MACC GA Toulouse GRG presentation by yaofenji


									    Global Reactive Gases
             Martin Schultz
IEK-8, Forschungszentrum Jülich GmbH
•   MACC G-RG (13 partners) combines heritage from GEMS
    (12 partners) and PROMOTE (5 partners)
•   PROMOTE heritage:
    •   level 2 satellite data products (GOME, GOME-2, Sciamachy, OMI)
    •   decentralized data assimilation for stratospheric (and total column)
        ozone (SACADA, BASCOE, TM3DAM)
•   GEMS heritage:
    •   quasi-operational monitoring of tropospheric and stratospheric
        composition with IFS-MOZART and coupling achieved for IFS-
        TM5 and IFS-MOCAGE
    •   reanalysis 2003-2008
    •   support of scientific field campaigns
    •   a lot of validation activities

                                                               Introduction   Slide 2
MACC G-RG comprises 4 work packages:
• WP 1: Satellite based monitoring of stratospheric ozone and
  tropospheric trace gas columns
• WP 2: Consolidation and improvement of integrated global
  stratospheric ozone service
• WP 3: Consolidation and improvement of integrated service for
  global tropospheric reactive gases
• WP 4: Development of fully integrated chemistry transport in

                                                    Introduction   Slide 3
WP 1 objective:
Continue existing decentralized services from
PROMOTE that deliver value-added satellite data
products related to stratospheric ozone and
tropospheric trace gas columns (ozone, NO2,
HCHO, CO, SO2) to end users

                                        WP1 Status   Slide 4
•   Task G-RG_1.1:
    Near-real time provision of ozone and NO2 data from OMI,
•   Task G-RG_1.2:
    Stratospheric ozone record and NRT service using SACADA 4D-Var
•   Task G-RG_1.3:
    Stratospheric ozone record and NRT service using BASCOE 4D-Var
•   Task G-RG_1.4:
    Total ozone record, monitoring and forecast service Assimilation and
    forecasts of global stratospheric ozone
•   Task G-RG_1.5:
    Validation of WP G-RG 1 trace gas services

                                                           WP1 Status   Slide 5
•   Main results of first reporting period
    •   Routine provision of satellite based data for European
        GOME-2, OMI and SCIAMACHY
    •   Quasi-operational application of independent
        assimilation systems:
        BASCOE, SACADA, TM3DAM for stratospheric
        ozone chemistry
    •   Initial validation of stratospheric (assimilation)

                                                   WP1 Status   Slide 6
                        Task GRG 1.1

Provision of satellite based data (level-2)
•   Most level-2 data on stratospheric ozone, BrO,
    tropospheric and total NO2, CH2O and SO2 is now
    available in NRT via MACC and/or dedicated access
•   Improvements of the trace gas retrieval w.r.t. speed,
    temperature data, new (standard) absorption spectra
•   Reprocessing for GOME(1995-2003) and SCIA(2002-

                                                  WP1 Status   Slide 7
                    Tasks GRG 1.2--1.4

Stratospheric ozone services
•   SACADA assimilation of SCIA nadir ozone observations
    since March 2010.
•   BASCOE assimilation of MLS AURA data since
    December 2009.
•   Multi-instrumental 30 year reanalysis based on TOMS,
    GOME, SBUV, SCIA, OMI and GOME-2 data.
•   SCIA ozone forecasts are now corrected for instrumental
    calibration issues, which is especially important for UV

                                                 WP1 Status   Slide 8

Assimilated total ozone record for the period 1978 – 2008 based on satellite
observations of TOMS, SBUV, GOME, SCIAMACHY, GOME-2 and OMI

R. J. van der A, M. A. F. Allaart, and H. J. Eskes (2010): Multi sensor reanalysis of total ozone,
Atmos. Chem. Phys. Discuss., 10, 11401-11448.
                                                                                      WP1 Status     Slide 9
                           Task GRG 1.5

Intercomparison and comparison to independent data
(focus on 2003 episode)
•   BASCOE and SACADA more similar than TM3DAM.
•   Results agree well in regions with good daily data coverage.
•   Best correlation with independent data for the middle and high
    northern latitudes. Worst during ozone hole conditions.
•   Most systematic deviations occur in data void regions.
•   CTMs are capable to reproduce the Antarctic ozone hole. Though,
    timing and intensity differs from observations.

