Atmospheric Photooxidants

Lead authors :

To download the text of this chapter in PDF format, click here, for the figures click here, and for the tables here.

2.1. Introduction
2.2. Primary Emissions of O3 Precursors 2.2.1. NOx
2.2.2. CO, CH4 , and NMHC 2.3. Global Distribution of O3 Precursors 2.3.1. NOx
2.3.2. CO, CH4 , and NHMC

3. Photochemistry in the Troposphere

3.1. Background
3.2. IGAC Activities related to HOx photochemistry. 3.2.1. Direct measurements of atmospheric OH and HO2
3.2.2. HO2 /OH Measurement Campaigns
3.2.3. Conclusions from modelling studies
3.2.4. Direct measurements of peroxy radicals
3.2.5. Measurements of ozone and peroxide climatologies.
3.2.6. Modelling of radical chemistry and ozone production and loss
3.2.7. The influence of water vapour on tropospheric free radical chemistry
3.2.8. Reactive nitrogen chemistry
3.2.9. Progress in modelling the global budget of OH
3.2.10. Novel free radical oxidation chemistry
3.2.11. Halogen chemistry in the Troposphere
3.2.12. Progress in Laboratory Kinetics 3.3. Synthesis - What have we learnt?

4. Meteorological Processes affecting Ozone

4.1. Linking the local and global scales
4.2. Chemical transport models, dependence on meteorological data
4.3. Mesoscale modelling
4.4. Improvements in numerical weather prediction
4.5. Cloud cover prediction
4.6. Boundary layer-free troposphere exchange rates
4.7. Regional, intercontinental, hemispheric
4.8. The capability to forecast ozone in the atmospheric boundary layer

5. A Climatology of Tropospheric Ozone

5.1. The Origin of Ozone in the Troposphere
5.2. Stratosphere-Troposphere Exchange
5.3. Simulation of Ozone in the Global Troposphere
5.4. Global Measurements of Ozone and Its Precursors 5.4.1. Aircraft Data
5.4.2. Vertical Profiles 5.5. Ground-based and Sonde Data

6. Current Research: Results from Recent IGAC Campaigns designed to study Tropospheric Chemistry on Various Scales

6.1. Chemistry of the continental atmosphere in developed regions.
6.2. Chemistry of continental outflow in the marine atmosphere
6.3. Studies of Aircraft Emissions
6.4. Chemistry of the remote atmosphere

7. References

1. Introduction

Approximately 20% of the Earth's atmosphere is made up of oxygen but this is only a reservoir for the more reactive molecules and free radicals which provide most of the oxidising power of the atmosphere. These are formed almost entirely by photochemistry and many involve a coupling between ozone and water vapour, the influence of which on the overall oxidising power of the atmosphere is often under-estimated by atmospheric scientists.
The main oxidants to be dealt with here include ozone (O₃), hydroxyl radicals (OH), peroxy radicals (both inorganic (HO₂) and organic (RO₂)) and peroxides (H₂O₂ and RO₂H). Other oxidants include nitrate radicals (NO₃) and halogen atoms; these however only play a subsidiary role and probably are unimportant over large parts of the atmosphere.
Most of the Earth's ozone (~90%) is in the stratosphere where it is formed from the direct photolysis of oxygen with a maximum wavelength of light of 242 nm. Since the minimum wavelength of light reaching the 15 troposphere is about 295 nm, ozone cannot made there by oxygen photolysis, and the only route to ozone PRODUCTION in the lower atmosphere is by photolysis of nitrogen dioxide, amplified greatly by a chain reaction including hydroxyl radicals and peroxy radicals formed from OH oxidation of many carbonaceous species. The same chain reaction also results in ozone loss and the switchover from loss to gain is mostly dependent on the concentrations of nitric oxide at levels of the order of 10 pptv. The measurement of the global distribution of this molecule therefore presents one of the major challenges of tropospheric chemistry.
Hydroxyl radicals are the main oxidising species in the gas phase troposphere. They react with almost all trace gases containing H, C, N, O, S and the halogens (with the notable exception of the CFCs) and N₂O which are mostly removed by photolysis in the stratosphere but they do not react with the principal components of the atmosphere such as N₂, O₂, H₂O and CO₂.
The chemistry of the lower atmosphere is thus much concerned with the production of OH radicals from ozone photolysis in the presence of water vapour, and their loss either by reaction with other trace gases, or their subsequent removal by reaction with oxides of nitrogen and peroxides. Other important species are peroxy radicals and peroxides, the former of which are intimately involved in ozone production and loss, and the latter of which are a principal sink for HO_x radicals and are also involved in reactions with sulphur dioxide (SO₂) in clouds leading to the wide-scale production of a sulphate aerosol (SO₄).
About 10% of the Earth's ozone is present in the troposphere, where its presence is vital for the production of hydroxyl radicals. An obvious source is transfer from the large reservoir of ozone in the stratosphere. However, efficient photochemical loss of ozone at high water vapour concentrations requires the presence of a large source within the troposphere which is many times larger than any reasonable estimate of the stratospheric source, but which is highly regional since it is dependent on the NO_x concentration. This in turn is highly dependent on the distribution of emission sources and on transport processes within the troposphere, i.e. global circulation.
All of this will be explained in more detail later in the text. One of the main consequences of tropospheric photochemistry is both the production and loss of large amounts of ozone in the troposphere, in much larger amounts than is transferred from the main reservoir in the stratosphere. The establishment of this fact has been one of the principal achievements in the field of tropospheric chemistry over the last decade. The subject of tropospheric chemistry and tropospheric ozone has until now been treated rather as an addendum to other main subject areas in recent reviews of stratospheric ozone by WMO/UNEP, and of greenhouse gases and their influence on climate by IPCC. Tropospheric chemistry and tropospheric ozone is intrinsically important in its own right and it is the object of this review to present it as such. It is important because it is the chemistry that controls the trace gas composition of the lower atmosphere including important greenhouse gases such as methane, and also the composition of air entering the stratosphere. It is responsible for controlling atmospheric composition over geological time by modifying emissions from the geosphere and the biosphere and allowing them to be removed in soluble forms into the hydrosphere. It is currently responsible for many forms of pollution including photochemical smog, sulphate and nitrate aerosol production, and also for the production of most of the acidity found in rain. Over the last five or ten decades it has almost certainly been responsible for a doubling in ozone concentrations throughout much of the troposphere. For many reasons it must make a large contribution to the extent of the greenhouse effect.
This review deals sequentially with the main processes affecting tropospheric photochemistry, commencing with emissions of source gases which are processed photochemically leading to their removal and the production of other gas-phase and particulate-phase products, then presenting a detailed update of our knowledge of the photochemistry concentrating on in- situ measurements. This is followed by a section outlining the meteorological processes which transport both the primary emissions and their secondary products. The section of meteorology is particularly concerned with the transport of ozone on regional and global scales. A section dealing with the global budget of ozone in the troposphere from both a modelling and a measurement perspective follows this. This section particularly attempts to elucidate our current depth of understanding of this complex problem and it is succeeded by a section which reports highlights from some of the many comprehensive measurement campaigns, mostly using aircraft which have been carried out under the auspices of IGAC.

2. O₃ Precursors

2.1. Introduction

Ozone (O₃) photochemical production in the troposphere occurs by oxidation of carbon monoxide (CO), methane (CH₄), and other hydrocarbons (generally referred to as NMHC) in the presence of nitrogen oxides (NO_x). A quantitative understanding of the budgets and distributions of these O₃ precursors is key to understanding the global budget and distribution of tropospheric O₃, as well as in predicting how tropospheric O₃ concentrations will likely change in response to future changes in emissions of these precursors. It is also worth noting that the chemical cycles of O₃ precursors are intricately coupled to the chemical cycling of O₃ itself (see discussion in Section 2). In this section, we briefly summarize the present state of knowledge of the sources and distribution of these precursors. For a more detailed review of these topics, the reader is referred to Brasseur et al. [1999] 8 and Ehhalt in Zellner [1999].

2.2. Primary Emissions of O₃ Precursors
2.2.1. NOx

The release of nitrogen oxides into the atmosphere occurs due to the fixation of nitrogen by human activities or by natural phenomenon. A breakdown of the estimated primary NO_x emissions by source category is provided in Table 3.2.1 [WMO, 1999]. The latitudinal variation of the zonally-averaged total column NO_x source is shown in Figure 3.2.1. Clearly, the predominant source of NO_x (~22 Tg N year^-1 ) is surface-based fossil-fuel combustion accounting for about half of the total global source. This source is concentrated in Northern Hemisphere midlatitudes. In the tropical boundary layer, the principal sources of NO_x are biomass burning (~8 Tg N year^-1 ) and emissions from soils (~7 Tg N year^-1 ). In the free troposphere, production of NO_x by lightning (~5 Tg N year^-1 ) is the dominant source, while emissions from aircraft and injection from the stratosphere (each about 0.5 Tg N year^-1 ) make smaller contributions. Table 3.2.1 also shows that, even from a global perspective, there are considerable uncertainties in the magnitudes of the biomass burning and lightning sources of NO_x. The uncertainty in the biomass burning source is related to uncertainties in the areal extent of biomass burning and in the associated NO_x emission factors. The uncertainty in the lightning source is related to uncertainties in the characteristics of individual lightning strokes as well as to uncertainties in extrapolating to the global-scale.

Figure 3.2.1 Latitudinal Distribution of Zonally-Averaged NO_x Emissions (COURTESY OF H. LEVY II, NOAA/GFDL)

Table 3.2.1. Estimated Primary Emissions of O₃ Precursors. Best estimates are listed, with ranges given in parentheses.

Sources	CH₄ (Tg/yr)	CO (Tg/yr)	NMHC (Tg C/yr)	NO_x (Tg N/yr)
Energy use Aircraft Biomass burning Vegetation Soils Lightning Ruminants Rice paddies Animal wastes Landfills NH3 oxidation N2O breakdown Domestic sewage Wetlands Oceans Freshwaters CH4 hydrates Termites Total	110(65-155) 40 (10-70) 85 (60-105) 80 (30-120) 30 (15-45) 40 (20-60) 25 (20-30) 145 (115-175) 10 (5-15) 5 (1-10) 10 (5-15) 20 (1-40) 600 (520-680)	500(300-900) 500 (400-700) 100 (60-160) 50(20-200) 1150 (780-1960)	70(60-100) 40 (30-90) 400 (230-1150) 50(20-150) 560 (340-1490)	22(20-24) 0.5 (0.2-1) 8 (3-13) 7 (5-12) 5 (2-20) 0.9 (0-1.6) 0.6 (0.4-1) 44 (30.73)

*NO_y produced in the stratosphere and transported to the troposphere.
(From 1998 WMO O₃ Assessment Report - IGAC Should Check On Copyright)

While global estimates of the magnitudes of the individual NO_x sources are useful, there are several issues worth considering in the context of assessing current and future impacts of anthropogenic activities on tropospheric O₃. It is important to note that NO_x sources of smaller magnitude can have disproportionately large impacts on NO_x budgets in specific parts of the atmosphere owing to the relatively short tropospheric lifetime of NO_x. It is also worth noting that while the biomass burning and soil sources of NO_x are partly natural in character, they have been impacted by human activities. For example, Crutzen and Zimmermann [1991] estimated that preindustrial biomass burning emissions were about 10% of present-day values owing to the lower human population in the tropics. In a similar vein, Yienger and Levy [1995] estimate that preindustrial emissions of NO_x from soils were about 35% lower than present-day values. The low preindustrial emission rate reflects in part the use of fertilizers and in part the conversion of forest with low NO_x emission rates to grasslands and pastures with higher NO_x emission rates. Another issue worth noting is that the ambient NO_x concentration at certain remote locations during certain times of the year may be largely due to decomposition of PAN (REFERENCES) rather than due to direct transport of primary NO_x. In addition, it has been hypothesized that rapid chemical conversion of HNO₃ to NO_x by a heterogeneous pathway can be a significant source of NO_x in the remote troposphere (REFERENCES).

2.2.2. CO, CH₄, and NMHC

Primary emissions of CO into the atmosphere occur principally due to surface-based fossil-fuel combustion and biomass burning (see Table 3.2.1). Each of these sources is estimated to contribute about 500 Tg CO year^-1 . In addition to these primary sources, there are significant secondary sources of CO in the troposphere owing to oxidation of CH₄ and NMHC. According to recent estimates by Wang et al. [1998a], the combined magnitude of these secondary sources (~800 Tg CO year^-1 from CH₄ oxidation and ~300 Tg CO year^-1 from NMHC oxidation) is comparable to the combined magnitude of the primary sources. The global CH₄ source is estimated to be about 600 Tg year^-1 . A significant portion of this total is believed to be associated with anthropogenic activities related to energy use, biomass burning, animal husbandry, waste management and agriculture. The global NHMC source is estimated to be 600 Tg C year^-1 , with the dominant component being emissions of isoprene and monoterpenes from vegetation and a smaller contribution from anthropogenic activities related to energy use and biomass burning.
As is the case with NO_x sources, there are significant uncertainties in the estimated sources of CO, CH₄, and NMHC. In the case of CO, the uncertainties arise among other things due to uncertainties in CO emission factors during combustion processes, the areal extent of biomass burning and the associated CO emission factors, and the yield of CO from CH₄ and NMHC oxidation processes (REFERENCES). It is worth noting also that though oxidation of isoprene is generally considered to be a natural source, the yield of CO depends on the ambient NO_x concentration [Miyoshi et al., 1994] which could have a significant anthropogenic component. In the case of isoprene and monoterpenes, uncertainties in the global emission rates are related to sensitive dependence of emissions on a variety of factors (e.g. light, temperature, tree species, etc.) which makes global extrapolation difficult.

2.3. Global Distribution of O₃ Precursors
2.3.1. NO_x

Our present understanding of the global distribution of NO_x is derived from sporadic measurements at a variety of locations complemented by results from global-scale model simulations. In general, the spatial NO_x distribution is highly variable as would be expected given its relatively short lifetime and geographically inhomogeneous source distribution. Ground-based measurements at rural locations in North America and Europe show mean NO_x levels ranging from 1-3 ppbv in polluted regions (REFERENCES) to a few tenths of a ppbv in cleaner air masses (REFERENCES). By contrast, measured median summertime NO_x concentrations at Sable Island are less than 0.1 ppbv (REFERENCES) and even lower median NO_x concentrations (< 0.05 ppbv) have been reported from shipboard measurements in the eastern North Atlantic [Carsey et al., 1997]. At more remote locations, such as in the boundary layer of the equatorial Pacific, NO_x concentrations as low as 5-10 pptv have been measured [Torres and Thompson, 1993].
Airborne measurements of NO and NO₂ have also been made during a variety of field campaigns. It is generally accepted that airborne NO observations are more reliable than NO₂ measurements [Crawford et al., 1996]. Emmons et al. [1999] have recently created a data composite of tropospheric NO measurements from 27 aircraft field campaigns in the 1983- 1996 time period, and have developed maps of average NO concentrations at different altitudes based on this data composite. Figure 3.2.2 shows the NO maps developed by Emmons et al. [1999] for 2-4 km and 6-8 km. In the lower troposphere, average NO concentrations of 20-60 are evident over the continents. The mean NO concentrations over biomass burning regions in South America and southern Africa during September-November are comparable to summertime concentrations over the United State. Evidence for outflow of NO_x from continents to adjacent ocean regions is seen in the measurements off China during March-May and over the tropical south Atlantic during September-November. Over more remote oceanic regions, NO concentrations in the lower troposphere are typically less than 20 pptv. Similar features are seen in the 6-8 km maps, though interestingly NO concentrations over the central pacific and over the south Atlantic are significantly higher at 22 6-4 km than at 2-4 km.

The overall features of the observed NO_x concentrations are simulated reasonably well by global-scale models [e.g., Wang et al., 1998 (a or b); Levy et al., 1999]. Based on an analysis of model results, Levy et al. [1999] have delineated the largest source contributing to NO_x in various regions, altitudes, and seasons. Their results (see Figure 3.2.3) illustrate the complex controls on tropospheric NO_x as well as the fact that human activities have substantially altered NO_x concentrations in large parts of the atmosphere. It is also worth noting that the model-based analysis of Wang et al. [1998b] provides only limited support for the hypothesis that there is a significant secondary source of NO_x in the remote troposphere associated with a rapid heterogeneous conversion of HNO₃ to NO_x.

FIGURE 3.2.3 Maps showing largest source contributor to NO_x by season and altitude (S=stratosphere; BG=biogenic; L=lightning; A=aircraft; BM=biomass burning; C=fossil-fuel combustion)
(FROM LEVY ET AL. 1999 - IGAC SHOULD CHECK ON COPYRIGHT)

2.3.2. CO, CH₄, and NHMC

There exists a well established monitoring program for surface CO and CH₄ operated by the Carbon Cycle Group (CCG) at NOAA's Climate Monitoring and Diagnostics Laboratory (CMDL). The geographical distribution of the measurement sites is shown in Figure 3.2.4. In addition to the surface sites, measurements using tall tower and aircraft platforms are carried out at a limited number of locations as part of this monitoring program. The surface measurements reveal distinctive seasonal and latitudinal patterns in CO and CH₄ which are related to the distribution and seasonality of the sources and sinks of these compounds (see Figures 3.2.5 and 3.2.6). As discussed in Novelli et al. [1998], background surface CO mixing ratios are largest (200-225 ppbv) in the high latitudes of the Northern Hemisphere, while lowest mixing ratios are found in the Southern Hemisphere during summer (35-40 ppbv). The absolute seasonal amplitude is also highest in the high northern latitudes and the seasonal cycles in the two hemispheres are out of phase by about 6 months. CH₄ concentrations generally range from about 1.85 ppmv at high northern latitudes during winter to about 1.68 ppmv at the South Pole during the austral summer.

Figure 3.4 The NOAA/CMDL Carbon Cycle Group Global Air Sampling Network
(FROM http://www.cmdl.noaa.gov/ccgg/figures/figures.html -IGAC SHOULD CHECK ON COPYRIGHT)

Figure 3.2.5 Three dimensional representation of the latitudinal distribution of atmospheric carbon monoxide in the marine boundary layer for the period 1993 through 1997. Data from the NOAA CMDL cooperative air sampling network were used. The surface represents data smoothed in time and latitude.
(FROM http://www.cmdl.noaa.gov/ccgg/figures/figures.html - IGAC SHOULD CHECK ON COPYRIGHT)

Figure 3.2.6 Three dimensional representation of the latitudinal distribution of atmospheric methane in the marine boundary layer for the period 1988 through 1997. Data from the NOAA CMDL cooperative air sampling network were used. The surface represents data smoothed in time and latitude.
(FROM http://www.cmdl.noaa.gov/ccgg/figures/figures.html - IGAC SHOULD CHECK ON COPYRIGHT)

The CMDL CO and CH₄ measurements also provide information on recent trends in global CO and CH₄ concentrations. The data of Novelli et al. [1998] indicate that background mixing ratios have shown a downward trend (~2.3 ppbv/year on a global- and annual-average basis) over the last decade (see Figure 3.2.7). While reasons for this trend are not entirely clear, Novelli et al. [1998] speculate that changes in OH and in sources such as biomass- burning have contributed to the overall decrease. While the globally-averaged atmospheric CH₄ concentration continues to increase, the rate of increase has decreased [Dlugokencky et al., 1998]. This decrease in the growth rate has been attributed to a slow approach to a globally-averaged steady-state CH₄ concentration of about 1.8 ppmv consistent with relatively constant CH₄ emissions and OH concentrations.

Figure 3.2.7: Hemispheric and Global time series and trend curves for CO
(FROM NOVELLI ET AL. 1998 - IGAC SHOULD CHECK ON COPYRIGHT)

Figure 3.2.8: Time variation of global average atmospheric methane (top) and calculated emissions and loss assuming a methane lifetime of 8.9 years (bottom).
(FROM http://www.cmdl.noaa.gov/ccgg/news.html- - IGAC SHOULD CHECK ON COPYRIGHT)

Unlike the case for CO and CH₄, there is no well established monitoring network for NMHC. Consequently, our knowledge of the distribution of NMHC is derived from sporadic field measurements. A summary of a limited set of measurements for selected NMHC compiled by Brasseur et al., [1999], shows background mixing ratios over continental regions of a few ppbv or less (see Table 3.2.2). Over oceans, lower concentrations of light NMHC are generally observed though there is evidence oceanic emissions of NMHC such as ethene and propene [e.g., Rudolph and Ehhalt, 1981; Singh and Salas, 1982; Rudolph and Johnen, 1990].

