trends of formaldehyde columns over China observed by satellites: increasing anthropogenic emissions of volatile organic

Transcription

1 trends of formaldehyde columns over China observed by satellites: increasing anthropogenic emissions of volatile organic compounds and decreasing agricultural fire emissions Lu Shen 1, Daniel J. Jacob 1, Lei Zhu 1,, Qiang Zhang, Bo Zheng, Melissa P. Sulprizio 1, Ke Li 1, Isabelle De Smedt, Gonzalo González Abad, Hansen Cao, Tzung-May Fu, Hong Liao 1 John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 01, USA Atomic and Molecular Physics Division, Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts, USA Ministry of Education Key Laboratory for Earth System Modeling, Department of Earth System Science, Tsinghua University, Beijing 0, China Laboratoire des Sciences du Climat et de l'environnement, CEA-CNRS-UVSQ, UMR1, Gifsur-Yvette, France Belgian Institute for Space Aeronomy (BIRA-IASB), Brussels, Belgium Department of Mechanical Engineering, University of Colorado, Boulder, USA Department of Atmospheric and Oceanic Sciences, School of Physics, Peking University, Beijing, China School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing 0, China Key Points satellite observations of formaldehyde show increases in Chinese anthropogenic emissions of volatile organic compounds. The increases are largest in the North China Plain and Yangtze River Delta, and are consistent with the MEIC emission inventory. Decreases are observed in rural eastern China since the early 0s that appear due to new restrictions on crop burning. Abstract We use observations of formaldehyde (HCHO) columns over China from the OMI, GOME-, and SCIAMACHY satellite instruments to infer long-term trends in emissions of volatile organic compounds (VOCs) that affect air quality. The trends are corrected for temperature to remove biogenic VOC influences. The observations show large increases in the North China Plain (NCP, averaging +1. % a -1 ) and the Yangtze River Delta region (YRD, +1. % a -1 ) over the period, consistent with the trend of anthropogenic VOC emissions in the Multi-resolution Emission Inventory for China (MEIC). Unlike other pollutants, VOC emissions have not been decreasing in recent years. An exception is the Huai River Basin (HRB) in rural eastern China where the satellite data show rapidly decreasing VOC emissions since the early 0s that appear to reflect bans on agricultural fires. These fire emissions were much larger than estimated from inventories. 1

2 Introduction Volatile organic compounds (VOCs) affect air quality by producing ozone and organic particulate matter (PM). They are emitted by anthropogenic, pyrogenic, and biogenic sources, and they are oxidized in the atmosphere by the hydroxyl radical (OH) with a range of lifetimes. Highly reactive VOCs with lifetimes of less than a day are particularly relevant for air quality (Wert et al., 00). Oxidation of these highly reactive VOCs produces formaldehyde (HCHO) near the emission sources. HCHO columns can be observed by satellites and from there the rate of VOC emissions can be inferred (Palmer et al., 00; Stavrakou et al., 00; Marais et al., 01; Barkley et al., 01; Zhu et al., 01, 01). Major efforts have been underway in China over the past decade to reduce pollution and this has led to large decreases in emissions, but inventory estimates suggest that VOC emissions have not decreased (Zheng et al., 01). Here we examine trends of HCHO in China as observed from space by the OMI, SCIAMACHY, and GOME- instruments, and we infer from there the long-term changes of VOC emissions over that period. Quantifying anthropogenic VOC emissions in process-based inventories is difficult because the processes involved include leaks, volatilization, and incomplete combustion, with highly uncertain emission factors. Crop residue burning may also be a large seasonal source of VOCs in parts of China (Huang et al., 01; Cao et al., 01). Biogenic emissions by vegetation, in particular of isoprene, are an additional source of VOCs (Guenther et al., 00; Fu et al., 01). A number of studies have used HCHO observations from space to infer VOC emissions and test inventories (Palmer et a., 00; Stavrakou et al., 00; Zhu et al., 01), including specifically for China (Fu et al., 00; Chan Miller et al., 01; Cao et al., 01). De Smedt et al. (0) found a % a -1 HCHO increase over China between 1 and 00 that they attributed to changes in VOC emissions. Our work extends the trend analysis to more recent years including the spatial distribution across eastern China with focus on separating anthropogenic, pyrogenic, and biogenic contributions. Data and Methods The core of our analysis is based on the OMI HCHO Level product (Collection ) from the Smithsonian Astrophysical Observatory (OMI-SAO), as described by González Abad et al. (01). OMI is a UV/Vis nadir solar backscatter spectrometer launched in 00 and provides daily global observations at a 1x km nadir pixel resolution with a local equator crossing time of 1: (Levelt et al., 00). HCHO slant columns are fitted using in the backscattered solar radiation within the.-. nm band, and converted to vertical columns using air mass factors based on vertical profiles from the GEOS-Chem model (Gonzalez Abad et al., 01). Similar to Zhu et al. (01), we only use observations that (1) pass all quality checks, () have a cloud fraction less than 0. and solar zenith angle less than 0, and () are not affected by the row anomalies (Kroon et al., 0). The daily HCHO columns are then averaged on a 0. x0. grid to yield a 1-year monthly product for There is a global background drift in the data that we correct following Marais et al. (01) and Zhu et al (01a, b) by subtracting a smoothed trend of HCHO column over the remote North Pacific (10 - W, - N). We focus on May-September when prompt oxidation of VOCs provides the strongest HCHO signal (Zhu et al., 01a, b). Lower oxidant concentrations in winter leads to smearing and dilution of the signal below the detection limit of 1x 1 molecules cm - (González Abad et al., 01).

