Benchmark/GEOS-Chem 12.5.0: Difference between revisions

From Geos-chem
Jump to navigation Jump to search
Line 504: Line 504:
*GCHP 12.5.0 currently has a slow memory leak due to updates to MAPL ExtData. NASA is aware of the issue and will be developing a fix soon. The memory leak is more pronounced for long simulations and can result in a simulation crashing due to exceeding memory available. Breaking up long runs into smaller consecutive runs avoids the issue but we find that doing so changes the output. Lizzie Lundgren (GCST) is investigating differences in GCHP output that happen when restarting a simulation, which may or may not be a problem with GEOS-Chem Classic as well. We are documenting all single versus multi-run differences at [http://ftp.as.harvard.edu/gcgrid/geos-chem/validation/multi_vs_single_run/ http://ftp.as.harvard.edu/gcgrid/geos-chem/validation/multi_vs_single_run/]. Look there to assess if the scale of the differences would be a problem for you.
*GCHP 12.5.0 currently has a slow memory leak due to updates to MAPL ExtData. NASA is aware of the issue and will be developing a fix soon. The memory leak is more pronounced for long simulations and can result in a simulation crashing due to exceeding memory available. Breaking up long runs into smaller consecutive runs avoids the issue but we find that doing so changes the output. Lizzie Lundgren (GCST) is investigating differences in GCHP output that happen when restarting a simulation, which may or may not be a problem with GEOS-Chem Classic as well. We are documenting all single versus multi-run differences at [http://ftp.as.harvard.edu/gcgrid/geos-chem/validation/multi_vs_single_run/ http://ftp.as.harvard.edu/gcgrid/geos-chem/validation/multi_vs_single_run/]. Look there to assess if the scale of the differences would be a problem for you.
*GCHP 12.5.0 uses online ESMF regridding weights rather than external tile files for regridding. Due to ESMF domain decomposition rules this can result in a MAPL error if an input grid is too coarse for a run’s configured core count. We have seen this problem for 4°x5° input files when using >600 cores. GMAO has fixed this problem in a more recent version of MAPL. All 4°x5° input files are replaced with higher resolution files for GCHP 12.5.0 to avoid this issue. However, users may still run into problems if running with thousands of cores.
*GCHP 12.5.0 uses online ESMF regridding weights rather than external tile files for regridding. Due to ESMF domain decomposition rules this can result in a MAPL error if an input grid is too coarse for a run’s configured core count. We have seen this problem for 4°x5° input files when using >600 cores. GMAO has fixed this problem in a more recent version of MAPL. All 4°x5° input files are replaced with higher resolution files for GCHP 12.5.0 to avoid this issue. However, users may still run into problems if running with thousands of cores.
*There is a positive bias in surface ozone over the southern ocean in GCHP. It shows in comparisons with both GCC 12.5.0 and GCHP 12.4.0. The issue is resolution-dependent and occurs in FV3. An issue has been created on GEOS github for this: https://github.com/GEOS-ESM/GEOSgcm_GridComp/issues/48. To evaluate this bias at different resolutions, we also ran and GCHP 12.5.0 benchmark at c24. Results from that simulation may be viewed at [http://ftp.as.harvard.edu/gcgrid/geos-chem/1mo_benchmarks/GC_12/12.5.0/GCHP_c24/plots/ this link].
*There is a positive bias in surface ozone over the Southern Ocean in GCHP. It shows in comparisons with both GCC 12.5.0 and GCHP 12.4.0. The issue is resolution-dependent and occurs in FV3. An issue has been created on GEOS github for this: https://github.com/GEOS-ESM/GEOSgcm_GridComp/issues/48. To evaluate this bias at different resolutions, we also ran and GCHP 12.5.0 benchmark at c24. Results from that simulation may be viewed at [http://ftp.as.harvard.edu/gcgrid/geos-chem/1mo_benchmarks/GC_12/12.5.0/GCHP_c24/plots/ this link].
|-bgcolor="#cccccc"
|-bgcolor="#cccccc"
!Approval
!Approval

Revision as of 17:18, 13 September 2019


This page includes information about all benchmark simulations for GEOS-Chem 12.5.0 (DOI: 10.5281/zenodo.3403111).

