Institute of Environmental Physics and Remote Sensing IUP/IFE-UB Department 1 Physics/Electrical Engineering TP-HTAP WMO Geneva, 25 January 2007 Using GOME and SCIAMACHY NO 2 measurements to constrain emission inventories potential and limitations Andreas Richter Institute of Environmental Physics and Institute of Remote Sensing University of Bremen
Introduction anthropogenic pollution biomass burning ships transport 2
Using SCIAMACHY NO 2 for Emissions I At first glance, GOME, SCIAMACHY and OMI NO 2 data seem perfect to constrain and determine emission estimates: near global coverage reasonable spatial resolution sensitivity down to the surface short lifetime of NO x => transport is not that important However NO 2 is not NO x there is no vertical resolution in the troposphere a priori assumptions have a significant impact on the results what you see is the combined effect of all emissions, chemistry, and transport (+ retrieval), not e.g. anthropogenic emission 3
Using SCIAMACHY NO 2 for Emissions II What would we like to get? absolute amounts of emissions their distribution on individual sources / source types their spatial distribution their seasonality their diurnal variation Possible approaches: select for regions with only one source switch source on and off use a model to distribute on source types look at seasonality of signal 4
Example: Detection of Shipping Emissions Ship emissions: large source of NO x, SO x and aerosols relevant input into marine boundary layer With estimate of NO 2 lifetime, NO x emissions can be estimated => agreement within error bars. But: error bars still large (mainly from lifetime) A. Richter et al., Satellite Measurements of NO2 from International Shipping Emissions, Geophys. Res. Lett., 31, L23110, doi:10.1029/2004gl020822, 2004 well defined NO 2 patterns in Red Sea and Indian Ocean in SCIAMACHY data consistent with pattern of shipping emissions 5
Example: Quantification of Lightning Emissions GOME trop. NO 2 SCD (10 15 molec/cm 2 ) Cloud fraction NLDN flashes (time of last lightning event) NO 2 columns retrieved from GOME satellite data coincident measurements of clouds, lightning and NO 2 in space and time no indication for pollution impact direct evidence without a priori assumptions Beirle et al., Estimating the NOx produced by lightning from GOME and NLDN data: a case study in the Gulf of Mexico Atmos. Chem. Phys., 6, 1075-1089, 2006 6
Example: Detection of NO 2 changes I GOME annual changes in tropospheric NO 2 1996-2002 7 years of GOME satellite data DOAS retrieval + CTM-stratospheric correction seasonal and local AMF based on 1997 MOART-2 run cloud screening NO 2 reductions in Europe and parts of the US strong increase over China consistent with significant NO x emission changes A. Richter et al., Increase in tropospheric nitrogen dioxide over China observed from space, Nature, 437 2005 7
Example: Detection of NO 2 changes II 1996 2000 2005 GOME NO 2 timeseries shows nonsignificant trend in USA after 2000, clear decrease (> 30%) in NO 2 in Ohio-valley area no change in urban areas size and geographical pattern consistent with model simulations Kim, S.-W et al., (2006), Satelliteobserved U.S. power plant NOx emission reductions and their impact on air quality, Geophys. Res. Lett., 33, L22812, doi:10.1029/2006gl027749. 8
Example: Detection of NO 2 changes III NO 2 columns in summer over the US measurement and WRF model run SCIAMACHY updated base emissions Kim, S.-W et al., (2006), Satelliteobserved U.S. power plant NOx emission reductions and their impact on air quality, Geophys. Res. Lett., 33, L22812, doi:10.1029/2006gl027749. 9
Example: Quantification of soil emissions I Observation: Large NO 2 pulse from GOME data over the Sahel at onset of rainy season Approach: anthropogenic emissions removed using GEOS- CHEM model biomass burning emissions removed using fire counts relation between emission and NO 2 column from GEOS-CHEM Jaegle, L. et al., (2004), Satellite mapping of raininduced nitric oxide emissions from soils, J. Geophys. Res., 109, D21310, doi:10.1029/2004jd004787. 10
Example: Quantification of soil emissions II Chouteau, Hill and Liberty Counties in North-Central Montana, USA harvested cropland, low population density, no large stationary NO x sources NO 2 columns retrieved from SCIAMACHY satellite data are large after fertilisation and subsequent precipitation Bertram, T. H., et al., (2005), Satellite measurements of daily variations in soil NOx emissions, Geophys. Res. Lett., 32, L24812, doi:10.1029/2005gl024640 11
Example: Improvement of Emission Inventories I Approach: GEOS-CHEM model GOME NO 2 columns linearized relation between NO x emission and NO 2 column determined for each grid cell from model error weighted combination of a priori (GEIA) and a posteriori emissions improved emission inventory with reduced uncertainties Martin, R. et al.,, Global inventory of nitrogen oxide emissions constrained by space-based observations of NO2 columns, J. Geophys. Res., 108(D17), 4537, doi:10.1029/2003jd003453, 2003. 12
Example: Improvement of Emission Inventories II a priori a posteriori Approach summer measurements of SCIAMACHY set of empirical models between NO x emissions and NO 2 columns based on CHIMERE model output 0.5 x 0.5 resolution error estimate from inversion process test with surface NO 2 data Konovalov et al., Inverse modelling of the spatial distribution of NOx emissions on a continental scale using satellite data Atmos. Chem. Phys., 6, 1747 1770, 2006 13
Example: Speciation of Emissions Idea: Analysis of seasonality of NO 2 columns different emission types have different seasonality this facilitates assignment of main emission source Results: good agreement with expectations Problems based on a priori assumptions on seasonality what about mixed cases? R. J. van der A, submitted to JGR 14
Limitations when using GOME & SCIAMACHY data Basic Limitations transport bias towards cloud free scenes top-down approach relies on good model possible correlation between emission and radiative transfer limited to NO 2 with some potential for HCHO, SO 2, and CHOCHO Specific Limitations lack of diurnal variation large and systematic uncertainties lack of vertical resolution lack of continuity of measurements 15
Potential for Improvements I OMI (7.2004) higher spatial resolution better coverage two measurements per day if combined with GOME-2 or SCIAMACHY GOME-2 (10.2006) better spatial resolution better coverage continuity for 15 years Geostationary (?) better spatial resolution several measurements per day less cloud contamination 16
Potential for Improvements II Models / Retrieval: higher spatial resolution for a priori (model) more validation of the vertical profile of NO 2 used more accurate surface reflectivity better aerosol treatment, possibly by synergistic use of other satellite data use of measurements at different wavelengths 17
Summary What we have global datasets of NO 2 (and SO 2 and HCHO) over more than one decade with high potential to constrain / validate emissions and models a number of studies which have already used these data to get information on emissions using different approaches, with and without models What we need continuity in the measurements (=> GOME-2) extension to several measurements per day (=> GEO + LEO) improved ancillary data (clouds, aerosols, albedo, vertical profiles) much more validation data (=> e.g. GEOMON) 18