CEOS ACC8 Meeting April 2012 Columbia The Atmospheric

CEOS ACC-8 Meeting April 2012, Columbia The Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP): Overview and the Role of Satellite Data in Model Evaluation Jessica L. Neu, JPL / Cal Tech With thanks to: Kevin Bowman, Duane Waliser, Joao Teixeira © 2012 California Institute of Technology. Government sponsorship acknowledged

What is ACCMIP? Ø An IGAC / SPARC Atmospheric Chemistry and Climate project Ø Impetus: CMIP 3 models had substantial differences in climate projections that could be attributed to short-lived species (Shindell et al. , 2007; Levy et al. , 2008; Shindell et al. , 2008) Ø Differences due to: Inclusion of species and processes (e. g. aerosol indirect affect, tropospheric O 3 chemistry) Stratosphere-troposphere exchange of O 3 Emissions projections Northern Hemisphere mean surface temperature trends due to short-lived species (aerosols and ozone) under an A 1 B scenario in the CCSP coupled models (Shindell etal. , 2008)

What is ACCMIP? Ø CMIP 5 will have uniform anthropogenic emissions, but more sophisticated biosphere models = diversity of climate-dependent emissions Ø May not have self-consistent chemistry and aerosol treatments Ø Diagnostics needed for radiative forcing by non-CO 2 gases and aerosols will not be archived Ø ACCMIP interactive chemistry simulations designed to meet these needs: Characterization of forcings imposed in CMIP 5 historical and future simulations Diagnostics to understand cause of differences in model forcing Constraints on uncertainties due to natural emissions, anthropogenic emission projections Observation-based evaluation of our understanding of chemistry-climate interactions “The ACCMIP group concluded that …collaboration between the global modelers and global satellite data teams would be beneficial to both, and that following on existing activities such as the NASA obs 4 MIPs program could help achieve substantial progress in bringing in more observational data for evaluations. ” (IGAC

What is ACCMIP? Simulations (* to be analyzed for AR 5): *1) CMIP 5 –Complementary Timeslices * 2) Emissions Sensitivities 3) Emissions from multiple Integrated Assessment Models 4) Standardized Composition 5) Climate Forcing by Emission Sector 6) 2000 -2010 with Observed SSTs – For Direct Comparison to Satellite Data Output archived at the BADC Satellite Data: TES O 3 and CO (NASA JPL) CEOS-ACC Activity? Modeling Centers: CCC (Canada) CICERO (Norway) ECHAM (Germany) Hadley Centre/Met Office (UK) LLNL (USA) LSCE/IPSL (France) Meteo France (France) MIROC/CCSR/NIES (Japan) NASA GISS (USA) NASA GSFC (USA) NCAR (USA) NOAA GFDL (USA) PNNL (USA) UKCA/NIWA (New Zealand)

Relationship Between ACCMIP and other MIPs, etc Ø Direct relationship to CMIP 5 – results from 1) and 2) will feed into AR 5 Ø SPARC Chemistry-Climate Model Validation (CCMVal) Project Process-oriented diagnostics of chemistryclimate models focused on the stratosphere Quantitative metrics similar to CMIP 5 Extensive use of satellite data Feeds into WMO/UNEP Ozone Assessment Joint IGAC/SPARC meeting in Davos (May 2012) will discuss synergies between CCMVal and ACCMIP, focus on observations for model evaluation Expectation that future activities will desegregate stratospheric and tropospheric composition and climate Model performance on diagnostics assessing transport processes. 13 out of 20 diagnostics used satellite data for evaluation (SPARC CCMVal Report, Chapter 5) Model Performance: Transport Diagnostics

Relationship Between ACCMIP and other MIPs, etc Ø SPARC Data Initiative Intercomparison of satellite climatologies of longand short-lived trace gases Broad international participation UT/LS O 3 evaluation is a pathfinder for comparing limb and nadir observations and will provide a welldocumented set of O 3 data for model evaluation Figure removed per SPARC DI Data Policy SPARC DI Instruments ACE-FTS Aura MLS GOMOS HALOE HIRDLS MIPAS OSIRIS POAM II, III SABER SAGE II, III SCIAMACHY SMILES TES UARS MLS Ozone at 100 h. Pa from the SPARC DI climatologies (J. Neu and the SPARC DI team, in prep) Ø obs 4 MIPs NASA/DOE activity through Program on Climate Modeling Diagnostics and Intercomparison (PCMDI) Recently established formal obs 4 MIPs Science Working Group Providing satellite data for evaluation of physical climate in CMIP 5 models

Obs 4 MIPs: An Infrastructure for Evaluating Climate Models with Satellite Observations How to bring as much observational scrutiny as possible to the IPCC process? How to best utilize the wealth of NASA Earth observations for the IPCC process? ~120 ocean ~60 land ~90 atmos ~50 cryosphere Current NASA Missions ~14 Total Missions Flown ~ 60 Many with multiple instruments Most with multiple products (e. g. 10 -100 s) Many cases with the same products Over 300 Variables in (monthly) CMIP Database Over 1000 satellitederived quantities