3 Validation with aircraft and ground-based measurements shows that the OMI-SAO HCHO columns have a uniform 0% low bias, and that other satellite HCHO products have similar biases (Barkley et al., 01; De Smedt et al, 01; Zhu et al., 01), possibly due to errors in spectral fitting, uncertainties in a priori vertical profiles and/or assumed surface reflectivity. Validation by Zhu et al. (01) over the Southeast US suggests that this bias is uniform and can be remedied with a correction factor. Figure 1 compares the OMI-SAO data with ground-based MAX-DOAS column measurements in four Chinese cities (Li et al., 01; Lee et al., 01; De Smedt et al, 01; Wang et al., 01), showing a mean % low bias consistent with past work. On that basis we apply a uniform correction factor of 1. to the OMI-SAO data. We compare the OMI-SAO HCHO trends with data from other satellite instruments and retrievals available from the Belgian Institute for Space Aeronomy (BIRA) at These include SCIAMACHY for 00-0 (Wittrock et al., 00; De Smedt et al., 00), GOME- on MetOp-A for and MetOp-B for 01-present (De Smedt et al., 01; Hewson et al., 01), and the OMI retrieval developed by BIRA for (OMI-BIRA, De Smedt et al., 01). We multiply the GOME- HCHO columns in MetOp-A by 1.1 to match the mean 01 HCHO columns observed by MetOp-B over eastern China (1 E-1 E, 1 N- N), and then we combine these two into a single GOME- dataset. We adjust the SCIAMACHY, GOME-, and OMI-BIRA products with constant scaling factors to have them match the corrected OMI-SAO data over the same eastern China domain during 00-0 (an overlapping period for all products). We use the Multi-resolution Emission Inventory for China (MEIC, Zheng et al., 01) as bottom-up estimate of anthropogenic VOC emissions. MEIC incorporates detailed provincial activity rates and emission factors for a variety of sources and for individual years.. For open fire emissions, we use the Fire INventory from NCAR (FINN version 1., Wiedinmyer et al, 0), which is based on fire counts from the MODIS satellite instrument. For biogenic emissions we use the MEGAN v.1 inventory of Guenther et al. (01) as implemented in GEOS-Chem by Hu et al. (01). All data are available for We perform a 1-year simulation (00-01) with the GEOS-Chem global model of atmospheric chemistry (v-0, to relate bottom-up trends of anthropogenic VOC emissions to trends of HCHO columns. The model has a horizontal resolution of. with pressure levels extending from surface to 0.01 hpa, driven by the MERRA (Gelaro et al., 01) assimilated meteorological data from the NASA Global Modeling and Assimilation System. Anthropogenic and open fire emissions for individual years are from MEIC and FINN respectively. Biogenic VOC emissions for all years are computed with MEGAN v.1 using 00 meteorological data in order to avoid interannual variability (IAV) of biogenic HCHO driven by temperature; we correspondingly correct the HCHO column data for temperature in trend analyses (Zhu et al., 01a). This allows us to focus on interpretation of anthropogenic and open fire trends.. Results and Discussion Figure shows the mean May-September OMI HCHO columns over China for The general patterns have been previously discussed by Jin and Holloway (01). Values are high in

