Climate Change 2001:
Working Group I: The Scientific Basis
Other reports in this collection Atmospheric measurements and modelling of photochemistry

Figure 4.9: (left panel) Observed versus modelled (1) HO2 abundance (ppt), (2) OH abundance (ppt), and (3) HO2/OH ratio in the upper troposphere (8 to 12 km altitude) during SONEX. Observations are for cloud-free, daytime conditions. Model calculations are constrained with local observations of the photochemical background (H2O2, CH3OOH, NO, O3, H2O, CO, CH4, ethane, propane, acetone, temperature, pressure, aerosol surface area and actinic flux). The 1:1 line (solid) and instrumental accuracy range (dashed) are shown. Adapted from Brune et al. (1999). (right panel) Observed (4) HO2 abundance (ppt), (5) OH abundance (ppt), and (6) derived O3 production rate (ppb/day) as a function of the NOx (NO+NO2) abundance (ppt). Data taken from SONEX (8 to 12 km altitude, 40° to 60°N latitude) and adapted from Jaeglé et al. (1999). All values are 24-hour averages. The lines correspond to model-calculated values as a function of NOx using the median photochemical background during SONEX rather than the instantaneous values (points).
Atmospheric measurements provide another cornerstone for the numerical modelling of photochemistry. Over the last five years direct atmospheric measurements of HOx radicals, made simultaneously with the other key species that control HOx, have been conducted over a wide range of conditions: the upper troposphere and lower stratosphere (e.g., SPADE, ASHOE/MAESA, STRAT; SUCCESS, SONEX, PEM-TROPICS A & B), the remote Pacific (MLOPEX), and the polluted boundary layer and its outflow (POPCORN, NARE, SOS). These intensive measurement campaigns provide the first thorough tests of tropospheric OH chemistry and production of O3 for a range of global conditions. As an example here, we present an analysis of the 1997 SONEX (Subsonic assessment program Ozone and Nitrogen oxide EXperiment) aircraft campaign over the North Atlantic that tests one of the chemical models from the OxComp workshop (HGIS).

The 1997 SONEX aircraft campaign over the North Atlantic provided the first airborne measurements of HOx abundances concurrent with the controlling chemical background: H2O2, CH3OOH, CH2O, O3, NOx, H2O, acetone and hydrocarbons. These observations allowed a detailed evaluation of our understanding of HOx chemistry and O3 production in the upper troposphere. Figure 4.9 (panels 1-3) shows a comparison between SONEX measurements and model calculations (Jaeglé et al., 1999) for OH and HO2 abundances and the ratio HO2/OH. At each point the model used the local, simultaneously observed chemical abundances. The cycling between OH and HO2 takes place on a time-scale of a few seconds and is mainly controlled by reaction of OH with CO producing HO2, followed by reaction of HO2 with NO producing OH. This cycle also leads to the production of ozone. As seen in Figure 4.9, the HO2/OH ratio is reproduced by model calculations to within the combined uncertainties of observations (±20%) and those from propagation of rate coefficient errors in the model (±100%), implying that the photochemical processes driving the cycling between OH and HO2 appear to be understood (Wennberg et al., 1998; Brune et al., 1999). The absolute abundances of OH and HO2 are matched by model calculations to within 40% (the reported accuracy of the HOx observations) and the median model-to-observed ratio for HO2 is 1.12. The model captures 80% of the observed variance in HOx, which is driven by the local variations in NOx and the HOx sources (Faloona et al., 2000, Jaeglé et al., 2000;). The predominant sources of HOx during SONEX were reaction of O(1D) with H2O and photodissociation of acetone; the role of H2O2 and CH3OOH as HOx sources was small. This was not necessarily the case in some of the other airborne campaigns, where large differences between measured and modelled OH, up to a factor of 5, were observed in the upper troposphere. In these campaigns the larger measured OH concentrations were tentatively ascribed to enhanced levels of OH precursors, such as H2O2, CH3OOH, or CH2O, whose concentrations had not been measured.

Tropospheric O3 production is tightly linked to the abundance of NOx, and Figure 4.9 (panel 6) shows this production rate (calculated as the rate of the reaction of HO2 with NO) for each set of observations as function of NOx during the SONEX mission. Also shown in Figure 4.9 (panels 4-5) are the measured abundances of OH and HO2 as a function of NOx. The smooth curve on each panel 4-6 is a model simulation of the expected relationship if the chemical background except for NOx remained unchanged at the observed median abundances. This curve shows the "expected" behaviour of tropospheric chemistry when only NOx is increased: OH increases with NOx abundances up to 300 ppt because HO2 is shifted into OH; it decreases with increasing NOx at higher NOx abundances because the OH reaction with NO2 forming HNO3 becomes the dominant sink for HOx radicals. Production of O3 is expected to follow a similar pattern with rates suppressed at NOx abundances greater than 300 ppt under these atmospheric conditions (e.g., Ehhalt, 1998). These SONEX observations indicate, however, that both OH abundance and O3 production may continue to increase with NOx concentrations up to 1,000 ppt because the high NOx abundances were often associated with convection and lightning events and occurred simultaneously with high HOx sources. By segregating observations according to HOx source strengths, Jaeglé et al. (1999) identified the approach to NOx-saturated conditions predicted by the chemical models when HOx sources remain constant. A NOx-saturated environment was clearly found for the POPCORN (Photo-Oxidant formation by Plant emitted Compounds and OH Radicals in north-eastern Germany) boundary layer measurements in Germany (Rohrer et al., 1998; Ehhalt and Rohrer, 2000). The impact of NOx-saturated conditions on the production of O3 is large in the boundary layer, where much of the NOx is removed within a day, but may be less important in the upper troposphere, where the local lifetime of NOx is several days and the elevated abundances of NOx are likely to be transported and diluted to below saturation levels. This effective reduction of the NOx-saturation effect due to 3-D atmospheric mixing is seen in the CTM modelling of aviation NOx emissions where a linear increase in tropospheric O3 is found, even with large NOx emissions in the upper troposphere (Isaksen and Jackman, 1999).

Other reports in this collection