                                                        WP1 Status   Slide 10
                                WP1 Deliverable Status

        Near-real time provision of
        ozone and NO2 data from          M4
D 1.1
        OMI, SCIAMACHY, GOME           onwards
        and GOME-2
        Stratospheric ozone record
D 1.2   and NRT service using                            Since March 2010
        SACADA 4D-Var
        Stratospheric ozone record
D 1.3   and NRT service using                            Since Dec 2009
        BASCOE 4D-Var
        Total ozone record,
D 1.4   monitoring and forecasting                       (TM3DAM) continued

        Validation report on
D 1.5                                   M18              In preparation
        stratospheric ozone services

                                                         Deliverable added during first MACC assembly.
        Unified web interface for                        Integration of stratospheric assimilation services
D 1.7   integrated MACC and former      M15      o       achieved (see Task 2.3), integration of level 2
        PROMOTE services                                 satellite products and tropospheric services

                                                                                   WP2 Status        Slide 11
WP 2 objective:
Consolidate, operate and improve the integrated
global reactive gases forecasting for stratospheric
ozone developed in the GEMS project with
products comprising of ozone, N2O, CH4, BrOx,
ClOx and others based on user-consultation,
including the extended validation with independent
data and through well-defined case studies

                                          WP2 Status   Slide 12
Task G-RG_2.1: Preparation of datasets for stratospheric model
Task G-RG_2.2: Quasi-operational monitoring and evaluation of
    MACC integrated stratospheric ozone service
Task G-RG_2.3: Development of improved web-based service
    products and documentation
Task G-RG_2.4: Improvement integrated global stratospheric
    chemistry model
Task G-RG_2.5: Non-operational validation of continued GEMS
    stratospheric ozone service (case studies)
Task G-RG_2.6: Technical and scientific documentation of the
    integrated global stratospheric chemistry model
Task G-RG_2.8: Validation of initial CT-IFS results

                                                        WP2 Status   Slide 13
                 Major accomplishments in WP2

•   Acquiring and maintenance of necessary datasets
    •   NRT groundbased and satellite observations
    •   NRT and historic model output
•   Creation of the Stratospheric Ozone Webpage:

    •   Centralized stratospheric ozone products:
        MACC, BASCOE, SACADA and TM3DAM shown side-by-
        side, allowing quick comparison
    •   Initial NRT evaluation of stratospheric services

•   Extensive improvement in automated evaluation software
    allowing for quasi-operational monitoring and evaluation

                                                      WP2 Status   Slide 14
                    Major accomplishments in WP2

Evaluation with statistical plots in observation space shows:
•   IFS-MOZART has not been able to simulate/forecast polar O3
•   Elsewhere: Both BASCOE CTM and IFS-MOZART overestimate
    (+20%) in the lower stratosphere and underestimate (-20%) in the
    upper stratosphere
•   Monitoring and reanalysis of total O3 columns: very successful
    ... but vertical distribution of the analyses is wrong (bias ~ 20%)
     •   in South Pole vortex where model is too biased
     •   when no profile is assimilated…

                                                               WP2 Status   Slide 16
                    Antarctic ozone hole problem in MOZART
         Monthly mean vertical profiles at Neumayer station, Antarctica

ez2m MOZART 3.1, ff0f wetdep bug fix, f3yj Analysis
                                                                 WP2 Status   Slide 17
                       Antarctic ozone hole problem in MOZART