4 megacity clusters, such as in the North China Plain (NCP), Pearl River Delta (PRD), Yangtze River Delta (YRD) and the Sichuan Basin. Also shown in Figure are the HCHO production rates from anthropogenic, biogenic, and open fire VOCs, as estimated by applying 1-day HCHO production yields under high-no x conditions (Chan Miller et al., 01) to emissions of individual VOCs from the bottom-up MEIC, MEGAN, and FINN inventories respectively. The 1-day HCHO yields of individual VOCs and their contributions to the mean HCHO columns over China in May-September are listed in Table S1. The principal anthropogenic VOC contributors to HCHO are alkanes (%), alkenes (%), and aromatics (%). The principal open fire VOC contributors are alkanes (0%), carbonyls (%), and alkenes (1%). The principal biogenic VOC contributors are isoprene (%) and monoterpenes (1%), and propene (%). We see from Figure that the observed OMI HCHO distribution follows that expected from the MEIC anthropogenic inventory, with high values found in megacity clusters. It does not follow the biogenic source from MEGAN, which contributes more HCHO over eastern China during May-September (. Tg vs.. Tg), but it is more broadly distributed. This contrasts with the US where OMI HCHO distribution closely matches the MEGAN biogenic source pattern because the anthropogenic source is considerably lower (Millet et al., 00). The open fire source from FINN follows the distribution of cropland (Xie et al., 01) but is much smaller (0. Tg) than the anthropogenic and biogenic sources. The biogenic source has a large temperature dependence that is apparent in the IAV of HCHO columns seen from satellite (Palmer et al., 00; Zhu et al., 01) and can obfuscate long-term trend analysis (Zhu et al., 01a). Here we remove this dependence following Zhu et al. (01a) by regressing the May-September monthly mean HCHO columns onto mid-day surface air temperatures for in each grid cell, and subtracting the fitted temperature dependency from the original data. Figure a shows the May-September trends of OMI HCHO columns after correcting for the temperature dependence. We find that HCHO columns increase at a mean rate of 0. 1 molecules cm - a -1 (+1.% a -1 ) in the North China Plain (NCP) and molecules cm - a -1 (+1.% a -1 ) in the Yangtze River Delta (YRD). In between these two regions, we find decreasing trends in the Huai River Basin (HRB) where emissions from agricultural fires are concentrated (Figure ). Also shown in Figure are the trends simulated by GEOS-Chem with fixed biogenic emissions at the 00 level on the o x. o grid. GEOS-Chem can capture the general increases observed by OMI over the NCP and YRD when smoothed over the o x. o model grid but fails at reproducing the decrease over HRB. The HCHO trends in GEOS-Chem are driven by MEIC anthropogenic emissions and FINN open fire emissions. Figure shows the linear trends of these emissions. MEIC shows the largest anthropogenic increases in the NCP and YRD, corresponding to the regions of maximum OMI HCHO increases in Figure a. FINN shows a decrease in open fire emissions in the HRB, matching closely the pattern of OMI HCHO decrease in that region, and reflecting recent bans on agricultural fires because of concerns over air quality (Zhang et al., 01; Zhuang et al., 01). The magnitude of the decreasing trend in FINN open fire emissions is however much smaller than the increase in MEIC anthropogenic emissions (note the different scales in Figure ), explaining why GEOS-Chem does not capture the localized HCHO decrease in the HRB

