Dutch OMI NO 2 product DOMINO www temis

Dutch OMI NO 2 product DOMINO (www. temis. nl) Folkert Boersma Collaborators: Daniel J. Jacob, Miri Trainic, Yinon Rudich, Ruud Dirksen, and Ronald van der A Comparison of NO 2 air pollution in Israeli cities measured from the ground and from SCIAMACHY and OMI

Motivation Trends in emissions • Can we use satellite measurements of NO 2 to obtain better estimates of NOx emissions? Boersma et al. , Atmos. Environ. , 2008 Monitoring of a target polluter • Is tropospheric NO 2 a proxy for near-surface NO 2?

Comparison of surface and satellite NO 2 Different quantities and error budgets: • near-surface concentrations vs. trop. columns • point measurements vs. spatial average • interference from NOz vs. AMF errors BUT: • Validation still sparse, especially on larger spatial and longer temporal scales • Allows for validation of diurnal cycle Schaub et al. , ACP, 2006 Interference • Ground-based instruments with molybdenum converter overestimate NO 2 • Interference from HNO 3, PAN, alkyl nitrates • Effect largest in summer (photochemistry) for regions downwind of sources

• 8 urban stations in Israel • Half-hourly NO 2 & O 3 concentrations • Molybdenum converters

Known issues with in situ NO 2 Summarize Dunlea et al. , et al. (ACP, 2007)

What does this mean for urban, in situ NO 2 ? • Very little interference at 10: 00 (fresh emissions, little photochemical processing yet) • Interference modest at 13: 30 • Interference correlates with ambient O 3 concentration, but O 3 itself does not represent an interference slope = 0. 1 Israeli cities: no specific NO 2 measurements available, but we have in-situ observed O 3

Interference correction based on in situ O 3 • O 3 = O 3[13: 30] – O 3[10: 00] • Dunlea et al. : interference = 0. 1 O 3 • Two limiting cases: slope = 0. 10 – no correction at all – correcting NO 2 at 13: 45 as follows: slope = 0. 09 slope = 0. 24

Δlat, lon < 0. 1° OMI VZA < 35° fclrad < 50%

r = 0. 61 (n=396) RMA: y = -1. 28 + 0. 90 x Δlat, lon < 0. 1° No correction OMI VZA < 35° fclrad < 50%

r = 0. 54 (n=396) RMA: y = -0. 68 + 0. 92 x Δlat, lon < 0. 1° Uniform No correction (-8%) OMI VZA < 35° fclrad < 50%

Boundary-layer columns from surface NO 2 • • • Assume well-mixed boundary-layer Extrapolate surface NO 2 throughout the BL depth Seasonal variation in noontime BL depths over Israel from Dayan et al. [1988, 2002] z (m) 1091 DJF 790 MAM, SON 608 JJA NO 2 (ppb) Summertime subtropical high is associated with subsidence of FT air, “capping” the BL.

Boundary-layer columns from surface NO 2 Surface – OMI vs. BL column - OMI No correction Column-column comparison No interference correction: y = -0. 86 + 0. 90 x Interference correction: y = -0. 35 + 0. 94 x No correction

Seasonal variation in surface and OMI NO 2 columns

Weekly cycle in surface and OMI NO 2 columns Beirle et al. (ACP, 2004)

Diurnal cycle in NO 2 columns? NO 2 : NOx E k NO 2+OH+M HNO 3+M Boersma et al. , JGR, 2008

Diurnal cycle in surface-based and satellite NO 2 n = 26 Δlat, lon < 0. 1° fclrad < 50%

Diurnal cycle in surface-based and satellite NO 2

MAM

Diurnal cycle in surface-based and satellite NO 2

Dutch OMI NO 2 product DOMINO (www. temis. nl) END

NOx emission inventories in need of verification New set of emission inventories Lin Zhang et al. (2008) • EDGAR (global) • Streets et al. 2006 • EMEP (Europe) GOME, SCIAMACHY (1996 -2005) EMEP (1996 -2005) Konovalov et al. , ACPD, 2008

NOx emissions show trends in the Middle East Trend 1996 -2006 from GOME, SCIAMACHY • Figure 1996 -2006 Cairo Tehran Calcutta Delhi + 1. 3%/yr + 6. 1%/yr + 2. 2%/yr + 7. 4%/yr

NOx emissions show trends in the Middle East Richter et al. , 2005 1996 -2002 van der A et al. , 2008 1996 -2006

Effect of averaging kernel a: scattering weights x. TM 4: TM 4 NO 2 profile b: forward model parameters GEOS-Chem direct GEOS-Chem with AK MI

Effect of averaging kernel GEOS-Chem direct NGC = ∑x. GC, l GEOS-Chem with AK NGC, AK = ∑Al·x. GC, l OMI NOMI = ∑Al·xtrue, l

Effect of averaging kernel GEOS-Chem direct GEOS-Chem with AK GEOS-Chem 0. 71 r (gc, omi) = direct r (gc_ak, omi) = 0. 75 Ngc = 0. 87 Ngc_ak = 0. 93 (+7%) Nomi = 1. 09 OMI

What does NGC, AK > NGC mean? Write NGC, AK > NGC differently: Definition of tropospheric air mass factor: Substitute tropospheric air mass factor: Or ( SGC = scatt. weights · GEOS-Chem NO 2 profile ):

What does NGC, AK > NGC mean? Mean NO 2 profiles over Middle East for 1 -7 February 2006 from TM 4 and GEOS-Chem

Correcting for background NO 2 Can we simply subtract UT NO 2 at mid-latitudes? Rynda Hudman et al. , JGR, 2007 Randall Martin, 2006 Lin Zhang et al. , 2008

Checking GEOS-Chem background NO 2 with OMI

• Qinbin paper • Duncan paper

Comparison of ground-based to GOME NO 2 Ordonez et al. (ACP, 2006) • Compares 7 years of data over Po Valley to GOME NO 2 • Regional comparison (~10. 000 km 2) • Struggles with interference for ground-based stations

Comparison of ground-based to SCIAMACHY NO 2 Blond et al. (2006) • Compares 1 year of data over northwestern Europe to SCIAMACHY NO 2 • Regional comparison (1. 800 km 2) • No interference correction for ground-based stations Annual mean 2003 • Best agreement for rural stations All stations Rural stations Differences attributed to (lack of) spatial representativity of stations

Boundary-layer columns from surface NO 2 Surface – SCIAMACHY vs. BL column - SCIAMACHY Δlat, lon < 0. 25° n = 135 No correction y = -0. 32 + 0. 82 x No correction y = -0. 35 + 0. 70 x If Δlat, lon < 0. 1° as for OMI: r = 0. 65, n = 32 y = -0. 47 + 0. 93 x