Answer changing tags - ESCOMP/CTSM GitHub Wiki

List of answer-changing tags on master since ctsm1.0.dev001

Purpose of this page

This page lists answer-changing tags on master since ctsm1.0.dev001, which is the first tag not on the release-clm5.0 branch. Some of these are significantly answer-changing, and others are just roundoff-level.

Tags in bold, with the additional designator [SIGNIFICANT] are expected to have greater than roundoff-level differences from the release-clm5.0 branch in standard configurations, in more than just a handful of diagnostic fields.

  • (2020-10-06) We are currently only marking a tag as significant if it changes answers for clm50 configurations: if it just changes answers for clm51, it is not labeled as significant.
  • (2022-03-25) We are now marking a tag as significant even if it just changes answers for clm51 configurations. This changes for tags after ctsm5.1.dev028.

See this landing page for diagnostics noted below: http://www2.cgd.ucar.edu/tss/clm/diagnostics/ctsm_answer_changing_tags/index.html

Some logistics:

  • Sam and Keith use a script for setting up, building, and submitting simulations; see here: /glade/work/$USER/Answer_Changing_Tags
  • Sam and Keith save model output here: /glade/campaign/cgd/tss/people/$USER/ANSWER_CHANGING_TAGS (was /glade/p/cgd/tss/people/$USER/ANSWER_CHANGING_TAGS)
  • When lmwg diag. pkg has completed all steps, update /web/web-data/tss/clm/diagnostics/ctsm_answer_changing_tags/index.html
  • When iLamb has completed successfully, do scp -r _build_newscore_<case_identifier> moffatt.cgd.ucar.edu:/project/diagnostics/external/I20TR/<case>/lnd

Tags that have been satisfactorily investigated

The following tags have been satisfactorily investigated, and either confirmed to be climate preserving or have answer changes in line with expectations.

  • ‼️ ctsm1.0.dev001 [SIGNIFICANT]: Begin separating SoilHydrology flux calculations

    • Larger than roundoff-level changes expected from the following:
      • Include qflx_h2osfc_surf in BGC code: expected to be same climate
      • Two changes just affecting VIC
      • A change just affecting the QOVER diagnostic field
    • Additional roundoff-level changes
    • Keith Oleson ran the 9-timestep verification test, with comparison against release-clm5.0.01. This passed. (See also https://github.com/ESCOMP/CTSM/issues/690)

  • ctsm1.0.dev002: Fix NFIX flux variables so special land-units are zeroed out, tools update, add some *_MAX fields on mksurfdata_map for transient cases

    • Changes answers for all clm5_0 BGC cases
    • This is on the release branch; no need to do verification of it

  • ctsm1.0.dev004: Add some new diagnostic fields, fix a few issues, update cmip6 output

    • Changes answers for year-2000 cases
    • This is on the release branch; no need to do verification of it

  • ctsm1.0.dev007: Avoid glacier dynamic landunit adjustments in first time step

    • Changes answers for startup / hybrid runs that include CISM (even NOEVOLVE) and do not use init_interp. Changes can be larger than roundoff.
    • This is on the release branch; no need to do verification of it

  • ctsm1.0.dev008: Update 1850 ndep file and last year for streams for Historical transient cases

    • Changes answers for 1850 and 20th century cases; expected to be similar climate.
    • This is on the release branch; no need to do verification of it

  • ctsm1.0.dev013: Fix the fact that transient Bgc and SP cases had constant crop area in time

    • Significant answer changes for these transient non-crop cases
    • This is on the release branch; no need to do verification of it

  • ctsm1.0.dev016: Update cime, fix FATES DEBUG token, add script to easily run system tests

    • Changes answers due to datm update: new presaero and CO2 datasets
    • No need to do verification of this

  • ctsm1.0.dev018: Water tracer updates for initial things in driver loop

    • Roundoff-level changes in sensible heat flux; everything else bfb
    • No need to do verification of this

  • ‼️ ctsm1.0.dev019 [SIGNIFICANT]: Rework cold start initialization of wa and zwt

    • Larger-than-roundoff changes for CLM50 cold start and transient (Hist) cases; expected to be same climate
    • It might be worth doing more thorough testing of this tag

  • ctsm1.0.dev020: New options for irrigation and crop fsat

    • No longer resets wa_col to aquifer_water_baseline every time step
    • Differences only observed for cold start tests or when using a CLM45 initial conditions file in a CLM50 case, but it's possible that other configurations would see differences rarely
    • It might be worth doing more thorough testing of this tag. This doesn't seem too important, but this could be combined with other tag(s) in doing some additional testing.

  • ctsm1.0.dev022: Set tracer version of irrigation fluxes

    • Roundoff-level changes with groundwater irrigation
    • Roundoff-level changes in the diagnostic field, QIRRIG_FROM_SURFACE
    • No need to do verification of this

  • ctsm1.0.dev028: Interpolate out-of-the-box initial conditions and remove expensive tests

    • Changes answers for cases that require interpolation of out-of-the-box initial conditions
    • I was a bit suspicious of the changes in this tag. However, given that this only affects cases that require interpolation of out-of-the-box initial conditions, this probably won't affect most scientifically-important configurations, and also wouldn't be picked up by standard diagnostics runs.
    • So no need to do verification of this

  • ctsm1.0.dev030: Update CIME; hookup expected test fails

    • cime update resulted in differences for f19_g17 (roundoff-level) and T31_g37
    • No need to do verification of this

  • ctsm1.0.dev031: Subtract virtual states to reduce dynbal fluxes for transient glaciers

    • Large answer changes for runs with transient glaciers
    • For runs with transient vegetation, roundoff-level changes in dynbal fluxes
    • Large changes in diagnostic variables ICE_CONTENT1, LIQUID_CONTENT1 and HEAT_CONTENT1, over grid cells containing some glacier area
    • No need to do verification of this

  • ctsm1.0.dev034: Change clm4_5: Use Justin Perket snow on vegetation

    • Just changes answers for clm45
    • No need to do verification of this

  • ctsm1.0.dev035: Change h2ocan to a purely diagnostic variable

    • Replaces uses of h2ocan with snocan + liqcan
    • Expected to be just roundoff-level, but some differences seen in the test suite raised suspicions
    • We should do more thorough testing of this tag

Opened https://github.com/ESCOMP/CTSM/issues/691 for tags up to this point

  • ctsm1.0.dev036: Fix carbon isotope bug that caused wrong answers for isotopes under transient land-use change

    • Just changes answers for ciso cases
    • This is also on the release branch (release-clm5.0.22)

  • ctsm1.0.dev037: Change year alignment for present-day I compsets

    • Just changes answers for I2000 cases

  • ‼️ ctsm1.0.dev039 [SIGNIFICANT]: Remove excess canopy liquid/snow regardless of temperature

    • We want to do more thorough testing of this tag

(2019-05-03) Keith ran the diagnostics package on the changes up to this point, with comparison against release-clm5.0.01 (comparing 1989-2014). He found that diffs are larger than roundoff but same climate: GOOD!