Challenges of Satellite Meaurement. Based Model Evaluation Ø Satellite observations have vastly different spatio-temporal sampling patterns, vertical sensitivity , retrieval methods, individual biases, uncertainty estimates, time scales Ø Climate models may not represent the “real” world but some statistical representation. Ø Gridded output / data may or may not be sufficient Climate Model Fields Simulated Fields Retrieved Fields Observation Simulator retrieval Algorithm Simulated Radiances Observed Radiances Satellite Instrument

Obs 4 MIPs: An Infrastructure for Evaluating Climate Models with Satellite Observations About 15 variables identified as being “safely” comparable in the first phase. Gridded monthly averages are used. CMIP Protocol Variables ta - Atm Temp Dataset AIRS (≥ 300 h. Pa) MLS ( < 300 h. Pa) TES (standard pressure levels) AMSR-E Time Period Comments 9/02 – AIRS +MLS needed to cover all 8/04 pressure levels 9/02 – 8/04 2004 - rlut, rlutcs, rsdt, rsutcs – TOA outgoing LW & SW Radiation, Incident SW Radiation clt – Total Cloud Fraction zos - Sea Surface Height Above Geoid CERES 3/00 - MODIS TOPEX/JASON series 2/00 10/92 - AVISO Product pr - Total precipitation uas, vas - Surface (10 m) zonal wind TRMM Quik. SCAT 1997 1999 – 2009 Monthly Ave + 3 hourly Oceans only. No land products. sea ice cover/fraction Land Surface products (TBD) Blended Microwave MODIS 1987 -2010 2/00 - Monthly Ave Perhaps 2 CMIP variables, TBD hus - Specific Humidity tro 3 – Mole Fraction of Ozone tos - Sea Surface Temperature 6/02 - SST science team recommends multiple products Match up of available NASA datasets to PCMDI priority list Orange datasets still in process

Obs 4 MIPs: An Infrastructure for Evaluating Climate Models with Satellite Observations Ø Use the CMIP 5 simulation protocol (Taylor et al. 2009) as guideline for deciding which observations to stage in parallel to model simulations. Target: monthly avg products on 1 ox 1 o grid Ø Convert Satellite Observations to be formatted exactly the same as CMIP Model output CMOR output, Net. CDF files, CF Convention Metadata Ø Includes a 6 -8 page Technical Note describing strengths/weaknesses, uncertainties, dos/don’ts regarding interpretations comparisons with models. (at graduate student level) Ø Hosted side by side on the ESG with CMIP 5 Ø Advertise availability of observations for use in CMIP 5 analysis.

Obs 4 MIPs: An Infrastructure for Evaluating Climate Models with Satellite Observations Earth System Grid Gateway - A side by side archive with CMIP obs 4 MIPS Project

CCM Evaluation Observational Needs How to reduce uncertainties in modeling of chemistry-climate interactions? How to best utilize the wealth of international composition measurements? Recommendations for (Stratospheric) Observations from the CCMVal Report: Ø Long-term vertically resolved data sets of constituent observations in the stratosphere are required to assess model behavior (ozone & other species) Ø The current set of GCOS Essential Climate Variables is not sufficient for process oriented validation of CCMs. Ø More global vertically resolved observations are required, particularly in the UTLS. Ø A systematic comparison of existing observations is required in order to underpin future model evaluation efforts Ø Need a coordinated international “data initiative” to support model intercomparison projects like CCMVal – The SPARC DI is a good start, but more to be done…… ACCMIP will establish additional needs for tropospheric observations. The first step is an obs 4 MIPs-type analysis of comparable quantities

The TES ACCMIP Archive Ø Resides side-by-side with model output at the BADC Ø The archived Level 2 files required adaptation for climate model evaluation Ø The data is processed leveraging the obs 4 MIPs activity to make data CF compliant, uses ACCMIP CMOR tables Ø Ensemble climate model analysis showed that the spatio-temporal sampling is sufficient for zonal, monthly means for ozone, CO, temperature (H 2 O exhibits bias) (Aghedo et al, 2011) Ø Retrieval Observation Operator, defined by the TES averaging Kernels and constraint vectors (Jones et al, JGR, 2003, Worden et al, 2007) is provided in TES product Ø Available from 2005 -2010

The TES ACCMIP Archive Ø Preliminary analysis has revealed large differences in O 3 radiative effects between the models and TES Ø These differences are primarily due to differences in UT O 3 abundances Ø Identifying a consistent tropopause definition is a major challenge Differences in radiative effect between TES and NASA GISS Model E for a nominal August (TES 2006, GISS seasonal cycle over decadal SST). Black line is WMO tropopause, Dotted line is TES chemical tropopause (<120 ppb) and dashed red is GISS chemical tropopause.

Summary Ø Future ACCMIP activities will be greatly strengthened by a systematic Ø Ø collaboration between the satellite and climate model communities Need infrastructure for providing satellite data for model comparison – obs 4 MIPs provides a template CF Compliance, CMOR tables, processing of data into grid, etc Need apples to apples comparison TES effort has shown feasibility Should be expanded to an international effort with additional satellites, trace gases, aerosols SPARC Data Initiative demonstrates importance of multiple data sets Also need mechanisms for developing comparable data sets, observation simulators, and doing comparisons Eventual goal is merging of CMIP/ACCMIP/CCMVal efforts into self-consistent troposphere-stratosphere-chemistry-climate modeling with a side-by-side obs 4 MIPs with observations of both physical climate and composition.