5 observed by OMI. When averaged across eastern China (domain of Figure ), the GEOS-Chem trend for of May-September HCHO columns is molecules cm - a -1, compared to 0. 1 molecules cm - a -1 in the OMI observations. The difference is within observational and model errors (Zhang et al., 00), thus we can conclude that the increasing HCHO trends seen by OMI and their patterns are consistent with the increase of anthropogenic VOC emissions reported by MEIC. The inability of GEOS-Chem to reproduce the observed OMI HCHO decrease over the HRB suggests that FINN open fire emissions may be too low, consistent with Cao et al. (01). Burning in that region is from small agricultural plots and this may be poorly detected by MODIS either as fire counts or burned area (van der Werf et al., 0; Huang et al., 01; Zhang et al., 01). The frequently used Global Fire Emissions Database version with small fires (GFEDs, van der Werf et al., 01), also relying on MODIS, is times lower than FINN over the HRB. An inverse analysis of OMI HCHO data by Stavrakou et al. (01) for eastern China finds open fire emissions - times higher than FINN. We have so far analyzed linear trends over the period, but the MEIC and FINN inventory trends show significant non-linearity (Figure ). Anthropogenic emissions grew at a weaker rate after 0 and leveled off in the NCP. Open fire emissions begin to decrease only after 01 and drop to nearly zero by 01. Figure addresses this non-linearity in the trend by showing the annual changes relative to 00 of HCHO columns in our three target regions, and including other satellite products in addition to OMI-SAO. For all products, we removed the trend of HCHO columns in the remote Pacific sector and also the temperature dependence, as described above. We find good consistency between satellite products in the long-term trends and IAV, even though the associated signals are relatively small (less than %), reflecting the robustness of temporal trends in the satellite data when averaged spatially on a regional scale (Zhu et al., 01). Even in NCP and YRD, where open fire emissions are low, the satellite data show significant and consistent IAV superimposed on the long-term trend. We find that this IAV in the satellite data is driven by the slant columns, not the air mass factor (AMF). The model represents well the long-term trends in NCP and YRD, including the leveling off after 0 apparent from the MEIC and FINN inventories (Figure ). It does not capture the observed IAV in these two regions, which suggests more year-to-year variation in VOC emissions than indicated by MEIC, though the amplitude is small (less than %). The relatively low values observed in 00 could reflect the economic recession. There could also be some IAV from agricultural fires not accounted in FINN. The observed IAV in the HRB is much larger than in NCP or YRD, as might be expected from fire emissions, and the decrease starts after 0 consistent with the ban on agricultural fires. The inability of the model to reproduce this decrease is consistent with the FINN fire emissions being too low. In summary, we have used temperature-corrected (May-September) trends of formaldehyde (HCHO) columns observed by satellites to infer trends of anthropogenic and open fire VOC emissions in China. We find increasing trends of HCHO columns across much of eastern China over that period, especially in the North China Plain (averaging +1.% a -1 for 00-01) and Yangtze River Delta (+1.% a -1 ). These trends are consistent with the MEIC

6 anthropogenic VOC emission inventory. The MEIC inventory finds a leveling of VOC emission trends after 0 and this is indeed seen in the satellite data. Unlike other pollutants like SO or NO x, VOC emissions in China have not been decreasing in recent years. An exception is the Huai River Basin, where the satellite instruments show a decrease over the period though with large interannual variability. This is a region where considerable crop residue burning used to take place until the early 0s but has since been banned because of air quality concerns. The decreasing trend seen in the HCHO satellite data for that region implies much higher open fire emissions in the pre-01 period than estimated from satellite fire counts and burned areas. Acknowledgments This work was funded by the Harvard Global Institute (HGI), the NASA Earth Science Division, and the Joint Laboratory for Air Quality and Climate (JLAQC) between Harvard and the Nanjing University for Information Science and Technology (NUIST). References Barkley, M. P., et al. (01), Top-down isoprene emissions over tropical South America inferred from SCIAMACHY and OMI formaldehyde columns, J. Geophys. Res. Atmos.,,, doi:.0/jgrd.0. Cao, H., et al. (01), Adjoint inversion of Chinese non-methane volatile organic compound emissions using space-based observations of formaldehyde and glyoxal, Atmos. Chem. Phys., 1, 1-, Chan Miller, C., et al. (01), Glyoxal yield from isoprene oxidation and relation to formaldehyde: Chemical mechanism, constraints from SENEX aircraft observations, and interpretation of OMI satellite data, Atmos. Chem. Phys. Discuss., doi:.1/acp-01-. Chan Miller, C, et al. (01), Glyoxal yield from isoprene oxidation and relation to formaldehyde: chemical mechanism, constraints from SENEX aircraft observations, and interpretation of OMI satellite data, Atmos. Chem. Phys., 1,, De Smedt, I., J.-F. Müller, T. Stavrakou, R. van der A, H. Eskes, and M. Van Roozendael (00), Twelve years of global observations of for- maldehyde in the troposphere using GOME and SCIAMACHY sensors, Atmos. Chem. Phys.,,, doi:.1/acp De Smedt, I., T. Stavrakou, J.-F. Müller, R. J. van der Abo, and M. Van Roozendael (0), Trend detection in satellite observations of for- maldehyde tropospheric columns, Geophys. Res. Lett.,, L10, doi:./0gl0. De Smedt, I., et al. (01), Diurnal, seasonal and long-term variations of global formaldehyde columns inferred from combined OMI and GOME- observations, Atmos. Chem. Phys., 1, 1,1 1,, doi:.1/acp Fu, T.-M., D. J. Jacob, P. I. Palmer, K. Chance, Y. X. Wang, B. Barletta, D. R. Blake, J. C. Stanton, and M. J. Pilling (00), Space-based formal- dehyde measurements as constraints on volatile organic compound emissions in east and south Asia and implications for ozone, J. Geophys. Res.,, D01, doi:./00jd00.