  • ctsm1.0.dev040: Move some hard-coded parameters from code to params.nc file

    • Single-precision roundoff-level changes to four parameters

  • ctsm1.0.dev041: Add water tracers to CanopyHydrologyMod

    • Just roundoff-level changes

  • ctsm1.0.dev042: Rename, correct, and simplify parameters dewmx and sno_stor_max

    • Just roundoff-level changes

  • ‼️ ctsm1.0.dev043 [SIGNIFICANT]: Fix FUN bug (frac_ideal_C_use was backwards in regard to delta_CN), and replace Ball-Berry mbbopt with the CLM4.5 version

    • FUN bug fix affects any case with use_fun=.true. (so Clm50Bgc compsets). Climate is similar but not identical: "Our assessment is that the impact is relatively small, but would affect climate; however it doesn’t appear to strongly affect transient C response." See ChangeLog entry for this tag for pointers to diagnostics run on just this change.
    • Ball-Berry parameter fix just affects Clm5.0 cases with Ball-Berry method (so, by default, just Clm50Nwp compsets)

  • ‼️ ctsm1.0.dev045 [SIGNIFICANT]: Recalculate h2osno for the sake of SnowCapping

    • Answer changes are expected to be small, since this should just change the timing of snow capping a bit, but I haven't looked closely at the magnitude of the answer changes.

  • ctsm1.0.dev046: Separate the two uses of h2osno

    • For standard configurations, just roundoff-level changes
    • Greater than roundoff-level changes for cases for which number of snow layers is less in the case than in its initial conditions file, due to fixing #749

  • ‼️ ctsm1.0.dev047 [SIGNIFICANT]: Fix negative snow compaction during snow melt

    • Sean Swenson and Keith Oleson looked at impacts via a run of the LMWG diagnostics package
    • Overall, this doesn't look climate changing, but could potentially have a significant impact on select grid cells, and on some variables in polar regions
    • For more details, see notes in ChangeLog entry for this tag

  • ctsm1.0.dev048: Updates for buildlib changes and cime and externals updates

    • Just roundoff-level changes for cases with RTM, due to update in the version of RTM

  • ctsm1.0.dev049: Update mosart and intel to intel-19 on cheyenne

    • Roundoff-level changes due to externals updates

  • ctsm1.0.dev050: dz --> dz_lake bug-fix in LakeTemperatureMod.F90 line 960

    • Just changes answers in diagnostic variable, WTGQ, for cases with use_lch4 = .true.

  • ctsm1.0.dev051: Update water tracers for remainder of first stage of hydrology

    • Just roundoff-level changes in FSH_PRECIP_CONVERSION, FSH_TO_COUPLER, l2x_Fall_sen due to refactored calculation of FSH_PRECIP_CONVERSION

  • ctsm1.0.dev053: Soil layer definition clean-up and user-defined option

    • Just changes answers for NWP configurations

  • ctsm1.0.dev056: Start adding water tracers to LakeHydrology, and related refactoring

    • Just roundoff-level changes

  • ‼️ ctsm1.0.dev057 [SIGNIFICANT]: Fix frac_sno bugs

    • Two bug fixes with significant answer changes:
      • Fix lake frac_sno always being 0 (https://github.com/ESCOMP/CTSM/issues/783). This fixes the albedo calculation (and possibly others) over snow-covered lake surfaces.
      • Fix threshold for explicit snow pack initiation to use frac_sno_eff, not frac_sno (https://github.com/ESCOMP/CTSM/issues/785). For standard runs (which have use_subgrid_fluxes = .true.), this just changes answers for urban columns. This also changes answers more widely for runs with use_subgrid_fluxes = .false.
    • In addition, a roundoff-level change: Rewrite Swenson & Lawrence 2012 frac_sno equation to be more straightforward and less sensitive to roundoff errors (https://github.com/ESCOMP/CTSM/issues/784)
    • Expected to be same climate in general, because the only impacts are on urban and lake points, which represent a small fraction of most grid cells. However, this needs further investigation.

  • ‼️ ctsm1.0.dev058 [SIGNIFICANT]: Soil texture interpolation bug fix

    • Can change which soil layer from the surface dataset is used for a given model layer

(2019-08-26) Keith Oleson ran the diagnostics package on changes up to this point.

  • ctsm1.0.dev060: In SnowWater, truncate small h2osoi residuals

    • Expected to be just roundoff-level changes
    • Contains new endrun calls to ensure that these states don't go significantly negative. These endrun calls weren't triggered in the test suite, but it's possible that they'll be triggered in rare circumstances in a production run. These endrun calls help ensure that the changes in this tag really are only roundoff-level: as long as these endrun calls aren't triggered, the differences really should be just roundoff-level (unless there was a bug in the refactoring in this tag)

  • ‼️ ctsm1.0.dev063 [SIGNIFICANT]: Two answer changing fixes (fire, DWT_SLASH)

  • ctsm1.0.dev064: User defined top-two snow layers

    • Just roundoff-level changes

  • ‼️ ctsm1.0.dev065 [SIGNIFICANT]: Add water tracers to SnowWater

    • Changes were expected to be just roundoff-level, but were recently found to lead to large differences in rare circumstances, due to https://github.com/ESCOMP/CTSM/issues/965 . Once the fix is in place for that bug (expected in ctsm1.0.dev088), the changes in this tag will once again be expected to be just roundoff-level.

  • ctsm1.0.dev069: Misc. code cleanup and minor bug fixes

    • Changes just in select diagnostic fields, which are now averaged over only vegetated landunits: BTRAN, BTRANMN, RSSHA, RSSUN, ROOTR, ROOTR_COLUMN, RRESIS; and possibly NFIRE

  • ctsm1.0.dev071: Split CanopyTemperature into separate pieces

    • Just changes answers for the diagnostic field QFLX_EVAP_VEG

  • ‼️ ctsm1.0.dev073 [SIGNIFICANT]: Fix bug in calculation of dqgdT

    • Expected to be greater than roundoff-level differences, but same climate
    • Based on limited investigation, changes in daily-average latent heat flux appear to be typically 0.01 W m-2 or less
    • This warrants greater investigation

  • ctsm1.0.dev074: For lakes: when reading finidat, set frac_sno=1 if h2osno_total > 0

    • Expected to be greater than roundoff-level differences, but same climate
    • The only change is in initial lake frac_sno. This should only affect the start of the simulation, and only over lake points.