   Simulation results with MOZART 3.5.02 showing ozone depletion down to
   ~140 DU in September 2003 (old version had a minimum of ~220 DU)

Offline results with NCAR settings very similar to MACC settings; integration into MACC-IFS ongoing

                                                                                  WP2 Status     Slide 18
                                 WP2 Deliverable Status
        Inventory of stratospheric
        composition datasets for
D 2.1                                  M6, M24
        validation in NRT and
        delayed mode
        Quasi-operational monitoring
                                                                Basic system continued from GEMS and side-
        and evaluation chain for         M6
D 2.2                                                           by-side comparisons with SACADA and
        MACC integrated                onwards
                                                                BASCOE (from WP1)
        stratospheric ozone service
        Service product catalogue
        and web documentation of
D 2.3                                  M12, M24       
        stratospheric ozone
                                                                Albeit the reason for the IFS-MOZART
                                                                deficiency to simulate Antarctic ozone depletion
                                                                are still unclear, a new model version (MOZART
        Updated stratospheric
D 2.4                                    M12                    3.5.02) which was received from NCAR in
        chemistry model
                                                                September 2010 shows much improved
                                                  (Delay 6 M)   simulation results. The new model is currently
                                                                integrated in the MACC-IFS system.
        Stratospheric case study
D 2.5   model results and evaluation     M18                    In preparation

                                                                                         WP2 Status       Slide 19
WP 3 objective:
Consolidate, operate and improve the integrated
global reactive gases forecasting for tropospheric
ozone, ozone precursors (NOx, CO, HCHO, SO2,
selected NMVOC and others) and oxidizing
capacity developed in the GEMS project, including
the extended validation with independent data and
through well-defined case studies

                                          WP3 Status   Slide 20
Task G-RG_3.1: Prepare datasets for tropospheric model validation
Task G-RG_3.2: Quasi-operational monitoring and evaluation of MACC
    integrated tropospheric trace gas service
Task G-RG_3.3: Improve integrated global tropospheric chemistry model
Task G-RG_3.4: Development of improved web-based service products and
Task G-RG_3.5: Adapt G-RG model to use new vegetation fire emission data
    and parameterisations from D-FIRE
Task G-RG_3.6: Adapt G-RG model to use new anthropogenic and natural
    emission data and parameterisations from D-EMIS
Task G-RG_3.7: Non-operational validation of continued GEMS tropospheric
    trace gas service (case studies)
Task G-RG_3.8: Technical and scientific documentation of the integrated
    global tropospheric chemistry model
Task G-RG_3.9: Negotiation of an SLA with a key user for tropospheric trace
    gas service post-MACC
                                                                WP3 Status    Slide 21
Main achievements:
•   4 NRT streams:
     •   IFS-MOZART with assimilation of CO and ozone
     •   IFS-TM5 with assimilation of CO and ozone
     •   IFS-MOZART without assimilation
     •   IFS with tagged CO-like tracers
•   Preparation of the MACC re-analysis with IFS-MOZART
     •   Code and emission update & resolution increase
     •   Optimisation of AN suite with coupled system
•   Tracer forecasts, plume modelling and analysis
     •   Eyjafjalla eruption in April 2010
     •   Russian fires in July 2010
•   Further development of validation metrics and web services

                                                           WP3 Status   Slide 22
1-2 slides with NRT stream results
(could also be tied in with Russian fires…)
        GEMS versus MACC reanalysis                     MACC

Ozone                                            CO

                  For O3, Fbov shows an improvement over f026, however the
                  O3 anomaly from 1-14 August still underestimated.
                  Fbov underestimates the CO concentration throughout the
                  atmosphere and both IFS runs fail to capture the increase in
                  CO near 5000m due to the urban emissions and forest fires in
                  southern europe.

