Atmospheric Inversion results

Inversion results

Atmospheric Inversion resultsPhilippe PeylinLSCE, FranceRachel Law CSIRO, AustraliaKevin Gurney, Xia Zhang Arizona State University/Purdue University, USAZegbeu poussi LSCE, Franceand model participantsViterbo RECCAPTutorial 1 : LogisticsResults accessible at: http://transcom.lsce.ipsl.frPlotted maps and time series (component fluxes X aggregated regions) no psswrdData files (component fluxes X aggregated regions and gridded) - psswrd11 state of the art inversions (some multiple submissions)All re-gridded to a common grid (1x1, monthly)Component fluxes = prior, fossil fuel, posterior (estimate)Regional fluxes = land/ocean regions + various aggregates (109)Region mask boundaries of the chosen regional breakdown and aggregates (regular and extended) We urge regional leads to contact inverters for additional clarificationThe complete 1x1, monthly files are also available roll your own regions

transcom1 of these is previous T3 averageThese exclude some submissions that were too shortTutorial 2 : ParticipantsNameTime periodTransport model WindsAtm DataFlux spatial res.Flux temp. Res.Inverse MethodLsce_an_v2.11996 - 2004LMDZ v4ECMWFMonthly meangridcellmonthlyBayesian MatrixLsce_var_v1.01990- 2008LMDZ v3 ECMWFRawgridcell8 day with night & day sepVariationalJena_s96_v3.21996 - 2008TM3NCEP reanalysisRawgridcellDailyVariationalCarbntrckr_US2000 - 2008TM5 zoomRaw156 ecoregionsweeklyKalman smootherCarbntrckr_EU2000 - 2007TM5 zoomRaw145 land + 30 oceanweeklyKalman smootherRigc_patra1993 - 2007NIES/FRCGCNCEP reanalysisMonthly mean64MonthlyBayesian MatrixT3 mean1995 - 200813 models13 models (climatology)Monthly mean22MonthlyBayesian MatrixJMA1985 - 2007Monthly mean22MonthlyBayesian MatrixNicam_NiwaNICAM-TMMonthly mean22MonthlyBayesian MatrixC13_MATCHRayner1992 - 2005MATCHNCEP 1999 (climatology)Monthly mean116MonthlyBayesian MatrixC13_CCAM Law1992 - 2005CCAMNCEP 1999 (climatology)Monthly mean146MonthlyBayesian MatrixWe are choosing a start date of 1995 for RECCAPTutorial 3: Miscellaneous notesFossil emissions differed between inversions:Attempted fix: Fossil adjustment with common fossil fuel CO2 fluxUncertainties: (in progress, Enting et al., Transcom)Bayesian errors will be provided for key regionsSpread induced by changing model components available for few cases (contact individual modeler)Validation against independent data (in progress)Atm. Vertical profile, campaign, independent C-cycle obs, etcRegridding & land/sea mask can create inaccuracies special care for regions with shaped coastal boundariesWebsite

Thank you Zegbeu!Tutorial 4: Proceed with CautionPoorly-constrained region (Africa)

Well-constrained region (Europe)

4 continued: Consequences of poor constraintsInfluence of priorSensitivity to an individual siteAustralian seasonal cycle from 4 inversions. Posterior (solid), prior (dashed). Different prior seasonality due to inclusion or not of biomass burning.

Jena (dark blue) only inversion without EIC and very different E-W flux split4 continued: Caution with big region estimates

Large flux dipoles : regional estimates should be interpretedwith great care depending on the boundary of the regions

EUROPENorth AsiaAnnual total flux (GtC/yr)spurious anti-correlations a form of representation error

Results 1 : Long term meansLSCE_an_v2.1JENA_s96_v3.2CTracker_EULSCE_var_v1.C13_MATCHCTracker_USTRCOM_meRIGC_patra JMA_2010C13_CCAMNCAM_NiwaNorthern land + oceanNatural fluxes (GtC/yr) 2000-2003 periodTropical land + oceanSouthern land + oceanThe following results are preliminary and based on current submissions. These may change at a future time.

Results 1 : Long term means LSCE_an_v2.1JENA_s96_v3.2CTracker_EULSCE_var_v1.C13_MATCHCTracker_USTRCOM_meRIGC_patra JMA_2010C13_CCAMNCAM_NiwaEuropeN. AmericaN. AtlanticN. AsiaNatural fluxes (GtC/yr) 2000-2003 periodResults 2: IAV (land)

LSCE_an_v2.1JENA_s96_v3.2CTracker_EULSCE_var_v1.C13_MATCHCTracker_USTRCOM_meRIGC_patra JMA_2010C13_CCAMNCAM_Niwa

Northern landTropical landSouthern landTotal landResults 2: IAV (ocean)

LSCE_an_v2.1JENA_s96_v3.2CTracker_EULSCE_var_v1.C13_MATCHCTracker_USTRCOM_meRIGC_patra JMA_2010C13_CCAMNCAM_NiwaNorthern oceanTropical oceanSouthern oceanTotal oceanResults 2: IAV (continental scale)

EuropeN. AmericaN. AtlanticN. Asia

LSCE_an_v2.1JENA_s96_v3.2CTracker_EULSCE_var_v1.C13_MATCHCTracker_USTRCOM_meRIGC_patra JMA_2010C13_CCAMNCAM_NiwaResults 3 : Seas. Cycle & IAV

Mean amplitude of all models Standard dev. of all models

Mean IAV for all models Standard dev. of all models Seas.CycleIAVResults 4 : N-S zonal mean integrated fluxes

SummaryNew set of 11 inverse results (which includes TRcom3 mean as 1)WEB-site available for downloadingUncertainties & Validation in progress (part of Transcom)Results may be updated with announcementProceed with caution together with inverse specialists

Differences in long term mean but coherence in IAVat large scalesSmall scales regional results sensitive to: Network, Priors, Flux resolution Thank YouGlobal Inversion Status/PlansResults accessible at: http://transcom.lsce.ipsl.fr11 results available at 1x1 & region (individual and aggregate: 109)1995 to 2008 (inc) but individuals will run back to 1980 monthly/annualFluxes include: prior, fossil, posterior, total, adjusted fluxesRegion mask and explanation availableUpdates will occur (we will announce)Uncertainty work & decisions ongoingDecision on a weighted mean ongoingProceed with caution (tentative inverter assignments made for ocean)TransCom meeting: December San Francisco AGU