(2019-10-30) Keith Oleson ran the LMWG diagnostics on ctsm1.0.dev074 vs. ctsm1.0.dev058: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d074_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d074_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d058_2deg_GSWP3V1_hist.1995_2014/setsIndex.html . His feeling was that it shows just minor differences. He added in an email (2020-04-06) "I don't see any significant changes in climate, but others should also look. There are some interesting patterns of change (although small) in FAREA_BURNED."

  • ‼️ ctsm1.0.dev075 [SIGNIFICANT]: Lakes: Adjust frac_sno after updating 0-layer snow pack for dew & sublimation

    • Expected to be greater than roundoff-level differences, but same climate
    • Changes are introduced only over lake columns, and these changes should be small, since they only relate to the setting of frac_sno when there is dew or sublimation on a very thin snow pack.
    • Although the direct changes should only affect lakes, they can cause answer changes in other columns in some circumstances, seemingly due to the nonlocal interactions caused by the dependence of methane on TWS (see https://github.com/ESCOMP/CTSM/issues/658#issuecomment-546740771 for more details).

  • ctsm1.0.dev076: Set frac_sno_eff=0 if frac_sno is 0; avoid unnecessary calls to QSat

    • Just changes the diagnostic field FSNO_EFF

  • ‼️ ctsm1.0.dev077 [SIGNIFICANT]: Consistently use frac_sno_eff rather than frac_sno in qg calculations

    • Expected to be greater than roundoff-level differences, but same climate; but not investigated carefully
    • For standard configurations with use_subgrid_fluxes true, just changes answers over urban pervious road (but this can affect other landunits, seemingly due to the nonlocal interactions caused by the dependence of methane on TWS; see https://github.com/ESCOMP/CTSM/issues/658#issuecomment-546740771 for more details).
    • Note that this might be a small step backwards in terms of the accuracy of surface humidity for urban pervious road, but was desired for consistency and to facilitate upcoming cleanup.

  • ‼️ ctsm1.0.dev078 [SIGNIFICANT]: Fix rootr calculation with use_hydrstress true

  • ctsm1.0.dev080: Update externals, minor fixes to work with latest cime, get nuopc cap working

  • ctsm1.0.dev082: Rename variables to avoid confusion; fix QSNOEVAP diagnostic

    • Just changes answers for the diagnostic field QSNOEVAP
    • Possibility of other answer changes in rare circumstances, not observed in testing

  • ctsm1.0.dev086: Updates to NUOPC cap and minor fixes to ncdio_pio

    • Differences only when using the NUOPC driver

  • ctsm1.0.dev087: Change hard-coded soil layers in phenology to use a target depth

    • Just changes answers for Non-SP cases with NWP or non-standard soil layer structures: no answer changes for standard CLM45 or CLM50 configurations

  • ‼️ ctsm1.0.dev088 [SIGNIFICANT]: Fix bug in snow aerosol numerics (causes occasional HUGE aerosol values)

    • Fixes bug in ctsm1.0.dev065
    • By itself, this tag can lead to significant answer changes, at least occasionally. However when combined with the changes in ctsm1.0.dev065, answer changes should be only roundoff-level (since this undoes an inadvertent change in ctsm1.0.dev065).

(2020-04-08) Keith Oleson ran the LMWG diagnostics, comparing ctsm1.0.dev089 with ctsm1.0.dev074 (the last tag for which he ran diagnostics to check on ctsm answer-changing tags). Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d089_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d089_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d074_2deg_GSWP3V1_hist.1995_2014/setsIndex.html

  • He notes:

    • This comparison includes the following "significant" tags listed on the answer-changing tags page:
      • Lakes: Adjust frac_sno after updating 0-layer snow pack for dew & sublimation
      • Consistently use frac_sno_eff rather than frac_sno in qg calculations
      • Fix rootr calculation with use_hydrstress true
      • Fix bug in snow aerosol numerics (causes occasional HUGE aerosol values)
    • I don't see any climate-changing effects here, differences are quite small in key fields.
    • Note that there was a separate set of simulations to assess the rootr fix. This was judged to be greater than roundoff-level differences, but same climate.

  • ctsm1.0.dev091: Increase tolerance on near-zero truncation for a snow state update

    • Potentially changes answers by roundoff, but no changes were observed in any test in the test suite

  • ctsm1.0.dev093: Bring changes on release-clm5.0 branch to master

(2020-05-06) Keith Oleson ran the LMWG diagnostics, comparing ctsm1.0.dev093 with ctsm1.0.dev089. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d093_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d093_2deg_GSWP3V1_hist-clm50_ctsm10d089_2deg_GSWP3V1_hist/setsIndex.html. He says, "My opinion is that it is not climate-changing, but I don't have a record of anyone else weighing in on that."

(2020-09-06) Keith Oleson ran the LMWG diagnostics, comparing ctsm1.0.dev104 with ctsm1.0.dev093. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d104_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d104_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d093_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. He says, "I think we expect dev104 to change climate compared to dev093 because of the LUNA changes".

  • ctsm1.0.dev105: Bring in some new SE grids for CAM as well as initial condition files for them

    • Standard I compsets are bit-for-bit
    • Changes answers for CAM because of new finidat files
    • Running with nuopc changes answers because of updates to nuopc drivers
    • Dry-deposition changes answers when turned on (but doesn't effect anything else), due to cime update

  • ‼️ ctsm1.0.dev107 [SIGNIFICANT]: Answer changes needed for CESM2.2.0

    • Very small answer changes in Hist, SSP, 2010 compsets, Clm45 and some f10, 1x1 because of IC files
    • Hist and SSP BGC compsets change answers because of the correction of the end year for ndep. You wouldn't think this would matter until you get to the end of the time-sequence -- but it does for the first time-step.
    • The 2010 compset changes answers because of using the 2000 finidat file rather than 2010
    • Some finidat files interpolated to the resolution they are designed for. This causes an apparent change of answers when these datasets are used for other resolutions (even though it does NOT for the resolution it's interpolated to)

  • ‼️ ctsm1.0.dev109 [SIGNIFICANT]: Allow for resorbtion in transition from live to dead wood N

    • Changes answers for Clm50 cases with Bgc or Cn
    • Changes expected to be larger than roundoff / same climate

(2021-02-20) Keith Oleson ran the LMWG diagnostics, comparing ctsm1.0.dev109 with ctsm1.0.dev104. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d109_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d109_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d104_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm1.0.dev111 [SIGNIFICANT]: Compsets don't use 2014 for GSWP3 forcing, LUNA prevyr changed back
    • Changes answers for (2010 and SSP compsets) and Clm50 with LUNA on (Bgc or Sp)
    • Expected to give similar climate