7 Fu, Y., and H. Liao (01), Simulation of the interannual variations of biogenic emissions of volatile organic compounds in China: Impacts on tropospheric ozone and secondary organic aerosol, Atmos. Environ., (0), 10 1, doi:.1/j.atmosenv Gelaro, R. et al. (01), The Modern-Era Retrospective Analysis for Research and Applications, Version (MERRA-), J. Climate, 0, 1, González Abad, G., X. Liu, K. Chance, H. Wang, T. P. Kurosu, and R. Suleiman (01), Updated Smithsonian Astrophysical Observatory Ozone Monitoring Instrument (SAO OMI) formaldehyde retrieval, Atmos. Meas. Tech.,, 1, doi:.1/amt Guenther, A., T. Karl, P. Harley, C. Wiedinmyer, P. I. Palmer, and C. Geron (00), Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature), Atmos. Chem. Phys.,, 11, doi:.1/acp Guenther, A. B., X. Jiang, C. L. Heald, T. Sakulyanontvittaya, T. Duhl, L. K. Emmons, and X. Wang (01), The Model of Emissions of Gases and Aerosols from Nature version.1 (MEGAN.1): An extended and updated framework for modeling biogenic emissions, Geosci. Model Dev.,, 1, doi:.1/gmd Huang, X., M. Li, J. Li, and Y. Song (01), A high resolution emission inventory of crop burning in field in China based on MODIS thermal anomalies/fire products, Atmos. Environ., 0, 1, doi:.1/j.atmosenv Hewson, W., M. P. Barkley, G. Gonzalez Abad, H. Bösch, T. Kurosu, and R. Spurr (01), Development and characterisation of a state- of-the-art GOME- formaldehyde air-mass factor algorithm, Atmos. Meas. Tech.,, 0 0, doi:.1/amt Hu, L., et al. (01), Global budget of tropospheric ozone: evaluating recent model advances with satellite (OMI), aircraft (IAGOS), and ozonesonde observations, Atmos. Environ., 1, -. Jin, X., and T. Holloway (01), Spatial and temporal variability of ozone sensitivity over China observed from the Ozone Monitoring Instrument, J. Geophys. Res. Atmos., 10,, doi:.0/01jd00. Kroon, M., J. F. de Haan, J. P. Veefkind, L. Froidevaux, R. Wang, R. Kivi, and J. J. Hakkarainen (0), Validation of operational ozone profiles from the Ozone Monitoring Instrument, J. Geophys. Res.,, D10, doi:./0jd010. Levelt, P. F., G. H. J. van den Oord, M. R. Dobber, A. Malkki, H. Visser, J. de Vries, P. Stammes, J. O. V. Lundell, and H. Saari (00), The Ozone Monitoring Instrument, IEEE Trans. Geophys. Remote Sens.,,. Li, X., T. Brauers, A. Hofzumahaus, K. Lu, Y. P. Li, M. Shao, T. Wagner, and A. Wahner (01), MAX-DOAS measurements of NO, HCHO and CHOCHO at a rural site in Southern China, Atmos. Chem. Phys., 1, 1, Lee, H., J. Ryu, H. Irie, S.-H. Jang, J. Park, W. Choi, and H. Hong (01), Investigations of the Diurnal Variation of Vertical HCHO Profiles Based on MAX-DOAS Measurements in Beijing: Com- parisons with OMI Vertical Column Data, Atmosphere,, 11-1, Millet, D. B., D. J. Jacob, K. F. Boersma, T. M. Fu, T. P. Kurosu, K. Chance, C. L. Heald, and A. Guenther (00), Spatialdistribution of isoprene emissions from North America derived from formaldehyde column measurements by the OMI satellite sensor,j. Geophys. Res.,, D00, doi:./00jd000

