1. What can we learn about the Earth system using space-based observations of tropospheric chemical composition? Paul Palmer, University of Leeds www.env.leeds.ac.uk/~pip

2. + UK/European missions + CMDL network = sparse data coverage

3. NO HO 2 OH NO 2 O3O3 hv HC+OH  HCHO + products NOx, HC, CO Tropospheric O 3 is an important climate forcing agent IPCC, 2001 Level of Scientific Understanding Natural VOC emissions (50% isoprene) ~ CH 4 emissions.

4. Bottom-up Isoprene emissions, July 1996 MEGAN 3.6 Tg C GEIA 7.1 Tg C [10 12 atom C cm -2 s -1 ] Guenther et al, JGR, 1995 EPA BEIS2 2.6 Tg C Pierce et al, JGR, 1998 Guenther et al, ACP, 2006 E = A ∏ i γ i Emissions (x,y,t); fixed base emissions(x,y); sensitivity parameters(t)

5. Global Ozone Monitoring Experiment (GOME) & the Ozone Monitoring Instrument (OMI) GOME (European), OMI (Finnish/USA) are nadir SBUV instruments Ground pixel (nadir): 320 x 40 km2 (GOME), 13 x 24 km2 (OMI) 10.30 desc (GOME), 13.45 asc (OMI) cross-equator time GOME: 3 viewing angles  global coverage within 3 days OMI: 60 across-track pixels  daily global coverage O3, NO2, BrO, OClO, SO2, HCHO, H2O, cloud properties Launched in 2004

6. GOME HCHO columns July 2001 [10 16 molec cm -2 ] 0 1 2 0.5 1.5 2.5 Biogenic emissions Biomass burning * Columns fitted: 337-356nm * Fitting uncertainty < continental signals Data: c/o Chance et al

7. MayJunJulAugSep 1996 1997 1998 1999 2000 2001 GOME HCHO column [10 16 molec cm -2 ] 0 1 2 0.5 1.5 2.5 Palmer et al, JGR, 2006.

8. Tropospheric chemistry: a highly non-linear system, e.g. isoprene chemistry Yield dependent on oxidant concentrations Models typically use a simplified (lumped) oxidation mechanism Assimilating HCHO possible but involves integrating many 100s of uncertain chemical reactions Neglected to mention other (more chemically uncertain) biogenic sources of HCHO! For now, we can extract lots of information from the HCHO using a simple method…

9. Relating HCHO Columns to VOC Emissions VOC HCHO hours OH hours h, OH Local linear relationship between HCHO and E k HCHO E VOC =  (k VOC Y VOC  HCHO )  HCHO ___________ VOC source Distance downwind  HCHO Isoprene  -pinene propane 100 km E VOC :  HCHO from GEOS-CHEM CTM and MEGAN isoprene emission model Palmer et al, JGR, 2003. Net

10. Seasonal Variation of Y2001 Isoprene Emissions Good accord for seasonal variation, regional distribution of emissions (differences in hot spot locations – implications for O 3 prod/loss). Other biogenic VOCs play a small role in GOME interpretation May Jun Aug Sep Jul 0 3.5 7 10 12 atom C cm -2 s -1 GOMEMEGAN GOME Palmer et al, JGR, 2006.

11. GOME Isoprene Emissions: 1996-2001 MayJunJulAugSep 1996 1997 1998 1999 2000 2001 [10 12 molecules cm -2 s -1 ] 0 5 10 Relatively inactive Palmer et al, JGR, 2006.

12. Isoprene flux [10 12 C cm -2 s -1 ] Julian Day, 2001 MEGAN Obs GOME Sparse ground-truthing of GOME HCHO columns and derived isoprene flux estimates Seasonal Variation: Comparison with eddy correlation isoprene flux measurements (B. Lamb) is encouraging Atlanta, GA May Jun July Aug Sep PAMS Isoprene, 10-12LT [ppbC] GOME HCHO [10 16 molec cm -2 ] 1996 1997 1998 1999 2000 2001 Interannual Variation: Correlate with EPA isoprene surface concentration data. Outliers due to local emissions. Atlanta, GA PROPHET Forest Site, MI

13. Surface temperature explains 80% of GOME- observed variation in HCHO NCEP Surface Temperature [K] GOME HCHO Slant Column [10 16 molec cm -2 ] G98 fitted to GOME data G98 Modeled curves Time to revise model parameterizations of isoprene emissions? Palmer et al, JGR, 2006.

