Atmospheric Modeling with Interac5ve Chemistry · Atmospheric Modeling with Interac5ve Chemistry Presented by Simone Tilmes ACOM/CGD • Running with fully-coupled chemistry, diﬀerent

CESMTutorial2016,Chemistry

AtmosphericModelingwithInterac5veChemistry

PresentedbySimoneTilmesACOM/CGD•  Runningwithfully-coupledchemistry,differentchemical

mechanismversions•  DescripJonofsecondaryorganicaerosolsasparameterizedin

CAM-chem•  EvaluaJon•  OtherapplicaJonsofCAMchem


TroposphericChemistryandAerosols

Photo-chemistryGas-phasechemistryHeterogeneouschemistryAerosolformaJonOHdeterminesthelifeJmeofMethane


ModelingwithChemistryAvailablechemicalmechanismsinCAM-ChemSuperfastChemistry(CAM4/5):

12species,simplechemistrymechanism,CH4prescribedLINOZ+Cariolleinstratosphere,fullycoupled

BulkAerosolModel(BAM)(CAM4/5):IncludesBlackCarbon,OrganicCarbon,SeaSalt,Dust(prescribedmonthlyfieldsofCO2,CH4,O3,OH,HO2,NO2,N2O,SO2/SO4)

Troposphericchemistry(trop_mozart)(CAM4/5):Troposphericmechanism,over100species(MOZART:Emmonsetal.,2010)Stratosphericchemistryisprescribedabout50hPa:(O3,HNO3,CH4,CO)Emissions,Dry/WetDeposiJonSecondaryOrganicaerosols

Plusstratosphericchemistry(trop-stratmozart)(CAM4/5):TroposphericandStratosphericmechanism(~122species)includingstratosphericheterogeneousreacJons,about300reacJons(similartoWACCM)(Lamarqueetal.,2012,Tilmesetal.,2015)


ChemistryinCESMCAM5

EmissionsofSO2AndAerosolsAnthropogenicBiomassburningBiogenic

Removal:Wet/DryDeposiJon

Land/OceanBiosphere

Radia5onClouds(indirecteffect)Dynamics/H2O

Snow/Ice

SimpleChemistry

AerosolformaJon(MAM)


ChemistryinCESMCAM5-Chem

ComplexChemicalReacJons

(chemicalmechanism)AerosolformaJon

EmissionsAnthropogenicBiomassburningBiogenic

Removal:Wet/DryDeposiJon

Land/OceanBiosphere

Radia5onClouds(indirecteffect)Dynamics/H2O

Snow/Ice

MEG

AN

CAM4-chemdoesnotincludecouplingbetweenaerosolsandclouds


1.SoluJons:listsallchemicalconsJtuenciesanddefineswhattheyare,e.g.•  IncludesFixedSpecies:listsofchemicalspeciesthatdonotchange•  AndNon-transportedComponents(veryshortlifeJme)2.SoluJonclasses,dividessoluJonspeciesintoexplicitandimplicit(shorterJmestepinthesolver)4.Chemistry:•  Photolysis:[jo3_a]O3+hv->O1D+O2•  Gas-phasechemistry:[O_O3]O+O3->2*O2;8.00e-12,-2060.•  HeterogeneousReacJon[usr_N2O5_aer]N2O5->2*HNO3

ModelingwithChemistryChemicalMechanism:includessetofequaJonsThechemistrypreprocessor:toolthatgeneratesCAMFortransourcecode;numericallysolveasetofdifferenJalequaJonswhichrepresentthechemicalreacJons->temporalevoluJonofthechemicaltracers

rates(temperaturedependenceetc.)

ISOP -> C5H8,BIGENE->C4H8,BIGALK->C5H12,MEK->C4H8O

UserdefinedreacJons:aredefinedinthemodelcode:mo_usrrxt.F90


Emissions:aredefinedintheuser_nl_camnamelistvariable•  Dependingonmechanism,differentspeciesneedtobeemiled

Emissions

•  Chemicalcomponentsarecurrentlyemiledatthesurface.

•  Includeanthropogenic,biomassburningandfireemissions.

•  BiogenicemissionscanbecalculatedbyMEGAN(makesuretonotdouble-count)

•  Comingsoon:fireemissionsmaybecalculatedbythelandmodel

->Carefulifchangingemissions,tomatchyourmodelsetup


Emissions:surfaceemissionsfields,fixedboundarycondiJons,calculatedusingvegetaJontype(biogenic,VOC)

Lamarqueetal.,2010


OrganicAerosols(simulatedinCAM4-Chem)

FromC.Heald,MITCambridge

Forma5onofSOA•  Emissionsof

volaJleorganiccarbons

•  FormaJonofSemi-VolaJles

•  EmissionsofPOM->SOA


WaterSolub

ility

log(H)(M

/atm

)

Vola=lity,log(C*),μg/m3

COCO2

Oxidative Chemistry

Fragmentation

SVOCi

SVOCn

SOA

dry & wet depo.