GEOS-Chem Classic 1-month benchmark

Benchmark plots

Enable PDF bookmarks1 in the files linked below to navigate between species. See the Species in GEOS-Chem wiki page for species definitions and details.

Species Category2 Plots Additional Information
Level Map Zonal Mean Emissions
Oxidants

O3, CO, OH, NOx

sfc 500hPa full column stratosphere sfc
Aerosols

NH4, NIT, SO4, DST1, DST2, DST3, DST4, BC, OC, AERI, BrSALA, BrSALC, ISALA, ISALC, NITs, SALA, SALC, SO4s, SOA_Simple, SOA_Complex

sfc 500hPa full column stratosphere sfc AOD
Bromine

Bry, BrOx, Br, Br2, BrCl, BrNO2, BrNO3, BrO, CH4Br, CH2Br2, CH2Br2, CH3Br3, HOBr, HBr

sfc 500hPa full column stratosphere sfc
Chlorine

Cly, ClOx, Cl, ClO, Cl2, Cl2O2, ClOO, ClNO2, ClNO3, CCl4, CFCs, CH3Cl, CH2Cl2, CH4CCl3, HOCl, HCl, Halons, HCFCs, OClO

sfc 500hPa full column stratosphere --
Iodine

Iy, IxOy, I, I2, IBr, ICl, IO, INO2, IONO2, CH3I, CH2I2, CH2ICl, CH2IBr, HI, HOI, OIO

sfc 500hPa full column stratosphere sfc
Nitrogen

NOy, NOx, HNO2, HNO3, HNO4, IMPN, NIT, NO, NO2, NO3, N2O5, MPN, PAN, PPN, N2O, NHx, NH3, NH4

sfc 500hPa full column stratosphere sfc
Primary Organics

ALK4, BENZ, CH4, C2H6, C2H4, C3H8, PRPE, TOLU, XYLE, ISOP, MTPA, MTPO, LIMO, GLYX, HCOOH, MAP, MBO, MOH, RCHO

sfc 500hPa full column stratosphere sfc
Secondary Organics

ACET, MEK, MVK, ALD2, CH2O, HPALD, MP, ISOPN, IEPOX, ACTA, EOH, MOH

sfc 500hPa full column stratosphere sfc
Secondary Organic Aerosols

SOAP, SOAS, TSOA0, TSOA1, TSOA2, TSOA3, ISOA1, ISOA2, ISOA3, ASOA1, ASOA2, ASOA3, ASOAN, TSOG0, TSOG1, TSOG2, TSOG3, ISOG1, ISOG2, ISOG3, ASOG1, ASOG2, ASOG3, INDIOL, ISN1OA, LVOCOA, SOAIE, SOAGX, SOAME, SOAMG

sfc 500hPa full column stratosphere sfc
Sulfur

SOx, DMS, OCS, SO2, SO4

sfc 500hPa full column stratosphere sfc
ROy

H2O2, H, H2, H2O, HO2, O1D, OH, RO2

sfc 500hPa full column stratosphere --
J-values sfc 500hPa full column stratosphere -- --
 
Emissions by HEMCO category

Total, Aircraft, Anthropogenic, Biomass Burning, Biogenic, Lightning, Natural, Ocean, Seabirds, Decaying Plants, Ship, Soil, Degassing Volcanoes, Erupting Volcanoes

Emission totals

Tables by species, Tables by inventory

Global mass

Entire atmosphere, Troposphere only

Budget Ox and CO budgets

1All plots created with GCPy have PDF bookmarks. If you do not see them in your browser try using a different browser (not Safari) or installing an Adobe Acrobat plug-in.

2Species category and lumped species definitions are simple to adjust. Updating them requires only editing the json files.