14. Tropical ecosystems represent 75% of biogenic NMVOC emissions 1996 1997 1998 1999 2000 2001 What controls the variability of NMVOC emissions in tropical ecosystems? Importance of VOC emissions in C budget? Kesselmeier, et al, 2002 GOME HCHO column, July

15. Challenges: Cloud cover, biomass burning, and lack of fundamental understanding of NMVOC emissions… OMI, 24/9-19/10, 2004 13x24 km 2 TES data @ 6km, 11/04 O3O3 CO MODIS Firecount O 3 -CO-NO 2 -HCHO- firecount correlations import to utilize when looking at the tropics Improved cloud- clearing algorithms and better spatial resolution data help. TES data c/o Bowman, JPL A more integrated approach to understanding controls of NMVOCs, e.g., surface data, lab data,

16. Africa Some aerosol- climate effects South AmericaAfrica Fe fertilization deposition Primary and secondary aerosol sources: biomass burning, biogenic, desert dust Internally or externally mixed? visibility CCN DZ N P Ocean Ecosystem SEVERI AOD

17. “Normal” airmass flow Stagnant airmass flow 0 200 400 600 800 1000 1200 1400 27-Jul29-Jul31-Jul 2-Aug 4-Aug 6-Aug8-Aug 10-Aug 12-Aug 14-Aug 16-Aug 18-Aug20-Aug22-Aug24-Aug26-Aug28-Aug30-Aug 0 5 10 15 20 25 30 35 40 Temperature (C) Isoprene (ppt) Estimated up to 700 extra deaths attributable to air pollution (O 3 and PM10) in UK during this period O 3 > 100 ppb on 6 consecutive days 2pm, 6 th Aug, 2003 Compiled from UK ozone network data Isoprene c/o Ally Lewis “Expect harmful levels of ozone and PM2.5 over the next couple of days; please keep small children and animals inside. Transatlantic pollution represents 20% of today’s UK surface ozone.” 2010

18. Resolution of new satellite data allows study UK AQ from space SCIA NO 2 @ 0.4x0.4 o, Aug 2004 GOME NO 2 @ 1x1 o Aug 1997 NAEI NO X emissions as NO 2, 2002 GOME and SCIA NO 2 c/o R. Martin; OMI (unofficial) NO 2 c/o T. Kurosu Length of day [hours] Day of Year Edinburgh, 56N Denver, 40N Cloud cover [%], ISCCP August 83-04 30 100 70 OMI NO 2 @ 0.1 o x0.1 o, Jul 2004

19. The increasing role of BVOCs: constraints from OMI HCHO? Stewart et al, 2003 Isoprene Monoterpenes BVOC fluxes for a “hot, sunny” day 10 16 [molec cm -2 ] OMI HCHO 2<0.3

20. Unclear what PM characteristics affect health Secondary PM is formed from: –Oxidation of organic compounds –Oxidation of SO 2 –Difficult to estimate offline – need models and data MISR Surface PM2.5 = Model Surface [PM2.5] x MISR AOT Model AOT 2003 roadside (primary) PM10 MISR AOT can help estimate total PM2.5: Space-based aerosol optical properties can help map emissions of particulate matter Liu et al, 2004

21. Case study of Oregon Fire using data from the Multi-angle Imaging SpectroRadiometer (MISR) 1 2 3 4 5 0.0 0.6 1.2 0.0 1.22.4 0 5000 10,000 1 2 3 4 5 Kahn, et al., JGR, submitted Plume Periphery: Higher AOT, Lower ANG, Lower SSA than Background Plume Core: Too Thick & Variable for Standard AOT Retrieval How do we optimally use this information operationally?

22. Current Development in Modelling UK AQ UK currently using MODELS 3 (MM5 + CMAQ) for AQ 1) UM mesoscale CTM UKMO Unified Model 2) UKCA gas-aerosol chemistry scheme AQ-climate links AQ Model Chem-Clim Model Similar equations for data assimilation and inverse modelling J(x) = ½(y o – H(x)) T (E+F) -1 (y o -H(x)) +½(x-x b ) T B -1 (x-x b ) Multi-species analyses – inter-species error covariance? Radiance versus retrieved products? Limit of linearization of non-linear oxidant chemistry? Global vs urban chemistry? Subgrid scale processes?

23. “First global space-based measurements of CO 2 with the precision and spatial resolution needed to quantify carbon sources and sinks” The Orbiting Carbon Observatory (OCO) Spectroscopic observations of CO 2 ( 1.61  m and 2.06  m) and O 2 (0.765  m) to estimate the column integrated CO 2 dry air mole fraction, X CO2 = 0.2095 x (column CO 2 ) / (column O 2 ) Precisions of 1 ppm on regional scales Global coverage in 16 days (nadir 1x1.5 km footprint) JPL-based instrument: PI D. Crisp; Deputy PI: C. Miller (Crisp et al, 2004) Launch in 2008 2-year mission