of SOA and soluble SVOC Emissions

S/IVOC,VOC

Removal by photolysis

+Ox

Simplis=cwaysoftrea=ngthecomplexSOAlifecycle

Fi0edtochamberdata

VOCSOASOAyield:netconversion

Vola;lityBasisSet VOC+OH,O3

2-productsBAM VOC+OH,O3

WithSolubilityaxis

VOCCAM5BASE SOG directlyemitsSOG

MorephysicalapproachDirectcouplingtobiogenicemissionschangesfromMEGAN->couplesSOAforma5ontolanduseandclimatechange->onlyworksinfullchemistryversionatthispoint

NewSOAapproachinCESM2CAM5-Chem/WACCM(nextyear)


Aerosols:CAM5ModalAerosolModel(MAM3/7)

Liuetal.,2011

a2 a1 a3

a2 a1

a3 a6

a5a4

a7

MAM3

MAM7

712 X. Liu et al.: Toward a minimal representation of aerosols in climate models

Table 1. Geometric standard deviations (�g) and dry diameter sizeranges for MAM3 and MAM7 modes. The size range values arethe 10th and 90th percentiles of the global annual average numberdistribution for the modes (from simulations presented in Sect. 3).

Mode �g Size range (µm)

MAM3

Aitken 1.6 0.015–0.053Accumulation 1.8 0.058–0.27Coarse 1.8 0.80–3.65

MAM7

Aitken 1.6 0.015–0.052Accumulation 1.8 0.056–0.26Primary Carbon 1.6 0.039–0.13Fine Sea Salt 2.0 0.095–0.56Fine Dust 1.8 0.14–0.62Coarse Sea Salt 2.0 0.63–3.70Coarse Dust 1.8 0.59–2.75

single coarse mode based on the recognition that sources ofdust and sea salt are geographically separated. Although dustis much less soluble than sea salt, it readily absorbs water(Koretsky et al., 1997) and activates similarly as CCN (Ku-mar et al., 2009), particularly when coated by species likesulfate and organic. So dust is likely to be removed by wetdeposition almost as easily as sea salt, and the merging ofdust and sea salt in a single mode is unlikely to introducesubstantial error into our simulations; (3) the fine dust andsea salt modes are similarly merged with the accumulationmode; (4) sulfate is partially neutralized by ammonium inthe form of NH4HSO4, so that ammonium is effectively pre-scribed and NH3 is not simulated. The total number of trans-ported aerosol tracers in MAM3 is 15. The transported gasspecies are SO2, H2O2, DMS, H2SO4, and a lumped semi-volatile organic species. The prescribed standard deviationand the typical size range for each mode are given in Table 1.

3 Aerosol distributions and budgets

All simulations are performed with the stand-alone versionCAM5.1, using climatological sea surface temperature andsea ice and anthropogenic aerosol and precursor gas emis-sions for the year 2000. The model is integrated for 6 yr, andresults from the last 5 yr are used in this study. In this sectionmodel-simulated global distributions and budgets for differ-ent aerosol species are analyzed and comparisons are madebetween MAM3 and MAM7.

65

Figure 2. Predicted species for interstitial and cloud-borne component of each aerosol

mode in MAM3.

Fig. 2. Predicted species for interstitial and cloud-borne componentof each aerosol mode in MAM3.

3.1 Simulated global aerosol distributions

Figures 3a and b show annual mean vertically integrated (col-umn burden) mass concentrations of sulfate, BC, POM, SOA,dust and sea salt from MAM3, and the relative difference ofthese concentrations between MAM7 and MAM3, respec-tively. These aerosol species concentrations are summationsover all available modes (e.g., the POM concentration inMAM7 includes contributions from the primary carbon andaccumulation modes). Sulfate has maximum concentrationsin the industrial regions (e.g., East Asia, Europe, and NorthAmerica). The distribution patterns and absolute values ofsulfate concentration are very similar (mostly within 10%)between MAM3 andMAM7 (Fig. 3b). This is expected sincemost of sulfate burden (⇠90%) is in the accumulation mode(see sulfate budget in Sect. 3.2). This is also the case forSOA, which has high concentrations over the industrial re-gions and tropical regions with strong biogenic emissions(e.g., Central Africa and South America). The differencesbetween MAM3 and MAM7 are generally small (mostlywithin 10%). POM column concentrations have spatial dis-tributions and magnitudes similar to SOA, but are lower inEurope, Northeastern US and South America, and higher inCentral Africa. The distribution patterns of BC burden con-centrations are similar to those of POM, but have relativelylarger contributions from the industrial regions because ofdifferent emission factors of BC/POM from different sectors.Dust burden concentrations have maxima over strong sourceregions (e.g., Northern Africa, Southwest and Central Asia,and Australia) and over the outflow regions (e.g., in the At-lantic and in the western Pacific). Sea salt burden concentra-tions are high over the storm track regions (e.g., the SouthernOcean) where wind speeds and emissions are higher, and inthe subtropics of both hemispheres where precipitation scav-enging is weaker.One major difference between MAM3 and MAM7 is the

treatment of primary carbonaceous aerosols (POM and BC).These aerosols are instantaneously mixed with sulfate andother components in the accumulation mode in MAM3 oncethey are emitted, and thus are subject to wet removal byprecipitation due to the high hygroscopicity of sulfate. InMAM7, carbonaceous aerosols are emitted in the primary