Benchmark assessment form

Description
New features added into GEOS-Chem
Feature Submitted by
Updates that will affect full-chemistry simulations
Retire obsolete emission inventories Eloise Marais (Leicester)
Emissions and Deposition Working Group
Updated offline biogenic emissions generated with GEOS-Chem 12.3.0 Hongjian Weng (PKU)
Updated volcano emissions to include eruptive emissions through April 2018 Christoph Keller (NASA/GMAO)
Remove CO from MEGAN HEMCO extension to avoid double counting Jenny Fisher (Wollongong)
Fixes for GFED4 emissions, including:
  • Change the HEMCO cycling flag so that the model crashes if it doesn't find the right GFED year
  • Remove the 27% scaling of POG to POA and emit 1:1 instead
Sidhant Pai (MIT)
Scale DICE-Africa emissions to address errors in inventory Eloise Marais (Leicester)
Separate anthropogenic PM2.5 dust source from the DustDead extension Melissa Sulprizio (GCST)
Corrections for CFCs in the chemistry mechanism Xuan Wang (Harvard)
Fix typo in CLD_PARAMS routine in gckpp_HetRates.F90 Stephen Steenrod (NASA GMAO)
Bug fix: make sure stratospheric BrY concentrations are read properly each month Lyatt Jaeglé (U. Washington)
Updates that will NOT affect full-chemistry simulations
Enable accurate cloud pH diagnostic calculation via post-processing Jonathan Moch (Harvard)
Bug fix for proposed Henry's law constants Jenny Fisher (Wollongong)
Updates that will only affect GCHP
Update GCHP to use MAPL v1.0.0 and ESMF v8.0.0 beta snapshot 28 Lizzie Lundgren (GCST)
Replace 4x5 input files with higher resolution equivalents to avoid ESMF domain decomposition issue at high core counts Lizzie Lundgren (GCST)
Version, resolution, met fields used: GEOS-FP (72L), 4x5, July 2016
1-month benchmark finished on: Wed Aug 28 03:28:34 EDT 2019
Performance statistics:
  • Compiled with ifort 17.0.4
  • Ran on 24 CPUs of holyjacob01.rc.fas.harvard.edu (Intel(R) Xeon(R) CPU E5-2680 v3 "Haswell" @ 2.50 GHz)
  • Wall time: 07:28:13
  • CPU time / wall time: 21.1711
  • % of ideal performance: 88.21%
  • Memory: 13.8379 GB
Compared to previous benchmark: 12.4.0
This update will impact:
(select all that apply with boldface)
Advection, BL Mixing, Convection, Met Fields, Dry Dep, Wet Dep, Stratosphere, Emissions, Photolysis, Chemistry, Other (please specify): Diagnostics
Unit test results may be viewed at: http://ftp.as.harvard.edu/gcgrid/geos-chem/1mo_benchmarks/GC_12/12.5.0/GCC/Plots/GC_12.5.0_Aug08.results.html
Metrics
Global mean OH from log file (% change): 12.3915975939893 x 105 molec/cm3 (-0.65%)
Methyl chloroform lifetime (% change): 4.9487 (1.06%)
Methane lifetime (% change): 8.3312 years (1.10%)