Table 3.2.2: Background Continental NMHC Mixing ratios in Several Rural and Remote Terrestrial Ecosystems (Surface and Boundary Layer)

Ecosystem	Ethane	Ethene	Propane	Benzene	Isoprene	_-Pinene
Sub-Alpine forest ^a Tropical rainforest ^b Tropical rainforest ^c Tropical rainforest ^d Wooded savanna ^e Wooded savanna ^f Mixed temperate forest ^g	2.2 0.98 1.10 0.73 0.65 0.26 x	0.46 0.97 0.70 0.29 0.33 0.09 x	1.27 0.16 0.10 0.11 0.05 0.88	0.24 0.08 0.17 0.07 0.16 0.04 0.13	0.63 2.04 5.45 1.21 0.04 <0.01 4.33	0.14 0.10 0.20 0.06 <0.01 <0.01 0.28

^a Niwot Ridge, CO, USA, August/September, surface (Greenberg and Zimmerman, 1984).
^b Amazon Tropical Forest, Brazil, dry season, boundary layer (Zimmerman et al., 1988).
^c Amazon Tropical Forest, Brazil, wet season, boundary layer (Zimmerman et al., 1988).
^d Nigerian Tropical Forest, wet season, surface (Zimmerman et al., 1988).
^e Kenya, dry season, surface (Greenberg et al., 1985).
^f Kenya, wet season, surface (Greenberg et al., 1985).
^g Jacquin, Alabama, July 1990, boundary layer (Guenther et al., 1995).
(FROM BRASSEUR ET AL. 1999 - IGAC SHOULD CHECK ON COPYRIGHT)

3. Photochemistry in the Troposphere
FREE RADICAL CHEMISTRY IN THE GLOBAL TROPOSPHERE - a Synthesis of Current Results

3.1. Background

Photochemical generation of ozone in the troposphere was first identified in the work carried on in California in the 1950s and until the 1970s was thought to be a local phenomenon associated with air pollution. Photons at wavelengths shorter than ~290 nm are not present in the troposphere so that O₂ cannot be photolysed and ozone cannot be produced via the Chapman mechanism i.e:

O₂ ^hn-> O + O (3.0)
O+ O₂ + M -> O₃ + M (3.1)
However NO₂ can be photolysed in the troposphere, yielding ground state O which recombines with O₂ to form ozone i.e.:
NO₂ ^hn-> NO + O (3.2)
NO also reacts with ozone reforming NO₂:
NO₂ + O₃ -> NO₃ + O₂ (3.3)
The coupled reaction cycle 3.1 - 3.3 leads to the establishment of a photostationary state between NO, NO₂ and O₃ in the sunlit atmosphere on a timescale of ~ 100s. Since 3.2 is effectively instantaneous we then obtain:

Although O atoms are reactive chemically, their concentration is suppressed by reaction with molecular oxygen (reaction 3.1), and they do not play a significant role in trace gas oxidation. This research pointed to the possible importance of free radicals of the HO_x family (H, OH, HO₂) and related radicals derived from organic species (e g CH₃O₂, CH₃O) in atmospheric chemistry. In 1968 Weinstock and Niki proposed that OH radicals led to the atmospheric oxidation of CO:
OH + CO -> H+ CO₂ (3.4)
and in 1971 H. Levy outlined a new theory which predicted significant OH concentrations in the normal sunlit troposphere and pointed out its significance for the chemical removal of many minor constituents. This theory involves production of OH from the small amount of highly reactive excited atomic oxygen, O(¹D), produced by photolysis of O₃ at wavelengths less than l~320 nm, which reacts with water vapour, present at high concentrations in the lower troposphere, in competition with quenching to the ground state:
O₃ + hn ( l < 310 nm) -> O( ¹D) +O₂ (3.5)
O( ¹D) + M -> O( ³P) + M (3.6)
O( ¹D) + H₂O -> OH + OH (3.7)
OH rapidly converts to HO₂, primarily (~70% of the time in unpolluted air) by reaction with CO:
OH+ CO -> H+ CO₂ H + O₂ + M -> HO₂ + M (3.9)
and HO₂ converted back to OH by reaction with O₃:
2 HO₂ + O₃ -> 2O₂ + OH (3.10)
About 30% of the OH radicals convert to HO₂ via a more complex chain of reactions initiated by:
OH+ CH₄ -> CH₃ + H₂O (3.11)
followed by oxidation of the CH₃ radical to HCHO and HO₂. A steady state between HO₂ and OH is rapidly established in sunlight, allowing us to define the odd hydrogen family HO_x = [OH] + [HO₂].
HO_x is lost through formation of H₂O₂ and HNO₃:
HO₂ + HO₂ -> H₂O₂ + O₂ (3.12)
OH + NO₂ + M -> HNO₃ + M (3.13)
These products are highly soluble, and are removed from the atmosphere fairly rapidly by absorption into cloud water and rain-out. Both H₂O₂ and HNO₃ can also be photolysed releasing HO_x (and NO_x). Using steady state approximation, a knowledge of the photolysis frequency of ozone, and rate coefficients for the radical loss reaction, Levy [1972] estimated the [OH] present in the sunlit surface atmosphere to be 3 x 10 -6 molecule cm^-3 .
Crutzen [1973] and Chameides and Walker [1973] first suggested that photochemical generation of ozone might be important in the global troposphere. They pointed out that the oxidation of CO and CH₄ initiated by OH radicals would lead to ozone production, whenever NO_x was present, due to the reactions of peroxy radicals with nitric oxide. NO competes with O₃ for the available HO₂, in an extended reaction cycle:
OH + CO(+O₂ ) -> HO₂ + CO₂ (3.8)
HO₂ + NO -> O H+NO₂ (3.14)
NO₂ can be photolysed yielding ozone i.e.:
NO₂ + hn -> NO+ O (3.2)
O+ O₂ + M -> O₃ + M (3.1)
Thus the net effect of incorporating reactions 3.8 - 3.14 is: net:
CO+ 2O₂ -> CO₂ + O₃
so that ozone is produced by the overall oxidation cycle. Note that both OH and NO are regenerated during the cycle.
The oxidation of methane is initiated by:
OH + CH₄ ( +O₂ ) -> H₂O + CH₃O₂ (3.11)
The principal fate of CH₃O₂ in a low NO_x environment is:
HOSUB>2 + CH₃O₂ -> CH₃ O O H+O₂ (3.15)
CH₃OOH is removed from the atmosphere in a similar way to hydrogen peroxide, leading to a net loss of HO_x. In this situation the oxidation of CO and CH₄ leads to a loss of ozone at a rate equal to the rate of the O[¹D] + H₂O reaction. The reaction of methyl-peroxy with NO leads to the formation of NO₂ (3.13), which leads to further ozone formation via 3.10) and to formaldehyde which also can be photolysed in the troposphere to liberate HO_x.
C H₃O₂ +NO -> CH₃O + NO₂ (3.16)
C H₃O₂ + O₂ -> HCHO+ HO<>SUB>2 (3.17)
HCHO + hn -> HCO + H (3.18)
HCO + O₂ -> CO + HO₂ (3.19)
Thus, the presence of NO completely shifts the balance of the oxidation process in favour of production of both ozone and HO_x.
These developments showed the potential importance of photochemistry in determining the budget of ozone in the troposphere. Thus there is a sink for ozone - the O[¹D] reaction with H₂O - and a source for ozone (from HO_x/NO_x coupling). The balance between these modifies the ozone fields, which would otherwise be determined only by transport and surface deposition. The mechanism also explains the anthropogenic perturbation of the ozone amounts by man-made emissions of NO_x and volatile organic compounds. Quantitative understanding of tropospheric ozone production has been largely dictated by the extent of knowledge of the amounts and distribution of tropospheric NO_x and VOCs.
Following the introduction of tropospheric free radical theory, most interest was directed in determining the lifetime of pollutant gases due to reaction with OH. Of particular interest were the halocarbons because it was realised that their potential for depletion of stratospheric ozone depended on their tropospheric lifetime, which in turn depended on the global mean OH concentration. In 1975 Singh and others used a simple box model of the troposphere together with observational data for methyl chloroform, and an inventory of its industrial production and release, to establish its tropospheric lifetime. This information combined with a knowledge of the rate coefficient for its reaction with OH, allowed an estimate of the global mean concentration of OH.
Subsequently more sophisticated models of tropospheric chemistry were developed, in which 2-dimensional time dependent OH fields were computed using chemical schemes with ozone photochemistry, hydrocarbon and NO_x gas phase chemistry, and physical scavenging by aerosols and rainout. Some of the first examples of these models were reported in 1977 by Crutzen and co-workers, and by Derwent et al. [1977]. They are now widely used tools for tropospheric research.
At the same time the first attempts to measure OH directly in the atmosphere were being made. Field instruments based on two techniques were developed utilising spectroscopic properties of OH in the ultraviolet region. The first used laser-induced fluorescence (LIF) where the emission was detected from OH following excitation by pulsed laser radiation near 280 nm. This method presented many difficulties for tropospheric conditions, most notably the unavoidable production of OH due to photolysis of O₃ by the probe laser. The second technique utilised long path absorption spectroscopy (LPAS), where the characteristic absorption features of OH in its UV spectrum were detected in a broad band laser beam propagating over a long atmospheric path, up to 10 km. This method lacks intrinsic sensitivity and is subject to interference due to absorptions by other atmospheric gases present at much higher abundance. However calibration of this measurement is straightforward since the absorption cross-sections for OH are well known. Both these methods have subsequently benefited from technical improvements and during the period of IGAC have provided a substantial amount of data which has demonstrated the basic validity of tropospheric photochemical theory. <
Thus by 1988 at the start of IGAC there existed a well developed theory of tropospheric OH which enable estimates to be made of local time-varying and global mean concentrations. This provided a working basis for determination of tropospheric lifetime of halocarbons, which were being considered as substitutes for CFCs, of climate gases e.g.CH₄ and DMS, and the budget of tropospheric ozone. The theory was supported by the limited amount of observational data on relevant atmospheric composition available at that time, such as measurements of CH₃CCl₃ and ¹⁴CO, and embodied the most up-to-date knowledge of atmospheric chemistry. The first reliable measurements demonstrating the presence of OH in the sunlit atmosphere using the LPAS method had been made [Perner et al., 19..], and indirect detection of peroxy radicals using a chemical amplification technique had been reported [Cantrell, et al., 19..]. However, these observations were neither extensive nor reliable enough to validate the photochemistry schemes being used in models. The nitrate radical had been detected in the nighttime boundary layer, but its reactivity was considered to be low. Oxidation by halogen radicals was not considered an issue in tropospheric chemistry.
More importantly there was no consensus as to the extent to which photochemical processes were active in determining tropospheric ozone amounts, and hence tropospheric oxidising efficiency. One aim of IGAC was to extend the knowledge of tropospheric free radical chemistry needed to formulate models with better diagnostic and predictive power for Global Change issues involving ozone and chemically reactive greenhouse gases.

3.2. IGAC Activities related to HOx photochemistry.

One of the main aims of IGAC was the creation and co-ordination of major field campaigns to provide observational data on atmospheric chemical species to test theory. The photochemical oxidant production in the troposphere was the major focus of a number of field projects worldwide. In recent years there have been some major improvements in the instrumental capacity for measurement of trace species involved in the photochemical ozone balance, and many important observational campaigns have been carried out in remote areas [Ridley et al., 1987; Cantrell et al., 1996a, b; Jaffe et al., 1996; Carsey et al., 1997]. In the first years the emphasis, was on precursor species (hydrocarbons, nitrogen oxides) and the reservoir species for HO_x and NO_x (peroxides, nitrates, HNO₃ and PAN). In recent years measurements of HO_x free radical species, oxygenated VOCs and other radicals such as NO₃ and halogen oxides have also been achieved. Measurements have been at meteorologically and climatologically well- characterized ground observation stations as well as from ships and aircraft. Airborne measurements have expanded recently, allowing concentrations of HO_x and related species to be determined throughout the free troposphere. These have provided new opportunities to test the photochemical theories of tropospheric free radical chemistry and its relation to ozone.

3.2.1. Direct measurements of atmospheric OH and HO2: Testing the theory of fast photochemistry.

Photochemical theory predicts that a steady state concentration of the free radicals OH and HO₂ exists in the sunlit atmosphere. Measurement of these concentrations provides a direct test of the theory. On factor that was recognised early was the sensitivity of the steady-state radical concentrations to local conditions of solar irradiation and composition of trace species. Consequently field campaigns have required the assembly of a suite of measuring instruments to measure not only the radicals, but other species and parameters influencing the radical concentrations. Fortunately there has been a significant advance in instrument capability for in-situ measurements of many species in recent years.
Improved techniques for OH. The LIF- based techniques have benefited from improvements of laser and detector technology allowing detection of emission from excitation of ground state OH near 308 nm, reducing the problem of laser generation of OH by photolysis of ambient ozone. LIF combined with the use of an expansion jet to prepare the air sample for measurement forms the basis of the FAGE (Fluorescence Assay by Gas Expansion) method which has been developed successfully by several groups. HO₂ can be measured in the same instrument after conversion to OH by reaction with NO. Progress in the development of long path absorption spectroscopy has been mixed, most progress having been made using folded path, multi-reflection optical configurations. However it can be noted that considerable improvement has been made in broad band DOAS (Differential Optical Absorption Spectroscopy) for the detection of several species including nitrate radicals and halogen oxides. A major achievement is the development of an entirely new technique for OH, based on SICIMS (Sulphur Isotope Chemical Ionisation Mass spectrometry). The sensitivity and time response of this method in the field exceeds that of all others, and allows diurnal profiles and even residual nighttime levels of OH to be detected. The sensitivity and time response of the SICIMS instrument are: less than 1x10 5 molecules cm^-3 for a few minutes averaging, and for 1 hour averaging the 2s detection limit is 2x10⁴ [Eisele et al., 1996]. Instruments using SICIMS have been used for both remote surface and airborne measurements. Alongside these developments for OH, the past 10 years has seen significant improvement in the peroxy radical chemical amplifier (PERCA) method, which can be used to measure total atmospheric HO_x radical concentration. In this method radicals present in a flowing air sample are allowed to react with a mixture of CO and NO. A chain reaction results leading to formation of NO₂ with a yield equal to the product of the initial radical concentration and the chain length, which is determined in the calibration procedure.

3.2.2. HO₂/OH Measurement Campaigns

The following summary traces the progress in HO_x radical measurements. 1991-2 saw successful direct measurement of HO_x (OH and HO₂) at a remote site at high altitude on the Island of Hawaii, MLOPEX 2. Maximum OH of 4 to 6x10⁶ molecule cm^-3 was measured using SICIMS [Eisele et al., 1996], which was within a factor of 2 of modelled values; an unmeasured reactive organic was proposed to explain low measured levels; these effects were much larger for upslope winds assumed to carry organics from vegetation emissions; model overpredicts by factor of 2 or 3 [McKeen et 11 al, 1997].
Two important intercomparison campaigns took place in the US in 1993 and in Germany in 1994. The TOHPE intercomparison at Idaho Hill involved measurement of OH by SICIMS, LIF (OH/HO2) and LPAS. Only two systems worked satisfactorily; a fourth OH technique was used which did not get published. This was the wet chemical salicylic acid method. Problems were encountered with the interpretation of long path OH absorption data. The calculated [OH] was typically a factor of 1.5 greater than observed. The calculated [HO₂] was up to a factor of 10 greater than observed. The HO₂/OH ratio was dependent on local NO_x; values of 15-18 observed for NO_x > 100 pptv agreed well with model calculations whilst for clean air the values (3-4) were lower than predicted [Eisele et al., 1997, Mather et al., 1997]. The POPCORN campaign involved intercomparison between a FAGE-LIF and LPAS in a multi-reflection probe, at a rural site in North Germany. Good agreement found between techniques was observed (see Figure 3.3.1). The recorded maximum [OH] was 1x10 7 molecule cm^-3 and [OH] correlated well with j _(O¹D) (see Figure 3.3.2), [Dorn et al., 1996; Hofzumahaus et al., 1996; 1998; Plass-D�lmer et al., 1998].

Figure 3.3.1. Correlation between OH measurements made by long path UV absorption and Fluorescence assay.

Fig.3.3.2. Relationship between diurnal variation of [OH/] and J[O¹D].

The first reported aircraft measurements of OH in lower atmosphere were conducted in 1995 as part of the ACE 1 campaign in the Southern Ocean in the vicinity of Tasmania and New Zealand (reference needed). A SICIMS instrument mounted in a aircraft was successfully deployed. An important observation was the much higher [OH] (8 to 15 x 10⁶ molecule cm^-3 ) above clouds compared to clear skies (3 to 5 x 10 6 molecule cm^-3 ). Comparison of results with model calculation showed greatest errors in clouds and in the marine boundary layer [Mauldin et al., 1997, 1998].
The first aircraft measurements of OH and HO₂ in the upper troposphere were made using the LIF technique in the STRAT campaigns in 1995- 6[Jaegl� et al., 1997]. The concentrations of OH and HO₂ were substantially higher than predicted by models based on the O(¹D) + H₂O source of radicals, leading to the suggestion that acetone photolysis was a major OH source in dry upper troposphere/lower stratosphere. Acetone has been found to be a widespread constituent of the troposphere and is believed to result from degradation of various volatile organic compounds, both natural and man- made.
Sources of OH other than ozone photolysis in mid/upper troposphere was also indicated by airborne measurements of OH and HO₂ by LIF in the SUCCESS campaign conducted in 1996 [Brune et al., 1998]. Observed midday [OH] was 2 to 10 x 10⁶ molecule cm^-3 , and [HO₂] was 6 to 30 x 10⁷ molecule cm^-3 . [OH] was sensitive to cloud and the HO₂/OH ratio, measured in the upper troposphere in clean air and in aircraft contrails, was observed to be a function of NO_x, and varied from >100 at low NO_x (<20 ppt) to ~2 at very high NO_x (~1ppb). The agreement of the ratio with a model was good. Further airborne measurements with a SICIMS instrument, conducted in 1996 on the PEM-TROPICS A campaign, showed midday boundary layer [OH] of 2 to 10 x 10⁶ cm^-3 with higher levels (x 2-3) seen above cloud [Mauldin et al., 26 1999].
In 1997 the Subsonic Assessment: Ozone and Nitrogen Oxide Experiment (SONEX) assembled the most complete measurement complement to date for studying HO_x (OH and HO) chemistry in the free troposphere. This mission included flying in contrails in the North Atlantic flight corridor. The measured HO_x, and also the measured HO₂, depended upon NO_x. Observed and modeled HO and HO₂ agree on average to within experimental uncertainties (see Figure 3.3.3). However, significant discrepancies occur as a function of NO and at solar zenith angles. Some discrepancies appear to be removed by model adjustments to HO-NO chemistry [Brune et al., 1999; Jaegl� et al., 1999].

Figure 3.3.3 Calculated and observed OH and HO₂ as a function of [NO], data from SONEX experiment (Brune et al., 1999). There are a number of fairly sophisticated box and trajectory models developed with improving mechanisms. For clean air the predictions of simple reduced models not much different to those using large mechanisms.
Further surface measurements of OH and HO₂ have been conducted recently, and these are summarised in Table 3.3.1. These results show that OH is indeed being measured. The quality of OH measurements has improved dramatically, giving much lower detection limits and freedom of interferences. Full diurnal profiles are now measured routinely. Several intercomparisons have now been made between physically distinct techniques, notably POPCORN (FAGE and LPA), but there have been no blind intercomparisons of HO_x measurements.

Table 3.3.1. Recent Surface Measurements of OH Radicals

Experiment	Location	Technique	[OH] and [HO₂] noon	Comment	Reference
ALBATROS S (1996)	Atlantic Ocean	FAGE/LPAS	2 - 10 x 10⁶	Shipborne; good latitude dependence	Hofzumahaus et al. [19xx]
EASE (1996)	Mace Head, Eire	FAGE	2 - 10 x 10⁶	Model over-prediction	Creasey et al. [1997]
EASE (1997)	Mace Head, Eire	FAGE	1 - 6 x 10⁶ 0.5 - 3 x 10⁸	Model over-prediction	Carslaw et al, JGR in press [1999]
AEROBIC (1997)	Greece	FAGE	3 - 10 x 10⁶ 1 - 8 x 10⁸	Heavily forested area
SOAPEX-2 (1999)	Cape Grim, Tasmania	FAGE	1 - 5 x 10⁶ 0.8 - 2 x 10⁸	Good agreement for OH
PUMA (1999)	Birmingham, UK)	FAGE	3 - 8 x 10⁶ 1.5 - 10 x 10⁸
PRIME 1999	Ascot, UK	FAGE	2 - 9 x 10⁶ 1.6 - 5 x 10⁸	Data during eclipse; OH does not peak with j_(O¹D)
BERLIOZ (1998)	Berlin	FAGE		OH and HO₂ overpredicted in models
PROPHET ()	Michigan, US	FAGE		Measured on towers; large OH observed at night	Brune [19..]
IZANA ()	Tenerife	LPAS	4 - 5 x 10⁶	Elevated OH during forest fire	Comes [1997
SCATE (1993)	Antarctica	CIMS	5 - 7x10⁵	Hourly averages	Eisele [19..]
1993	Los Angeles	FAGE	5 x 10⁶ 2 x 10⁸	Good agreement with simple model	Hard and O Brien, JGR 1999

There is now a considerable database for OH at many surface locations with differing characteristics and also in the free troposphere from aircraft. Recently the campaigns have provided comprehensive supporting data e g CO, CH₄, CH₂O, j _(O¹D) , j _(NO₂) , other photolysis rates from spectroradiometric measurements, aerosol size and surface area, enabling case study modelling to be performed in order to validate theory.