(2021-02-20) Keith Oleson ran the LMWG diagnostics, comparing ctsm1.0.dev111 with ctsm1.0.dev109. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d111_2deg_GSWP3V1_hist/lnd/clm50_ctsm10d111_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d109_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm5.1.dev002 [SIGNIFICANT]: Reduce duplication between caps
    • Changes answers for all runs; expected to be larger than roundoff / same climate
    • Answer changes arise from using the same QSat approximations everywhere, rather than using different formulations for QSat in different places

(2021-02-20) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev002 with ctsm1.0.dev111. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d002_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d002_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm10d111_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ctsm5.1.dev003: Add capability for dynamic lakes

    • Changes answers for transient cases
    • Changes are basically roundoff level; exceptions are:
      • Big differences in the diagnostic fields HEAT_CONTENT1, ICE_CONTENT1, LIQUID_CONTENT1
      • Methane fields can grow to greater than roundoff-level different over time, but they seem to arise from differences that are originally only roundoff-level different
      • Cases with transient glacier areas, where glaciers increase, can lead to significant differences, probably where glaciers overrun lakes

  • ctsm5.1.dev004: Improve robustness of onset and offset counters when changing dt

    • Only changes answers when the model run uses a different time step than the initial conditions file

  • ctsm5.1.dev005: Answer changing bug fixes for clm51: fire and organic_frac_squared

    • Only changes answers for CLM51

  • ctsm5.1.dev006: Call correct routine to calculate btran2 for CNFireLi2021

    • Only changes answers for Clm51Bgc

  • ‼️ ctsm5.1.dev007 [SIGNIFICANT]: CNFire: btran2 fixes and general cleanup

    • Changes answers for all CN/BGC configurations
    • Answer changes expected to be same climate, but not investigated yet
    • Answer changes arise from fixing some bugs in the calculation of btran2. These fixes can impact fire occurrence; from some initial, quick investigations, it appears that the impact should be pretty small - but this needs confirmation. So, in investigating the impact of this tag, it could be good to specifically check differences in fire statistics.
    • Specific causes of the answer changes, for standard configurations, are:
      • using updated h2osoi_vol rather than the one earlier in the time step when calculating btran2
      • only considering points in the exposed veg filter when averaging btran2 from patch to column, rather than using stale values from no-longer-exposed patches (which also means having fire_m = 0 if there are no currently-exposed veg patches in a column) (fix for ESCOMP/CTSM#1153)
      • treating btran2 values that are slightly greater than 1 as 1, rather than ignoring them completely (fix for ESCOMP/CTSM#1170)
    • There are additional large changes in just the BTRAN2 diagnostic field, due to setting this to 0 over currently-non-exposed veg patches

(2021-02-21) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev007 with ctsm5.1.dev002. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d007_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d007_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d002_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm5.1.dev008 [SIGNIFICANT]: Two answer changes: Clm45/50/51 with crop, and for 2000Clm51
    • Turn dribble_crophrv_xsmrpool_2atm when crop is on for clm4_5, clm5_0, and clm5_1 physics
    • Fix the finidat files being selected for 2000_control conditions for clm5_1
    • Changes answers for Clm45/50/51 with crop (except with prognostic CO2) and 2000Clm51

(2021-03-23) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev008 with ctsm5.1.dev007. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d008_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d008_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d007_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

(2021-03-23) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev011 with ctsm5.1.dev008. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d011_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d011_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d008_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm5.1.dev014 [SIGNIFICANT]: Grass/crop snow burial fraction fix, and some other fixes
    • Crop/grass snow burial changed based on height and lodging (Danica Lombardozzi) - resolves ESCOMP/CTSM#516. Updates grass and crop snow burial ('fb' in SatellitePhenologyMod.F90) from being 0.2m to change with PFT height and accounts for 20% lodging (falling over). (Large change.)
    • In BGC code, make elai and esai depend on frac_sno (Danica Lombardozzi, Keith Oleson & Bill Sacks) - resolves ESCOMP/CTSM#1116. (Large change.)
    • Fix to quadratic solution error bug caused by negative shaded photosynthesis (Danica Lombardozzi & Keith Oleson) - resolves ESCOMP/CTSM#756. (Small change.)
    • The change in snow burial parameterizations results in changes to albedo and water & energy fluxes.
    • For analysis of diffs, see https://github.com/danicalombardozzi/ctsm_py/blob/7543b0f3b413bae9974c11b467fdbc0413c3b7fa/notebooks/SnowBurial.ipynb

(2021-03-23) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev014 with ctsm5.1.dev011. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d014_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d014_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d011_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm5.1.dev019 [SIGNIFICANT]: Fix ndep from coupler

    • There was a bug in ndep forcing (forc_ndep_grc) when receiving ndep from CAM (ndep_from_cpl true). This first appeared in ctsm5.1.dev002. NDEP forcings have been garbage since then when receiving ndep from CAM.
    • Bit-for-bit for I compsets, but significant changes, with potentially new climate, for coupled configurations where CTSM receives NDEP from CAM

  • ctsm5.1.dev020: Potential roundoff changes in preparation for bio-mass heat storage option

    • Only a few tests changed answers from baseline
    • Earlier commit tested the new terms versus the old and showed they were order e-11 for absolute difference and e-14 relative difference.

  • ctsm5.1.dev021: Add option for biomass heat storage (BHS) to clm5_1 physics

    • Climate changing for clm5_1; bit-for-bit for earlier physics versions

  • ctsm5.1.dev022: Merge fates_main_api into ctsm master

    • Larger than roundoff changes for FATES configurations; bit-for-bit for non-FATES configurations

  • ‼️ ctsm5.1.dev023 [SIGNIFICANT]: Calculate leaf biomass for non-woody PFTS, and a few other small answer changes

    • Replace hard code constant 0.25 for leaf mass per area with calculation based on parameter slatop (specific leaf area, top of canopy). Also move num_iter into loop over patches; currently it sits outside a loop, so p index is incorrect.
    • Also do some small answer changes in terms of new parameter files, and some other existing issues that have mild answer changes.
    • Answer changes: clm5_1 for all, clm50 for non-crop (Sp and Bgc), clm4_5 two diagnostics
    • Magnitude of change: clm51--BGC climate, others similar climate

(2021-03-24) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev023 with ctsm5.1.dev014. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d023_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d023_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d014_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ‼️ ctsm5.1.dev028 [SIGNIFICANT]: Change limitation of top layer evaporation/sublimation
    • Sublimation from top soil layer and evaporation/sublimation from top snow layer needs to be limited to ensure moisture states do not become negative. The original formulation did not always work, so we added a new limitation to SoilFluxesMod.
    • Also removes a limitation in SoilHydrologyMod that seemed not to conserve energy and should no longer be necessary with the reworked limitation in SoilFluxesMod.
    • Changes expected to be greater than roundoff but same climate

(2021-03-24) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev028 with ctsm5.1.dev023. Results are here: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm51d028_2deg_GSWP3V1_hist/lnd/clm50_ctsm51d028_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d023_2deg_GSWP3V1_hist.1995_2014/setsIndex.html.