List all SURFACE species, J-Values, and COLUMN AOD values that changed by 10% or more:
  • Oxidants: O3 CO OH NOx
  • Aerosols: NH4 NIT SO4 DST1 AERI BrSALA BrSALC ISALA ISALC NITs SO4s SOA_Complex SOA_Simple
  • Nitrogen: NOy NOx HNO2 HNO3 HNO4 IPMN NIT NO NO2 NO3 N2O5 MPN PAN PPN NHx NH3 NH4
  • ROy: H2O2 H HO2 O1D OH RO2
  • Sulfur: SOx DMS SO2 SO4
  • Iodine: Iy IxOy I I2 IBr ICl IO INO2 IONO2 HI HOI OIO
  • Bromine: Bry BrOx Br Br2 BrCl BrNO2 BrNO3 BrO HOBr HBr
  • Chlorine: Cly ClOx Cl ClO Cl2 Cl2O2 ClOO ClNO2 ClNO3 HOCl HCl HCFCs OClO
  • Primary_Organics: ALK4 BENZ C2H6 C3H8 PRPE TOLU XYLE ISOP MTPA MTPO LIMO GLYX HCOOH MAP RCHO
  • Secondary_Organics: ACET MEK MVK ALD2 CH2O HPALD MP ISOPN IEPOX ACTA EOH SOAIE LVOCOA
  • J-Values: CCl4 CFC11 CFC113 CFC114 CFC115 CFC12 CH2Cl2 CH3CCl3 CH3Cl HCFC123 HCFC141b HCFC142b N2O O2
  • Column AOD: Dust OCPI SO4 SOA_from_Aqueous_ISOP Total
List all 500 hPa species and J-Values that changed by 10% or more :
  • Oxidants: OH NOx
  • Aerosols: NH4 NIT SO4 DST1 AERI BrSALA BrSALC ISALC NITs SO4s SOA_Complex SOA_Simple
  • Nitrogen: NOx HNO2 HNO4 IPMN NIT NO NO2 NO3 N2O5 MPN PAN PPN NHx NH3 NH4
  • ROy: OH
  • Sulfur: SOx DMS SO2 SO4
  • Iodine: I2 ICl INO2 IONO2 CH2I2 HI OIO
  • Bromine: BrOx Br Br2 BrCl BrNO2 BrNO3 BrO HOBr HBr
  • Chlorine: ClOx Cl ClO Cl2 Cl2O2 ClOO ClNO2 ClNO3 HOCl HCl OClO
  • Primary_Organics: ALK4 BENZ C2H6 C3H8 PRPE TOLU XYLE ISOP MTPA MTPO LIMO GLYX HCOOH MAP RCHO
  • Secondary_Organics: ACET MEK MVK ALD2 HPALD MP ISOPN IEPOX ACTA EOH SOAIE LVOCOA
  • J-Values: None
List all ZONAL MEAN species and J-Values that changed by 10% or more:
  • Oxidants: OH NOx
  • Aerosols: NH4 NIT SO4 DST1 BrSALA BrSALC NITs SO4s SOA_Complex SOA_Simple
  • Nitrogen: NOy NOx HNO2 HNO3 IPMN NIT NO2 NO3 N2O5 MPN PAN PPN NH3 NH4
  • ROy: OH RO2
  • Sulfur: SOx SO2 SO4
  • Iodine: I2 IBr ICl INO2 IONO2 CH2I2 HI OIO
  • Bromine: BrOx Br Br2 BrCl BrNO2 BrNO3 BrO HOBr HBr
  • Chlorine: Cly ClOx Cl ClO Cl2 Cl2O2 ClOO ClNO2 ClNO3 HOCl OClO
  • Primary_Organics: ALK4 C2H6 C3H8 PRPE TOLU XYLE ISOP MTPA MTPO LIMO GLYX HCOOH MAP RCHO
  • Secondary_Organics: ACET MEK MVK ALD2 CH2O HPALD ISOPN IEPOX ACTA EOH SOAIE LVOCOA
  • J-Values: None
List all EMISSIONS that changed by 10% or more:
  • Anthro: ALD2 ALK4 BCPI BCPO C2H6 C3H8 CO HAC MEK NH3 NO pFe PRPE SO2 SO4 SOAP
  • Biogenic: ACET ALD2 EOH ISOP LIMO PRPE SOAP SOAS
  • Ship: CO HNO3 NO2 NO O3 SO2 SOAP
  • Total: ACET ALD2 ALK4 BCPI BCPO C2H6 C3H8 CO EOH GLYX ISOP MEK MGLY MTPA NH3 NO pFe PRPE SO2 SO4 SOAP
  • VolcDegas: SO2
  • VolcErupt: SO2
Comments on observed differences: Below we summarize the notable changes caused by specific updates.