3.2.3. Conclusions from modelling studies (a) Ground. In general the models overpredict observed [OH] and [HO₂] (e.g. TOPHE, (is it TOPSE) ACSOE). A number of possible reasons have been suggested, including heterogeneous losses, missing gas phase sinks, poor co-location of instruments, etc. In SOAPEX-2, preliminary modelling shows that under baseline conditions, model and experiment agree to within 20%, but agreement deteriorates (overprediction of [OH] and [HO₂]) as airmasses influenced by continental sources move over the site at Cape Grim. This suggests that the primary production of HO_x in clean air is accurate, but the secondary production of radicals, e g via HO₂+NO, is not correctly modelled. The opposite behaviour was found in AEROBIC, a forested site, where the model underpredicts measured OH and HO₂, which is thought to be due to additional HO_x sources from O₃-terpene reactions, which may be underestimated in models. However the mechanism is likely to be incomplete, as all terpenes are converted to equivalent alpha-pinene. The modelling of OH and HO₂ observed in the polluted air of the Los Angeles basin seems remarkably good considering the simple parameterised treatment of VOC.
(b) Airborne. Initial model/experiment comparisons showed deficiency in OH and HO₂. Now that acetone/peroxide sources of HO_x are included agreement is pretty good in the remote UTLS, but in and around clouds there are discrepancies. This is attributed to poorly-measured actinic flux above and missing in-cloud heterogeneous processes for HO₂ loss. In SONEX, measurements were made in contrails of aircraft in North Atlantic Flight Corridor. A huge variation in was NO_x observed, and OH and HO₂ was observed to be a strong function of NO_x.

3.2.4. Direct measurements of peroxy radicals

Measurement of the atmospheric steady state concentrations of peroxy radicals provides a direct insight into the local production or loss rate of ozone, as well as a diagnosis for HO_x photochemistry. Improved field measurements of RO₂ has been has been an objective of several projects in IGAC. Two techniques are used; the PERCA and MIESR (Matrix isolation electron spin resonance), and much effort has been made in obtaining reliable field intercomparison measurement using these techniques. PERCA makes no distinction between organic peroxy radicals and HO₂, whilst the MIESR method can give information on specific peroxy radicals, although with low resolution. Stand-alone deployments suffer from uncertainty in absolute calibration. Table 3.3.2 shows a summary of measurements campaigns for peroxy radicals. To date only terrestrial surface measurements have been reported for RO₂. There are still a very limited HO₂ measurements mainly using FAGE. The measurement is not as straightforward as for OH.

Table 3.3.2 Summary of measurements campaigns for peroxy radicals

Experiment Location, Year Technique(s) [HO₂] or S[RO₂] pptv Comments Reference(s)

Oregon, 1986, 1987 LIF (HO₂) 8) One coastal site, one urban site; similar maxima at both sites for clear-sky Hard et al., [1992]

Schauinsland, 1990 MIESR 40 Forested location; anti-correlation between RO₂ and NO₃ at night Mihelcic et al., [1993]

ROSE 1 Alabama, 1990 PERCA 200 Forested location; measured levels compared well with PSS calculations Cantrell et al., [1992, 1993]

MLOPEX 2 Mauna Loa, 1991-2 PERCA 25 Free troposphere; upslope/downslo pe Effects; measurements low compared with PSS calculations; peroxy radicals observed at night Cantrell et al., [1996a, 1996b, 1997]

SONTOS Ontario, 1992 PERCA 23 Rural site Arias and Hastie, [1996]

OCTA Iza�a, Tenerife, 1993 3 PERCA instruments and 1 MIESR instrument 80-100 2 PERCA instruments within 25% of MIESR instrument, 1 systematically low Zenker et al., [1998]

Denver, Colorado, 1993 PERCA 20-80 Urban site; reactions of O₃ with alkenes Claimed as important source of peroxy radicals at night Hu and Stedman, [1995]

FIELDVOC Brittany, 1993 PERCA 60 Measurements much lower than PSS calculations; low night-time levels Cantrell et al., [1996]

LAFRE Los Angeles, 1993 LIF (HO₂) 8 Measurements in smog; poor agreement with simple model at midday George et al., [1999 ]

TOHPE Idaho Hill, Colorado, 1993 LIF(HO₂), PERCA 6_(HO₂), 60 RO₂/HO₂ ratio 4-15 times larger than predicted; HO₂/OH ratio in range 15-80: agreed well with theory for [NO] > 100 pptv; factor of 3-4 too low under clean conditions Cantrell et al., [1997]; Stevens et al., [1997]

PRICE I Schauinsland, 1994 PERCA MIESR i 25 (PERCA), 50 (MIESR) No systematic difference between PERCA instruments; MIESR instrument measured systematically higher Carpenter, [1996]

WAO-TIGER Weybourne, UK, 1993-5 PERCA 12 Positive correlation of peroxy radicals and NO₃ observed at night Carpenter, [1996], Clemitshaw et al. [1997]; Carslaw et al., [1997]; Carpenter et al., [1998]

BESSE Bush Estate, Edinburgh, 1994 PERCA 15 Measurements systematically low compared with PSS Calculations Carpenter, [1996]

SOAPEX I Cape Grim, Tasmania, 1995 PERCA 11 Relationships of peroxy radicals with j_(O¹D) in clean and polluted air; O₃ production /destruction compensation point ca. 25 pptv NO_x; CH₃O₂ at night; winter / summer comparison; measurements systematically low by factor of 2-3 Monks et al., [1996]; Penkett et al., [1997]; Cox, [1999]; Monks et al., [2000]

ATAPEX Mace Head, Ireland, 1995 PERCA 5 Compensation point ca. 50 pptv NO Carpenter et al., [1997]

STRAT Aircraft, 1995-6 LIF (HO₂) Measurements in upper troposphere; Acetone photolysis as major HOx source in dry upper troposphere / lower stratosphere Jaegl� et al., [1997]

SUCCESS Aircraft, 1996 LIF (HO₂) 3-15 HO₂/OH ratios 30 % higher than modelled Brune et al., [1998]

Aircraft, 1996 SICIMS (HO₂) 10-40 4-8 km altitude; good agreement between measurements and steady-state calculations for clear skies Reiner et al., [1997; 1999]

ACSOE-EASE Mace Head, Ireland, 1996-7 LIF (HO₂), PERCA 8 _(HO2), 25 Effect of NO_x on radical levels; control of peroxy radicals at night by NO₃ / O₃ reactions Creasey et al., [1997]; Monks et al., [1999]

SONEX Aircraft, 1997 LIF (HO₂ Effect of aircraft emissions in troposphere Brune et al., [1999]; Jaegl� et al., [1999]

FREETEX Jungfraujoch, Switzerland, 1996, 1998 PERCA 18 Radicals shown to be dependent on V(j(O¹D)) in low-NO_x ozone- production regime; seasonal comparison; photochemical control of diurnal ozone variation Zanis et al., [1999; 2000]

Oki Island, Japan, 1998 LIF (HO₂) 14 HO₂ observed at night Kanaya et al., [1999]

BERLIOZ Berlin, 1998 LIF (HO₂ No results yet published

PROPHET Michigan, LIF (HO₂) No results yet published

SOAPEX 2 Cape Grim, Tasmania, 1999 LIF (HO₂) PERCA No results yet published

PUMA Birmingham, 1999 LIF (HO₂) No results yet published

PRIME Ascot, 1999 PERCA Instrument LIF (HO₂) No results yet published

It will be seen from the remarks in the table that the body of information is broadly consistent with the theory, although difficulty arises when the detailed amounts of OH, HO₂ and RO₂ (total peroxy radicals) are examined. RO₂ concentrations in the surface atmosphere are mostly lower than calculated in the models. Other points are:
1. RO₂ radicals are present at nighttime (MLOPEX,1992; FIELDVOC,1993)
2. Total peroxy radical concentration show a square root dependence on j_(O¹D) in clean air, as expected from theory (Figure 3.3.4); this relationship changes towards linear dependence as NO_x increases ATAPEX;SOAPEX1)

Fig. 3.3.4 Diurnal variation in total peroxy radicals and their dependence on J[O¹D] 1/2 for baseline conditions at Cape Grim, Tasmania.

3.2.5. Measurements of ozone and peroxide climatologies in clean air - evidence for ozone photochemistry.

Long term measurements and intensive measurement campaigns at meteorologically and climatologically well characterized ground observation stations have provided new opportunities to test the photochemical theories of tropospheric ozone. Sites in the southern hemisphere (Cape Grim, Tasmania, 40.7�S, 144.7�E) and in the northern hemisphere (on the Oki Islands, 36�N; 133 �E) in the Sea of Japan and at Mace Head, Ireland, (53.3�N, 9.9�W) have been exploited in this way [Ayers, et al., 1992; 1996; Penkett et al., 1997; Carpenter et al., 1997, Jaffe et al., 1996]. These sites are exposed to unpolluted maritime air for long periods, interspersed with periods when more or less polluted continental air advects into the region. The homogeneity of the marine air masses in terms of chemical composition provides optimum conditions for testing atmospheric photochemistry in the surface atmosphere, and the contrast between the behavior of maritime and continental air at the sites provides information on the perturbation of the unpolluted atmosphere by terrestrial emissions, especially man-made pollutants.
This approach has been particularly successful at Cape Grim where Ayers et al. [1992] have observed diurnal modulation of ozone and hydrogen peroxide concentrations which are consistent with a simple model in which O₃ photolysis depletes ozone in the marine boundary layer (MBL) during daytime with associated production of peroxides (H₂O₂ and organic peroxides, mainly CH₃OOH) through the well known reaction sequence (Figure 3.3.5). Changes in patterns when polluted air advected over the Cape Grim site, and also at Mace Head in Ireland are quantitatively consistent with photochemical theory [Carpenter et al, 1997].

Fig. 3.3.5 Diurnal variation in ozone and total hydroperoxide (ROOH = CH₃OOH + H₂O₂ ) for baseline conditions at Cape Grim, Tasmania.

3.2.6. Modelling of radical chemistry and ozone production and loss

Because the time constants for radical photochemistry are short compared with transport times, photochemical theory can be readily tested by comparison to OH and RO₂ measurements using simple zero-dimensional models, providing the boundary condition imposed by the concentrations of relevant stable constituents can be defined. Numerous examples of modelling studies associated with the OH and HO₂ measurements have been reported (see above Tables). The more recent measurements with fast time resolution generally compare well with the general features of the photochemical theory (see for example Hofzumahaus et al., 1998; Carslaw et al, 1999a]. Ayers et al. [1997] were able to describe the peroxy radicals, ozone and hydrogen peroxide observed at Cape Grim with a simple 1 dimensional model of the marine boundary layer.
Box modelling has also been used to investigate the critical nitric oxide concentration NO crit , the so called 'compensation point', where ozone loss and production due to reactions:

HO₂ + O₃ -> 2O₂ + OH (3.10)
HO₂ + NO -> O H+NO₂ (3.14)
is balanced. Cox [1999] investigated the influence of NO_x in the clean marine boundary layer in a box model with simple chemistry. Figure 3.3.6 shows the mean hourly rate of change of ozone concentration during the period 09:00-14:00 hours as a function of local NO concentration calculated for Cape Grim. Net loss of ozone at [NO]<15 ppt changing over to net production above this level. On the basis of their observations of O₃ and NO₃ at Cape Grim, Galbally et al. [19..] estimate a compensation point of 20�5 ppt. NO. Similar model calculations and observations at Mace Head give a higher value of [NO] at the compensation point of 30-55 ppt [Cox, 1999; Carpenter et al., 1997], which is consistent with the control being primarily due to a competition between the above reactions, taking account of the higher climatological ozone concentrations at the northern hemisphere site.

Figure 3.3.6 Net ozone production as a function of [NO], in clean oceanic air at Cape Grim, Tasmania, (Cox, 1998)

Davis et al., [1996] have conducted a comprehensive box modelling study of ozone photochemistry in the north west Pacific region, making use of aircraft measurements from the PEM-West A mission. They calculated the ozone photochemical tendency term, P(O₃), i.e. the difference between the total gas phase production and loss terms for ozone) using measurements as input to the model. P(O₃) values were generally negative below ~6 km altitude and positive above 8 km throughout the study region, although significant positive values were found in fresh boundary layer air of continental origin. Davis et al., [1996] also estimated the critical nitric oxide concentration NO crit for which P(O₃) was zero (equivalent to the compensation point), based on median values of all model input parameters in 13 altitude/latitude bins. In the lowest 2 km, NO crit was between 9 and 17 ppt, while O₃ was in the range 10-30 ppb. Above 1 km, NO crit tended to decrease with increasing altitude. These results indicate somewhat lower compensation points than in the boundary layer, where physical loss processes have an influence. Overall these studies confirm the basic theory for NO- mediated ozone production in clean tropospheric air Observations of free radicals have also been used to calculate ozone budgets in the troposphere. Ozone production in the upper troposphere has been studied by Jaegle et al, [1999]. Simultaneous observations of NO, HO and other species were obtained as part of the SONEX campaign in the upper troposphere over the North Atlantic (40-60N). These were used to derive ozone production rates, P(O), and to examine the relationship between P(O₃) and the concentrations of NO_x (= NO + NO₂) and HO_x (= OH + peroxy) radicals. A positive correlation is found between P(O₃) and NO but after filtering out transport effects, P(O) is nearly independent of NO for NO>70 pptv, showing the approach of NO-saturated conditions (Figure 3.3.7).

Fig. 3.3.7. OH, HO₂ and P(O₃) from Jaegle et al., 1999.

3.2.7. The influence of water vapour on tropospheric free radical chemistry

The crucial role of water vapour in determining ozone loss is revealed in the airborne measurements of peroxides in the troposphere, at least up to altitudes of 8 km. The concentration of hydrogen peroxide in a NO_x-free and CO dominated troposphere is given by the expression derived in Penkett et al. [1998] which is dominated in most situations by the water vapour term.

In atmospheric vertical profiles where measured ozone and peroxide are anti-correlated due to the preponderance of ozone loss processes, as seen at the surface in Figure 3.3.5, the actual concentration of peroxide can be calculated with considerable accuracy from the surface up to 8 km using expression (C). Figure 3.3.8 shows a high degree of correlation between calculated and measured peroxide above 1 km with small differences being caused by the flow response of the peroxide instrument during rapid ascent of the aircraft. This suggests that for the lower and mid troposphere at least, the free radical chemistry involving hydroxyl and peroxy radicals is simply determined by light levels, by ozone, but particularly by water vapour which is totally correlated with the peroxide in the profile and which changes by almost two orders of magnitude between the bottom and the top [Penkett et al. 1995; 1998].

Figure 3.3.8. A comparison of measured and calculated peroxide between the surface and 7.5 km. The measurements were collected in the OCTA campaign in September 1994, which was a component of IGAC/NARE.

Figure 3.3.8 also shows that below 1 km there is a large divergence between calculated and measured peroxide. This occurs in a portion of the profile when the measured peroxide and measured ozone are positively correlated, and it is caused by a large proportion of the peroxy radicals, which are formed photochemically by reacting with nitric oxide leading to ozone production rather than with themselves. The chemistry described above leading to both ozone production and ozone destruction is summarised in the two parts of Figure 3.3.9 which is taken from [WMO 1994]. The top part of the figure shows the cycling of NO_x driven by the HO_x free radical chemistry either with the production of nitrate sinks (HNO₃) and reservoirs (organic nitrates of various types). The bottom part of the figure shows the chemistry which produces the HO_x radicals and the peroxides whilst at the same time destroying ozone. The switch from one to the other is determined almost entirely by the levels of NO_x, which in the absence of a very limited database, makes a quantitative prediction of the distribution of ozone in the troposphere very difficult. Water has a direct influence on the production of OH, HO₂, and H₂O₂ from ozone destruction; it also plays a role in the switch-over in the chemistry shown in the lower part of the figure to that shown at the top [Levy Ref].

3.2.8. Reactive nitrogen chemistry

Figure 3.3.9 shows that in the course of generating ozone in a chain reaction, many different oxidised nitrogen compounds are formed from reactions with NO and NO₂. These are commonly denoted NO_x, (NO+NO₂), NO_y and NO_z where NO_z=NO_y-NO_x. NO_y is really the sum of all forms of oxidised nitrogen compounds in the atmosphere with the notable exception of N₂O. The oxidised nitrogen in these compounds is highly inter-related and involves reaction of individual nitrogen oxides, NO and NO₂, either with inorganic or organic free radicals. Thus nitric oxide (NO) reacts with ozone to produce nitrogen dioxide (NO₂) which can then undergo a variety of processes, including photolysis back to NO, which produces ozone. Some of the NO₂ is converted to PAN (peroxy acetyl nitrate).

Figs. 3.3.10 Global concentration fields of OH, in July and January for the whole earth surface layers

Figs. 3.3.11 Global concentration fields of HO₂ , in July and January for the whole earth surface layers

CH₃COO₂ (from acetyaldehyde reaction) + NO₂ -> CH₃CONO₂ (PAN) (3.20)
Other PAN-like molecules are also formed from high analogues of CH₃COO₂, but PAN greatly predominates. PAN is in fact in a thermal equilibrium with its precursors such that its lifetime varies with temperature. This tendency for PAN to be stable at low temperatures typical of the mid to upper free troposphere, and labile at temperatures found in the boundary means that it can act as a carrier of NO_x out of polluted regions to more remote regions. It is certainly widespread in the atmosphere as is shown by data in a recent publication by Singh et al. [19..] where the concentration of PAN often exceeds that of any other form of reactive nitrogen including NO, NO₂, HNO₃ and organic nitrates (RONO₂). PAN is in equilibrium with its components and it is destroyed by reaction of the peroxy acyl radical with NO.
CH₃COO₂ + NO -> CH₃ + CO₂ + NO₂ (3.21)
Organic nitrates are formed in the atmosphere by reaction of alkyl peroxy radicals with nitric oxide:
RO₂ + NO -> RONO₂ (3.22)
This reaction is in competition with the reaction which converts NO to NO₂ and a higher percentage of RONO₂ is made the larger the size of R. The short chain alkyl nitrates are also emitted directly into the atmosphere from the ocean, particularly methyl nitrate (CH₃ONO₂). It has therefore been suggested that the organic nitrates could be a significant source of the small amount of NO_x formed in the remote marine boundary layer [Atlas et al., 19..], but there are also other possible sources including direct emission from the ocean in an inorganic form, or transfer of NO_x from the free troposphere. This is certainly a possibility as indicated by NO_x measurements in the southern hemisphere free troposphere summarised recently [Singh et al., 19..].
Formation of nitric acid is a sink both for OH and NO₂ since its lifetime is relatively long and it is very soluble. This means that it is rapidly lost in the boundary layer by deposition to the surface, and it can be taken up very efficiently in clouds. Uptake in clouds does not in itself constitute a final loss process since the fate of most cloud droplets (90%) is to evaporate rather than precipitate. Some HNO₃ will therefore be converted back to its component parts (OH and NO₂) by photolysis in the upper troposphere before it can be rained out.

3.2.9. Progress in modelling the global budget of OH

OH concentration responds almost instantaneously to variations in sunlight and local trace gas composition changes and the OH field varies by orders of magnitude in space and time. Observations can be used to test photochemical theory in specific circumstances but are not capable of providing globally averaged values of [OH]. We must therefore rely on numerical models and surrogates to provide the global distributions of OH which are required to calculate, for example, trace gas lifetimes and ozone production and loss rates. The development of such models has advanced significantly in the past decade. An example of the output of such models, calculated global concentration fields of OH and HO₂ in July and January for the whole earth surface layers, are given in Figures. 3.3.10 and 3.3.11. Broad features of seasonal and latitudinal variation are superimposed by smaller scale variations caused by multitude of factors such as UV penetration, trace gas variations, which affect HO_x radicals. Discussion of these model developments is beyond the scope of this synthesis and we concentrate here on recent progress in determination of globally averaged [OH] based on new observations of methyl chloroform and ¹⁴C carbon monoxide, obtained in IGAC, and using more sophisticated models than those used earlier.

