The central role of OH and HO₂ radicals (collectively known as HO_X) in driving tropospheric chemistry has long been recognized. OH oxidation of SO₂ in the upper troposphere initiates the nucleation of new particles, and HO_X radicals together with nitrogen oxides (NO_X) are key catalysts in the production of ozone, an effective greenhouse gas. Attention has recently focused on the upper troposphere due to the potential role of NO_X emissions from commercial aircraft in changing ozone and its climate forcing. Knowledge of the chemistry of HO_X is essential to assessing the effect of aircraft on ozone. Over the last five years direct atmospheric measurements of HO_X radicals and their critical controlling species have been conducted in the upper troposphere, providing the first tests of HO_X photochemistry in this region.

Initial surprises of upper tropospheric HO_X observations

The first measurements of HO_X radicals in the upper troposphere during the ASHOE/MAESA (1994), STRAT (1995-1996) and SUCCESS (1996) aircraft missions revealed a more photochemically active upper troposphere than had been anticipated [Folkins et al., 1997; Wennberg et al., 1998; Brune et al., 1998]. Observed HO_X levels were frequently 2–4 times higher than expected based on the commonly assumed primary source:

By contrast, observations in the lower stratosphere had been generally reproduced well by models. To explain the elevated levels of HO_X in the upper troposphere, a number of new sources have been proposed. Acetone is ubiquitous in the troposphere and, at the low levels of water vapor in the upper troposphere, its photolysis can largely dominate (R2) as a source of HOx [Singh et al., 1995; Arnold et al., 1997]. Inclusion of this source improved the predictions of models [McKeen et al., 1997; Brune et al., 1998], but in many cases observed HO_X levels were still much higher than expected. These cases were often associated with recent deep convection. How can deep convection affect upper tropospheric HO_X concentrations? Prather and Jacob [1997] found that convection turns over the upper troposphere at rates comparable to the photochemical processes controlling the abundance of HO_X. Thus, injection of air from the surface, carrying high levels of HO_X precursors, could be an additional HOx source to the upper troposphere. Such precursors might include H₂O₂ [Chatfield and Crutzen, 1984], CH₃OOH [Prather and Jacob, 1997; Cohan et al., 1999], and aldehydes [Müller and Brasseur, 1999]. Photolysis of convected peroxides and CH₂O, together with acetone, seemed indeed to provide a strong enough source to account for the observed levels of HO_X [Jaeglé et al., 1997]. However, it was not possible to test the importance of these new sources directly because they were not measured during the initial three campaigns.

Recent observations of HOx and its precursors during SONEX

The 1997 SONEX aircraft campaign over the North Atlantic provided the first measurements of HO_X concentrations concurrent with the ensemble of species thought to control HO_X production and loss: H₂O₂, CH₃OOH, CH₂O, O₃, H₂O, acetone and hydrocarbons. These observations allowed a detailed evaluation of our understanding of HOx chemistry in the upper troposphere. Figure 1 shows a summary comparison between SONEX measurements and model calculations for HO₂, OH, and HO₂/OH.

The cycling between OH and HO₂ takes place on a time scale of a few seconds and is mainly controlled by CO+OH (R3) and HO₂+NO (R4). This cycle is very important because it also leads to the production of ozone:

As seen in Figure 1c, the HO₂/OH ratio is reproduced by model calculations to within the combined uncertainties of observations (±20%) and rate coefficients (±100%). Thus, the photochemical processes driving the cycling between OH and HO₂ appear to be well understood [Wennberg et al., 1998; Brune et al., 1998, 1999]. How well are the absolute values of OH and HO₂ reproduced by model calculations constrained with observations of HO_X precursors? Figure 1 illustrates that observed levels of HO₂ and OH are predicted to within about 40% (the reported accuracy of the HO_X observations). The median model-to-observed ratio for HO₂ is 1.12. The model captures 80% of the observed variance in HO_X, which is driven by the local concentration of NO_X and the strength of the primary HO_X sources [Jaeglé et al., 1999b].

Despite this overall good agreement, some systematic differences between observed and modeled HO_X as a function of NO have been noted and could reflect flaws in our understanding of the coupling between HO_X and NO_X chemistry involving HO₂NO₂ in particular [Brune et al., 1999; Faloona et al., 1999]. The observations shown in Figure 1 are for cloud-free, daytime conditions only. The chemistry of HO_X at night and close to the terminator is not fully understood [Wennberg et al., 1999; Jaeglé et al., 1999b]. Model overestimates of HO_X inside cirrus clouds could be due to the rapid uptake of HO₂ on ice particles.

