Move some hardcoded parameters to the paramsfile #1039
Conversation
BFB according to following tests: ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput Parameters moved: fnps theta_psii theta_ip theta_cj theta_cj vcmaxha jmaxha tpuha lmrha kcha koha cpha vcmaxhd jmaxhd tpuhd lmrhd lmrse denitrif_nitrateconc_coefficient denitrif_nitrateconc_exponent denitrif_respiration_coefficient denitrif_respiration_exponent k_nitr_max_perday (renamed from k_nitr_max)
BFB according to following tests: ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput Parameters moved: jmaxb0 wc2wjb0 enzyme_turnover_daily
BFB according to following tests: ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput Parameters moved: relhExp minrelh wimp ssi drift_gs eta0_anderson (was eta0 in OverburdenCompactionAnderson1976) eta0_vionnet (was eta0 in OverburdenCompactionVionnet2012) n_baseflow perched_baseflow_scalar (was hardcoded 1.e-5 in SoilHydrologyMod) tkd_sand (was hardcoded 8.8 in SoilStateInitTimeConstMod) tkd_clay (was hardcoded 2.92 in SoilStateInitTimeConstMod) tkd_om (was om_tkd in SoilStateInitTimeConstMod) tkm_om (was om_tkm in SoilStateInitTimeConstMod) pd (was hardcoded 2.7e3 in SoilStateInitTimeConstMod) csol_om (was om_csol in SoilStateInitTimeConstMod) csol_sand (was hardcoded 2.128 in SoilStateInitTimeConstMod) csol_clay (was hardcoded 2.385 in SoilStateInitTimeConstMod) e_ice
I ran the CLM test suite (aux_clm) on the entire set of changes up to this point (all commits). All pass as expected and are BFB with the baseline (ctsm1.0.dev089). This includes CLM4.5 tests as well. I've created a new CLM4.5 parameter file that produces BFB results and included it in namelist_defaults_ctsm.xml Parameters moved: tau_l1 (now expressed as 1/year) tau_l2_l3 (now expressed as 1/year) tau_s1 (now expressed as 1/year) tau_s2 (now expressed as 1/year) tau_s3 (now expressed as 1/year) tau_cwd (now expressed as 1/year) cryoturb_diffusion_k som_diffus
BFB according to the following tests: ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERS_D_Ld6.f10_f10_musgs.I1850Clm45BgcCrop.cheyenne_intel.clm-clm50CMIP6frc clm_short test suite Parameters moved: prh30 (was hardcoded as 0.7 in CNFireLi2016Mod) ignition_efficiency (was hardcoded as 0.22 in CNFireLi2016Mod) tpu25ratio (was hardcoded as 0.167 in PhotosynthesisMod) kp25ratio (was hardcoded as 20000 in PhotosynthesisMod) bsw_adjustfactor (set to 1.0, did not previously exist in SoilStateInitTimeConstMod) hksat_adjustfactor (set to 1.0, did not previously exist in SoilStateInitTimeConstMod) sucsat_adjustfactor (set to 1.0, did not previously exist in SoilStateInitTimeConstMod) watsat_adjustfactor (set to 1.0, did not previously exist in SoilStateInitTimeConstMod)
BFB according to the following tests: ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERS_D_Ld6.f10_f10_musgs.I1850Clm45BgcCrop.cheyenne_intel.clm-clm50CMIP6frc clm_short test suite Parameters moved: wind_min (was hardcoded as 1.0 in CanopyFluxesMod.F90, LakeFluxesMod.F90, BareGroundFluxesMod.F90, UrbanFluxesMod.F90) bsw_adjustfactor, hksat_adjustfactor, sucsat_adjustfactor, watsat_adjustfactor are now also applied to soil beneath lakes
|
In running testing on cheyenne, these two tests show difference in answers... ERS_Lm20_Mmpi-serial.1x1_smallvilleIA.I2000Clm50BgcCropQianGs.cheyenne_gnu.clm-monthly_noinitial Tests of five years or more that passed... ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput |
|
That's interesting. I ran the aux_clm test suite with gnu and the only failure was an expected one. However, those two tests that failed for you don't appear to be part of the aux_clm test suite? |
|
@olyson this is after the update to ctsm1.0.dev100, and those two tests are part of the test suite there. They might not have been in the older ctsm1.0.dev089 version the branch came from. |
…lue of 0.8, this is needed because of the update to get these two biofuels working
|
OK, I looked through to figure this out. In the older baseline version ctsm1.0.dev089 these tests don't complete, but I was able to run one for 460 days. I ran that in ctsm1.0.dev089 and in a few of the olyson branch versions and they don't change answers. I realized the problem might be due to PFT changes in the params file due to the miscanthus,switchgrass update. It made changes to the paramsfile to convert them from C3 to C4 types. When I fix the new paramsfile for this change on the newly added PFT variable of theta_cj. -- answers go back to being identical. |
|
That's great Erik, thanks for solving this. |
…n be used by all versions, and use ignition_efficiency in CNFireLi2014Mod as well
|
OK, this has passed testing on cheyenne and izumi, and I'll update the Change files and bring it to master. |
Description of changes
Change some hardcoded parameters to go on the parameter files. This is needed in preparation of running the Perturbed Parameter Ensemble.
Specific notes
Contributors other than yourself, if any: This is primarily work of @olyson
CTSM Issues Fixed (include github issue #): #995 #138, #206
Fixes #995
Fixes #138
Fixes #206
Some work on #290
Are answers expected to change (and if so in what way)? No, expected to be bit-for-bit
Any User Interface Changes (namelist or namelist defaults changes)? New params files
for both clm4_5 and clm5_0
Testing performed, if any: Ran test suite on cheyenne for gnu and intel