Retire obsolete emission inventories

Retiring the EMEP, Bond, RETRO, and Yevich & Logan biofuel inventories results in large fractional differences in anthropogenic emissions and species concentrations over Europe, South America, Central America, Russia, and Australia, where the global anthropogenic, biofuel, and ship emissions from the CEDS inventory are now used. The impact on emission totals are included below.
  ###############################################################################
  ### Emissions totals                                                        ###
  ### Ref = GC_12.4.0; Dev = GC_12.5.0                                        ###
  ###############################################################################
                                   Ref            Dev            Dev - Ref
  ACET Anthro               :      0.171664       0.069874      -0.101790 Tg C
  ALD2 Anthro               :      0.042310       0.042140      -0.000169 Tg C
  ALK4 Anthro               :      2.407840       2.572299       0.164460 Tg C
  BCPI Anthro               :      0.106725       0.106747       0.000021 Tg
  BCPO Anthro               :      0.426900       0.426986       0.000086 Tg
  C2H6 Anthro               :      0.473419       0.664990       0.191571 Tg C
  C3H8 Anthro               :      0.399813       0.595361       0.195548 Tg C
  CH2O Anthro               :      0.263482       0.235037      -0.028445 Tg
  CO Anthro                 :     42.984464      40.778317      -2.206146 Tg
  HAC Anthro                :      0.212160       0.042432      -0.169728 Tg
  HCOOH Anthro              :      0.577087       0.571744      -0.005343 Tg
  MEK Anthro                :      0.043368       0.064733       0.021364 Tg C
  NH3 Anthro                :      4.430414       4.592343       0.161929 Tg
  NO Anthro                 :      5.543591       5.293092      -0.250499 Tg
  OCPI Anthro               :      0.628814       0.610537      -0.018278 Tg
  OCPO Anthro               :      0.628815       0.610537      -0.018278 Tg
  pFe Anthro                :      0.006943       0.006628      -0.000315 Tg
  PRPE Anthro               :      0.490600       0.448865      -0.041735 Tg C
  SO2 Anthro                :      6.974655       7.092453       0.117797 Tg
  SO4 Anthro                :      0.218581       0.210372      -0.008208 Tg
  SOAP Anthro               :      2.978745       2.826521      -0.152224 Tg
  TOLU Anthro               :      0.584716       0.553756      -0.030960 Tg C
  CO Ship                   :      0.079328       0.059695      -0.019634 Tg
  HNO3 Ship                 :      1.484222       1.449563      -0.034659 Tg
  NO Ship                   :      0.081681       0.082875       0.001195 Tg
  NO2 Ship                  :      0.771094       0.738918      -0.032176 Tg
  SO2 Ship                  :      0.805693       0.773045      -0.032648 Tg
  SOAP Ship                 :      0.005474       0.004119      -0.001355 Tg