  • ctsm5.1.dev029: Rework domain initialization for nuopc

    • Just changes answers for NUOPC and limited changes for LILAC (note that NUOPC was not yet the default at this point)
    • Roundoff-level changes
    • For nuopc: changes in area (relatively large differences in the f10 test in the test suite, but Mariana saw only very small changes in an f09 case), landfrac; these influence l2r fields, which in turn influence TWS and methane fields.
    • For lilac: just changes in area

  • ctsm5.1.dev031: Update externals and fixes for nuopc threading

    • Just changes answers for cases with the NUOPC driver (not yet the default) with CISM
    • Roundoff-level changes due to the new externals

  • ‼️ ctsm5.1.dev032 [SIGNIFICANT]: Fix bugs in co2 from atmosphere

    • There was a bug in co2 from atmosphere; for the default mct driver, this only impacted results if co2 was spatially-varying.
    • This bug was introduced in ctsm5.1.dev002
    • This appears not to impact I compsets, but could impact coupled configurations where CTSM receives spatially-varying CO2 from CAM. In this case, the bug resulted in all CTSM grid cells on a given processor being given the co2 value from the last grid cell on that processor.

  • ‼️ ctsm5.1.dev034 [SIGNIFICANT]: Bring in Arctic changes to LUNA from Leah Birch

    • Changes to improve arctic plants
    • Changes answers for clm5_1
    • New climate
    • Diagnostics were run to evaluate the new climate

(2022-04-12) Keith Oleson ran a full CTSM5.1 AD/pAD spinup and a full CTSM5.1 historical with ctsm5.1.dev034 and then ran the LMWG diagnostics, comparing ctsm5.1.dev034 with ctsm5.1.dev028. Note that this is a comparison of CTSM5.1 with CLM5.0.
Results are here: http://webext.cgd.ucar.edu/I20TR/ctsm51_ctsm51d034_2deg_GSWP3V1_hist/lnd/ctsm51_ctsm51d034_2deg_GSWP3V1_hist.1995_2014-clm50_ctsm51d028_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. The differences here are very similar to those obtained previously using the PPE tag:

https://webext.cgd.ucar.edu/I20TR/ctsm51_cesm23a02cPPEn08ctsm51d030_1deg_GSWP3V1_hist/lnd/ctsm51_cesm23a02cPPEn08ctsm51d030_1deg_GSWP3V1_hist.1995_2014-clm50_cesm23a02cPPEn08ctsm51d030_1deg_GSWP3V1_hist.1995_2014/setsIndex.html

  • ctsm5.1.dev036: FATES tag update and snow occlusion of LAI fix

    • Larger than roundoff changes for FATES configurations; bit-for-bit for non-FATES configurations

  • ‼️ ctsm5.1.dev037 [SIGNIFICANT]: Correct vertical interpolation in init_interp when soil thicknesses change

    • Answers change for all cases that interpolate from a soil profile found in finidat to a different soil profile used by the model
    • Expected to be larger than roundoff / same climate

  • ctsm5.1.dev039: Add NEON sites, update externals

    • Just changes answers for cases with CISM on (compsets that end with G)
    • Expected to be similar climate

  • ctsm5.1.dev040: Replace masked with "nomask" SRC files for use with mkmapdata and mksurfdata_map

    • Will change answers for any new surface datasets, but doesn't change answers for out-of-the-box configurations with existing surface datasets

  • ctsm5.1.dev041: Updates to coupling interface to enable FATES-CH4, and to be compatible with FATES-CNP

    • Just changes answers for FATES configurations

  • ctsm5.1.dev042: Small answer changes for double precision constants and limit on organic soil

    • Changes answers for all configurations
    • Mostly single-precision roundoff-level: many constants that were single precision are now double
    • Some limits put on soil organic that weren't there previously

  • ctsm5.1.dev044: New stream functionality when using NUOPC or LILAC

    • Just changes answers for NUOPC & LILAC configurations (note that NUOPC was not yet the default at this point)
    • Differences larger than roundoff / same climate. Diffs arise due to changes in the interpolation used in cdeps streams vs. the old mct-based streams. Diffs generally keep 3-4 digits of equivalence or better.

  • ctsm5.1.dev046: Update FATES

    • Just changes answers for FATES configurations: minor changes related to a bug fix

  • ctsm5.1.dev051: New LUNA-specific ozone stress option

    • Just changes answers for LUNA cases with the pgi compiler; roundoff-level changes

  • ctsm5.1.dev053: Add Antarctica compset and change min_snow_to_keep

    • Changes answers rarely, when there is a large snow capping flux - removing the entire bottom layer of the snow pack in one time step due to snow capping
    • Changes are slightly larger than roundoff when they occur

  • ctsm5.1.dev054: New parameters on paramsfile, many bit-for-bit changes, new history fields for IWUELN,VPD_2M, allow Medlyn Photosynthesis for non-PHS cases, fixes to HumanStressIndex

    • Just changes diagnostic fields when HumanStress turned on

  • ctsm5.1.dev056: FATES Satellite Phenology mode implemented

    • Just changes answers for FATES FixedBiogeo mode: larger than roundoff due to corrected hlm-fates PFT mapping

  • ‼️ ctsm5.1.dev058 [SIGNIFICANT]: Some miscellaneous small answer changes

    • Significant answer changes for clm5_1; some answer changes for other configurations
    • All clm5_1 physics change answers because of change in ceta parameter.
    • When PHS off for clm5_1 and clm5_0 changes answers because of Medlyn photosynthesis over Ball-Berry.
    • Photosynthesis potentially changes all the time because a maximum parameter was off.
    • Answer should be close to roundoff for Sp and Bgc cases without Crop because we now use crop surface datasets.
    • FATES changes answers since user_nitrif_dentrif is now on by default
    • COLDSTART BGC cases will change answers because of a change in initial soil temperature
    • clm4_5 and clm5_0 vichydro results are identical
    • clm4_5 and clm5_0 single point urban tests are identical
    • Some tests are still identical mostly crop-cases non-cold-start
    • In some cases GSSUN and GSSHA will be different.