The primary sources of HO_X during SONEX were O(1D)+H₂O and acetone photolysis. The role of H₂O₂, CH₃OOH and CH₂O as net sources of HO_X was of much smaller importance compared to what had been inferred from previous campaigns. Although the concentrations of these HO_X precursors were enhanced in fresh convective outflows, the enhancements were driven primarily by the high levels of water vapor in these outflows. Most of the observations during SONEX took place at lower altitudes and warmer temperatures, such that water vapor levels were 3–10 times higher than during STRAT, SUCCESS, and ASHOE/MAESA, swamping out the effect of convected peroxides and aldehydes.

However, the concurrent observations of HO_X, H₂O₂, CH₃OOH and CH₂O allowed a constrained analysis of the photochemical interactions between HO_X and these species. In particular, observed H₂O₂ concentrations were reproduced by model calculations if heterogeneous conversion of HO₂ to H₂O₂ on aerosols was included [Jaeglé et al., 1999b]. Observations of CH₃OOH were systematically underestimated by factors of 2 or more, possibly indicating poor knowledge of CH₃OOH reaction kinetics at cold temperatures. Observations of CH₂O were generally below the 50 pptv detection limit of the instrument, consistent with model results; however frequent occurrences of high values in the observations were not captured by the model. These high CH₂O values were associated with high methanol, so heterogeneous conversion of methanol to CH₂O on aerosols might provide an explanation [Singh et al., 1999].

Sensitivity of HOx and ozone production to changes in NOx

The chemistry of HO_X and ozone production is tightly linked to the concentrations of NO_X: NO_X controls the cycling within HO_X, which leads to ozone production, and NO_X also regulates the loss of HO_X through reactions of OH with HO₂, HO₂NO₂ and NO₂. This results in the well known non-linear dependence of HO_X and ozone production on NO_X, which we can now test directly in the atmosphere with measurements of HO_X and NO_X species.

Figure 2a and Figure 2b show the variations of OH and HO₂ as a function of NO_X concentration during the SONEX mission. Photochemical model calculations for median background conditions, shown by the solid lines, illustrate the expected behavior of OH and HO₂ as a function of NO_X. In the model, OH increases with NO_X up to ~300 pptv NO_X due to the shift in the HO₂/OH ratio towards OH (reaction R4), and decreases with increasing NO_X at higher NOx concentrations due to the loss of HO_X promoted by NO_X. The concentration of HO₂ decreases with increasing NO_X as a result of the shift in the HO₂/OH ratio towards OH and thus a greater efficiency of HO_X sinks which depend on the OH concentrations.

These model dependencies of HO_X on NO_X are also generally found in the observations (Figures 2a and 2b). The scatter around the model lines can be often explained by variations in the magnitude of the local HO_X source [Jaeglé et al., 1999a]. However, for high NO_X concentrations (>300 pptv), the model constrained with background conditions during SONEX underestimated observed HO_X by a factor of 2 or more. These high levels of NO_X were often associated with convection and lightning. Using the locally observed concentrations of HO_X precursors (H₂O, acetone, peroxides and CH2O) for these points improves the agreement but still falls short of the observed levels of HO_X (see Figure 1). This discrepancy at high NO_X could be due to the presence of unmeasured HO_X precursors transported by convection, such as higher aldehydes, or it could indicate an incomplete understanding of the coupling between HO_X and NO_X chemistry at very high NO_X [Faloona et al., 1999; Brune et al., 1999].

Ozone production is initiated by OH oxidation of CO (R3), followed by reaction of the resulting HO₂ with NO (R4) and photolysis of NO₂ (R5). The rate-limiting step in this ozone production cycle is reaction of HO₂ with NO, such that the ozone production rate can be expressed as P(O₃) ~ k4 [HO₂] [NO]. Concurrent observations of HO₂ and NO can thus lead to a direct determination of P(O₃) in the upper troposphere (reaction of organic peroxy radicals RO2 with NO generally contributes less than 15% of P(O₃)). As shown in Figure 2c, the expected dependence of P(O₃) on NOx is similar to that of OH: the model predicts that P(O₃) should increase with increasing NO_X (NO_X-limited regime) up to a turnover point of ~300 pptv NOx, beyond which further increases in NO_X cause P(O₃) to decrease (NO_X-saturated regime). The bulk of the observations (NO_X<300 pptv) indeed shows a leveling off of the dependence of P(O₃) on NO_X as NO_X increases above 70 pptv, in accordance with the expected behavior. However at the highest NOx concentrations, P(O₃) computed from observed HO₂ and NO continues to increase with increasing NOx, suggesting a consistently NOx-limited regime which is at odds with model results. This directly results from the model underestimates of HO₂ at high NO_X (Figure 2a).