8 Marais, E. A., et al. (01), Isoprene emissions in Africa inferred from OMI observations of formaldehyde columns, Atmos. Chem. Phys., 1, 1, doi:.1/acp Palmer, P. I., D. J. Jacob, A. M. Fiore, R. V. Martin, K. Chance, and T. P. Kurosu (00), Mapping isoprene emissions over North America using formaldehyde column observations from space, J. Geophys. Res., (D), 10, doi:./00jd001. Palmer, P. I., et al. (00), Quantifying the seasonal and interannual variability of North American isoprene emissions using satellite observations of the formaldehyde column, J. Geophys. Res., 1, D, doi:./00jd00. Stavrakou, T., J.-F. Müller, I. De Smedt, M. Van Roozendael, G. R. van der Werf, L. Giglio, and A. Guenther (00), Global emissions of non- methane hydrocarbons deduced from SCIAMACHY formaldehyde columns through 00 00, Atmos. Chem. Phys.,,, doi:.1/acp Stavrakou, T., et al. (01), Substantial Underestimation of Post-Harvest Burning Emissions in the North China Plain Revealed by Multi-Species Space Observations, Sci. Rep.,, 0, van der Werf, G. R., et al. (0), Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1 00), Atmos. Chem. Phys.,, 0-, van der Werf, et al. (01), Global fire emissions estimates during 1 01, Earth Syst. Sci. Data,, -0, Wang, Y., et al. (01), Validation of OMI, GOME-A and GOME-B tropospheric NO, SO and HCHO products using MAX-DOAS observations from 0 to 01 in Wuxi, China: investigation of the effects of priori profiles and aerosols on the satellite products, Atmos. Chem. Phys., 1, 00 0, Wert, B. P., et al. (00), Signatures of terminal alkene oxidation in airborne formaldehyde measurements during TexAQS 000, J. Geophys. Res., (D),, doi:./00jd000. Wittrock, F., A. Richter, H. Oetjen,J. P. Burrows, M. Kanakidou, S. Myriokefalitakis, R. Volkamer,S. Beirle, U. Platt, and T. Wagner (00), Simultaneous global observations of glyoxal and formaldehyde from space, Geophys.Res. Lett.,, L10, doi:./00gl0. Wiedinmyer, C., S. K. Akagi, R. J. Yokelson, L. K. Emmons, J. A. Al Saadi, J. J. Orlando, and A. J. Soja (0), The Fire INventory from NCAR (FINN): A high resolution global model to estimate the emissions from open burning, Geosci. Model Dev.,, 1, doi:.1/gmd---0. Xie, H., L. Du, S. Liu, L. Chen, S. Gao, S. Liu, H. Pan, and X. Tong (01), Dynamic monitoring of agricultural fires in China from 0 to 01 using MODIS and GlobeLand0 data, ISPRS Intl. J. Geo-Info., (), 1. Zhang, Q., et al. (00), Asian emissions in 00 for the NASA INTEX-B mission, Atmos. Chem. Phys.,, -1, Zhang, L., Y. Liu, and L. Hao (01), Contributions of open crop straw burning emissions to PM. concentrations in China, Environ. Res. Lett.,, doi:./1- //1/0101.

9 Zheng, B., et al. (01), Trends in China's anthropogenic emissions since 0 as the consequence of clean air actions, Atmos. Chem. Phys., 1, 10-, doi:.1/acp Zhuang Y., R. Li, H. Yang, D. Chen, Z. Chen, B. Gao, and B. He (01), Understanding Temporal and Spatial Distribution of Crop Residue Burning in China from 00 to 01 Using MODIS Data. Remote Sens.,, 0; doi:.0/rs000 Zhu, L., D. J. Jacob, L. J. Mickley, E. A. Marais, D. S. Cohan, Y. Yoshida, B. N. Duncan, G. González Abad, and K. V. Chance (01), Anthropogenic emissions of highly reactive volatile organic compounds in eastern Texas inferred from oversampling of satellite (OMI) measurements of HCHO columns, Environ. Res. Lett.,, 00, doi:./1- ///00. Zhu, L., et al. (01), Observing atmospheric formaldehyde (HCHO) from space: Validation and intercomparison of six retrievals from four satellites (OMI, GOMEA, GOMEB, OMPS) with SEAC RS aircraft observations over the southeast US, Atmos. Chem. Phys., 1, 1, 1,0, doi:.1/acp Zhu, L., L. J. Mickley, D. J. Jacob, E. A. Marais, J. Sheng, L. Hu, G. G. Abad, and K. Chance (01a), Long-term (00 01) trends in formaldehyde (HCHO) columns across North America as seen by the OMI satellite instrument: Evidence of changing emissions of volatile organic compounds, Geophys. Res. Lett.,, doi:.0/ 01GL0. Zhu, L., et al. (01b), Formaldehyde (HCHO) as a hazardous air pollutant: Mapping surface air concentrations from satellite and inferring cancer risks in the United States, Environ. Sci. Technol., 1, 0, doi:.1/acs.est.b01.