                                                          WP3 Status     Slide 24
                                          MACC reanalysis

CO – long-range transport over Atlantic (30W)

     Zonal CO Flux = U * MMR_CO * ρ

                                           WP3 Status   Slide 25
                 Eyjafjalla eruption: plume modelling

•   GEMS&MACC developments allowed for quick
    implementation of tracer forecast within 24 h after
    eruption using different injection height assumptions
•   Good agreement in shape with forecast from VAAC -
    Metoffice and others
•   Large uncertainty in emission source strength and
    injection height
•   Ongoing experiments with data assimilation of SO2
•   Ongoing inter-comparison of plume forecast within
    ENSEMBLE framework (Dispersion models)

                                                        WP3 Status   Slide 26
Eyjafjalla eruption: plume modelling

                                       WP3 Status   Slide 27
  Iceland -

                   The potential use of SO2 column data
                   to assimilate volcanic plumes

                   MACC models currently don‘t account for
                   volcanic emissions in NRT

   Congo -                    Banks Islands -
 Nyamuragira                      Gaua

                                                WP3 Status   Slide 28
                 Russian forest fires 1-15 August 2010:
                       Assimilation of IASI CO


Agreement between IASI and MOPITT is good; IASI slightly higher.

Mean of IASI data used in the assimilation underestimates, because high values
get first-guess and varqc rejected

                                                                   WP3 Status    Slide 29
  CO column          Russian forest fires 1-15 August 2010

TM5-semi-oper               TM5-GFASv0                   MOPITT-V4

 • No CO assimilation for    •Assimilation of MOPITT CO
   current period            •MACC emissions
 • RETRO/REAS emissions      •GFASv0
 • GFEDv2 climatology

          Model is drawn towards observations
                                                             WP3 Status   Slide 30
  NO2 column       Russian forest fires 1-15 August 2010

TM5-semi-oper             TM5-GFASv0                   OMI
     1      2

 •No NO2 assimilation    •Assimilation of OMI NO2
 •RETRO/REAS emissions   •MACC emissions,
 •GFEDv2 climatology     •GFASv0

         1. Model is drawn towards observations
         2. Artificial spots of wildfires are suppressed
                                                             WP3 Status   Slide 31
WP3 Status   Slide 32
IFS-TM5 model
Assim uses IASI CO columns

 GFAS doesn‘t capture burning
 events or emission magnitude
 leading to „observed“ CO
        Development of tropospheric GRG services

MACC pages at ECMWF                    BC service at Jülich

SCIAMACHY val. at IUP            MOZAIC/IAGOS val. at Toulouse
             Use of global boundary conditions for regional
                             AQ modeling

GEMS-RAQ model: MM5/CAMx
Climatic Boundaries vs. MOZART-GRG f026 boundaries
Comparison with MOZAIC

                                                              WP3 Status   Slide 35
                                WP3 Deliverable Status

        Inventory of tropospheric composition
D 3.1   datasets for validation in NRT and
        delayed mode
        Quasi-operational monitoring and
D 3.2   evaluation chain for MACC integrated
        tropospheric reactive trace gas service
        Improved tropospheric chemistry
D 3.3   model based on GEMS validation              M6
        Service product catalogue and web
D 3.4   documentation of tropospheric reactive
        trace gases evaluation
        Updated tropospheric chemistry model
D 3.5   code for use with vegetation fire          M12
        emissions from D-FIRE

                                                                       • Definition of upgrades in D-FIRE
        Updated tropospheric chemistry model
D 3.6   code for use with anthropogenic and        M18                   products
        natural emissions from D-EMIS                       Delay 3M   • Needed to fix stratospheric ozone issue

        Tropospheric reactive gases case
D 3.7   study model results and evaluation         M18                 In preparation

                                                                                         WP3 Status       Slide 36
WP 4 objective:
Begin the development of a fully coupled chemistry
transport model based on the ECMWF integrated
forecasting system in order to eliminate
inconsistencies arising from the coupled set-up in