3.2.9.1 CH₃CCl₃

Methyl chloroform is destroyed primarily by through its reaction with OH in the troposphere. In the past years improved modelling techniques have been used to determine the global average OH from the increasingly comprehensive data on the burden and distribution of this gas in the atmosphere. Most calculations use an inverse technique in which emissions and the atmospheric burden are reconciled with an appropriate loss rate, calculated from OH concentration in a seasonally and spatially varying field. Spivakovsky et al., [1990] used a 3-D chemical transport model to generate global fields of [OH]. The computed distribution of OH implied a lifetime of 5.5 yr for methyl chloroform. This agrees with the empirical methyl 30 chloroform lifetime of 5.4�0.6 yr from an optimal fit to the observed concentrations of methyl chloroform, obtained from five ALE-GAGE surface sites over the period 1978-1990 (as reported by Kaye et al., in their 1994 assessment). The error arises mainly from calibration uncertainty in the measurements. Recent work by Prinn et al., (1995) used three different 'inverse' methods. The 'trend' method, which matches the time dependent concentration and emissions, is independent of the absolute calibration of CH₃CCl₃ concentration. The other methods based on total burden and spatial distribution are sensitive to absolute calibration, which has been revised in recent years. The latest estimates of [OH] are consistent in all three methods, yielding a tropospheric average [OH] of (9.7�0.6) x 10⁵ molecule cm^-3 and a global lifetime of CH₃CCl₃ of 4.8�0.3 yr, taking into account small ocean and stratospheric sinks. This currently accepted lifetime is lower than in previous assessments owing to revised calibration standards for the ALE-GAGE programme. Lifetimes of other trace gases have been accordingly revised downward.

3.2.9.2 ¹⁴CO

Atmospheric carbon monoxide exists in several isotopic forms, one of which, ¹⁴CO, has been useful in understanding tropospheric OH, which is the main sink for CO. ¹⁴C carbon monoxide is produced in the atmosphere from cosmic rays, mainly in the stratosphere and upper troposphere, and with a reasonably well known source strength. A substantial amount of new data for the global distribution of ¹⁴CO has been obtained in recent years using a new technique based on accelerator mass spectrometry [Mak et al., 1992; Brenninkmeier et al., 1992]. Comparison of the observed ¹⁴CO distributions with those obtained using a photochemical-transport model calculations [Mak et al., 1994; Derwent, 1994] shows a good general level of agreement, with an implied global [OH] of ~ 9 x 10⁵ molecule cm³ . However detailed differences of up to 30% are apparent in the seasonal and latitudinal distributions, which reduce the accuracy in the global [OH] to ~�30%. A larger mismatch is seen between model fields and observations for ¹²CO.

3.2.9.3 OH trends

The question of temporal trends of OH has been discusses by several workers. This would be an important diagnostic for model performance, particularly of interest for prediction of the oxidising efficiency of future atmospheres. Trends reported by Prinn et al (19 ) based on inverse modelling of CH₂CCl₃ data are not now considered significant in view of the uncertainty in calibration of the halocarbon data. The issue of trends in the contemporary atmosphere remains unresolved, although changes in tropospheric ozone since pre-industrial times imply an increase in OH production.

3.2.10. Novel free radical oxidation chemistry

3.2.10.1 The Nitrate Radical
NO₃ is formed in the atmosphere by the reaction:

NO₂ + O₃ -> NO₃ + O₂ (3.23)
In daylight it is removed instantly by photolysis but at nighttime it is removed from the atmosphere either directly (eg by reaction with DMS and unsaturated organic molecules such as terpenes), or indirectly through its equilibrium with di-nitrogen pentoxide (N₂O₅), which reacts directly with H₂O on aqueous aerosol surfaces to form HNO₃.
N₂O₅ + H₂O(surface) -> 2HNO₃ (3.24)
The reactions of NO₃ with organic molecules generate peroxy radicals (RO₂) as secondary products, and a correlation between NO₃ and RO₂ has been observed at night [Carslaw et al., 1997b]. Box models of NO₃ chemistry in the marine boundary layer, when constrained by a comprehensive suite of measurements of all these loss processes, are able to reproduce the observed NO₃ levels satisfactorily (Allan et al., 1999). Such models show that night- time NO₃ chemistry provides a route for converting to NO_x to HNO₃ that is a sizeable fraction (perhaps 50% or more) of the daytime route via OH + NO₂.
Atmospheric measurements of the nitrate (NO₃) radical have mostly been made by the technique of differential optical absorption spectroscopy (DOAS), using the strong absorption band at 662 nm. The exception to this is a data set obtained using Matrix Isolation Electron Spin Resonance (MIESR) [Mihelic et al.,1993]. DOAS instruments operating with an artificial light source in the lower boundary layer over optical path lengths up to 10 km have now achieved detection limits of 1 ppt. DOAS measurements using either the moon or scattered pre-dawn sunlight as a light source have also been employed to measure NO₃ columns in the free troposphere [Weaver et al., 1996; Aliwell and Jones, 1998]. There has also been one direct measurement of free tropospheric NO₃ at a high altitude site, Iza�a de Tenerife [Carslaw et al., 1997a]. Table 3.3.3 shows a summary of observations of nitrate radical over the last 20 years.
Although there have been a few published reports on measurements of continental NO₃ during the past 10 years [Platt and Heintz, 1994; Smith et al., 1995], most field campaigns have focused on the behaviour of the radical in the marine boundary layer. These have been driven by the close coupling of the NO_x and sulphur cycles through the rapid reaction between NO₃ and dimethyl sulphide (DMS) [Yvon et al., 1996; Carslaw et al., 1997b; Heintz et al., 1996; Allan et al., 1999]. All these studies, as well as more recent measurements in the remote marine boundary layer (Tenerife and Mace Head, Ireland), have shown that NO₃ is present at up to approx. 5 ppt even under low NO_x conditions ([NO_x] < 150 ppt), when the turnover lifetime of the radical is well over an hour. By contrast, the lifetime can be less than a minute in semi- polluted air where reactive organics and aerosol loadings are higher. It should be noted that when the NO₃ concentration is larger than about 1 ppt, DMS is oxidised more rapidly at night than during the day by OH. This seems to be the case even in the quite remote marine boundary layer [Allan et al., 1999].

Table 3.3.3. Summary of observations of nitrate radical.

Location	Measurement Dates	Technique	[NO₃] / ppt	Reference
Loophead, Ireland (BL)	April 1979	DOAS	< 3	Platt and Perner, [1980]
Colorado Mts., USA (BL)	August September 1979	DOAS	<3 - 65/TD>	Noxon et al., [1980]
L.A. Basin, USA (BL)	August September 1979\	DOAS	<6 - 315	Platt and Perner, [1980]
Deuselbach, Germany (BL)	April 1980	DOAS	<6 - 280	Platt et al., [1981]
J�lich, Germany (BL)	May August 1980	DOAS	< 8 - 69	Platt et al., [1981]
Mauna Loa, Hawaii (BL+FT)	November 1981	DOAS	0.3	Noxon, [1983]
Whitewater, USA (BL)	October December 1981	DOAS	<2 - 164	Platt et al., [1984]
Phelan, CA., USA (BL)	April 1982	DOAS	4 47	Platt et al., [1984]
Death Valley, CA., USA (BL)	April May 1982	DOAS	8 - 40	Platt et al., [1984]
Edwards AFB, CA., USA (BL)	May 1982	DOAS	17 - 88	>Platt et al., [1984]
Brittany, France (BL)	July 1988; June 1989	DOAS	<0.3 - 29	Barnes et al., [1991]
Biscayne Bay, Miami, USA (BL)	May November 1989	DOAS	<1 - 20	Plane and Nien, [1991]; Yvon et al., [1996]
San Joaquin Valley, CA., USA (BL)	July August 1990	DOAS	<1 - 80	Smith et al., [1995]
Schauinsland, Germany (BL)	August 1990	MIESR	<3 - 9.5	Mihelcic et al., [1993]
R�gen, Germany (BL)	April 1993 June 1994	=DOAS	0.5 - 90	Heintz et al., [1996]
Weybourne, Norfolk, UK (BL)	April 1994; October November 1994; June July 1995	DOAS	<1 - 60	Carslaw et al., [1997a]; Allan et al, [1999]
Iza�a, Tenerife (FT)	May 1994	DOAS	<5 - 20	Carslaw et al., [1997b]
Mace Head, Ireland (BL)	July August 1996; April May 1997	DOAS	1 - 40	Results not yet published
Taganana Bay, Tenerife (BL)	June July 1997	DOAS	1 - 20	Results not yet published

BL = Boundary Layer
FT = Free Troposphere

3.2.11. Halogen chemistry in the Troposphere

Observational data on tropospheric ozone over the past 40 years shows that there are a number of anomalies in tropospheric ozone distribution which are not easily explained by current theory. One example is the episodic and rapid depletion of ozone in the surface atmosphere in polar regions, particularly in the boreal and austral spring [Bottenheim, et al, 1990]. Another example is the observation of almost complete absence of ozone in the troposphere in the tropical oceanic regions of the eastern Pacific Ocean [Kley et al., 1996]. Recent observations from Mace Head, Eire, indicate a seasonal correlation of low ozone with phytoplankton activity [Carpenter et al., 1999].
Of the known species which could destroy ozone catalytically in the troposphere, the halogens and particularly bromine are potentially the most efficient. It is now fairly clear that rapid loss of ozone in the Arctic boundary layer is due to catalytic cycles involving BrO radicals [Barrie et al., 1988; Hausamann et al., 1996].

BrO + BrO -> 2Br + O₂ (3.25)
Br + O₃ -> BrO + O₂ (3.26)
--------------------
net 2O₃ -> 3O₂
Catalytic destruction of ozone by reactions involving IO radicals has also been postulated to account for low ozone in the marine boundary layer [Davis et al., 1997].
IO + HO₂ -> HOI + O₂ (3.28)
HOI + hn -> I + OH (3.29)
OH + O₃ -> HO₂ + O₂ (3.30)
I + O₃ -> IO + O₂ (3.31)
______________________
net 2O₃ -> 3O₂ (3.27)
Although there is no direct evidence, a recent model study [Vogt et al, 1999] based on observed IO concentrations indicates that iodine could have removed up to 12% of the ozone per day, making it more important that dry deposition or odd hydrogen photochemistry. Recent work shows that OIO may be important in iodine chemistry [Bloss et al., 2000].

3.2.11.1 Release of halogens into the troposphere
Halogens are released into the troposphere from the photochemical breakdown of organic halides like CH₂I₂, CH₂ICl, CH₃I and CHBr₃ [Schall and Heumann, 1993], which is probably the major source of reactive iodine species in the marine boundary layer. Modulation of tropospheric ozone and oxidising capacity can thus be achieved by the release of biogenic source gases Cl and Br can also be released from the oxidation of Cl^- and Br^- in sea salt. The processes leading to loss of halide from sea salt have only recently been diagnosed, but it is clear that this source of reactive Br and Cl is likely to be more significant than thought hitherto. Currently there are two theories:
1) Auto catalytic release of Br₂ (and probably BrCl and IBr) from sea-salt bromide [Vogt, et al., 1994] the 'bromine explosion' mechanism);
2) Formation of BrNO₂ and ClNO₂ by the reaction of gas-phase N₂O₅ with sea-salt halides [Vogt and Finlayson-Pitts, 1994), which could be important in polluted regions.
However, to date there is rather circumstantial evidence for these mechanisms and there appear to be - yet unknown - sources for Cl. The role of iodine in the release of Cl (and Br) requires study.

3.2.11.2 Reactive halogen in the troposphere
Free tropospheric BrO has been observed both in the Arctic as well as at mid-latitudes [Hebestreit, et al., 1999] by the technique of Differential Optical Absorption Spectroscopy (DOAS). Satellite observations using the same technique indicate that BrO is rather ubiquitous in the free troposphere. IO is the only halogen oxide so far observed in the low to mid-latitude marine boundary layer. IO was measured for the first time in 1997, by the DOAS technique [Alicke et al., 1999; Allan et al., 1999]. Figure 3.3.12 illustrates measurements off the north coast of Tenerife, where IO displays a clear diurnal cycle. It has now been observed in both hemispheres [Allan et al., 1999]. . However the presently available data are much too sparse to allow firm conclusions about the geographical and temporal distribution of reactive halogen species in the troposphere, which is needed for assessment of their impact on tropospheric chemistry.

Figure 3.3.12. Observed IO concentration profiles obtained at Tenerife during the period of 04 - 17 July 1997. The IO concentrations (broken symbols, courtesy of B.J. Allan, UEA) are plotted with the measured jNO2 values (solid dark lines, courtesy of G. McFadyen, ITE). The average error for the IO concentrations is 0.2 ppt (2s) for this period with an average detection limit of 0.2 ppt (2s).

Halogen atoms especially Cl, are efficient oxidising species in their own right and can lead to the degradation of volatile organics. There is definite evidence for oxidation by Cl in the observed concentrations of volatile organics (VOC) in Arctic regions, and also in CH₄ isotopes [Ariya,et al, 1998; Keene,et al, 1990], and there are observations indicating a significant role for this process in the oceanic and coastal environment [Volpe et al., 1998]. However quantitative knowledge of the concentration of halogen atoms in the relevant air masses is needed.

3.2.12. Progress in Laboratory Kinetics

The photochemistry and reaction kinetics and mechanisms needed to describe the chemical processes controlling OH and other radical concentrations, and the associated production and loss of ozone in the troposphere relies ultimately on accurate data measured in the laboratory. The original theories drew on relevant data available in the past, which contained significant uncertainties and gaps. Although there were no IGAC projects specifically focussed on the provision of improved data for tropospheric free radical and O₃ chemistry, a large amount of experimental effort has been devoted to this aspect over the past decade. In Europe studies co-ordinated within the EUREKA EUROTRAC project (LACTOZ report) have led to a major extension of knowledge in several specific areas such as: improved rate coefficients for elementary reactions of peroxy radicals involved in degradation of VOCs; extended knowledge of the kinetics and mechanism of the formation and removal of organic nitrates and oxygenates in VOC degradation; these are important respectively in determining NO_x distribution and secondary HO_x production in the troposphere; this is directly relevant to the production of O₃ through coupling with the NO_x photochemistry. Kinetics and mechanisms of reactions NO₃ with organic compounds including DMS, which is central to nighttime oxidation chemistry. Similar work has been undertaken in US laboratories leading to a consolidated data base for VOC oxidation mechanisms for surface conditions.
Very important new work on the detailed photodissociation pathways of ozone in the UV region has been carried out in laboratories in the UK, Japan and the US. This has revealed that the production of electronically excited O(¹D) atoms occurs over a wavelength region much extended beyond the cut- off at ~ 310 nm which was previously regarded as the limit for this process. This has a significant effect on OH radical production through the O(¹D) H₂O reaction, particularly at high latitudes and in winter months at mid-latitudes.
At the start of IGAC there was virtually no reliable quantitative data concerning heterogeneous and multiphase reactions in the troposphere, although work on acidification of precipitation had long pointed to important processes occurring in cloud droplets. A major effort to understand liquid phase oxidation process for sulphur, nitrogen and organic species has been undertaken and the knowledge base in this area is much improved, for example for the oxidation of SIV to SVI in marine and terrestrial cloud systems. Less attention has been directed to the kinetics of reactions on aerosol particles, although definitive data has been obtained for some heterogeneous reactions, for example the hydrolysis of N₂O₅, which is important for the NO_x budget under many atmospheric regimes.
The transfer of the knowledge base from the laboratory experiments to atmospheric modellers who attempt to simulate the atmospheric chemical composition to interpret data and predict changes relies on reviews, assessments and evaluations of the laboratory data. These provide recommendations for the optimum kinetic, thermodynamic and photochemical parameters to be used in models. The NASA Panel for Data Evaluation and the IUPAC Subcommittee for Gas Kinetic Data Evaluation have produced regular updated evaluations for stratospheric chemistry and for a limited number of tropospheric reactions.

3.3. Synthesis - What have we learnt?

The basic model for fast photochemistry in the troposphere has been validated by measurements in clean air at the surface. OH is present in sunlight and its concentration depends linearly on its rate of production from ozone photolysis, j_(O¹D), indicating a controlling role of primary versus secondary production. Recent airborne measurements of OH indicate the upper troposphere have much more OH than expected from O(¹D) production; acetone and other oxygenates seem to be responsible.
The coupled nature of HO_x chemistry is demonstrated; peroxy radicals, including HO₂ are present in daytime and, at lower concentration, at night, and are associated with OH. In general though, closure is not demonstrated. Even in very clean air, with a detailed model, 20-25% systematic differences appear for OH. In most cases modelled OH is a factor of 1.5 too high on the ground.
Complexities are noted in terrestrial situations due to VOC; the ratio OH/HO₂ is rarely what we expect and there are indications of additional sinks for radicals. Something is learnt each time we compare model versus experiment. Even for constant sources and sinks, OH is not smooth, but fluctuates, perhaps due to effects of mixing, turbulence of the atmosphere, this may place a limit on the level of agreement between model and experiment for short averaging times. There has now been experimental verification of theoretical OH versus NO_x relationships. This confirms the importance of NO_x concentrations in determining the concentrations of OH and RO₂ and the coupling between them.
There is now general agreement in the broad features of the global distribution of HO_x radicals, as calculated in 3-dimensional models. The subtropical regions sustain the highest concentrations of OH with suppressed levels over tropical terrestrial regions. OH in the upper troposphere is also augmented due to additional photochemical sources from oxygenated organics.
There is also general closure on the global mean OH concentrations needed to estimate global lifetimes of long lived trace gases implicated in greenhouse gas forcing and stratospheric ozone depletion. A consensus global mean [OH] of ~ 9 x 10⁵ molecule cm^-3 has been reached. The global or regional significance of the role played by other radical species (Cl, NO₃) in trace gas oxidation has not yet been quantified.
The relationship between local ozone production and loss and the HO_x photochemical processes has been demonstrated in several representative 8 atmospheric domains. Ozone is photochemically destroyed throughout the lower atmosphere as a result of the O(¹D) + H₂O reaction, and under some local condition catalytic destruction by bromine and iodine chemistry has been demonstrated. The latter provides potential biological modulation of ozone and oxidising capacity in the troposphere since many of the halocarbon gases are of biological origin. The dependence of ozone production on NO has been demonstrated from field measurements and from modelling studies, covering large atmospheric domains. Thus the so-called photochemical smog process is now known to have global significance. Closure on the budgets for NO_x and ozone production is hampered by the inhomogeneity of the troposphere in terms of its constituent air masses. The solution to the problem of averaging the HO_x photochemistry in the diverse chemical-transport regimes remains a major challenge. However the have been great improvements in knowledge of globally average oxidation efficiency of the troposphere and life times of key species (CO, CH₄, halocarbons). Further work will be required to unravel the more complex nighttime chemistry of NO₃ and the oxidation processes in the marine boundary layer involving halogens.

4. Meteorological Processes affecting Ozone

Changes in surface tracer emissions contribute to global atmospheric change . To assess the magnitude of global change, the handover of emissions and their transformation products from the source to the local, mesoscale, continental and global scales has to be followed, and the effects of changed emissions have to be quantified as they propagate through the same sequence of scales. The handover from the small to the larger spatial scales is a function of location and weather situation, and with very large differences from region to region.
Ozone is handed over on all scales. It is formed in the atmospheric boundary layer and handed over to the larger scales, or the precursors are handed over to the larger scales and ozone is formed in the free troposphere. Dry deposition is the main removal mechanism for tropospheric ozone and takes place in the atmospheric boundary layer. Stratospheric-tropospheric exchange of ozone also occurs efficiently on the local scale in tropopause folds.
Time series of ozone concentrations can be separated into short-term, seasonal and long-term variations. Variations at different frequencies are caused by different processes. The synoptic-scale variability is short-term and can be attributed to weather fluctuations or changes in precursor emissions. Seasonal variations follow changes in the solar flux over the year, while the long-term variations are climate related or triggered by policy changes influencing the intensity of the precursor emissions [Rao et al., 1997]. On the 17 basis of statistical analysis of more than 10 years, ozone measurements at over 400 sites in the United States, Rao et al. [1997] concluded that baseline ozone (defined as the sum of the seasonal and long-term components) retains global information on the scale of more than 2 months in time and about 300 km in space. Short-term ozone is highly correlated in space, retaining 50% of short- term information at distances from 350-400 km. The correlation structure of short-term weather-related ozone permits the prediction of ozone concentrations up to distances of 600 km from a monitor. In the high ozone cases, defined as exceeding 120 ppbv, the information from a monitor extends only out to about 15 km, mainly because of the inhomogeneity in the NO_x levels in an urban area. Based on ozone data from the German measurement network, Tilmes and Zimmermann [1998] made an even stricter conclusion; they found that measurements at sites less than about four km apart, represent each other.
In situations with favourable meteorology and precursor emissions in the atmospheric boundary layer, ozone can be formed on local and regional scales at a net rate of 10 or more ppb/h during the day, indicating that ozone concentrations then can change significantly in the matter of a few hours. In the free troposphere the net formation or destruction rate of ozone is typically of the order of 0.1 ppb/h during the day, indicating a chemical lifetime of ozone there of the order of 50 days.