(2022-04-12) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev058 with ctsm5.1.dev034. Results are here: http://webext.cgd.ucar.edu/I20TR/ctsm51_ctsm51d058_2deg_GSWP3V1_hist/lnd/ctsm51_ctsm51d058_2deg_GSWP3V1_hist.1995_2014-ctsm51_ctsm51d034_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. There are small but significant increases in TLAI and GPP in forested regions (due to change to rf_cwdl2_bgc & rf_cwdl2_bgc).

  • ctsm5.1.dev060: Izumi aux_clm test fixes for fates testmods

    • Just changes answers for FATES configurations because of a FATES tag update

  • ctsm5.1.dev061: Changes needed to get aux history output with cmeps

    • Changes answers for NUOPC configurations only (note that NUOPC was not yet the default at this point)
    • Just changes missing value patterns for lnd to mediator fields

  • ‼️ ctsm5.1.dev062 [SIGNIFICANT]: Make NUOPC the default driver rather than MCT

    • Changes answers due to new driver/mediator
    • Similar climate, as verified with diagnostic runs

  • ctsm5.1.dev063: Provide access to the fraction of canopy covered by snow variable to fates

    • Changes answers for FATES configurations only due to FATES tag update

  • ctsm5.1.dev064: Updates to facilitate FATES history variable overhaul

    • Changes answers for FATES configurations only, just in diagnostic fields

  • ctsm5.1.dev066: API change with FATES to enable running means inside fates, includes passing in of model timestep

    • Changes answers for FATES configurations only

  • ctsm5.1.dev067: NEON UI update, externals updates, small miscellaneous fixes

    • Just changes answers for NEON sites due to new surface datasets and settings

  • ctsm5.1.dev071: Small changes to enable new fates dimension and update fates tag

    • Just changes answers for some FATES configurations

  • ctsm5.1.dev073: Some fixes for Gregorian calendar

    • Just changes answers for some runs with a Gregorian calendar
    • Larger than roundoff, same climate - only very small changes

  • ‼️ ctsm5.1.dev075 [SIGNIFICANT]: Small answer changes: urban ventilation, fire-emission, irrigate off when not crop, fix two SSP ndep files

    • Similar climate
    • Diagnostic runs were done to validate the new climate (at least for some of these changes)

  • ctsm5.1.dev077: Updates to FATES API, including removal of patch dimensions from fates history and using soil instead of ground layers for fates history.

    • Just changes answers for FATES-hydro

  • ‼️ ctsm5.1.dev079 [SIGNIFICANT]: Changes to CropPhenology timing

    • Changes answers for all Crop cases
    • Changes expected to be larger than roundoff / same climate, but not investigated carefully
    • Possibility of greater changes in F / B compsets (where doalb isn't true every time step), but from looking through the code, nothing was seen that suggested that this would be the case

  • ctsm5.1.dev080: Use avg days per year when converting param units

    • Changes answers for Gregorian cases with BGC
    • Changes are larger than roundoff / same climate, due to changing a few BGC-related parameters by a small amount (< 1/365 in a relative sense) for Gregorian cases

  • ctsm5.1.dev081: Do not subtract irrigation from QRUNOFF diagnostic

    • Just changes answers in a few diagnostic fields for Crop cases with irrigation: QRUNOFF, QRUNOFF_R, QRUNOFF_TO_COUPLER

  • ctsm5.1.dev084: FATES parameter file updated to align with clm pft optical parameters

    • Just changes answers for FATES configurations (larger than roundoff)

  • ctsm5.1.dev085: Expand crop reproductive pools; remove some unused options

    • Just changes answers in a few C iso diagnostic fields for Clm45 cases (or any cases with use_grainproduct false) with crop and C isotopes

  • ‼️ ctsm5.1.dev088 [SIGNIFICANT]: Add outputs for annual crop sowing and harvest dates

    • Changes answers for Crop configurations
    • Answer changes are due to fixing issues / oddities in the old code; these changes are greater than roundoff but are expected to be same climate. (Differences may be significant at some times for certain patches, but should probably be small for regional and global averages, and probably even most grid cell averages.)

  • ctsm5.1.dev089: For CLM4.5 apply peaklai to aleaf in grainfill

    • From the test suite Bill Sacks found differences when using CLM45 together with old hydrology (h2osfcflag = 0 and use_subrid_fluxes = .false.). Otherwise no differences in our typical configurations, though this surprised him.

(2022-04-14) Keith Oleson ran the LMWG diagnostics, comparing ctsm5.1.dev090 with ctsm5.1.dev058. Results are here: http://webext.cgd.ucar.edu/I20TR/ctsm51_ctsm51d090_2deg_GSWP3V1_hist/lnd/ctsm51_ctsm51d090_2deg_GSWP3V1_hist.1995_2014-ctsm51_ctsm51d058_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. I don't see much going on here other than some changes in the seasonal cycle of runoff, which presumably is due to the change in ctsm5.1.dev081 where irrigation is no longer subtracted from runoff (e.g., apparent in crop-intensive India). I don't see any changes in the seasonal cycle of LAI in crop-intensive areas (such as might be caused by ctsm5.1.dev079). I do see some increases in fractional burned area - crop in crop-intensive regions (e.g., Central Africa, not sure what the cause of that is).

  • ctsm5.1.dev092: Refactor NutrientCompetition / CNAllocation to provide hooks for AgSys

    • From the test suite Bill Sacks found no differences, though he expects answers to change in the atypical situation of using FUN without FlexibleCN.

  • ctsm5.1.dev095: Turn soil BGC off for FATES-SP mode, externals updates, FatesSP user-mod directory and compset

    • Roundoff diffs in cpl history files from this test: SMS_D_Ld1_Mmpi-serial.f45_f45_mg37.I2000Clm50SpRs.izumi_gnu.clm-ptsROA

  • ‼️ ctsm5.1.dev098 [SIGNIFICANT]: Correct perched water table calculation

    • Larger than roundoff changes to surface climate due to diffs originating in soil moisture in all clm50 and clm51 cases

  • ctsm5.1.dev101: Fix winter wheat sowing window bugs

    • Potentially larger than roundoff/same climate changes in areas with winter wheat
    • Winter wheat is non-standard and is included only in smallville tests, so these do change answers

  • ctsm5.1.dev103: Fix accumulation variables when changing model time step

    • Changes answers for configurations with a non-standard (i.e. non-30-minute) time step
    • Nature of change not investigated carefully, though further discussion appears in the ChangeLog

  • ctsm5.1.dev104: Update to fates history names and machine configuration

    • Changes answers due to a correction only for a subset of fates testmods for the fates_levcdam dimension

  • ‼️ ctsm5.1.dev107 [SIGNIFICANT]: Update Externals

    • Larger than roundoff changes in almost all configurations due to major change on CDEPS side to coszen angle time interpolation
    • Diffs in single-point cases from update to how we set ZBOT
    • Diffs in NLDAS cases from a fix for zero pressure in nldas2 configurations

  • ‼️ ctsm5.1.dev108 [SIGNIFICANT]: Connect ozone from atmosphere

    • Larger than roundoff changes when o3_veg_stress_method = 'stress_falk' or 'stress_lombardozzi2015

(2023-06-22) slevis ran the LMWG diagnostics, comparing ctsm5.1.dev110 with ctsm5.1.dev090. Results are here: http://webext.cgd.ucar.edu/I20TR/ctsm51_ctsm51d110_2deg_GSWP3V1_hist/lnd/ctsm51_ctsm51d110_2deg_GSWP3V1_hist.1995_2014-ctsm51_ctsm51d090_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. Nothing of concern stands out for me in this comparison.