                                         WP4 Status   Slide 37
Task G-RG_4.1: Design study for the integrated CT-IFS
Task G-RG_4.2: Analysis of IFS transport parameterisations for use
    with reactive gases
Task G-RG_4.3: Implementation of simplified linear chemistry
    schemes for CO and its adjoint code
Task G-RG_4.4: Preparation and implementation of chemistry
Task G-RG_4.5: Preparation and implementation of emission
Task G-RG_4.6: Preparation and implementation of deposition
Task G-RG_4.7: Testing and optimizing of the integrated CT-IFS

                                                        WP4 Status   Slide 38
                        C-IFS Development Status

•   Expanded IFS-code to run with 100+ tracers
•   Scripts to run C-IFS and to archive results (not in mars yet)
•   Global mass, source and sink diagnostic
•   Global tracer mass fixer (same relative change in MMR at all grid
    points to ensure conservation)
•   Implementation of TM5 chemistry package for troposphere (provided
    by KNMI)
•   Cariolle-scheme for stratospheric ozone
•   Integration of wet-deposition and lightning modules
•   Successful completion of first one-year run with good results

                                                              WP4 Status   Slide 39
                  222Rn   simulation with C-IFS
                                                  Area-averaged 222Rn
C-IFS                TM5                          profiles at 12 UTC…

                                                   … and at 24 UTC.

        900 hPa

                                                          WP4 Status    Slide 40
                       Surface ozone simulation with C-IFS


Differences to be expected, because of different wet deposition/dry deposition schemes

                                                                        WP4 Status   Slide 41
                                                                    IFS Tracer Transport

•Species emitted at surface are increased by non-conservation of semi-lagrange advection
•Ozone (and other stratospheric species) tend to be decreased

                                                                            WP4 Status     Slide 42
                 C-IFS physical chemistry parameterisations
•   NO lightning emissions
    •   Three different parameterisations for flash rate density using cloud
        height (Price and Rind, 1993) , convective precipitation (Meijer et al,
        2001) or updraft velocity & ice cloud height (P. Lopez) implemented
•   Wet deposition
    •   Simple parameterisation based on precipitation fluxes and clouds
    •   Re-evaporation and in-cloud scavenging in convection routine
•   Dry deposition
    •   Constant surface flux in vertical diffusion
    •   More explicit treatment
•   Photolysis rates
    •   Look up-table with corrections for cloud optical depth
    •   Use (extended) SW radiation scheme
                                                                   WP4 Status     Slide 43
      Lightning NOx: Flash frequency parameterisations
   Observations                           Meijer 2001 (TM5)
    LIS OTD                                Conv. Precip.

Price and Rind, 1993                          Lopez p.c.
 Conv. Cloud height                   Updraft & Ice Cloud height

                                                         WP4 Status   Slide 44
          Lightning NOx: Flash frequency parameterisations
        LIS OTD

     Grewe et al., 2001                           Lopez p.c.
Updraft & Conv. Cloud height              Updraft & Ice Cloud height


                                                             WP4 Status   Slide 45
                       C-IFS development plans for P2

•   Implement MOZART and MOCAGE chemistry modules
•   Consolidate input/output data handling for C-IFS
•   Continue work on mass diagnostics and simple mass fixers
     •   Family advection to reduce gradients
     •   Test different interpolation options
•   Improve wet-deposition scheme and lightning
•   Implement and test linear CO scheme
•   Prepare C-IFS for data assimilation

                                                        WP4 Status   Slide 46
                                 WP4 Deliverable Status
        Planning document on
D 4.1   design outline and interface    M4
        standards of CT-IFS
                                                            • Acute work on Eyjafjalla eruption/plume
D 4.2   IFS transport study results     M12
                                                    o         modeling
                                                            • Testing more extensive due to use of more
                                              (Delay 6 M)
                                                              realistic tracers (TM5 chemistry)
        Simplified linear chemistry
D 4.3   scheme for CO and adjoint       M16         o       Code delivered from CERFACS, but not yet
        code integrated