4.1. Linking the local and global scales

Both synoptic scale advection processes and convection play an important role in vertical exchange in the troposphere. Observations have shown that trace gases are efficiently redistributed in deep convective clouds [Dickerson et al., 1987, Pickering et al., 1988], and model studies have indicated that the vertical redistribution of chemical trace species, in rapid convective up- or downdraughts, is important for their chemical processing [Chatfield and Crutzen, 1984; Pickering et al., 1992].
In some regions vertical exchange from the boundary layer to the free troposphere may be particularly important. One such region for the linking of the local emissions with global change may be over the Mediterranean, where the density of emissions is high at the same time as the land mass-ocean differences and the topography in the region give rise to circulations with a significant potential for rapid surface-mid troposphere exchange. This has been studied in some detail by the combined use of aircraft measurements of trace species and mesoscale meteorological calculations.
Mill�n et al. [1996; 1997] summarize middle to upper troposphere injection mechanisms in the Mediterranean basin from available evidence and postulated processes, see Figure 3.4.1. In the lower to middle troposphere (2.5- 3 km altitude) polluted layers are produced by thermally driven land-sea processes and the corresponding return flows. This occurs along south or east sloping coastal mountains. During the day the layer system circulates over the coastal areas in the course of 2-3 days, and during the night these layers travel along the coast uncoupled from the surface until the next day. These processes are favoured along the east and south Spanish coasts, Italian coasts, south Turkish coasts, and the Lebanese and Israeli coasts, perhaps also over the northwestern African coasts.

Figure 3.4.1 Schematic summary of observed and postulated circulations for the Mediterranean basin in summer (Mill�n et al., 1997)

Over the Alps, along the Apennines and over the Spanish and Turkish Plateaux and perhaps over the Atlas mountains in Morocco direct injection into the middle troposphere (up to 5 km) is documented to take place from diurnal convective cycles over large and dry land areas and from orographic injection. Condensation can occur. The stratified layers move away with the synoptic flow at their height, in particular during the night. The advection speed can be quite high (10-20 m/s), and if the layers drift over the Mediterranean Ocean, large-scale subsidence is likely.
Mill�n et al. [1997] also propose as a working hypothesis (Figure 3.4.1) that injection could occur up to 10-14 km as a result of convective pumping with cumulus congestus developing in the late afternoon. Chemical transformation and removal due to the formation of precipitation will be important. The injected pollution is advected by the upper level flow. Evening storms develop regularly in sea breeze or upslope wind systems in mountain ranges like the Apennines in Italy. Over North Africa sea breeze circulation systems may flow toward the Intertropical Convergence Zone and be pumped into the upper troposphere.
These flow regimes are likely to be very important for the exchange rate of pollutants between the boundary layer and the free troposphere throughout the Mediterranean Basin. Their dimension is limited, however, and to assess their impact on the global scale models need to resolve the most important features of each event. The local scales up to the global scales are interlinked, and the handover processes need to be treated in a consistent way. This is an important conclusion from the Mediterranean Basin studies. Another important conclusion is that the mesoscale circulations and convective events in the Mediterranean Basin may transport ozone and its precursors to the free troposphere so efficiently that the contribution to global change is significant. Some evidence for this is found in the tropospheric ozone residuals derived by Fishman et al., [1990] from satellite data. The averaged tropospheric ozone residual for June-August indicate a significant enhancement of the residual ozone column over the Mediterranean Basin 5-10 Dobson Units above the adjacent regions (40 ppb over a layer of 1 km near the surface corresponds to 4 Dobson Units).

4.2. Chemical transport models, dependence on meteorological data

The quality of chemical transport model calculations depends on the following main groups of data:
1. Model formulation (grid resolution, numerical methods, parameterizations)
2. The physical data for winds, clouds incl convective clouds, and precipitation, as calculated by some kind of state-of-the-art mesoscale, synoptic scale or global scale weather prediction or global circulation models
3. The quality of the boundary conditions for the chemical tracers (both lateral and upper boundary), as calculated by a chemical transport model which has a larger domain
4. The quality of the emission data and the description of the chemical transformation and removal processes
The importance of error propagation from group (2) into chemical tracer calculations is to a large extent overlooked in the literature, because it is generally believed that modern numerical weather prediction models has advanced far and that the trust placed in the quality of the weather analysis and forecast is inherited when chemical tracer calculations are analysed. Also good observations for the validation of clouds and boundary layer mixing processes are not generally available, contributing to the weak emphasis on the validity of these parameters.
For chemical tracers water vapour, shear in wind speed (giving very different transport times from a source region to a receptor volume in the free troposphere), convective activity and cloud cover and type are very important parameters. Water vapor is a key factor in the determination of the hydroxyl radical concentration, which determines the lifetime of most atmospheric trace species. Cloud cover and type influenced photolysis rates significantly.
Also data group (3) propagate serious problems into the interpretation of chemical transport model calculations in that ill defined upwind boundary conditions can be a dominant component in the distribution of long lived species like CO and NO_y far into the interior of the model domain. The method used is not really well defined if it turns out that factors outside of the domain can be dominant in explaining the concentration distributions of chemical tracers in the model. For data group (4) the surface emissions are the driving force for the concentration of chemical tracers.
It is difficult to balance a model development appropriately. The sophistication of the various model elements should approximately reflect the quality of the underlying physical or chemical data (resolution in time and space for instance). Also the model should resolve the spatial and temporal distribution of the phenomenon to be analysed. This means that the model results should not depend too much on ill defined initial or boundary conditions.

4.3. Mesoscale modelling

The modelling of ozone requires observational basis and theoretical skills in treating physical and chemical processes on all spatial scales, starting on the scale of the city plumes and convective cells, and ending with the global scale.
In the United States ozone plumes from major conurbations have been observed with a lateral dimension of the order of 10 km [Imhoff et al., 1995; Chameides and Cowling, 1995]. The rate of formation of ozone is a nonlinear function of the precursor concentrations, therefore this observation implies that a model calculation of ozone needs to resolve the 10 km scale, or at least allow testing and parameterization to assess the errors when a coarser resolution is used. If the objective is to make a careful assessment of the exposure of the population or crops to ozone levels above certain thresholds, a grid resolution of 10 km or finer in the horizontal can be required.
Such models, often denoted meso-scale chemical transport models, have been developed for many locations both in the United States and in Europe, and they depend on meso-scale hydrostatic or non-hydrostatic meteorological models to quantify the physical fields, and on synoptic scale weather prediction and chemical transport models for boundary values and initial concentrations.
Mesoscale nonhydrostatic models have to be applied when the size of the topographic elements is typically one order of magnitude or less than the required horizontal grid resolution. To generate appropriate boundary conditions for a small grid domain with high grid resolution, the small domain can be nested once or several times in coarser models in succession.
The capability of mesoscale meteorological models to calculate the physical parameters which are essential for chemical transport models is generally believed to be good. However, in the German Tropospheric Research Programme [Schaller and Wenzel, 1999] extensive physical and chemical measurements are applied in model validation. Mixing height and specific humidity are two essential parameters in a chemical transport model calculation, but not so important in weather forecasting. In the German programme these mesoscale parameters show systematic and large biases with a too high mixing height in the model calculation, while the specific humidity is too low. When the mixing height is overestimated all precursors, notably CO and NO_x, will be underestimated, which is confirmed by the chemical measurements, while the underestimated specific humidity will cause the calculated hydroxyl concentration to be on the low side, causing the calculated chemical lifetimes of trace species like the hydrocarbons, CO and NO₂ to be too long compared to what a measurement would have shown.
These findings are very important because they are somewhat unexpected. Their implication is that it is hard to validate the chemical aspects of a mesoscale chemical transport model, because the physical parameters are so important for the result of the chemical calculations. It is required to constrain the error bounds on parameters like mixing height and cloud cover before it is possible to state with confidence that discrepancies between for instance measured and modelled NO_x. CO or nonmethane hydrocarbons are due to an inferior emission inventory or not properly understood chemical transformation or removal processes.

4.4. Improvements in numerical weather prediction

Chemical transport model calculations depend critically on the validity of the weather prediction system. Numerical weather prediction has improved over the last one or two decades, at the same time there are aspects of weather prediction which are particularly important for chemical tracer calculations and not so much for the weather forecast. This is for example the distribution and optical thickness of clouds, the mixing properties of the atmospheric boundary layer including predictions of the mixing height, and the distribution of water vapor. These physical properties determine in turn the rate of photodissociation that drive atmospheric photochemistry and the concentration profile in the atmospheric boundary layer and hence the dry deposition rate as well as the rate of exchange between the atmospheric boundary layer and the free troposphere.
The improvement in numerical weather prediction over the last decade is in parts due to better observations, fewer simplifying assumptions are invoked when solving the fundamental equations, numerical solution techniques have been advanced, model resolution has increased, physical parameterizations have been advanced and data assimilation improves model initialization.
Bengtsson [1999] discussed the importance of these factors, and concluded that useful predictive skills over most of the northern hemisphere extratropics has been extended to more than a week at present from about three days through the 1960s and 1970s. Improved observation systems was the main factor until the late 1970s, since then improved forecasting systems including model physics, dynamics and numerical methods have been the main factors. Data assimilation has been of fundamental importance in order to realize fully the model improvements. The combination of an accurate model and an advanced data-assimilation scheme makes it possible to reduce the error of the initial state significantly. Bengtsson [1999] further concluded on the basis of results from ECMWF that using today s models and data- assimilation systems, past forecasts are significantly improved even by only using observations available at the time. This finding justifies the ongoing reanalysis at major weather prediction centres of past observations by today's forecasting systems. In that way the full value of past global observations is exploited.
The forecast skill varies considerably from day to day, and according to Bengtsson [1999] the length of useful forecasts can vary by more than 6 days for 90% of the predictions. The remaining 10% have an even larger span. Ensemble prediction studies (ECMWF) suggest that the major part of this variability is related to the inaccuracy of the initial state and the way these inaccuracies are exposed to rapidly growing perturbations. The growth of errors is for example much faster in active baroclinic zones than in the inner regions of large anticyclones.
Chemical tracer calculations can be very sensitive to errors in advection and physical processes which affect the transformation and hence the lifetime of chemical trace species. Forecast errors can propagate into the chemical tracer calculation and in some cases amplify errors that arise due to for example poorly known emissions or chemical transformation mechanisms and rates.
In Figure 3.4.2 is shown the improvement in forecast skill in the northern hemisphere 20�N-90�N in the winter since 1980 [Bengtsson, 1999].

Figure 3.4.2 The improvement in forecast skill for the northern hemisphere winter (20 o -90 o N) since 1979. The root mean square (RMS) error in the calculated 1000 hPa height is shown as a function of forecast day for the operational ECMWF model in 1980, 1988 and 1998. The ECMWF models were global in extent and have undergone major improvements in physics, resolution and data assimilation. It is seen that a 2.5 days forecast from ECMWF in 1998 has the same RMS error as a 1 day operational forecast in 1980. The speed of improvement is decreasing, however, indicating that the predictability limit is being approached (Bengtsson, 1999).

4.5. Cloud cover prediction

The calculation of photolysis rates requires an accurate description of the distribution and optical thickness of clouds. When deep convective clouds occur, they affect the vertical transport of trace species in a dramatic way. Cloud sensors on satellites are providing datasets that are gradually been used for model validation. The International Satellite Cloud Climatology Project (ISCCP, a part of World Climate Research Programme WCRP) was established in 1983 to produce good global datasets on radiative properties of the atmosphere, datasets that are needed to derive cloud parameters, and to contribute to the parameterization of clouds in climate models [Schiffer and Rossow, 1983; Rossow and Schiffer, 1991]. An atmospheric model intercomparison project (AMIP) was established in 1989, also as a part of WCRP, with the purpose to carry out systematic and comprehensive intercomparison of atmospheric climate models [Gates, 1992].
In the IPCC 1995-report [Houghton et al., 1996] it is concluded that current global circulation models portray the large-scale latitudinal structure and seasonal change of the observed total cloud cover with only fair accuracy, and there is an apparent systematic underestimate of the cloudiness in low and middle latitudes in both winter and summer. In the higher latitudes the models appear to overestimate the observed cloudiness, especially over Antarctica. Here aggregated information is compared, where measurements and model results are averaged in time and space. It is then expected that the agreement is much better than when comparisons are made on an event basis.
In chemical transport models cloud cover and type of cloud are important parameters because of their direct influence on the photolysis rates and convective exchange. Lin et al. [1994] developed a parameterization of subgrid scale convective cloud transport in a mesoscale regional chemistry model for Eastern and Central North America, and showed that the model calculations of ozone and precursors with the cloud transport parameterizations were in much better agreement with aircraft observations than those without. Flat�y et al. [1995] showed how convective processes parameterized in a mesoscale chemical transport model over Europe significantly reduced the maximum daytime concentrations of ozone in the atmospheric boundary layer in anticyclonic weather over polluted continental areas.
Discrepancies in cloud cover and type of cloud propagate into the chemical transport model results and bias the calculated distribution of the chemical trace species in quite unpredictable ways. If a convective event over a region with high emissions was calculated to take place in a neighboring grid square where the emissions are much lower, a significant bias is introduced in the calculation of the formation of ozone in the atmospheric boundary layer.
Summer time boundary layer formation of ozone usually takes place in weather situations which are also conducive to the formation of thunderstorms. Convective events occur over continents where also anthropogenic emissions can be significant. Events that can lead to photooxidant episodes in the boundary layer may therefore coincide much more frequently with convective activity than the frequency of convective activity itself indicates. In convection over polluted continents, the boundary layer is often a net source to the free troposphere both of ozone and precursors. Convection therefore leads to fewer episodic peak concentrations of ozone near the surface, but the precursors vented into the free troposphere cause formation of ozone there, and in addition more ozone is formed in the free troposphere, for the same amount of precursors, than in the boundary layer. The ozone lifetime in the free troposphere is also much longer than in the atmospheric boundary layer.
In conclusion, when comparing the calculated distribution of clouds from numerical weather prediction models or in global circulation models, the agreement with measurements is fair at best. The discrepancies can cause significant errors in chemical transport model calculations. Often errors in calculated chemical tracer fields as evaluated by measurements, are thought to be due to erros in the emissions estimates or in the parameterization of chemical transformation or removal. It turns out, however, that errors in calculated chemistry can often be due to errors in the data for cloud cover, advection or mixing properties in the atmospheric boundary layer.

4.6. Boundary layer-free troposphere exchange rates over Europe and the North Atlantic

Precursor emissions of ozone are mainly removed in the atmospheric boundary layer and do not enter the free troposphere. The fraction of NO_x entering the free troposphere is essential for the formation of long lived ozone there. A 3-dimensional chemical transport model where a numerical weather prediction model was used to generate the meteorological data needed to run a comprehensive chemical dispersion model was described for Europe and the North Atlantic by Flat�y and Hov [1995]. Several such models exist (e.g. Atherton et al., 1996; Lin et al., 1998]. Some results will be shown from a calculation by Hov and Flat�y [1997] of the rate of exchange of species between the atmospheric boundary layer and the free troposphere in comparison with other processes, as indicated by the continuity equation which for each species can be expressed as:
d[c]/dt = emission + advection + diffusion - dry deposition + convection + chemical production chemical loss
Here d[c]/dt is the local rate of change in the concentration of compound c. The terms for advection, diffusion and convection on the right hand side can be both positive and negative. The difference (chemical production chemical loss) is the net rate of formation of the species �C.
The model domain covered Europe, the North Atlantic and parts of Russia. Results are shown here for two sub regions for a 10 day period in July 1991 when there was a weak anticyclonic circulation over north Europe, northwesterly flow off the continent over the British Isles and an easterly flow over central and southeast Europe. A maximum in the ozone concentration in the atmospheric boundary layer was calculated over northern Italy with a ridge of high ozone all the way to Iceland caused by European emissions. Results for two sub regions, one covering 10x15 grid cells in the Eastern Atlantic west of the British Isles, and one covering 5x5 grid cells over Germany, The Low Countries, The Alps and Northern Italy (one grid cell covers approximately 150x150 km² ) are shown in Table 3.4.1 for the atmospheric boundary layer defined as the three lowest layers in the model (confined below 865 hPa).

Table 3.4.1. Mean concentration of NO_x, NO_y (sum of NO_x and species generated when NO_x is oxidized, except that nitrate aerosol was not included in NO_y) and ozone for two sub regions in a 3d chemical transport model calculated for 10 days in July 1991 by Hov and Flat�y [1997]. The average of the tendency terms for chemistry (chem) in the atmospheric boundary layer (ABL), vertical advection (v.adv) and convection (conv) out of the ABL, average emissions, ratio between the rate of vertical transport of NO_x and NO_y out of the ABL and the rate of emissions into the ABL (v.adv+conv)/emis. The numbers in parentheses are for a similar calculation for October 1994 with westerly winds over Europe and frequent precipitation in North Europe and shorter residence times of the air masses than in July 1991.

Eastern North Atlantic Continental Europe

NO_x NO_y ozone NO_x NO_y ozone

[c], ppt
(dc/dt)chem, ppt/h
(dc/dt)dep, ppt/h
(dc/dt)v.adv, ppt/h
(dc/dt)conv, ppt/h
emission, ppt/h
(v.adv+conv)/emis

56
-2.6
small
-0.0
-0.0
0.7
0.09
(0.32)
474
small
-7.4
-2.0
0.1
0.7
2.9
(0.15)
31.4
-36
-29
-57
3.5

1391
-214
small
-5.3
-18
290
0.08
(0.07)
3373
small
-113.5
-23.4
-47
290
0.24
(0.11)
47.6
1075
-351
-194
-128

Table 3.4.1 shows the mean tendency terms over 10 days in July 1991 for chemistry, (dry+wet) deposition, vertical advection, convection and emissions of NO_x, NO_y and ozone. Over the European continent 8% of the NO_x emissions are brought to the free troposphere as NO_x, while 16% of the NO_x emissions are brought to the free troposphere as NO_y-NO_x, i.e. as PAN or HNO₃. A similar calculation was made for a part of October 1994 when Europe was exposed to westerly winds. For this period the numbers were not significantly different from the July period numbers
The convective exchange of ozone is not so important for the redistribution of ozone. It takes 44 h to generate the concentration of ozone chemically (47.6 ppb/1.075 ppb/h), while the lifetime due to dry deposition which is 136 h (47.6 ppb/0.351 ppb/h). Convection is a significant factor in reducing the maximum ozone concentration in episodes, however, and this is important from a policy point of view [Flat�y et al., 1995]. Vertical advection is on the average more important in the redistribution of ozone than convection. The term (d[c]/dt)conv for the ABL can be both negative and positive in a convective event, depending on the concentration of ozone in the ABL compared to the mid or upper troposphere, and when averaging over several days and large areas the individual terms tend to cancel.
In Table 3.4.2 is shown the average ratio (d[O₃]/dt)chem/(d[NO_x]/dt)_{emis+v.adv+h.adv+conv} over Continental Europe for the July 1991 calculation as a function of height. This ratio is one way of expressing the ozone producing efficiency of NO_x, an alternative ratio would be to put the transformation rate of NO_x to NO_y-NO_x in the denominator. The ratio used here expresses the net rate of chemical formation of ozone per NO_x molecule delivered by a physical process (emission, advection or convection) to the volume considered. This ratio in general increases with height and is typically 15-20 in the upper half of the troposphere over Continental Europe and around 5 in the atmospheric boundary layer (ABL). As an average for the total domain there is a net chemical removal of ozone at some altitudes where the concentration of NO_x is so low that the flux through chemical loss reactions of ozone (notably the O₃+HO₂ and O(¹D)+H₂O reactions) is larger than the flux through ozone production reactions like NO+HO₂ _ NO₂+OH followed by NO₂ photolysis. The ozone producing efficiency increases with altitude because the lifetime of NO_x is longer in the free troposphere. The conversion to HNO₃ and removal of NO_y in precipitation or by dry deposition are slower there than in the ABL.
Liu et al. [1987] estimated that around 10 molecules of ozone is formed per NO_x molecule emitted in polluted air masses and around 100 for very clean conditions. See also discussion in Hov [1997].