Updated offline biogenic emissions generated with GEOS-Chem 12.3.0

The offline biogenic emissions were updated to include recent fixes to the MEGAN extension. With this update, emissions for MTPA and MTPO are now included, as opposed to the individual components (APIN, BPIN, BCAR, etc.) included in the previous version. CO from monoterpenes has been removed to avoid double-counting that source. In addition, the SOAP and SOAS emissions from the offline biogenic VOC files are ignored, and SOAP and SOAS emissions are instead computed by scaling isoprene and monoterpene emissions. The impact on biogenic emission totals are included below.
  ###############################################################################
  ### Inventory totals for OFFLINEBIOVOC                                      ###
  ### Ref = GC_12.4.0; Dev = GC_12.5.0                                        ###
  ###############################################################################
                                   Ref            Dev            Dev - Ref
  OFFLINEBIOVOC ACET        :      2.677313       2.982478       0.305164 Tg C
  OFFLINEBIOVOC ALD2        :      1.032241       1.037183       0.004942 Tg C
  OFFLINEBIOVOC CO          :      5.504973            nan            nan Tg C
  OFFLINEBIOVOC EOH         :      0.987091       0.991817       0.004726 Tg C
  OFFLINEBIOVOC ISOP        :     35.759849      33.723910      -2.035939 Tg C
  OFFLINEBIOVOC LIMO        :      1.022397       0.976173      -0.046225 Tg C
  OFFLINEBIOVOC MTPA        :           nan       8.644930            nan Tg C
  OFFLINEBIOVOC MTPO        :           nan       4.000710            nan Tg C
  OFFLINEBIOVOC PRPE        :      2.126864       2.030293      -0.096571 Tg C
  OFFLINEBIOVOC SOAP        :      0.621433       1.467334       0.845901 Tg C
  OFFLINEBIOVOC SOAS        :      0.621433       1.467334       0.845901 Tg C
While it appears that SOAP and SOAS emissions have a large increase from 12.4.0, these emission totals are now more in-line with 12.3.0 emission totals using online emissions from the MEGAN extension. In 12.3.0, emission totals of SOAP and SOAS were each 1.294232 Tg C.
Jeff Pierce pointed out that we see a much larger increase in the complex SOA concentrations as compared to the simple SOA concentrations (see the last two pages of this file). However, when we do the comparison of 12.5.0 vs 12.3.0, the differences are much lower altogether and have varying signs, consistent with the retirement of several emission inventories (see the last two pages of this file). The 12.5.0 vs 12.3.0 are the comparisons to focus on because the 12.4.0 biogenic emissions were generated with and older version of MEGAN from GEOS-Chem v11-02. An outcome of the 12.4.0 benchmarking was that we should apply scaling of isoprene emissions to match the 12.3.0 annual totals. That scaling would impact the complex SOA species but not the simple SOA species which were read directly from the offline biogenic VOC files. In 12.5.0, the Aerosol WG decided not to read the simple SOA species directly from the offline biogenic VOC files after finding that the emissions in those files were too low compared to 12.3.0. Instead we now compute the SOAP and SOAS emissions by scaling the isoprene and monoterpene emissions from the offline biogenic VOC files. For the complete set of plots comparing 12.5.0 to 12.3.0, please visit http://ftp.as.harvard.edu/gcgrid/geos-chem/1mo_benchmarks/GC_12/12.5.0/GCC/Plots_vs_12.3.0/.
In this benchmark, we also introduced the Secondary Organic Aerosols benchmark category and global mass tables to assist with evaluating this update. These plots will be included in future benchmarks.


Updated volcano emissions to include eruptive emissions through April 2018

This update increased both eruptive and degassing volcanic emissions, as summarized below. As a result, localized increases in SO2 and SO4 can be seen in the species concentration plots.
  ###############################################################################
  ### Inventory totals for VOLCANO                                            ###
  ### Ref = GC_12.4.0; Dev = GC_12.5.0                                        ###
  ###############################################################################
                                   Ref            Dev            Dev - Ref
  VOLCANOerupt SO2          :      0.000000       0.035000       0.035000 Tg
  VOLCANOdegas SO2          :      1.509524       1.951957       0.442434 Tg


Scale DICE-Africa emissions to address errors in inventory

This update contributes to changes in BC and OC concentrations over and just downwind of Africa. Slight differences can also be seen over Africa in the emission maps. The impact on the inventory emission totals are included below.
  ###############################################################################
  ### Inventory totals for DICEAfrica                                         ###
  ### Ref = GC_12.4.0; Dev = GC_12.5.0                                        ###
  ###############################################################################
                                   Ref            Dev            Dev - Ref
  DICEAfrica BCPI           :      0.004131       0.004153       0.000022 Tg
  DICEAfrica BCPO           :      0.016526       0.016612       0.000086 Tg
  DICEAfrica CH2O           :      0.044793       0.016348      -0.028445 Tg
  DICEAfrica CO             :      5.651943       3.648702      -2.003242 Tg
  DICEAfrica HAC            :      0.212160       0.042432      -0.169728 Tg
  DICEAfrica HCOOH          :      0.006679       0.001336      -0.005343 Tg
  DICEAfrica NH3            :      0.294221       0.284477      -0.009744 Tg
  DICEAfrica NO             :      0.153533       0.151820      -0.001713 Tg
  DICEAfrica OCPI           :      0.035185       0.016907      -0.018278 Tg
  DICEAfrica OCPO           :      0.035185       0.016907      -0.018278 Tg
  DICEAfrica PRPE           :      0.057572       0.019730      -0.037842 Tg
  DICEAfrica SOAP           :      0.389982       0.251758      -0.138224 Tg
  DICEAfrica TOLU           :      0.054817       0.020857      -0.033960 Tg


