The oxidation photochemistry and transport of hydrogen peroxide and formaldehyde at three sites in Europe : trends, budgets and 3-D model simulations
Date issued
Authors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
License
Abstract
The photochemistry of hydrogen peroxide (H2O2) and formaldehyde (HCHO) mediates the budget of HOx (= OH + HO2) and thus has a strong impact on ozone (O3) and NOx (= NO + NO2) in the troposphere. Comprehensive ground-based field measurements at three different sites in Europe were performed: DOMINO from Nov 20–Dec 9 2008 (El Arenosillo, Southern Spain), HUMPPA from Jul 12–Aug 12 2010 (Hyytiälä, Southern Finland) and PARADE from Aug 15–Sep
10 2011 (Kleiner Feldberg, Germany). Mixing ratios of gas-phase H2O2 and HCHO were measured in-situ between 8 and 21 m above ground level via two customized instruments (Model AL2021 and AL4021, Aero-Laser GmbH, Garmisch-Partenkirchen, Germany). Average daytime levels of H2O2 for DOMINO, HUMPPA and PARADE were 82 pptv, 639 pptv, and 323 pptv, respectively. Night-time mixing ratios reached 59 pptv (DOMINO), 99 pptv (HUMPPA) and 486 pptv (PARADE). Mean diurnal profiles of H2O2 showed a strong diurnal pronounciation for DOMINO and HUMPPA with maximum values in the afternoon, while an inverse profile was observed during PARADE. In case of HCHO, daytime HCHO averages of 569 pptv, 465 pptv and 1.9 ppbv were measured for DOMINO, HUMPPA and PARADE, respectively. Nighttime mixing ratios were 505 pptv (DOMINO), 383 pptv (HUMPPA) and 1.9 ppbv (PARADE) and showed smooth diurnal variations. The average deviation from photostationary state (PSS) for all campaigns range from 1.2 to 1.5. Simple steady-state calculations reveal that the major amount of daytime H2O2 during DOMINO and PARADE can be explained by photochemistry, while for HUMPPA the levels are overestimated by a factor of 6. Further, ambient HCHO was under photochemical control during DOMINO and HUMPPA. The production can be expressed by background chemical pathways, namely, the oxidation of methane, isoprene and methanol by OH radicals. During PARADE, over 80% of ambient HCHO were primarily emitted from anthropogenic sources and transported. HUMPPA allows challenging the current understanding of sources and sinks of H2O2 and HCHO in the boreal forest. The study focuses on calculations of the PSS, [ROx], [OH] and of the H2O2 and HCHO budgets. Four regimes can be identified and used for discussion: stressed boreal (R1, Jul 12-22 2010), cold and clean (R2, Jul 23-25), transported pollution (R3, Jul 26-29) and normal boreal with some short pollution events (R4, Aug 1-12). The calculated [HO2] agrees with the observations (r2 = 0.71), while the OH time series are reproduced reasonably well. The budget calculations excluding dry deposition show an average overestimation of H2O2 by a factor of 5 up to an order of magnitude (r2 = 0.24), while HCHO exceed the measurement by 3.6 times (r2 = 0.26). A simple linear regression method yield median values of deposition velocities of 3.03 cm/s for H2O2 and 1.08 cm/s for HCHO. Daytime transport plays a major role in boreal summer: entrainment of H2O2-rich air significantly enriches the ambient mixing ratios, while HCHO-poor air decreases the HCHO levels. Including deposition and transport, the budget of H2O2 is reproduced rather well (slope: 0.9520 ± 0.0834, intercept: 0.2330 ± 0.0543 ppbv, r2 = 0.30), while that for HCHO was reasonable (slope: 0.2320 ± 0.0258, intercept: 0.2980 ± 0.0224 ppbv, r2 = 0.35). The classification of the NOx sensitivity concerning the net H2O2 production rate showed evidence for a maximum in the NOx interval ranging from 0.24 to 0.41 ppbv. The net HCHO production, followed a linear trend until the NO interval from 0.05 to 0.07ppbv. Higher NO mixing ratios (70 to 140 pptv) resulted in the formation of a plateau. Tree model simulations (reference R; two sensitivity studies S1 and S2) were performed with the 3-D model EMAC for HUMPPA excluding terpene chemistry. The parameters for S1 were 50% NOx emissions and for S2 50% NOx emissions and double deposition velocities for H2O2 and HCHO). Evaluation of NOx led to using a vicinal box due to pollutant transport from a nearby town. The resulting average daytime levels about 350 pptv agreed with the observations. Large-scale meteorology, long-lived trace chemical species (carbon monoxide and methane) and the transport of biomass burning plumes from Russia were reproduced reasonably well. The model reproduced the observed downward transport of secondary HCHO from biomass burning plumes in the morning hours. As expected, R showed deficits in radical chemistry due to lacking terpene chemistry. H2O2 and HCHO levels exceeded measured data by a factor of 4 and 3.4, respectively. Lower NOx (S1 and S2) had an insignificant effect on the mixing ratios of the modeled trace chemical species including H2O2 and HCHO. Higher deposition velocities (S2) resulted in 2.2-fold (H2O2) and 3-fold (HCHO) increased levels, respectively. Sensitivity simulations with higher deposition velocities and basic terpene chemistry remain future research objects.