Table 3.4.2. Ozone forming efficiency as a function of the local source strength of NO_x given as the ratio between the local net chemical formation rate of ozone and the local net change with time of the local concentration of NO_x due to emissions, advection and convection for eight pressure heights for the sub region of 5x5 grid cells covering Continental Europe (d[O₃]/ dt)_chem/(dNO_x]/dt)_{emis+v.adv+h.adv+conv}

Pressure height (hPa)	Continental Europe
209 319 428 537 646 756 865 960	21.7 15.0 15.7 11.2 6.4 9.8 5.4 3.6

4.7. Regional, intercontinental, hemispheric and interhemispheric chemistry and transport

There is a strong hemispheric component in the tropospheric distribution of ozone. This means that emissions over one continent cause ozone formation which appears as background ozone over other continents or as ozone in the free troposphere. Several measurement campaigns in the Northern Hemisphere have provided data on how anthropogenic and biogenic emissions over distant upwind continents have affected the chemical composition of the troposphere over the Pacific and the Atlantic. A summary of the major findings of a number of campaigns is given in the 1998 WMO assessment of ozone [WMO, 1999, Table 8-4 therein]. The field experiments have been accompanied by a number of model developments set up to calculate the regional and global budgets of tropospheric ozone. These models are usually driven by meteorological data generated by a global circulation model, which means that the weather data used do not correspond to the weather situation for a particular date. The grid resolution is coarse which means that the description of the processes for the handover of precursors and secondary species from the sources to the global troposphere is heavily parameterized.
The effect of rising Asian emissions on surface ozone in the United States was investigated by Jacob et al. [1999]. Using the Harvard-GISS global 3-d model of tropospheric chemistry they found that with zero emissions of NO_x over North America, and emissions elsewhere at present-day levels, long-range transport from outside of the North American continent would maintain 20-40 ppb O₃ in surface air over the United States in the summer [see also Liang et al., 1998]. They found that a three-fold increase of the Asian emissions of oxides of nitrogen and hydrocarbons could increase the monthly mean ozone concentration by 2-6 ppb in the western US, and by 1-3 ppb in the eastern US, the maximum effect being in April-June.
Increasing emissions over the continents augment the likelihood of pollution episodes in the atmospheric boundary layer, and may also cause an increase in surface ozone concentrations over other continents. From a global change point of view, however, the free tropospheric impact of increasing emissions is of particular importance since as already discussed, the ozone producing efficiency per molecule of NO_x is higher in the free troposphere than in the boundary layer. In addition ozone in the free troposphere has a longer lifetime than closer to the ground, and its radiative forcing is higher per unit change than nearer to the surface. Using a global 3-d chemical transport model driven by meteorological data from the GISS model, Berntsen et al. [1996] calculated that a doubling of the NO_x emissions over Asia gives ozone increases of up to 30% in the upper troposphere and with a resulting radiative forcing change of up to 0.5 W/m² which is 30-50% of the negative radiative forcing due to sulphate aerosols in the region.
At the UK Meteorological Office a global three-dimensional Lagrangian tropospheric chemistry model has been developed where 50000 air parcels are followed. The meteorological data including the wind field is calculated in the UK Met Office UM global circulation model, and to make long calculations practical the meteorological data are smoothed as 18 days means in the chemical transport calculations. In Collins et al. [1997] is reported a global calculation of tropospheric ozone, and the chemical production and loss processes are quantified and aggregated on a global basis. For the month of August it was calculated a net photochemical production rate of ozone of 5.7x10³⁴ molecules/day, compared with a stratospheric input at 100 hPa of 8x10³³ molecules/day, or less than 20% of the in situ net production. If the European NO_x emissions were reduced by 50% it was found that ozone over Europe in summer dropped by 10-20% while in the winter ozone went up by 40% in the same area. This illustrates how NO_x reduces surface near ozone in the winter over polluted continents, while in the summer decreasing NO_x in general reduces ozone over the continents.
The transport of pollution from North America across the Atlantic into Europe and Eurasia has been the subject of intensive field campaigns and theoretical studies, North Atlantic Regional Experiment (NARE) in IGAC being one of them. Trajectories were used extensively to plan the field measurements and for the interpretation afterwards [Draxler, 1996, Wild et al., 1996]. Three-dimensional chemical transport model calculations were carried out by several groups in NARE. Atherton et al. [1996] based their chemical transport model calculations on meteorological data from the NCAR Community Climate Model Version 1, while Kasibhatla et al. [1996] carried out model calculations based on monthly averaged meteorological fields.
The model objectives determine its formulation. Coarse grid hemispheric or global models of the troposphere which are based on global circulation model data are suited for climatological investigations rather than event studies. Processes on horizontal scales shorter than about two grid elements are parameterized, which means that the dynamical influence of land-ocean interfaces, topography and convection is indirect and of a diagnostic character in the calculation. On the other hand, such models can be run for time periods of sufficient length to estimate the global impact of slowly changing parameters like emissions.
The CTM developed by Flat�y et al. [1995] on a circumpolar grid and using hourly meteorological data from a reanalysis with a numerical weather prediction model, was run for a two week period during the NARE intensive field campaign in August 1993 [Flat�y et al., 1996]. A CTM run in this way is most suitable for episodic studies because the resource requirements in terms of meteorological data, storage and computer power is substantial. One interesting aspect of these calculations was the budget established for the horizontal fluxes of trace species coming into and out of North America and Europe, integrated from the surface to the tropopause. As an example, translated into annual numbers assuming that the two weeks were representative for the whole year, about 15% of the European NO_x emissions were exported out of the troposphere over Europe as NO_y, i.e. as NO_x or oxidation products. This means that 15% of the emissions were left to react further or be removed in the troposphere outside of Europe, and this figure gives an impression of the spatial scale on which NO_x is transformed and removed.
Finally an example will be shown of a process study based on field measurements and photochemical steady state model calculations in the rural southeastern United States (during the 1990 ROSE campaign, Frost et al., [1998]). Based on the comprehensive set of aircraft based measurements and photochemical steady state model calculations, relationships were derived between the concentrations of NO_x and the sum of peroxy radicals XO₂ and of the net chemical formation rate of ozone P(O₃). In Figure 3.4.3 the results are shown, and the sum of peroxy radicals is seen to fall off with increasing concentrations of oxides of nitrogen, with a steeper fall off above 1 ppb than below. There is a maximum in the net formation rate of ozone for NO_x at approximately 5 ppb in the experimental data. In the theoretical calculations the fall off is found for about 3 ppb of NO_x when biogenic volatile organic species are neglected and about 7 ppb when the biogenic VOCs, which are quite important ozone precursors in the Southeastern US, are included.

Figure 3.4.3 The sum of peroxy radicals XO2 and the chemical formation rate of ozone P(O3) as a function of the concentration of NOx in the 1990 Rural oxidants in the southern United States environment (ROSE) programme, Frost et al., 1998. The dots, squares, solid and open triangles are measured or derived from measurements, while the solid line is a photochemical steady state model calculation using all volatile organic compounds and the dotted line is a similar calculation neglecting biogenic volatile organic compounds.

4.8. The capability to forecast ozone in the atmospheric boundary layer

Modelling of atmospheric chemistry has been advancing with the progress in numerical weather prediction. One of the objectives of this chapter has been to discuss how errors in weather prediction may propagate into a chemical transport model calculation. In particular parameters like mixing efficiency in the atmospheric boundary layer, cloud cover and water vapour concentrations may not attract so much attention in the weather forecast, while they are essential for a reliable chemical transport calculation. Also more measurement data are becoming available for CTM validation, in particular from instrumented aircraft in the Northern Hemisphere [see for example, Hov et al., 1998]. 3-D CTM forecasts are still in their infancy, not least because the initialization possibilities are still quite poor. Almost no chemical measurement data are available online to improve the determination of good initial conditions. Data assimilation has been a major break through for the skills in numerical weather prediction [Bengtsson, 1999], while the theoretical formulation has started in chemical transport calculations [Elbern and Schmidt, 1999]. Due to the lack of an appropriate infrastructure which can provide the necessary measurements in near true time, the implementation of routine chemical assimilation cannot be realized. Ozone sonde observations are slowly becoming available with only one hour or so time delay, and ozone is used as a proxy for potential vorticity in the upper troposphere and lower stratosphere and in that way used to improve weather prediction model assimilation (ECMWF activity). The infrastructure which is set up in that way may become a channel for other chemical measurements as well.
The increasing awareness of the negative impact of ozone on the population and for crops has created a demand for online information on expected ozone levels near the ground for the coming days. A legal infrastructure is being implemented in Europe which requires such information to be available to the public. An online Internet service of this kind has been available for Europe as well as the northern hemisphere north of approximately 30�N for some time, and an example of the interactive web interface is shown in Figure 3.4.4 [see also Hov et al., 1998, Flat�y et al., 2000].

Figure 3.4.4 Snapshot of an interactive web interface used for data serving used for experiment planning in aircraft field studies over Europe. The field shown is NO_x (ppb) at 270 hPa (� 10 km altitude) 21 September 1997 12 UT, from a simulation of the 3-d CTM coupled to a NWP-model (Flat�y et al., 2000)

5. A Climatology of Tropospheric Ozone

5.1. The Origin of Ozone in the Troposphere

Ozone (O₃) was first discovered in the troposphere in the 1840's and an intensive study was made of its distribution in the latter part of the nineteenth century because of its supposed health-giving properties as a germicide. The classical explanation of ozone in the troposphere was that it was made in the stratosphere by direct photolysis of oxygen at wavelengths less than 242 nm.
This was then followed by transfer from stratosphere to troposphere, ultimately to be destroyed by contact with the Earth's surface, particularly land surfaces. Whilst it became known that there was a direct source of ozone in the troposphere in the form of photochemical smog, first discovered in Los Angeles in the 1940's, it was not thought to be significant on a global scale. This classical view of the origin of ozone in the troposphere has undergone a complete about face during the course of the IGAC project, and it is now recognised that in situ chemical processes combined with vertical and horizontal mixing are the most important factors controlling its origin, distribution and fate (see earlier). This is a necessary conclusion to explain the observed trends over the twentieth century which have been well described in WMO [1999], and which suggest a doubling of the average ozone concentration in large parts of the troposphere [Marenco et al., 1994, Figure 3.5.1).

Figure 3.5.1.Ozone evolution in the free troposphere over western Europea during the XX� century

5.2. Stratosphere-Troposphere Exchange

Previous to the discovery in the 1970s that photochemistry is of global scale importance, it was widely assumed that stratosphere-troposphere exchange (STE) controls the distribution and abundance of tropospheric ozone [Danielsen, 1968]. This assumption was based on the observed decrease of O₃ with altitude at extra-tropical latitudes, being suggestive of downward O₃ transport across the tropopause, and removal by dry deposition on the surface. As indicated earlier, field campaigns and global modelling studies within the framework of IGAC have underscored the significance of photochemical processes. Nevertheless, STE provides an O₃ background , required to initiate OH formation. In the past decades many studies have addressed the dynamical processes that control STE [Tuck et al., 1985; Holton et al., 1995]. Field campaigns and mesoscale modelling studies have been performed to quantify tropopause folding events and cross-tropopause transports associated with cut-off lows [Davies and Schuepbach, 1994]. Recently, fully coupled dynamical-chemical models that simulate STE on a global scale have become available [Roelofs and Lelieveld, 1997].
Stratospheric ozone is transferred to the troposphere in the Brewer- Dobson circulation, being driven by wave disturbances that propagate upward from the troposphere [Haynes et al., 1991]. This wave forcing causes a mass flux that carries O₃-rich air from the tropical upper stratosphere downward and poleward, ultimately through the extra-tropical tropopause. The mass balance of the stratosphere is maintained by suction of tropospheric air across the tropical tropopause [Holton et al., 1995]. The wave forcing is most efficient in winter, since meridional temperature gradients are strongest, and thus wave disturbances are most frequent during this time of year. In addition, entrainment of relatively O₃-rich lower stratospheric air into the troposphere is strongest in spring, associated with increasing tropopause altitudes in the winter-summer transition [Appenzeller et al., 1996]. The combined effect of these processes is that STE at extra-tropical latitudes has a maximum during late winter and early spring. Since observations at background measurement stations show an O₃ maximum during this time of year, it might appear that O₃ transport from the stratosphere is a controlling factor. In fact, this is an erroneous inference that persists even in some of the more recent literature.

5.3. Simulation of Ozone in the Global Troposphere by Chemistry/Transport Models

The art of atmospheric modelling has advanced substantially in recent years, aided by enormous increases in computer power. It has led to significant improvement in weather forecasting, particularly over 3-5 day periods, and to the ability to predict chemical fields for a large range of species in the form of 3D synoptic charts. This is shown earlier in Section 3.... on transport processes.
Three-dimensional models are very useful for helping us visualise how the global ozone distribution should appear. They are not perfect however and are in an evolutionary phase, which has been aided recently by a major intercomparison exercise sponsored by the European IGAC Project Office (EIPO). The extent of their imperfections is perhaps emphasised by the range of values for the various terms in the integrated ozone budget, which were identified in previous discussions.
Table 3.5.1 shows estimates of total chemical production, total chemical destruction, deposition to the surface in both hemispheres, the amount of ozone transported from the stratosphere, and the ozone burden of the troposphere. The ranges in the individual terms are large; also the net chemical flux changes in sign from model to model with 7 models showing a net positive chemical term and 5 showing a net negative term. Whereas this is disappointing all the models show individual chemical terms which are much larger than the term for input from the stratosphere, or the physical removal term which together dominated the classical view of ozone in the troposphere.

Table 3.5.1. Ozone budget (Tg-O₃/yr) below 300 mb, except mentioned differently, as computed by the thirteen global three-dimensional models in the IGAC/3-D CTM O₃ intercomparison exercise

MODEL O3
budget IMAU-3 IMAGES HARVARD
(*) UKMETO ECHAM CTMK MATCH MOZART TM3 MOGUNTI UIO GFDL (**)

chemical
production
NH
SH
chemical
destruction
NH
SH
deposition
NH
SH
cross 300
mb flux
NH
SH
burden
NH
SH 2495

2821

724

1074

179
106
73 4763

3225
1538
4898

3125
1773
1253
857
396
1388

238
142
96 4100

2620
1480
3680

2290
1390
820
530
290
400

240
160
310
180
130 3890

2557
1333
3276

1958
1318
1199
797
402
401

140
261
207
125
81 3060

1935
1125
3191

188.
1308
559
391
168
5.82(!)

5.76(!)
5.79(!)
220
126
94 4168

2615
1553
4690

2719
1971
1419
970
449

3580

2138
1442
3550

2166
1384
641
449
192
594

466
128
280
158
122 3018

2023
995
2511

1540
971
898
625
273
391

170
221
193
114
79 3979(1)

2448
1531
4065(1)

2366
1699
681
441
240
768(1)

245
140
105 4061

2471
1590
3923

2312
1611
1017
661
356
878 (2)

240
141
99 295(net)

1178

846

4528

2737
1791
4379

2504
1875
898
592
306
748

336
190
146

(*) below 150 mbar; (**) poleward of 30�S and 30�N =udgets are calculated below 241 mb, between 30�N and 30�S budgets are calculated below 150 mb. (!) in 1E10 molecules/cm² /s;
(net): net chemical production, (1) below 100 mb in the tropics and below 200 mb poleward 30�, (2) 460 Tg-O3 through the tropopause level.

The size of the cross tropopause term incidentally has been derived by Murphy and Fahey [199..] purely from consideration of the ozone to NO_y relationship in the stratosphere where there is a strong correlation between NO_y and ozone, and the chemical budget of NO_y formed from reaction of N₂O with O(¹D) which must be lost each year out of the stratosphere to maintain an equilibrium concentration. This amounts to ~400 Tg ozone per year in good agreement with three of the models shown in Table 3.5.1, all of which calculate a net positive ozone production term due to tropospheric chemistry.
The independently derived cross tropopause term is much smaller than the total chemical destruction term, which should in principle be similar amongst the models since it is driven mostly by the seasonal and spatial distribution of water vapour and light levels, which are reasonably well known. This observation supports the overall conclusion of the preceding discussion that the behaviour of ozone in the troposphere is dominated by in situ processing, and necessitates a large in situ production term.
The differences shown in Table 3.5.1 are mostly associated with the sources and distribution for nitrogen oxides which determine ozone production, and with the inclusion or exclusion of additional ozone loss processes, including those involving halogen oxides (see earlier). This latter is a very new area of research which requires many more observations of the halogen radicals BrO and IO (see Section 3...).
In addition to the overall budget calculations shown in Table 3.5.1 the model groups in the intercomparison exercise also provided global maps of the surface ozone distribution.
Figure 3.5.2 shows the July global distribution for surface ozone calculated by 6 different groups as displayed in the EIPO intercomparison report. There are obvious differences, as in the comparison table, but there are many similarities including the low ozone concentrations calculated over the remote oceans, the high average ozone concentrations over the continents particularly in the northern hemisphere, and the spread of plumes of ozone- rich air thousands of kilometres downwind from the source. A very clear demonstration of this is shown in some calculations of ozone distribution in the northern hemisphere in the middle of August 1997 made by a model operated by the University of Bergen (REF). The calculations show in graphic detail the manner in which ground-based ozone formed from release of precursors in the USA spreads across the Atlantic to arrive over Europe 3 days later at an altitude of 7 km (Figure 3.5.3). Evidence of these outflows from the USA to Europe have recently been reported in LIDAR studies in the Alps whereby layers of ozone-rich air are frequently found over Garmisch with an origin not in the stratosphere but at ground level several thousand kilometres upwind (REF).

Figure 3.5.2. Ozone concentrations (ppbv) as computed by 6 3-D global models for the month of July at the surface.

Figure 3.5.3. Calculated ozone fields over the North Atlantic in August 1997 showing the spread of air pollution with a North American origin

5.4. Global Measurements of Ozone and Its Precursors

5.4.1. Aircraft Data

A full comparison with global data can of course only be obtained using satellites. These do show some interesting features including Fishman's famous maps of the ozone distribution (REF) obtained by subtracting the stratospheric burden from the total (Figure 3.5.4), but this is not ideal since the preponderance of ozone (~90%) is located in the stratosphere.

Figure 3.5.4. Composite seasonal distribution of ozone as determined from TOMS and SAGE

Gross features shown by the satellites of both the ozone destruction regimes over the equatorial Pacific, and build-up of large concentrations of ozone between the African and South American continents caused by biomass burning are also shown well by aircraft LIDAR for example (Figure 3.5.5) (REF). The low ozone concentrations produced mostly by photochemical (loss) in low NO_x conditions at high water vapour concentrations, rise in equatorial convection up to altitudes of 18 km and then move northwards before encountering the falling tropopause at about 25�N. In contrast a plume high in ozone rises from the continents to fill up the troposphere between South America and Africa over the latitude range 0 to 20�S. Unfortunately there are no equivalent measurements of ozone precursors from the aircraft LIDAR experiments. Measurement of large concentrations of ozone precursors such as NO_x and CO in concert with elevated ozone is shown in aircraft experiments carried out both in the South Atlantic (REF) and the North Atlantic (REF) there is no doubt that the high ozone concentrations shown spreading round the whole of the Northern Hemisphere in summer and across from Africa to Australia in the Southern Hemisphere spring are tropospheric in origin.

Figure 3.5.5. LIDAR depiction of large-scale ozone features over the South Atlantic (upper panel) and Western Pacific (lower panel)

The previous discussion has just shown that the global distribution of ozone in the troposphere is now being simulated in detail by current 3-D chemistry and transport models with output being produced in a synoptic manner. In contrast to the stratosphere however it is not currently possible to confirm details of the model output in the troposphere on a global scale using satellite data. The situation is improving however for ozone as demonstrated in recent presentations by Fishman's group concerning regional pollution in the Southeast USA (Ref). Also it is possible to represent the global distribution of emissions of precursors such as NO₂ from the GOME satellite (Figure 3.5.6), whilst snapshot global distributions of carbon monoxide have been produced in the MAPS experiment using the Space Shuttle. This situation is bound to improve over the next few years with the launching of several new satellites with much better height resolution in the troposphere (see Table 3.5.2). The GOME NO₂ tropospheric columns are very interesting though since they can verify some of the estimates of the distribution of surface ozone by models shown in Figure 3.5.2.

Figure 3.5.6. Global tropospheric columns of NO₂ determined by the GOME satellite in June 1997

Table 3.5.2. Space-borne measurements in the Draft GTOP Science Plan

Instruments	Platforms	Planned operation period	Remarks
(1) tropospheric ozone total ozone HIRS/2 GOME TOMS TOMS SBUV/2 SCIAMACHY TOMS-like instrument ODUS OMPS stratospheric profile SAGE II SBUV/2 SAGEIII SAGEIII GOMOS MIPAS SCIAMACHY H!RDLS MLS OMPS (2) infrared technique IASI TES ATRAS (3) ozone precursors MOPITT SCIAMACHY ASI TES ATRAS	residual technique NOAA ER-2 Earth Probe Meteor-3M? NOM ENVISAT-l EOS-chem1 ADEOS III NPOESS ERBS NOM Meteor3M SpaceStation ENVISAT-l ENVISAT-l ENVISAT-l EOS-Chem EOS-Chem 1 NPOESS METOP EOS-Chem 1 ADEOS III? EOS-Chem 1 ENVISAT-l METOP EOS-Chem1 ADEOSIII?	December 1994 April 1995 July 1996 1999? August 1997 1999 2002 2005? 2007? 1984 August 1997 1999 2002 1999 1999 1999 2002 2002 2007? 2002 2002 2005? 1998 1999 2002 2002 2005?	CO, CH4 CO, CH4 NO, CO, CH4 CO, CH4

Presently the situation is much more limited and the available data comes mostly from aircraft experiments. This has been summarised very recently in an effort coordinated from NCAR that has been referred to earlier and which includes data mostly from the NASA GTE experiments, but also from programmes which use in-service subsonic aircraft such as NOXAR, CARIBIC and MOZAIC, which is worth examining in more detail.
Since September 1994, a new dataset of ozone observations has been collected using commercial aircraft as a platform. The EU MOZAIC project consists of automatic instrumentation, measuring O₃ and water vapor, installed on five long-range Airbus A340 aircraft in normal airline operation. Details about the MOZAIC project and ozone dataset, in particular, are given by Marenco et al. [1998] and Thouret et al. [1998a, b].
In the Northern Hemisphere midlatitudes, at cruise altitudes, MOZAIC O₃ observations show a relatively well-defined seasonal variation with maximum concentrations in the late winter and spring (Figure 3.5.7). Thouret et al. (1998) [a or b] and Law et al. (2000) suggest that the magnitude of the maximum concentration is dependent, to a large extent, on the position of the particular flight level where the measurements were collected, relative to the height of the tropopause which is governed by the position of the large planetary scale ridge and trough pattern in the upper troposphere. For example, peak concentrations over Europe are lower (a ridge region) than over eastern North America and eastern Asia (trough regions).