10 Comparison of OMI-SAO HCHO to MAX-DOAS 1 HCHO column( 1 cm ) Aug-Sep 00 MAX-DOAS OMI-SAO Beijing Guangdong Wuxi Xianghe Jul 00 May-Aug 0-01 May-Sep 0-01 Figure 1. Evaluation of the OMI-SAO HCHO column data with mean ground-based MAX- DOAS measurements at four cities in China. The ground-based measurements are from Li et al. (01), Lee et al. (01), Vlemmix et al. (01) and Wang et al. (01), as first summarized by Cao et al. (01). The OMI-SAO data are means for the observation period.

11 OMI HCHO columns and estimated production from different sources (May-Sep, 00-01) g m - month -1 g m - month -1 1 molec cm - g m - month -1 1 Figure. Mean formaldehyde (HCHO) over China in May-September and its sources as estimated from bottom-up emission inventories of volatile organic compounds (VOCs). (a) HCHO columns from the OMI-SAO satellite retrieval with bias correction applied (Section ). (b) Mean anthropogenic production rate derived from MEIC VOC emissions with 1-day HCHO yields. (c) Same as (b) but for biogenic emissions from MEGAN v.1 driven by MERRA meteorology. (d) Same as (b) but for open fires from FINN. Note the lower scale for open fire emissions. The HCHO 1-day yields from oxidation of different VOC species are from Chan Miller et al. (01), as given in Table S1. The May-September total HCHO production amounts during from each source type in eastern China (east of E, covering the map domain) are shown inset.

12 Figure trends of formaldehyde (HCHO) columns over China. Values are fitted linear trends for May-September -month averages of each year. OMI trends are corrected for temperature as described in the text. (a) Observed OMI trends on the 0. o x0. o grid. From top to bottom, the three rectangles delineate the North China Plain (NCP), Huai River Basin (HRB), and Yangtze River Delta (YRD). (b) Same as (a) but on a x. grid. (c) trends simulated by GEOS-Chem at o x. o resolution with fixed biogenic sources at the 00 level. The averaged trends in eastern China are shown inset. 1 Figure trends of May-September VOC emissions from anthropogenic (MEIC) and open fire (FINN) sources, plotted on the same 0. o x0. o grid as OMI HCHO in Figure a. Note the smaller scale for open fire emissions. 1

13 HCHO production from anthropogenic and fire emissions 1 HCHO source (g month -1 m - ) ((a) North China Plain (b) Huai River Basin (c) Yangtze River Delta Open fire Anthropogenic Year Figure timeseries of May-September HCHO production rates from anthropogenic (MEIC) and open fire (FINN) emissions in the North China Plain (NCP), Huai River Basin (HRB), and Yangtze River Delta (YRD). The three regions are defined in Figure a. The HCHO production rates are derived from the emissions of individual VOCs and 1-day HCHO yields as described in the text. Change relative to 00 -% % 1% % (a) North China Plain OMI-SAO SCIAMACHY HCHO trends relative to 00 (b) Huai River Basin GOME- OMI-BIRA (c) Yangtze River Delta mean satellite data GEOS-Chem Figure timeseries of mean May-September HCHO columns from different satellite products over the North China Plain (NCP), Huai River Basin (HRB), and Yangtze River Delta (YRD). HCHO columns are shown as the changes relative to 00 (grey horizontal line) after correcting the temperature dependence of biogenic sources and the Pacific background, as described in the text. The three regions are defined in Figure a. Individual satellite products are shown as dashed lines and the mean of all products is the black bold line. Also shown is the trend in the GEOS-Chem simulation. 1