  • ctsm5.1.dev111: Fixes for NEON cases

    • Larger than roundoff for NEON cases only, due to fixes in the leap year handling when there are discrepancies between the model and the DATM forcing data
    • This involved updates to the share and cdeps externals from Jim Edwards and other fixes for NEON cases from Will Wieder

  • ctsm5.1.dev112: Rework fates test definitions and add new fates tests

    • Changes answers for fates simulations only due to answer-changing updates in multiple fates tags: both bug fixes and science updates

  • ‼️ ctsm5.1.dev113 [SIGNIFICANT]: Fix some compsets; add only clauses for ESMF use statements

    • Larger than roundoff/maybe new climate for cases with I1850Clm51BgcCrop and cases with DATM%CPLHIST

  • ctsm5.1.dev114: Some NEON updates fixing AG sites, update MOSART, small fixes

    • Larger than roundoff diffs for NEON cases or if bypass_routing_option==direct_to_outlet in MOSART

  • ctsm5.1.dev115: API compatability with FATES V2 nutrient dynamics

    • Only FATES cases affected
    • Carbon-only FATES simulations will not have qualitatively different results since the last API update but will have bitwise differences. Nutrient enabled FATES simulations (not fully coupled to CLM, only via prognosed plant N,P boundaries), and FATES-Hydro simulations (bug fix) will be different.

  • ‼️ ctsm5.1.dev116 [SIGNIFICANT]: Small answer changes with bug fixes, zetamaxstable=2 for biomass heat storage, new single-point fsurdat files

    • Adjusted climate
      • clm5_1 changes because zetmaxstable = 2. A long enough simulation may hit this max and answers will change; else answers can be identical.
      • clm5_0 and clm4_5 change if biomass heat storage is turned on
      • single-point resolutions (i.e. 1x1_smallvilleIA, 1x1_brazil, 1x1_mexicocityMEX) have differences in maintenance respiration
    • Keith Oleson ran experiments with changing zetamaxstable; some slides showing this are here: https://docs.google.com/presentation/d/1u6ycr7F97QYYRcRfEdD9yIxH75diUx2r

(2023-06-26) slevis ran the LMWG diagnostics, comparing ctsm5.1.dev116 with ctsm5.1.dev110. Results are here: http://webext.cgd.ucar.edu/I20TR/ctsm51_ctsm51d116_2deg_GSWP3V1_hist/lnd/ctsm51_ctsm51d116_2deg_GSWP3V1_hist.1995_2014-ctsm51_ctsm51d110_2deg_GSWP3V1_hist.1995_2014/setsIndex.html. Globally TSA cooler, snow deeper, snow fraction reduced, net radiation increased, FSH increased, carbon pools reduced, soil ice reduced, snow ice up, snow liq down. Keith Oleson saw the same temperature change patterns in a simulation with the zetmaxstable = 2 change in the context of the roughness length PR #1596.

  • ctsm5.1.dev117: Updates to facilitate running FATES at NEON sites

    • Only NEON sites show changes, due to updated fsurdat files

  • ctsm5.1.dev119: Allow gross unrepresented land use transitions (PR #309)

    • Changes answers only when do_grossunrep = .true. and surface dataset includes non-zero gross unrepresented land use transitions; the nature of the changes was not investigated.

  • ctsm5.1.dev120: Update externals and minor fixes

    • Changes answers on certain compilers, though nature of change not investigated:
      • nvhpc on cheyenne in non-debug cases: can be explained from differences in compilation flags for non-debug cases, and also some module differences
      • intel on izumi in debug cases: there were updates in ESMF modules, though that's the same for other izumi compilers; not seeing other relevant diffs in ccs_config so not sure why we're getting diffs here. Further investigation failed when trying to compile with the old ccs_config.

  • ctsm5.1.dev121: Changes soil moisture initialization logic for FATES

    • Changes answers in fates suite for all non-hydro fates tests since the soil moisture initialization matches that of fates hydro now
    • Changes answers for all fates testmods in the aux_clm suite as the science tag has iterated forward by 4 minor version updates
    • All diffs accounted for with prior fates suite tests

  • ctsm5.1.dev122: Rework handling of evaporation constraint in SoilFluxes

    • Roundoff changes
    • Two tests had > roundoff diffs and Bill Sacks introduced temporary code that helped verify that these differences were fundamentally no greater than roundoff-level

  • ctsm5.1.dev128: Update FATES tests to double precision

    • Only FATES cases affected
    • Diffs due to changing hist_ndens to 1 (double precision)

  • ctsm5.1.dev129: NEON fixes for TOOL and user-mods, add SP for NEON, some history file updates, black refactor for buildlib/buildnml

    • Diffs appear at NEON TOOL site and other NEON sites with namelist changes, else bit-for-bit

  • ctsm5.1.dev130: FATES parameter file and test definition update

    • Larger than roundoff changes for FATES tests only

  • ctsm5.1.dev131: Enable prescribed crop calendars

    • Roundoff changes for Clm45BgcCrop compsets only, due to order-of-operations changes

  • ctsm5.1.dev133: FATES API update to facilitate fates refactor

    • Changes answers for FATES tests only

  • ctsm5.1.dev134: Migration of FATES to share normal soil BGC call sequence and functionality

    • Roundoff changes in two diagnostic fields (TOTCOLC and TOTCOLN) relative to baseline
    • The diffs appear in many tests: all non-FATES tests had, at most, roundoff level (<e-12) diffs in TOTCOLC and TOTCOLN
    • FATES tests have diffs due to the subtle but unavoidable differences in how the models are coupled

  • ctsm5.1.dev135: Rename hist fields to track them down more easily

    • Field lists differ. In some tests, the namelists differ.

  • ctsm5.1.dev141: Change small snocan to zero

    • Answer changes are roundoff-level because we now truncate roundoff-level greater-than-zero snocan states to exactly zero that most likely needed to be zero anyway.
    • Answer changes grow to greater than roundoff, but the cprnc.out file from a 20-year izumi test-suite case does not contain differences of concerning magnitude.