Separate anthropogenic PM2.5 dust source from the DustDead extension

This update reintroduces the anthropogenic PM2.5 fine dust source and increases concentrations of DST1. The impact on DST1 emission totals are included below.
  ###############################################################################
  ### Inventory totals for AFCID                                              ###
  ### Ref = GC_12.5.0; Dev = GC_12.5.0                                        ###
  ###############################################################################
                                   Ref            Dev            Dev - Ref
  AFCID DST1                :           nan       1.087876            nan Tg


The following updates produced small changes, primarily over the southern hemisphere oceans and at high altitudes:

Comments on additional differences in GCHP:

GCHP 12.5.0 uses MAPL version tag "gchp/v1.0.0" released in early 2019. The MAPL library, which is developed by NASA, is a wrapper for the Earth System Modeling Framework (ESMF), and handles not only I/O but also passing of data between components. With this version update GCHP input speed executes roughly 3-4 times faster than previous versions! This speed-up is due to parallelization within input read across one node, and the MAPL updates in this version are just the beginning of improvements to GCHP I/O bottlenecks.

While we are encouraged by the improvements to GCHP in 12.5.0, a few issues remain open until the next MAPL update later this year.

  • GCHP 12.5.0 currently has a slow memory leak due to updates to MAPL ExtData. NASA is aware of the issue and will be developing a fix soon. The memory leak is more pronounced for long simulations and can result in a simulation crashing due to exceeding memory available. Breaking up long runs into smaller consecutive runs avoids the issue but we find that doing so changes the output. Lizzie Lundgren (GCST) is investigating differences in GCHP output that happen when restarting a simulation, which may or may not be a problem with GEOS-Chem Classic as well. We are documenting all single versus multi-run differences at http://ftp.as.harvard.edu/gcgrid/geos-chem/validation/multi_vs_single_run/. Look there to assess if the scale of the differences would be a problem for you.
  • GCHP 12.5.0 uses online ESMF regridding weights rather than external tile files for regridding. Due to ESMF domain decomposition rules this can result in a MAPL error if an input grid is too coarse for a run’s configured core count. We have seen this problem for 4°x5° input files when using >600 cores. GMAO has fixed this problem in a more recent version of MAPL. All 4°x5° input files are replaced with higher resolution files for GCHP 12.5.0 to avoid this issue. However, users may still run into problems if running with thousands of cores.
  • There is a positive bias in surface ozone over the Southern Ocean in GCHP. It shows in comparisons with both GCC 12.5.0 and GCHP 12.4.0. The issue is resolution-dependent and occurs in FV3. An issue has been created on GEOS github for this: https://github.com/GEOS-ESM/GEOSgcm_GridComp/issues/48. To evaluate this bias at different resolutions, we also ran and GCHP 12.5.0 benchmark at c24. Results from that simulation may be viewed at this link.
Approval
Requires further investigation: No
Approved by: Jeff Pierce, Daniel Jacob
Date of approval: 06 Sep 2019

GCHP 1-month benchmark

The following sections contain validation plots for GCHP 12.5.0.

Comparison of GCHP versions

This table contains plots comparing GCHP 12.4.0 (Ref) and GCHP 12.5.0 (Dev) for a 1-month simulation.

