POLARCAT Model Intercomparison Project (POLMIP)

Transcription

1 Institute for Climate and Atmospheric Science UNIVERSITY OF LEEDS POLARCAT Model Intercomparison Project (POLMIP) S.R. Arnold 1,2, L.K. Emmons 2, S.A. Monks 1, K.S. Law 3, S. Tilmes 2, S. Turquety 4 J. Thomas 3, I. Bouarar 3, J. Flemming 5, V. Huijnen 6, J. Mao 7, B.N. Duncan 8, S. Steenrod 8, Y. Yoshida 8, H. Schlager 9, H. Sodemann Institute for Climate and Atmospheric Science, University of Leeds 2. Atmospheric Chemistry Division, NCAR, USA. 3. LATMOS / IPSL, Paris. 4. Laboratoire de Meteorologie Dynamique, Paris. 5. ECMWF, Reading, UK. 6. KNMI, Netherlands. 7. GFDL / Princeton, USA. 8. NASA Goddard, USA. 9. DLR, Oberpfaffenhoffen, Germany. 10. ETH, Zurich, Switzerland.

2 IPY 2008 Campaigns Monks et al., GRL, (2012) Spring: NASA ARCTAS (Alaska) NOAA ARCPAC (Alaska) POLARCAT-France (Sweden) {May-June: START08 (NCAR-GV, N.Amer.)} Summer: NASA ARCTAS (Canada) DLR GRACE (Greenland) POLARCAT-France (Greenland) Fire source contributions to Arctic CO IASI satellite CO columns Matthieu Pommier, LATMOS, Paris

3 Model Resolution Meteorology Chemistry TOMCAT 2.8 x2.8, 31 levels ECMWF ERA oper. trop: 82 species MOZART x2.5, 56 levels GEOS 5 trop: 103 species, bulk aerosols; photolysis options: FTUV: online; LUT: lookup table CAM4 chem 1.9 x2.5, 56 levels GEOS 5 MOZART 4, bulk aerosols CAM5 chem 1.9 x2.5, 56 levels GEOS 5 MOZART 4, modal aerosols LMDZ INCA 1.9 x3.75, 19 levels ECMWF trop: 89 species C IFS T159 (~1 ), 60 levels ECMWF trop: CB05, strat: linear. O3 (Cariolle) TM5 2 x3, 60 levels ECMWF trop: CB05 NASA GMI 2 x2.5, 72 levels GEOS 5 strat&trop (154 species), GOCART aer. GEOS Chem 2 x2.5, 47 levels GEOS 5 trop: ~100 species WRF Chem 100, 50, 25 km NCEP GFS MOZART GOCART Emissions Same* for all models: Anthropogenic: Streets ARCTAS v1.2 Fires: FINN v1 Biogenic, Ocean, etc: MACCity *GEOS chem used slightly different anthro emissions and includes increased HO 2 aerosol uptake [Mao et al., ACPD, 2012] Focus here on gas phase chemistry an tropospheric ozone Artificial tracers 25 day lifetime, based on CO anthro and fire emissions

4 Model performance in the Arctic HTAP Models [Shindell et al., 2008] POLMIP Models

5 Evaluation with aircraft profiles (Spring, DC8 Canada) ethane propane O 3 CO PAN NO NO 2 HNO 3 H 2 O 2 CH 2 O CH 3 CHO acetone

6 Evaluation with aircraft profiles (Summer, DC8 Canada) ethane propane O 3 CO PAN NO NO 2 HNO 3 H 2 O 2 CH 2 O CH 3 CHO acetone

7 Artificial 25 day tracers with CO emissions from 3 regions Allow comparison of purely dynamics between the models, without chemistry Averages over poleward of 66N ANTHRO Asia N. America Europe FIRE Asia N. America Europe CO Anthropogenic Tracer Emissions All models show same general patterns: Anthro emissions dominate in winter Asia fires significant in spring and summer Europe anthro is major source in DJF lower trop In summer Asia is largest anthro source in UT Largest differences between models in fire tracers

8 Arctic Ozone Budget Surface 300 hpa totals P(O3) terms NO+HO2 NO+CH3O2 L(O3) terms O1D+H2O O3+OH O3+HO2 Production Loss MOZART-4 model, Louisa Emmons Arctic ozone budget is weakly net ozone loss, but much closer to net balance than other remote regions (e.g. oceans). Summertime eastern Arctic weakly ozone producing. How robust is this understanding?