  • ctsm5.1.dev143: Zeroing of wood product fluxes on fates columns

    • Answer changes only on FATES tests, and only on the specific wood product fluxes modified. These values are now zeros, instead of being uninitialized. Everything else is b4b.

  • ‼️ ctsm5.1.dev145 [SIGNIFICANT]: SNICAR snow albedo scheme updates

    • Namelist defaults are such that phys="clm5.0" and phys="clm4.5" give different answers only due to the changed fsnowoptics file.
    • Namelist defaults are such that phys="clm5.1" changes answers as a result of new parameterizations.

(2023-10-25) slevis ran the LMWG diagnostics, comparing ctsm5.1.dev145 with ctsm5.1.dev116. Simulation description and results are here: https://github.com/NCAR/LMWG_dev/issues/28. In summary, Cenlin He finds the simulation consistent with his expectations.

  • ctsm5.1.dev146: FATES cross-grid seed dispersal

    • Only fates testmods in aux_clm are answer changing due to science updates associated with externals update

  • ctsm5.1.dev149: Rearrange leaf/stem "harvest" and fix soil gas diffusivity

    • Roundoff-level differences were observed for the rearrangement of leaf/stem "harvest" code.
    • Notable differences were observed for the soil gas diffusivity bugfix, but only for output variable diffus.

  • ctsm5.1.dev152: Mv tools to /python and add tests; add snow_thermal_cond_method; a few fixes / refactors

    • Answers differ in these two NEON tests due to the one-line bug-fix in cime_config/usermods_dirs/NEON/defaults/shell_commands in https://github.com/ESCOMP/CTSM/pull/2044
    • SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Bgc.izumi_nag.clm-default--clm-NEON-NIWO
    • SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm51Bgc.izumi_nag.clm-NEON-MOAB--clm-PRISM

  • ‼️ ctsm5.1.dev154 [SIGNIFICANT]: New params files for Meier roughness, MIMICS, and SNICAR, and changes to leafcn and k*_nonmyc

    • Cumulatively, answers change more than roundoff for all three CLMs (see PR2258). Issues addressed:
    • Start using existing new params file for Meier roughness, include bug-fix 2219, update forcing heights per 2071.
    • Update params file for MIMICS per 1845.
    • Make leafcn for pfts 15 and 16 the same per 2184.
    • Switch the values of params kc_nonmyc and kn_nonmyc per 2120 (FUN bug-fix).
    • Move SNICAR parameters to ctsm51, clm50, and clm45 params files per 2247.

(2023-11-22 to 2023-12-06) slevis ran simulations and LMWG diagnostics comparing (a) ctsm5.1.dev154 with dev145 and (b) with dev156 that turns Meier2022 back off: dev154 vs. dev145 includes the FUN bug-fix plus Meier2022, which results in bigger LAI diffs and confuses the source of the diffs; dev154 vs. dev156 diffs are as expected, mostly in the partitioning of SH and LH fluxes. Simulation descriptions/results/discussion: https://github.com/NCAR/LMWG_dev/issues/38.

  • ctsm5.1.dev156: Do not use Meier roughness by default

    • Changes ctsm51 only: more than roundoff, possibly new climate
    • Reverts to the pre-dev154 default, so not labeling as "significant", but we will compare to dev154 to isolate the effect of Meier2022

  • ctsm5.1.dev160: FATES landuse version 1

    • All FATES configurations, more than roundoff, explained as follows: The fates tag update incorporates bug fixes and hydraulic mortality fixes, as well as the restructured disturbance code necessary to accommodate the new landuse transitions capability. As such, small differences were observed in testmods that engaged specific disturbance modes or were long enough to trigger other default disturbances (e.g. fire).

Tags that have NOT yet been satisfactorily investigated

  • ‼️ ctsm5.1.dev161 [SIGNIFICANT]: Refactor 20-year running means of crop GDD accumulation

    • Answers change (larger than roundoff/same climate) for all crop configurations.
    • GDD020, GDD820, and GDD1020 differ most strongly in the first few years after a crop becomes active, with downstream effects on many variables, since the GDD variables help determine sowing date and maturity requirements.

  • ctsm5.1.dev164: Compatibility and tests for FATES 2-Stream

    • One fates variable is not BFB with baselines (expected): FATES_RAD_ERROR.
    • This variable was changed to report the maximum of VIS and NIR, instead of just VIS.
    • In a follow-up we will change the dimension of this variable.

  • ‼️ ctsm5.1.dev165 [SIGNIFICANT]: Turn Meier2022, tillage, and residue removal on for ctsm5.1, fix #2212

    • Possibly climate changes for ctsm5.1 cases by setting Meier2022 to on, tillage to low, and residue removal to 0.5
    • More than roundoff changes for clm45/clm50 by including the #2212 fix

  • ctsm5.1.dev168: Remove a source of negative snocan in CanopyFluxesMod

    • Roundoff in short tests, diffs grow in longer tests

  • ctsm5.1.dev170: Add hillslope hydrology

    • FSDS answers change due to rounding differences, since the history field now uses a column-level variable instead of a gridcell-level one. Note that this is JUST the history field that's affected, which is why there are no diffs in any other variable. (Confirmed using branch at https://github.com/samsrabin/CTSM/tree/hillslope-revert-fsds-diffs.)
    • The origflag parameter (used to reproduce CLM4 behavior) was removed, so anything using that will break. This includes the oldhyd test.

  • ctsm5.1.dev171: Set initial t_soisno=272 for soils and 274K for urban road

  • ctsm5.1.dev173: New FATES namelist variable, fates_history_dimlevel

    • Answer changes for FATES tests were detected. All diffs were small enough to be consistent with order of operations changes, with the exception of some variables that were updated to have ignore values used for non-vegetated patches instead of zero.

  • ‼️ ctsm5.1.dev174 [SIGNIFICANT]: Improve vegetation health at high latitudes

    • Changes answers relative to baseline for clm51
    • New climate at high latitudes
    • See LMWG_dev discussion #3 and simulations discussed therein, including (though possibly not limited to) LMWG_devissues #51, 52, 54, 57

  • ‼️ ctsm5.2.0 [SIGNIFICANT]: New surface datasets and new mksurfdata_esmf tool to create them

  • ctsm5.2.002: FATES default allometry parameter file update

    • The FATES externals update incorporates a number of bug fixes and the new allometry default parameters result in a new scientific baseline (diffs larger than roundoff) for FATES cases only

Wait for new spin-up before running new Answer-changing test.

⚠️ **GitHub.com Fallback** ⚠️