Geosci. Model Dev., 5, 709–739, 2012 www.geosci-model-dev.net/5/709/2012/


DifferencesinAerosols,CAM4andCAM5Annual SummerWinter

Sulfate:SO4SO4(a1,a2,a3)SO4(a1,a2,a4-7)

BC:CB1,CB2BC(a1)BC(a1,a3)

OC:CB1,CB2,SOApom(a1,a3),soa(a1,a2)pom(a1,a3),soa(a1,a2) SurfacelayerconcentraJon(mg/m3)

IMPROVECAM4-BAMCAM5-MAM3CAM5-MAM7


Scien5ficallyValidatedChemistryVersionshZp://www.cesm.ucar.edu/models/scien5fically-supported.htmlTilmes,Setal.,2015:DescripJonandevaluaJonoftroposphericchemistryandaerosolsintheCommunityEarthSystemModel(CESM1.2),Geosci.ModelDev.,8,1395-1426,doi:10.5194/gmd-8-1395-2015,2015.(hlp://www.geosci-model-dev.net/8/1395/2015/gmd-8-1395-2015.html)


Evalua5onoftheModel

AMWGDiagnos5cPackageincludesChemistryEvalua5on


Evalua5onoftheModelAMWGDiagnos5cPackageincludesChemistryEvalua5on


CCMISimula5onsinComparisontoOzonesondeData:Mid-La5tudes,500hPa

EasternUS900hPa

50hPa250hPa

500hPa

WACCMREFC1CAM-ChemREFC1 1995-2004average


PerformanceoftheModels:CarbonMonoxideinComparisontoMOPITT


ChangingChemistryChemicalMechanism:addorchangetracersandreacJons(e.g.Taggingoftracers)Emissions:includerequiredemissionsDeposi5on:adddryandwetdeposiJoncomponentstothemodelcode

RunningwithMEGANEmissions:hlp://www.cesm.ucar.edu/working_groups/Chemistry/running_CESM1_MEGANNv0408.pdfRoadmaptorunandchangechemistry(CESM1.2.0)hlp://www.cesm.ucar.edu/working_groups/Chemistry/roadmap_cesm120.pdf


TheModelforOzoneandRelatedchemicalTracers(MOZART4)Emissions:StreetsARCTASemissions+dailyfires(C.Wiedinmyer)VerJcalInjecJonofFireEmissionbetween0-6km

South Asia

O3,CO,BCtagswithOfflineMeteorologyEmmonsetal.,2012,GMD


ImportanceofAnthropogenicCOEmissionswithout South Asia and SH South Asia and SH only

April

COaveragedcolumnbetweensurf.and200hPa


without South Asia and SH South Asia and SH only April

ImportanceofAnthropogenicCOEmissions



without South Asia and SH South Asia and SH only April

ImportanceofAnthropogenicCOEmissions



ObservationsØ  MOPITT-CO, IASI-CO, MODIS AODØ  Meteorological observations

Observations, plus errors

Ensemble of optimized initial conditions every 6 hours

DARTAssimilation

-> update CO concentrationAnd Meteorology

Weighted mean of observations and model knowing respective errors

CESMCommunityAtmosphereModel(CAMandCAM-Chem)Ensembleofatmosphericstatewithchemistry

Ø  Ensemble of emissions (+CO tags)

Ø  Ensemble of transportØ  Ensemble of

deposition (land model)

Ø  Ensemble of Chemistry


CAM-chemforecas5ng

FortheNASAKORUS-AQexperiment,CAM-chem/DARTforecastswererun

MOPITTCOretrievalswereassimilatedeachday(30-memberensemble)

GEOS-5meteorologywasusedtodriveasingleCAM-chemforecastsimulaJon

TaggedCOtracerswereincludedtotracksourcecontribuJonstoKorea

hlp://www.acom.ucar.edu/acresp/korus-aq/


WACCMandCAM-ChemCustomerSupport

CGD Forum: http://bb.cgd.ucar.edu/ Mike Mills WACCM Liaison [email protected] (303) 497-1425 Simone Tilmes CAM-Chem Liaison [email protected] (303) 497-1445

Documents

Atmospheric Modeling with Interac5ve Chemistry · Atmospheric Modeling with Interac5ve Chemistry Presented by Simone Tilmes ACOM/CGD • Running with fully-coupled chemistry, diﬀerent