Category Plots Additional Information
Level Map Zonal Mean Emissions
Oxidants sfc 500hPa full column stratosphere full column
Aerosols sfc 500hPa full column stratosphere full column Column AOD
Bromine sfc 500hPa full column stratosphere full column
Chlorine sfc 500hPa full column stratosphere --
Iodine sfc 500hPa full column stratosphere full column
Nitrogen sfc 500hPa full column stratosphere full column
Primary Organics sfc 500hPa full column stratosphere full column
Secondary Organics sfc 500hPa full column stratosphere full column
Sulfur sfc 500hPa full column stratosphere full column
Emissions by HEMCO category

Aircraft, Anthro, Bioburn, Biogenic,
Lightning, Monoterp, Natural, Ocean, Seabirds, DecayingPlants,
Ship, Soil, Total, VolcDegas, VolcErupt,

Emission totals

Comparison of GCHP and GEOS-Chem Classic

This table contains plots comparing GEOS-Chem Classic 12.5.0 (Ref) against GCHP 12.5.0 (Dev) for a 1-month simulation.

Category Plots Additional Information
Level Map Zonal Mean Emissions
Oxidants sfc 500hPa full column stratosphere full column
Aerosols sfc 500hPa full column stratosphere full column Column AOD
Bromine sfc 500hPa full column stratosphere full column
Chlorine sfc 500hPa full column stratosphere --
Iodine sfc 500hPa full column stratosphere full column
Nitrogen sfc 500hPa full column stratosphere full column
Primary Organics sfc 500hPa full column stratosphere full column
Secondary Organics sfc 500hPa full column stratosphere full column
Sulfur sfc 500hPa full column stratosphere full column
Emissions by HEMCO category

Aircraft, Anthro, Bioburn, Biogenic,
Lightning, Monoterp, Natural, Ocean, Seabirds, DecayingPlants,
Ship, Soil, Total, VolcDegas, VolcErupt,

Emission totals

GCHP and GEOS-Chem Classic difference of differences

This table contains plots comparing the difference [GCHP 12.5.0 - GCHP 12.4.0] against the difference [GC Classic 12.5.0 - GC Classic 12.4.0], taken from 1-month benchmark simulations.

Category Plots
Level Map Zonal Mean
Oxidants sfc 500hPa full column stratosphere
Aerosols sfc 500hPa full column stratosphere
Bromine sfc 500hPa full column stratosphere
Chlorine sfc 500hPa full column stratosphere
Iodine sfc 500hPa full column stratosphere
Nitrogen sfc 500hPa full column stratosphere
Primary Organics sfc 500hPa full column stratosphere
Secondary Organics sfc 500hPa full column stratosphere
Sulfur sfc 500hPa full column stratosphere

NOTE: At present, we have only created the diff-of-diffs plots for species concentration. This can be extended to other categories in the future.

Validation of netCDF diagnostics

In this version, we did a complete comparison of the BPCH diagnostics to the netCDF diagnostics in preparation for retiring the old BPCH diagnostics. Comparison plots can be accessed via the links below. A summary of the results may be found on this Google spreadsheet.

NetCDF collection BPCH equivalent(s) Plots
Level Map Zonal Mean
AdvFluxMerid ND25 sfc full column
AdvFluxVert ND26 sfc full column
AdvFluxZonal ND24 sfc full column
AerosolMass ND42 sfc full column
Aerosols ND21 sfc full column
CloudConvFlux ND14 500hPa full column
ConcAfterChem ND43 sfc full column
DryDep ND44 sfc
Emissions
(i.e. HEMCO diagnostics)
ND06, ND07, ND08, ND11,
ND13, ND28, ND29, ND32,
ND34, ND36, ND46
sfc

PlantDecay

JValues (24-hr avg) ND22 sfc full column
JValuesLocalNoon ND22 sfc full column
LevelEdgeDiags ND31, ND66 850hPa full column
ProdLoss ND05, ND07, ND65 sfc full column
SpeciesConc ND45 sfc full column
StateChm ND73 sfc full column
StateMet ND57

ND31, ND55, ND66, ND67, ND68, ND69

sfc

sfc

full column

full column

WetLossConv ND38, ND37 850hPa full column
WetLossLS ND39 850hPa full column