Figure 3.5.7. Quaterly climatologies of O₃ established from MOZAIC data collected between September 1994 and August 1996 (source : Marenco, Laboratoire d'A�rologie Toulouse, France).

In the tropics, the MOZAIC data is more sparse but nevertheless has given some new insights into the distribution of O₃ around the tropopause. Observed O₃ shows increasing characteristics of the troposphere toward the Equator, that is, lower mean concentrations. However, in the subtropics, very high concentrations (greater than 400 ppbv) have sometimes been found in the MOZAIC data. These were first noted by Suhre et al. [1997] and were found to be small scale features existing in the upper troposphere. Further study by Cammas et al. [1998] showed that these events are most likely to be due to transport across the subtropical tropopause. These features have been observed over the Atlantic and the Indian Ocean and sub-continent.
In fact, O₃ observations over the northern India and neighbouring countries exhibit a spring maximum probably produced by incursions of stratospheric air across the subtropical tropopause. Data collected at these locations also show a summer minimum with monthly mean concentrations as low as 40 ppbv in the upper troposphere. Further south (e.g. over Burma), O₃ levels are lower throughout the year and there is no apparent influence from the stratosphere. Occasionally, O₃ concentrations, as low as a few ppbv, were observed in the upper troposphere in this region. Very low levels of O₃ have also been observed in ozonesonde data collected on ship cruises [Kley et al., 1996] over the Indian Ocean. Examination of MOZAIC data shows that this phenomenon is far more widespread and low concentrations have been observed over all tropical oceanic regions. Even so, the lowest concentrations (less than 1 ppbv) do appear to occur over the Indian Ocean. The mechanism leading to such low concentrations is still unclear, as yet. MOZAIC data have also been collected over continental regions in the tropics and in the Southern Hemisphere at cruise altitudes. Over Africa, it is possible to see the influence of the seasonal movement of the Intertropical Convergence Zone (ITCZ) on the distribution of O₃ at cruise altitudes. Generally, O₃ concentrations decrease toward the Equator and show less seasonal variation. Moving away from the Equator, it is clear that seasonal variation in biomass burning emissions is also an important factor. For example, over southern Africa, O₃ concentrations peak in the austral spring in the upper troposphere. This has also been observed in ozonesonde data and attributed to convective uplift of O₃ and its precursors from biomass burning regions into rather stable haze layers which can exist for many days and be transported large distances [Garstang et al., 1996; Diab et al., 1996). Similar features can also be seen in the MOZAIC data collected over South America.
In summary, MOZAIC data collected at cruise altitudes shows a pattern of higher O₃ concentrations further away from the Equator as air either originating from the stratosphere or in the lower stratosphere is sampled. Uplift of emissions (e.g. from industrial or biomass burning sources) which have subsequently led to O₃ production is also a factor affecting the observed O₃ distribution and seasonal variations. Seasonal variations in large-scale meteorological patterns such as the ITCZ and the monsoon circulation are also important.

5.4.2. Vertical Profiles

The MOZAIC vertical profile data collected over cities also show many interesting features. The airline centres (Paris, Frankfurt, Vienna, Brussels) are obviously the most documented. The large number of profiles collected over several years has enabled, for the first time in some cases, examination of day to day variations, seasonal variability and interannual variability of the O₃ distribution over locations where previously data was sparse or non-existent.
An interesting presentation is the 2D distribution (altitude versus time) formed by combining the daily vertical profiles obtained in one location over several years. Figure 3.5.8 shows an example of the 2D ozone distribution recorded at Frankfurt between August 1994 and February 1999. The vertical structure of the atmosphere appears clearly with: (i) the boundary layer (0-1.5 km) characterized by low O₃ concentrations (10-30 ppb); (ii) the free troposphere (1.5 to about 10 km) where O₃ concentrations vary between 40 and 90 ppb; (iii) the stratosphere, above 10-12 km, corresponding to O₃ mixing ratios higher than 100 ppb (100-400 ppb). This diagram depicts very well the evolution of tropospheric chemistry and tropopause characteristics along the year. Thus, O₃ concentrations in free troposphere are the highest in spring and summer (50-90 ppb), due to variable combinations between stratospheric contributions (mainly in higher troposphere) and an efficient photochemical formation from precursors (mainly of anthropogenic origin).

Figure 3.5.8. Vertical and temporal climatologies near Frankfurt established from MOZAIC data collected between 1994 and 1999 (source : Marenco and Zbinden, Laboratoire d'A�rologie Toulouse, France)

The O₃ concentrations progressively decline after September to be the lowest in autumn and winter (30-40 ppb), corresponding to the decrease of photochemistry. One can notice that, in certain occasions (May, August), high O₃ concentrations are observed even down to ground level. A feature of MOZAIC data is its ability to provide the fine structure of vertical distribution and temporal evolution, thus allowing more detailed studies and interpretations.
The greatest number of vertical profiles have been collected over Northern midlatitude cities. Many features, seen in ozonesonde data are confirmed by the MOZAIC data. However, what the MOZAIC data do show is considerable day to day variability in the O₃ distribution from the ground up to 12 km. This is illustrated in the example shown here for Frankfurt (Figure 3.5.8). There are large variations in the tropopause height which affects the distribution of O₃ in the upper and sometimes, the midtroposphere. There are also days when deep convective activity is active, transporting boundary layer air into the upper troposphere (see Law et al., [1998]). On seasonal and yearly timescales, the MOZAIC O₃ data clearly show the existence of summer maxima in O₃ up to an altitude of 9-10 km at many northern midlatitude locations although the magnitude varies considerably from city to city [Law et al., 2000; other refs?).
Data collected over several coastal locations are affected by the large scale high pressure systems which exist over the Pacific and Atlantic Oceans during the summer months. Clean maritime air masses with low levels of O₃ (less than 10-20 ppbv) are transported to locations such as Miami and Tokyo resulting in a summer minimum in the lower troposphere [Thouret et al., 1998 (a or b); Law et al., 2000]. This has also been observed at other surface coastal/oceanic sites influenced by clean maritime air such as Mace Head, Ireland [Simmonds et al., 1997] and Bermuda [Oltmans and Levy, 1994]. This phenomenon has also been observed in data collected over tropical locations such as Madras and Caracas where the advection of air masses with low levels of O₃ depends on the meteorological circulation patterns such as the summer monsoon (Figure 3.5.9). Evidence for vertical uplift of these air masses into the upper troposphere is also found in the profile data over these locations and also the cruise data, as discussed above.
Vertical profile data collected over continental regions in the tropics and the Southern Hemisphere over South America and southern Africa clearly show the influence of biomass burning and industrial emissions on the O₃ distribution. In particular, the seasonal influence of biomass burning produces layers of high O₃ which can be transported many thousands of kilometers from source regions. The MOZAIC data show that this is a feature which exists from year to year (Figure 3.5.10). It confirms earlier results from field campaigns such as TROPOZ I/II [Jonqui�res et al., 1998], TRACE-A and SAFARI. For example, over Sao-Paulo/Rio there is a clear austral spring maximum throughout the entire depth of the troposphere when deep convection combined with northwesterly flow out of the South American continent transports O₃ and precursors away from areas where burning is taking place over the Amazon forest and savannah regions to locations further south [Thompson et al., 1996, Singh et al., 1996; Law et al, 2000; ***other refs ?***].

Figure 3.5.10 Vertical and temporal climatologies near Johannesburg established from MOZAIC data collected between 1994 and 1999 (source : Marenco and Zbinden, Laboratoire d'A�rologie Toulouse, France)

MOZAIC data has also been collected over tropical Asia in locations where previously data was sparse. As stated above, these sites often have a summer minimum from July to September with mean concentrations as low as 15 ppbv. However, individual profiles occasionally show O₃ levels less than 1 ppbv. Conversely, during the winter and early spring, O₃ concentrations are higher over cities such as Bangkok, Saigon, Hanoi and Madras. It is possible that this seasonal maximum is due to long-range transport of pollutants from the Asian continent to the north [Thouret et al., 1998 (a or b); Law et al., 2000]. What is clear is that the O₃ distribution over these locations is affected by long-range transport of O₃ and/or its precursors.

5.5. Ground-based and Sonde Data

Another major resource has been reviewed recently REF. The major features of this ground-based and sonde record are now discussed in some detail since they represent one of the best resources currently available for describing the global behaviour of ozone. The record is only for ozone however, not for other important species, although there is quite an extensive ground-based record that can be used for seasonal studies; also the true nature of the smaller scale spatial and temporal behaviour of ozone is not captured well.
Figure 3.5.11 displays observed surface ozone seasonalities at several sites (references). The data are monthly averaged values from measurements over periods between a few and twenty years. The seasonalities are grouped according to region.
In remote regions at high latitudes in the NH, surface ozone displays a distinct maximum in late winter and spring (Figure 3.5.11a). The spring maximum is associated in the first place with downward transports from the stratosphere that carry relatively ozone-rich air into the troposphere, and in the second place with photochemical formation of ozone in the troposphere due to increasing insolation and the presence of ozone precursors which have accumulated during the winter. An exception is Barrow, where the presence of halogen compounds causes destruction of boundary layer ozone in spring. At most of the sites, the spring maximum is followed by a summer minimum. In summer, ozone in remote regions is dominatedp by transports from more polluted regions. Generally, these transports cross ocean surfaces where water vapor concentrations are high and, in combination with the strong insolation in summer, ozone destruction is efficient.
At relatively polluted locations on the European and North American continents a summer maximum is found between April and September (Figure 3.5.11b). This is attributed to photochemical ozone production due to ozone precursor emissions in the region, most of which are anthropogenic. The photochemical formation of ozone maximizes in summer when the insolation is strongest.
At remote locations in the NH middle latitudes (Figure 3.5.11c) a similar seasonality of surface ozone is observed as at more northern sites, although the amplitude is generally larger because of stronger insolation. This is, however, not the case at Mauna Loa and Izana, which are relatively high altitude measurement locations where the influence of high water vapour concentrations near the ocean surface on ozone destruction is less pronounced. At Niwot Ridge, a site at relatively high altitude on the North American continent, a summer minimum is also not observed.
The background ozone seasonality apparently does not vary much for different subtropical regions of the globe. In the NH tropical locations of Barbados and Venezuela, surface ozone maximizes in winter and minimizes in summer (Figure 3.5.11d). In the tropics the direction of STE is from the troposphere into the stratosphere, and the influence of ozone originating from the stratosphere is very small throughout the year. The seasonality is determined predominantly by the insolation so that photochemical destruction minimizes in winter and maximizes in summer in tropical remote areas (REF). At Samoa, a clean tropical site on the SH, a similar seasonality is observed but is shifted six months compared to the NH.
Surface ozone concentrations display a distinct maximum between August and November at Brazzaville (Congo) and Cuiaba and Natal (Brazil), which are located in the SH (sub)tropics (Figure 3.5.11e). Spring in the SH is the dry season, when large scale burning of biomass which releases ozone precursors into the atmosphere occurs in South America, Africa and Indonesia. Due to the meteorological conditions the biomass burning pollution is transported over large parts of the SH. Between December and July, surface ozone is about 30 ppbv at Brazzaville, which is a continental site, but smaller concentrations are observed at Natal where surface ozone is influenced by transports from the Atlantic Ocean, and at Cuiaba where ozone levels are strongly influenced by higher hydrocarbons emitted from vegetation. Finally, the ozone seasonality at locations in the SH middle and high latitudes displays a winter maximum and summer minimum (Figure 3.5.11f). Ozone precursor emissions are not significant so that photochemical destruction of ozone, which is most efficient during the summer months, dominates. The seasonality is influenced further by stratospheric-tropospheric exchange (STE). During winter STE maximises, at a time when photochemical destruction minimises, so that significant amounts of ozone of stratospheric origin can be transported to lower parts of the troposphere.

Figure 3.5.11. Seasonal variation of ozone at the surface at various locations from 71�N to 90�S.

Figure 3.5.12abcd. Seasonal variation of ozone in the free troposphere at 500 hPa and 300 hPa from 83�N to 71�S.

Figure 3.5.12efgh Continued. Seasonal variation of ozone in the free troposphere at 500 hPa and 300 hPa from 83�N to 71�S.

Figure 3.5.12 shows surface ozone seasonalities in the middle troposphere (500 hPa) and the upper troposphere/lower stratosphere (300 hPa), taken from ozone sonde measurements (references). Figure 3.5.12a, 12b and 12c show observed middle (500 hPa) and upper tropospheric (300 hPa) ozone seasonalities for locations north of 60�N, between 50�-60�N, and between 35�- 50�N, respectively. Although surface ozone seasonalities may differ strongly for these locations, as shown in the previous section, the observed ozone seasonalities at 500 hPa are quite similar, with an ozone maximum between April and September. The maximum is caused by efficient photochemical ozone formation in summer. NO_x emissions from lightning, which are most efficient over continents in summer, may contribute significantly. The actual peak occurs somewhat earlier at higher than at lower latitudes, which is probably due different contributions of STE, maximizing in spring, and photochemical formation, maximizing in summer. The ozone levels at 500 hPa are between 35-50 ppbv in winter and 50-75 ppbv in summer. In the upper troposphere, at 300 hPa, an ozone maximum occurs in spring, associated with the downward transport of ozone from the stratosphere which is most efficient in spring. There is a distinct latitudinal gradient with higher ozone concentrations at higher latitudes. The maximum peaks between April and June at high latitudes, and between March and July at somewhat lower latitudes due to the influence of photochemical formation.
At NH (sub-)tropical sites the ozone seasonalities at 500 and 300 hPa display a spring maximum at Kagoshima, Naha and Hilo, and a summer maximum at Bermuda, Grand Turk, Kennedy and Palestine (Figures 3.5.12d, 3.5.12e). The latter are significantly influenced by pollution from the adjacent continents. Exceptions are the seasonalities at Panama with a peak in autumn probably due to biomass burning activities during the SH spring, and New Delhi where ozone is relatively independent of season at 500 hPa but peaks in spring at 300 hPa.
The ozone seasonalities at 500 and at 300 hPa are influenced by biomass burning between August and November at SH tropical latitudes, as shown in 4 Figure 3.5.12f. This is also noticeable at Pretoria (South Africa). Except for Brazzaville, ozone minimizes between April and June when the photochemical activity decreases. The biomass burning influence is also noticeable at 500 and 300 hPa for Aspendale and Laverton, at middle latitudes, as well as at 500 hPa for Samoa in the tropical Pacific (Figure 3.5.12g), where ozone is low between January and April but maximizes between September and November. Due to the influence of STE, ozone levels are higher at 300 hPa at Aspendale and Lauder, but about the same as 500 hPa for Samoa. At SH high latitudes (Figure 3.5.12h) ozone displays a winter maximum associated with STE at 500 hPa, but a reverse seasonality at 300 hPa, which is probably related to the chemical destruction of the ozone layer in the Antarctic (see Figure 3.5.4).

IST Ch3 Analysis of ozone seasonality with global chemistry models
(to be written)

Figure 3.5.13. (SurfaceO3.ps). Observed (solid lines) and model simulated (dotted lines) surface O3 at several background monitoring sites [Lelieveld and Dentener, 2000]. Cape Grim (Australia) is located at 41�S, 145�E, Cape Point (South-Africa) at 42�S, 18�E, Jungfraujoch (Switzerland) at 47�N, 8�E (3.6 km altitude), Mace Head (Ireland) at 53�N, 10�W, Mauna Loa (Hawaii) at 19�N, 155�W (3.4 km altitude), and the Pacific island Samoa at 14�S, 171�W. Average seasonal cycles are shown on the right. Model calculated ozone of stratospheric origin is indicated with the dashed lines, O₃ from in situ tropospheric formation by the dash-dotted lines. Measurements by S. Oltmans, E. Brunke, H. Scheel and J. St�helin.

Figure 3.5.13 presents annual and seasonal observations and model calculations of surface O₃ at six background locations. The model includes a tracer of stratospheric O₃ in the troposphere that indeed shows an early spring maximum (dashed lines). Nevertheless, O₃ that has been photochemically produced within the troposphere (dash-dotted lines) clearly dominates the stratospheric contribution to surface O₃, although especially at remote locations the seasonal O₃ cycle is influenced by STE. Figure 3.5.14 shows the model calculated zonal and annual average contributions of the five main tropospheric ozone source categories [Lelieveld and Dentener, 2000]. Photochemical O₃ has been marked according to the NO_x source it derives from. Panel A shows that STE contributes significantly in those parts of the troposphere that are photochemically least active, i.e. the upper troposphere at high latitudes. As soon as the STE derived O₃ is transported deeper into the troposphere it is destroyed, and photochemistry becomes dominant. The model calculations presented in Figure XX also suggest that lightning NO_x strongly contributes to tropical O₃ (panel B), whereas industrial and fossil energy related NO_x predominates at middle latitudes in the northern hemisphere (panel E). It should be emphasised, however, that especially the estimates of NO_x production from lightning are associated with important uncertainties, so that continued studies are called for [Lelieveld et al., 1999].

Figure 3.5.14. (O3tracers.ps). Zonal and annual mean fractional contribution of stratospheric O₃ (STE) to total O₃ in the troposphere (A), the contribution from natural NO_x production by lightning (B) and soil exhalation (C), and that of biomass burning (D) and industrial sources (E), calculated with a global chemistry-transport model [Lelieveld and Dentener, 2000].

Table 3.5.1. Ozone budget (Tg-O₃/yr) below 300 mb, except mentioned differently, as computed by the thirteen global three-dimensional models in the IGAC/3-D CTM O₃ intercomparison exercise

MODEL O₃ budget	IMAU-3	IMAGES	HARVARD *	UKMETO	ECHAM	CTMK	MATCH	MOZART
chemical production NH SH chemical destruction NH SH deposition NH SH cross 300 mb flux NH SH burden NH SH	2495 2821 724 1074 179 106 73	4763 3225 1538 4898 3125 1773 1253 857 396 1388 238 142 96	4100 2620 1480 3680 2290 1390 820 530 290 400 240 160 310 180 130	3890 2557 1333 3276 1958 1318 1199 797 402 401 140 261 207 125 81	3060 1935 1125 3191 1883 1308 559 391 168 5.82 (!) 5.76 (!) 5.79 (!) 220 126 94	4168 2615 1553 4690 2719 1971 1419 970 449	3580 2138 1442 3550 2166 1384 641 449 192 594 466 128 280 158 122	3018 2023 995 2511 1540 971 898 625 273 391 170 221 193 114 79

(*) below 150 mbar; (**) poleward of 30�S and 30�N budgets are calculated below 241 mb, between 30�N and 30�S budgets are calculated below 150 mb. (!) in 1E10 molecules/cm² /s;
(net): net chemical production, (1) below 100 mb in the tropics and below 200 mb poleward 30�, (2) 460 Tg-O₃ through the tropopause level.

6. Current Research: Results from Recent IGAC Campaigns designed to study Tropospheric Chemistry on Various Scales

Experimental studies of atmospheric photochemistry from airborne, ship-based, and ground platforms are naturally limited in temporal and geographic scope. Yet the larger understanding of the chemistry of tropospheric oxidants and their precursors evolves from measurements, observations and modeling derived from basically regional scale and process studies. During the past decade or so, experimental programs dealing with regional and global photochemistry have evolved into increasingly comprehensive and complex missions that attempt to assess atmospheric photochemistry under different meteorological regimes or in unique atmospheric conditions. Many of the more prominent experiments for the northern and southern hemispheres were discussed in more detail in WMO 1999, chapter 8. Numerous publications and journal issues have described results and conclusions from these studies, and some of these results have been discussed earlier in this chapter. This section will highlight a few major results obtained from these tropospheric chemistry missions during recent years to indicate the manner in which much research is being carried out under the auspices of IGAC. A limited number of examples are given including the chemistry of the atmosphere in developed continental regions, outflow of airmasses from continental regions to the marine atmosphere, and chemistry of remote regions.

