Presentation on theme: "Dylan Millet Harvard University with"— Presentation transcript:
1Spatial Distribution of Isoprene Emissions from North America: New Constraints from OMI Dylan MilletHarvard UniversitywithD.J. Jacob and K.F. Boersma (Harvard), C.L. Heald (UC Berkeley),A. Guenther (NCAR), T.P. Kurosu and K. Chance (Harvard-Smithsonian)thanks toNASA-ACMAPNOAA C&GC Postdoctoral ProgramOMI Science TeamINTEX Science TeamsAmerican Geophysical Union Fall Meeting 2006 San Francisco, CA
2VOCs Affect Atmospheric Chemistry and Climate IsopreneMost important non-methane VOCGlobal emissions ~ methane(but > 104 times more reactive)~ 6x anthropogenic VOC emissions
3Mapping Isoprene Emissions from Space HCHO vertical columns measured by OMI(T.P. Kurosu & K. Chance)OH, hki, YiOH, hkHCHOVOCsHCHOVOC sourceDistancedownwindΩHCHOIsoprenea-pinenepropane100 kmdetectionlimitLocal ΩHCHO-Ei RelationshipPalmer et al., JGR (2003,2006).
4Testing the Approach: ΩHCHO = SEisoprene+ B What drives variability in column HCHO?Error CharacterizationMeasured HCHO production rate vs. column amountMean biasPrecisionPHCHO (1012 molec cm-2 s-1)isoprene25–31% error (1σ) in HCHO satellite measurementsINTEX-A/BΩHCHO (1016 molec cm-2)clouds primary source of errorHCHO Column ComparisonΩHCHO variability over N. America driven by isoprene25x101525x1015Other VOCs give rise to a relatively stable background ΩHCHO Not to variability detectable from spaceURI (DC8)OMI25x101525x1015Millet et al., JGR (2006).NCAR (DC8)Aircraft
5Defining Spatial Distribution of Eisoprene Using OMI HCHO HCHO columns from OMI satellite instrument (summer 2006)Isoprene emissions from the MEGAN biogenic emission inventory (summer 2006)?test bottom-up inventories against top-down constraints from OMImismatch in hotspot locationsimplications for OH, O3, SOA production
6Model of Emissions of Gases and Aerosols from Nature Vegetation-specific baseline emission factorsEnvironmental drivers (T, h, LAI, leaf age, …)Guenther et al., ACP (2006)Drive MEGAN with 2 land cover databasesG2006Guenther  (MODIS3)CLMCommunity Land Model(AVHRR)MEGAN Isoprene Emissionsw/ G2006 vegetationMEGAN Isoprene Emissionsw/ CLM vegetation14.6 TgC12.2 TgC[1013 atomsC cm-2 s-1]
7OMI vs. GEOS-Chem with MEGAN Emissions OMI HCHO: June-August 2006Similarity in broad spatial distribution (r2 = 0.86)GEOS-Chem HCHO: June-August 2006But important discrepancies regional & landscape scale emission hotspots[1015 molecules cm-2]
10Spatial Patterns in Isoprene Emissions MEGAN higher than OMI over dominant emission regionsOMI – MEGAN Isoprene EmissionsJune-August, 2006MEGAN w/ G2006 Land CoverOzark PlateauUpper SouthWisconsin-MinnesotaLarge sensitivity to surface database usedMEGAN w/ CLM Land CoverCLM-driven MEGAN lower than OMI over deep South & Atlantic coast
11Why are Emissions Overestimated in Dominant Source Regions? Bottom-up emissions are too high in Ozarks, upper South, MN/WIOMI – MEGAN Isoprene EmissionsJune-August, 2006MEGAN w/ G2006 Land CoverAverage Bias:Ozarks: 30-70%Upper South: 20-45%Large emissions driven by oak tree cover, high temperaturesBroadleaf tree isoprene emissions overestimated?MEGAN w/ CLM Land CoverG2006 Broadleaf Trees
12Why Are CLM Emissions Too Low in Deep South? Average Bias: 50 to >100%OMI – MEGAN Isoprene EmissionsJune-August, 2006MEGAN w/ G2006 Land CoverUnderestimate of broadleaf tree or shrub coverage?CLM – G2006Broadleaf Tree CoverCLM – GuentherShrub CoverMEGAN w/ CLM Land CoverModeled emissions from evergreen trees or crops too low?CLMFineleaf EvergreensCLMCrops
13ConclusionsOMI HCHO columns show similar magnitude & broad spatial patterns as state-of-the-art bottom-up emission inventories (R2 = 0.86) But reveal major discrepancies at the regional scaleBottom-up isoprene emission estimates are 20 to >70% too high in the dominant emission regions Ozarks, upper South, MN/WI Overestimate of broadleaf tree emissions?CLM-driven isoprene emission estimates are 50 to >100% too low over the deep South and along the Atlantic coast Underestimate of broadleaf tree and shrub cover? Underestimate of fineleaf evergreen or crop emissions?