Similarly to results from the SONEX mission, P(O₃) versus NO_X relationships during ASHOE/MAESA, STRAT and SUCCESS also indicated a much greater prevalence of NOx-limited conditions than expected from models. Part of this dependence could be due to the association of elevated NO_X with elevated HO_X precursors transported by convection. However, problems remain for very high NO_X. Are these problems due to missing HO_X sources, or due to flaws in our understanding of the chemistry? Determining the cause for this discrepancy is essential because it greatly affects the sensitivity of P(O₃) to increasing NO_X. Current levels of NO_X (~50-100 pptv) place the upper troposphere in the NO_X-limited regime. Will future increases in NO_X due to anthropogenic activities lead to a leveling off of P(O₃) in the NO_X-saturated regime as predicted by models, or will P(O₃) continue to increase as suggested by observations? This issue clearly needs to be addressed, and calls for further studies of upper tropospheric chemistry.

References

1. Arnold, F., et al., Observations of upper tropospheric sulfur dioxide and acetone pollution: Potential implications for hydroxyl radical and aerosol formation, Geophys. Res. Lett., 24, 57-60, 1997.
2. Brune, W.H., et al., Airborne in situ OH and HO₂ observations in the cloud-free troposphere and lower stratosphere during SUCCESS, Geophys. Res. Lett., 25, 1701-1704, 1998.
3. Brune, W.H., et al., OH and HO₂ chemistry in the north Atlantic free troposphere, Geophys. Res. Lett., 26, 3077-3080, 1999.
4. Chatfield, R.B., and P.J. Crutzen, Sulfur dioxide in remote oceanic air: Cloud transport of reactive precursors, J. Geophys. Res., 89, 7111-7132, 1984.
5. Cohan, D.S., et al., Convective injection and photochemical decay of peroxides in the tropical upper troposphere: Methyl iodide as a tracer of marine convection, J. Geophys. Res., 104, 5717-5724, 1999.
6. Faloona, I., et al., Observations of HO_x and its relationship with NO_x in the upper troposphere during SONEX, J. Geophys. Res., in press, 1999.
7. Folkins, I., P.O. Wennberg, T.F. Hanisco, J.G. Anderson, and R.J. Salawitch, OH and HO₂ in two biomass burning plumes: Sources of HO_x and implications for ozone production, Geophys. Res. Lett., 24, 3185-3188, 1997.
8. Jaeglé , L., et al., Observations of OH and HO₂ in the upper troposphere suggest a strong source from convective injection of peroxides, Geophys. Res. Lett., 24, 3181-3184, 1997.
9. Jaeglé , L., et al., Ozone production in the upper troposphere and the influence of aircraft during SONEX: Approach of NO_x-saturated conditions, Geophys. Res. Lett., 26, 3081-3084, 1999a.
10. Jaeglé , L., et al., Photochemistry of HO_x in the upper troposphere at northern midlatitudes, J. Geophys. Res., in press, 1999b.
11. McKeen, S.A., et al., The photochemistry of acetone in the upper troposphere: A source of odd-hydrogen radicals, Geophys. Res. Lett., 24, 3177-3180, 1997.
12. Mü ller, J.-F., and G. Brasseur, Sources of upper tropospheric HO_x: A three-dimensional study, J. Geophys. Res., 104, 1705-1715, 1999.
13. Prather, M.J., and D.J. Jacob, A persistent imbalance in HO_x and NO_x photochemistry of the upper troposphere driven by deep tropical convection, Geophys. Res. Lett., 24, 3189-3192, 1997.
14. Singh, H.B., et al., High concentrations and photochemical fate of oxygenated hydrocarbons in the global troposphere, Nature, 378, 50-54, 1995.
15. Singh, H.B., et al., Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlantic, J. Geophys. Res., in press, 1999.
16. Wennberg, P.O., et al., Hydrogen radicals, nitrogen radicals and the production of ozone in the middle and upper troposphere, Science, 279, 49-53, 1998.
17. Wennberg, P.O., et al., Twilight observations suggest unknown sources of HO_x, Geophys. Res. Lett., 26, 1373-1376, 1999.