and then the hybrid start, and then the branch. Since it's failing at the branch step that shows the hybrid part was fine. So it really is a problem with the branch startup. I checked and these fail at the branch step. I think the other two fail because of run length.

RUN_TYPE: branch

ERI_D_Ld120.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850.GC.slim-n1_cesm21chlist RUN_TYPE: branch ERI_D_Ld60.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850.GC.slim-n1_cesm21chintelasl RUN_TYPE: branch ERI_D_Ld90.f19_g17.I1850Clm45BgcGs.cheyenne_intel.clm-realistic_fromCLM5_1850.GC.slim-n1_cesm21chlist RUN_TYPE: branch

@marysa is a branch startup type something that you care about at all? The main reason for it is if you want to reproduce a simulation you already ran -- but change history output for a segment of it.

marysa commented 3 years ago

Hmm. Yes it would be nice, but not absolutely necessary. If it isn't clear why it doesn't work / will be too hard to fix, then just having a note that "hey, you can't do a branch run, make sure you've asked for all the history variables you want otherwise you'll have to run from the start again" would be fine!

ekluzek commented 3 years ago

I think we might put that warning in place to start off with. We will need to fix it eventually. But, it's good to know it isn't critical right now.

slevis-lmwg commented 1 year ago

We will fix #14 first.

ekluzek commented 1 year ago

The branch part of this is working now, we have a related issue with the hybrid part