D 4.4   Chemistry module integrated     M16     (       )   TM5 module is integrated and tested

                                                            C-IFS interfaced with inventories and GFAS
D 4.5   Emission module integrated      M20     (       )   data

                                                                                    WP4 Status       Slide 47
Outstanding issues:
• Harmonisation („one-stop access“) of tropospheric GRG
• Underestimation of CO got worse in MACC
• Testing and use of additional/new satellite observations
• Some elements of validation work have not functioned
  very efficiently
   new VAL sub project in MACC-2
• Further development of C-IFS remains challenging (but
  also exciting)

                                            Outstanding Issues   Slide 48
Additional slides
                                            Joint work CNRM-BIRA on strat. chemistry
                                   BIRA wanted to upgrade PSC chemistry representation in BASCOE. In the
                                   process, an error was found in MOCAGE PSC routine (from the REPROBUS
                                   original scheme), impacting specially HNO3 in the polar vortex :
                                   sedimentation and thus removal was previously much underestimated.
HNO3, zonal monthly mean 09/2001


                                                     old                          new
                                                                                              GAW NRT data delivery

GAW site list for NRT validation (CO and O3)

     Station                 NRT interval     lat            lon           alt

 1   Hohenpeissenberg        1 day                   47.8          11.02           985

 2   Jungfraujoch            1 day (12h)            46.55           7.99          3580    Offline validation performed for
 3   Monte Cimone            1 month                44.18          10.70          2165    following runs:
 4   Moussala                1 month                 42.2          25.40          2925

 5   Ryori                   1 month                39.03     141.82             260.00
                                                                                          f93i: 09/2009 – 07/2010
 6   Waliguan                1 week                 36.28     100.90              3842
                                                                                          f1kd: 10/2008 – 08/2009
 7   Santa Cruz (Tenerife)   1 day                   28.5      -16.30               50

 8   Izana (Tenerife)        1 day                   28.3      -16.50             2367    f9nd: 11/2009 – 07/2010
 9   Yonagunijima            1 month                24.47     123.02              30.00

10   Minamitorishima         1 month                24.29     153.98               8.00
                                                                                          fdrl: 05/2010 – 07/2010
11   Assekrem/Tamanraset     1 month                23.17           5.42          2728

12   Cape Point              1 month                -34.35         18.48           230

13   Ushuaia                 1 month                -54.85     -68.32             18.00

14   Neumayer                1 month                -70.65         -8.25            42

     New sites since GEMS               Submission via Email
      Submission via FTP              Currently no data transfer

                                                                                                         WP3 Status   Slide 51
                                                    NRT validation with GAW data

Comparison of F9nd (IFS TM5) and F93i (IFS MOZ) for Antarctica (Neumayer):
         SH Summer                                  SH Winter

                              f9nd does capture
                              the level of O3,
                              however, in the
                              winter time the

                               underestimation of
                               surface O3 for
                               Neumayer in winter
                               and summer!
Jan 2004
           Anthropogenic CO emission
           ratio MACC/GEMS

Jul 2004
•   Emission and deposition preprocessor SUMO
    •   Reggrid original emission datasets to working domain and convert to a reduced set of activity
        sector (optionally apply month/season/day temporal profiles)

          •   Global or regional datasets accepted
          •   1 file per specie and per activity sector (NetCDF format) at domain resolution

    •   Aggregate emission to model species (optionally apply hourly profile) and calculate deposition

          •   Meteorological fields from ECMWF or Météo-France for deposition velocities
          •   Wesely Ganzeveld-modified parameterization
          •   DV and emissions at domain resolution in 1 or 2 separate files (NetCDF format)
          •   !!! Output fields are in lat-lon coordinates !!!
          january                              july



          O3 deposition velocities in cm.s-1
                 (monthly means)

To top