9 Model ozone vs obs for fire dominated period Model Obs 25-d lifetime model source tracers allow us to separately evaluate models in air most influenced by fires.

10 Model O 3 / CO relationships (July) 25-d lifetime model source tracers allow us to separately evaluate models in air most influenced by fires. Summer Lat > 50N ppbv/ppbv - All model points north of 50N -Coloured points show fresh and aged air where fire tracer > 70% total tracer. - Don t have tracers for GEOS-Chem & CIFS. Models demonstrate +ve O 3 / CO slopes in firedominated air, increasing with age since emission. Lines show slopes from other models with tracers

11 PAN enhancements in POLMIP models (July) North of 50N Fire Lines show slopes from other models with tracers Model PAN / CO in fire air masses Same chemistry CAM-Chem 2.02 pptv ppbv -1 MOZART pptv ppbv -1 GMI 3.02 pptv ppbv -1 TM pptv ppbv -1 TOMCAT 4.62 pptv ppbv -1 GEOS5 ECMWF

12 POLMIP model O 3 / CO summary O 3 / CO ratios in aged (> 2 days since emission) boreal fire plumes from observations and models POLMIP Model-derived O 3 / CO [%] Observation-derived

13 Asian anthropogenic / Siberian fire plume (5-9 Jul 2008) FLEXPART IASI CO Column CO from POLMIP models Sodemann et al., (2011) molec cm -2

14 Lagrangian model plume simulations (initialised 6 th July) CiTTyCAT Lagrangian chemistry model (Pugh et al., GMD, 2012) initialised with plume maximum composition from each POLMIP model and integrated forwards for 4 days. Interplay between differences in NOy partitioning and vertical transport key driver of diversity in P(O 3 ).

15 Lagrangian model ovoc & NOy sensitivities Standard [ovoc] = 0. Large PAN fraction means that P(O 3 ) shows strong sensitivity to ovocs.

16 Lagrangian model transport sensitivities Single average trajectory applied to all model plumes. Standard [ovoc] = 0.

17 Summary Transport of mid-latitude sources into Arctic is broadly consistent between models. POLMIP models show +ve O 3 / CO slopes in fire emission-dominated air, with values consistent with observational estimates from boreal fire plumes. GEOS-Chem demonstrates lower O 3 / CO enhancement than other POLMIP models, likely due to new heterogeneous HO 2 loss. Models vary widely in their simulation of ozone precursors in the Arctic esp. for NOy and ovoc species. This has significant implications for insitu Arctic ozone production and loss.

18 Arctic analysis in HTAP Prescribed lifetime tracers useful for partitioning source influence. Emissions how much evaluation before implementation? 2008 is critical year for model evaluation in Arctic. Source perturbations for fires in addition to anthropogenic sources? Separate source perturbations for aerosol, NOx, VOCs, methane,... Arctic as a source? e.g vs present. Emissions only for 2030? Arctic surface changes? Focus for HTAP analysis: - Inter-model sensitivities of Arctic NOy / ozone / aerosol to source perturbations. - Differences in export efficiency between models & 2030/present. - Changes in sensitivities to source regions between 2030 and present. - Sensitivity of Arctic climate response to SLCF source regions local vs remote forcing (online simulations with CESM, others?).

19 Acknowledgements NCAR ACD and ASP Faculty Fellowship program Natural Environment Research Council (UK) ARCTAS & POLARCAT science teams

20

21

22

23

24 Arctic tropospheric ozone source contributions Boreal fires N. America Europe Asia Spring Spring Summer Monks (2011) [TOMCAT] Summer Philippe Rekacewicz, UNEP Wespes et al., (2012) [MOZART-4] - Frontal export stretches and stirs together plumes of differing origin, which slowly mix together. - Tough test for global models...