6.1. Chemistry of the continental atmosphere in developed regions.

As indicated earlier ozone production in the continental boundary layer constitutes a major source of ozone in the global atmosphere. The chemistry is very complex and includes many more carbonaceous species than were indicated in Section 3.2. Also oxygenated hydrocarbons play a major role as do organic nitrogen compounds. This complex chemistry will not be dealt with in any detail but some findings from recent experiments are relevant. While the historical development of our understanding of much of the oxidative chemistry of the atmosphere was based on smog chemistry of polluted urban regions, there also has been a continuing emphasis on studying urban emissions and photochemistry in the context of regional ozone pollution episodes. These studies are driven by a need to understand the causes of high ozone periods and to develop mitigation strategies for controlling ozone exposure to humans and agriculture. While the impact of anthropogenic emissions of NO_x and VOCs was well known, one of the major findings was that reactive biogenic hydrocarbons (such as isoprene), from urban and surrounding landscapes combined with urban emission of ozone precursors, could be driving ozone production on urban and regional scales [Chameides et al., 1988; Trainer et al., 1987]. [NO_x/VOC chemistry described earlier?]. Some of the studies that have addressed the problem issues of the complex chemistry of the continental atmosphere impacted by anthropogenic emission include the Southern Oxidant Study (SOS), the Tropospheric Ozone Research (TOR) project of EUROTRAC, the Biogenic Emissions in the Mediterranean Area (??)(BEMA) project, and the Photochemical & (PROPHET) study (Others??)
[Discussion of ozone, NO_z, peroxide relationships eg from Sillman, 1998 or other? Trainer, etc&..Discussion of radical sources from hydrocarbons, formaldehyde etc.? I need some time for this, or is it included earlier?]
To better simulate and understand regional atmospheric chemistry, models of regional/urban scale oxidant production and loss must incorporate an improved and more detailed knowledge of anthropogenic and biogenic emissions. Indeed, providing improved emission estimates of both biogenic and anthropogenic trace gases has been a major accomplishment over the past decade (GEIA refs, other). Meagher et al. [1998] document improvements in emission inventories applied to the SOS campaigns in the SE United States. For biogenic emissions, 1) combined use of satellite imagery and aerial photography produce land use and vegetative classifications to improve estimates of VOC emission; and 2) direct flux measurements of VOC, and estimates from indirect methods, relate the land-use data to more quantitative estimates of emission. Anthropogenic emissions inventories from traffic data, automotive VOC characterization, and industrial point sources were added to the total emissions inventories to develop a subregional domain inventory of VOC and NO_x emissions for use in larger scale models. Careful evaluation of the field data demonstrated the need and value of improved emission inventories to characterize the chemistry over the urban/rural source regions (ref).
One of the interesting diagnostic measurements related to the impact of biogenic hydrocarbons relative to anthropogenic VOC is the use of characteristic peroxyacyl nitrates formed from different sources [Roberts et al., 1998]. For example, one of the degradation products of isoprene is peroxymethacrolyl nitrate (or MPAN) that is formed from:

CH₂C(CH₃)CHCH₂ + OH + O₂ + NO₂ -> CH₂C(CH₃)C(O)OONO₂ (3.32)
(write as sequence)
Similarly PAN (peroxyacetyl nitrate), CH₃C(O)OONO₂, can be formed from numerous >C₃ hydrocarbons, while peroyxpropionyl nitrate (CH₃CH₂C(O)OONO₂) is characteristic of the oxidation of more reactive >C₄ anthropogenic hydrocarbons [Ridley et al., xxxx; Singh et al., xxxx, Bertman et al., xxxx]. As noted earlier (Eq ?), peroxyacyl nitrate formation occurs along with ozone production. Thus, ratios of the PAN:MPAN can be related to the proportion of isoprene relative to other HC to produce ozone. Simlarly, PAN:PPN ratios can indicate the relative contribution of reactive anthropogenic HC to ozone production. For example, Roberts et al. [1998] found that ratios of PAN:MPAN in the range of 6 - 10 were characteristic of isoprene dominated chemistry, while PAN:PPN ratios of 5.8 to 7.4 characterized chemistry dominated by anthropogenic hydrocarbons. Using these ratios, different ozone events were analyzed to determine the relative importance of biogenic to anthropogenic HC in the ozone production process. For the cases examined, biogenic hydrocarbons accounted for <25 50% of the ozone production above background levels.
In continental regions more remote from direct urban sources, new insights have been gained on the oxidative chemistry of the continental lower troposphere. [This will be some of the info from Mary Anne Carroll representing PROPHET.]

6.2. Chemistry of continental outflow in the marine atmosphere

Transport of continentally derived emissions extends the regional issues of urban/regional chemistry to a larger scale. The issues that have been examined are:

What are the pathways that transport ozone and ozone precursors from their continental source to the adjacent marine region?,
What is the magnitude of the direct transport of ozone or other oxidants to the broader atmosphere?, and
What are the physical and photochemical processes and rates in the continental effluent?

6.3. Studies of Aircraft Emissions

Global air traffic has more than doubled in the last 15 years and this trend is projected to continue at least into the early part of the twenty-first century. Aircraft exhaust emissions of NOx and particles have the potential to modify the chemistry and microphysics of the upper troposphere and lower stratosphere. The SASS (Subsonic Assessment) Ozone and NO_x Experiment (SONEX), was an airborne field campaign conducted in October-November 1997 in the vicinity of the North Atlantic Flight Corridor. A key objective of SONEX was "to assess the effect of emissions from subsonic aircraft on nitrogen oxides and ozone". A fully instrumented NASA DC-8 aircraft was used as the primary SONEX platform. SONEX activities were closely coordinated with the European POLINAT-2 (Pollution from Aircraft Emissions in the North Atlantic Flight Corridor) program, which used a Falcon-20 aircraft. Upper troposphere/"lowermost" stratosphere (UT/LS) was the region of greatest interest. Results from the SONEX/POLINAT-2 campaign are being published in Special Sections of GRL (October, 1999) and JGR (February, 2000) and are summarized here.
Measurements during SONEX showed that nearly 90% of the reactive nitrogen is accounted for over a large dynamic range of measured NO_y. PAN, HNO₃ and NO_x formed the main components of this budget and the fraction of HNO₃ progressively increased from the lower troposphere towards the LS. NO_x and CN spikes from aged exhaust plumes were frequently observed in the flight corridor region. There was evidence for NO_x (and CN) accumulations on regional scales in the upper troposphere and lower stratosphere, respectively. No corresponding O₃ accumulations could be detected. Direct measurements of HO_x, NO_x, and their precursors, were used to determine production and destruction rates of O₃. Net O₃ production rates increased dramatically with increasing NO_x. These observations also demonstrated the presence of a regime where O₃ production was insensitive to NO_x concentrations for distinct HO_x levels.
SONEX/POLINAT-2 successfully collected a comprehensive body of data over the north Atlantic that can be used to further test and validate global models of transport and photochemistry, and identified areas of uncertainty. The conclusion was reached that increased aircraft NO_x emissions in the future will likely lead to additional O₃ formation but the rate will vary greatly depending on the state of the atmosphere.
The overall objective of the POLINAT (Pollution from Aircraft Emissions in the North Atlantic Flight Corridor) projects was to determine by field measurements, analysis, and modelling studies the relative contribution from subsonic air traffic emissions to the composition of trace species at altitudes between 9 and 13 km within and near the major transatlantic air traffic routes and to assess potential effects of these emissions on ozone, aerosols, and clouds. Air traffic emissions of concern include nitrogen oxides (NO_x), sulphur oxides (SO_x), soot (black carbon or organic carbon), carbon dioxide (CO₂), and water vapour (H₂O). Measurements were performed using the Deutsches Zentrum f�r Luft- und Raumfahrt (DLR) Falcon research aircraft and an in-service Swissair B-747 over the North Atlantic in November 1994 and June/July 1995 and from August to November 1997 based from Shannon, Ireland. Details on POLINAT and the key partner mission SONEX are given by Schumann et al. [2000] and Singh et al. [1999], respectively.
During survey flights across the North Atlantic flight corridor, the emissions from air traffic were clearly measurable in terms of increases in the concentrations of NO_x, SO₂, and CN. Often, the measured plumes result from a superposition of several aircraft exhaust plumes. The measurements and results from plume dispersion models show that air traffic emissions cause a very inhomogeneous distribution of the NO_x and CN concentration fields in the flight corridor. It takes about 3 to 10 hours before individual plumes get diluted to background NO_x concentrations.
Based on the chemical forecasts made by the University of Bergen and NILU, selected flights during POLINAT-2 were specifically designed to measure in regions with predicted high impact of aircraft emissions to search for the corridor effect. These flights identified large-scale enhancements of NO mixing ratios inside the corridor of about 50-150 pptv. These enhancements were attributed to aircraft emissions by correlation s with simultaneous tracer measurements, back trajectory analyses, air traffic distribution, and model predictions with and without aircraft emissions [Schlager et al., 1999].
The POLINAT investigations showed for the first time that aircraft emissions lead to measurable NO_x changes not only in exhaust plumes up to many hours but also at the scale of the whole corridor during stagnant flow conditions over the North Atlantic. However, high abundance of trace species was also often found in the upper troposphere resulting from lightning and transport from continental sources regions. The role of these sources and the impact of cirrus clouds on the trace gas composition near the tropopause over the North Atlantic require future investigation.

6.4. Chemistry of the remote atmosphere

While it is virtually impossible to locate a region of the atmosphere that is not subject to anthropogenic perturbations, there are large areas, e.g., over the Pacific Ocean basin, that receive only episodic and seasonal inputs of air influenced by man-made emissions or similarly large perturbations, such as biomass burning. Instead, the oxidative chemistry in remote regions can be controlled by the relatively small residual concentration and transformations of continentally derived trace gases (particularly NO_x), by in-situ photochemical reactions of methane, CO, water vapor and ozone, and, potentially, by local emissions from the surface ocean or ice. At higher altitudes in the troposphere, convective redistribution of trace gases with surface sources and subsequent rapid transport can have a strong influence on the oxidizing power of the upper troposphere. Some examples of the chemistry of the remote regions follow next.
In the remote marine environment, the Mauna Loa Observatory Photochemistry Experiments (MLOPEX 1 and 2) were conducted to test and constrain models of photochemistry representing the mid-troposphere [Ridley and Robinson, 19xx; Atlas and Ridley, 19yy]. The MLOPEX 2 experiment was among the first to attempt a comprehensive set of measurements of photochemically related trace gases, radical species, and photolysis rates that would be sufficient to adequately characterize local photochemistry in the remote atmosphere. A detailed evaluation of the photochemistry and budgets of trace species at Mauna Loa was conducted by Hauglustaine et al. [1999], and other model and experimental studies were published as special sections in the Journal of Geophysical Research. Mesoscale regional studies based on the MLOPEX results also probed the factors controlling trace concentrations, variations and photochemistry over the Pacific Basin (Hess et al., submitted).
Even under the relatively simple chemical conditions found at MLOPEX, Hauglustaine et al. [199--] report only moderate success at comparing measurements to photochemical model results. One of the important conclusions was that all photochemically related species could not be reconciled over the entire seasonal range within the constraints of a photochemical box model. For example, one of the well-reproduced photochemical factors during much of the year-long experiment was the NO/NO₂ ratio (see Equation XXX, Figure _ NO/NO₂ comparison). However, this good agreement is based on model-calculated levels of peroxy radical that were significantly higher than those measured during the experiment. Also, disagreement between the calculated and measured NO/NO2 ratio during summertime suggested the presence of additional oxidants that were not among the suite of compounds measured at MLO. Other analysis of measurements from the remote marine atmosphere (PEM TROPICS A, [Schultz et al., 1999]) also have found reasonably good agreement between measured and modeled NO/NO₂ ratios, but large disagreements have also been noted (PEM WEST A, [Bradshaw et al., 1998]). Whether the noted disagreements are related to measurement artifacts or to missing chemistry is not fully understood.
Similarly, tests of hydroxyl radical measurements versus model results at MLO showed an excellent agreement during the spring experimental period, but the model results overestimated measured OH during the summer by a factor of 1.5 - 2. Though different hypotheses have been suggested to account for the differences in OH radical between spring and summer, no explanation has proved satisfactory. A simple box model also has difficulty in reconciling the levels of peroxides (mostly hydrogen peroxide and methyl hydroperoxide) with the measured formaldehyde. If peroxide concentrations are constrained to measured levels, then formaldehyde is reasonably well predicted by the photochemical model. Still, this leaves uncertain the factors controlling the concentration and variation of hydrogen and methyl hydroperoxide at 3 km over the mid-Pacific.
In spite of the uncertainties suggested above, the MLOPEX results provide insight into the major processes controlling the budget of ozone, nitrogen oxides, and radical species and their seasonal variations. During all seasons, ozone destruction is nearly balanced by ozone production, with a maximum ozone loss of 1.39 ppbv/day calculated for the summer. Thus, on average, the mid-Pacific troposphere is a sink for ozone during the year, with maximum photochemical activity in the summer. The budget of NO_x, though, is unbalanced in the model throughout the year. Sources of NO_x from 18 up to 56 ppt/day are required to balance the processes that produce and remove active NO_x from the troposphere at MLO. Reconciling the budget of NO_x, and understanding the sources, transport, and transformations of NO_x to the remote atmosphere is clearly a critical factor in evaluating oxidant budgets, because of the role of NOx in catalyzing ozone production.
It should be noted, though, that other budget studies of NO_x in the remote atmosphere do not indicate an imbalance in sources and sinks. The PEM TROPICS A study [Schultz et al., 19--], for example, evaluated the budget of ozone and NO_x in the tropical troposphere and found that transport and decomposition of PAN was most important in balancing the NO_x budget. Other NO_x sources, such as lightning and continental outflow, were suggested from the analysis of PEM campaigns in the Western Pacific Basin (PEM WEST A/B, refs&). The SOAPEX series of experiments conducted in 1991-92, 1995 and 1999 have yielded much valuable date on the free radical chemistry of the remote marine boundary layer, some of which is discussed earlier in Section 3&. The acronym stands for "Southern Ocean Atmospheric Photochemistry Experiment" and they are designed to study the self-cleansing capacity of the remote atmosphere and also to study the budget of ozone in the clean marine boundary layer. Recent papers have shown the importance of photochemistry in controlling ozone in this situation both in summer and in winter with additional impact of processes such as deposition into the sea and exchange of air across the top of the marine boundary layer [Ayers et al., 1992; 1996; Monks et al., 19--; Penkett et al., 19--; Carpenter et al., 19--]

7. References Still need to be worked on

Alicke B, K. Hebestreit, J. Stutz, and U. Platt, Iodine oxide in the marine boundary layer, Nature, 397, 572-573, 1999.
Aliwell, S.R., and R.L. Jones, Measurements of tropospheric NO₃ at midlatitude, J. Geophys. Res., 103, 5719-5727, 1998.
Allan, B.J., R.A. Burgess, N. Carslaw, H. Coe, and J.M.C. Plane, Observations of the Nitrate Radical in the Marine Boundary Layer, J. Atmos. Chem., 33, 129-154, 1999. Identify the two Allan refs
Allan, B.J., J. Plane and G. McFiggans, presentation at EGS symposium, Den Haag, 1999.
Appenzeller, C., J.R. Holton, and K.H. Rosenlof, Seasonal variation of mass transport across the tropopause, J. Geophys. Res., 101, 15071-15078, 1996.
Arias, M.C., and D.R. Hastie, Radical chemistry at the SONTOS site in rural Ontario, Atmos Environ., 30, 2167-2175, 1996.
Ariya, P.A., B.T. Jobson., R. Sander, H. Niki, G.W. Harris, J.F. Hopper and K.G. Analauf, J. Geophys. Res., 103, 13619-13180, 1998.
Atherton, C.S., S.Sillman and J.Walton, Three-dimensional global modeling studies of the transport and photochemistry over the North Atlantic Ocean. J.Geophys.Res., 101, 29289-29304, 1996.
Atkinson, R., D.L. Baulch, R.A. Cox, R.F. Hampson Jr., J.A. Kerr, and J. Troe, Evaluated kinetic and photochemical data for atmospheric chemistry; Supplement VI, J. Phys. Chem. Ref. Data, 26, 1329-1499, 1997. Does not appear in text
Atlas and Ridley, P61, L32
Atlas et al., -section 3.1.1.8
Ayers, G.P., S.A. Penkett, R.W. Gillett, B. Bandy, I.E. Galbally, C.P. Meyer, C.M. Elsworth, S.T. Bentley, and B.W. Forgan, Evidence for photochemical control of ozone concentrations in unpolluted marine air, Nature, 360, 446-448, 1992.
Ayers, G.P., S.A. Penkett, R.W. Gillett, B.J. Bandy, I.E. Galbally, C.P. Meyer, C.M. Elsworth, S.T. Bentley, and B.W. Forgan, Annual cycle of peroxides and ozone in marine air at Cape Grim, Tasmania, J. Atmos. Chem., 23, 221-252, 1996.
Ayers, G.P., H. Graneck and R. Boers, Ozone in the marine boundary layer at Cape Grim: Model simulation, J. Atmos. Chem., 27, 179-195, 1997. IS IT P25, L15?
Barnes, I., B. Bonsang, T. Brauers, P. Carlier, R.A. Cox, H.-P. Dorn, M.E. Jenkin, G. Le Bras, and U. Platt, In Laboratory and field studies of oxidation processes occurring in the atmospheric marine boundary layer, Air Pollution Research Report 35, Commission of the European Community, 1991.
Barrie, L.A., J.W. Bottenheim, R.C. Scnell, P.J. Crutzen and R.A. Ramussen, Ozone destruction and photochemical-reactions at polar sunrise in the lower Arctic, Nature, 334, 138-144, 1988.
Bengtsson, L., From short-range barotropic modelling to extended-range global weather prediction: a 40-year perspective. Tellus 51A-B, 13-32, 1999.
Berntsen, T., I.S.A. Isaksen, W.-C. Wang and X.-Z. Liang, Impacts of increased anthropogenic emissions in Asia on global tropospheric ozone and climate. Tellus 48B, 13-32, 1996.
Bertman et al. (Section 3.6)
Bloss, W., D.M. Rowley, R.L. Jones, and R.A. Cox, Geophys.Res.Lett., in press (2000)
Bottenheim, J.W., L.A. Barrie, E. Atlas, L.E. Heidt, H. Niki, R.A. Rasmussen and P.B. Shepson, Depletion of lower tropospheric ozone during Arctic spring - the polar sunrise experiment, J. Geophys. Res., 95, 18535-18568, 1990.
Bradshaw J, S. Sandholm S, and R. Talbot R, An update on reactive odd-nitrogen measurements made during recent NASA Global Tropospheric Experiment programs, J. Geophys. Res., 103: 19,129-19,148, 1998.
Brasseur, G. P., J. J. Orlando, and G. S. Tyndall (eds.), Atmospheric Chemistry and Global Change, 654 pp., Oxford University Press, 1999.
Brenninkmeijer, C.A.M., M.R. Manning, D.C. Lowe, G. Wallace, R.J Sparks, and A. Volzthomas, Interhemispheric asymmetry in OH abundance inferred from measurements of atmospheric (CO)-C-14, Nature, 356, 50- 52, 1992.
Brune WH, Faloona IC, Tan D, Weinheimer AJ, Campos T, Ridley BA, Vay SA, Collins JE, Sachse GW, Jaegle L, Jacob DJ, Airborne in-situ OH and HO2 observations in the cloud-free troposphere and lower stratosphere during SUCCESS, Geophys. Res. Lett., 25, 1701-1704, 1998.
Brune, W.H., D. Tan, I.F. Faloona, L. Jaegle, D.J. Jacob, B.G. Heikes, J. Snow, Y. Kondo, R. Shetter, G.W. Sachse, B. Anderson, G.L. Gregory, S. Vay, H.B. Singh, D.D. Davis, J.H. Crawford, D.R. Blake, OH and HO2 chemistry in the North Atlantic free troposphere, Geophys. Res. Lett., 26, 3077-3080, 1999.
Cammas, J.-P., S. Jacoby-Koaly, K. Suhre, R. Rosset, and A. Marenco, Atlantic subtropical potential vorticity barrier as seen by measurements of ozone by Airbus In-Service Aircraft (MOZAIC) flights, J. Geophys. Res., 103, 25,681-25,693, 1998.
Cantrell et al. references need to be sorted out.
Cantrell et al., 19.. (page 13, line 10)
Cantrell et al., 1992, 1993
Cantrell et al., 1996
Cantrell et al., 1996a, 1996b, 1997
Cantrell, C.A., page 12, line 32
Cantrell, C.A., R.E. Shetter, T.M. Gilpin, and J.G. Calvert, Peroxy radicals measured during MLOPEX II: The data and first analysis, J. Geophys. Res., 101, 14,643-14,652, 1996a.
Cantrell, C.A., D.H.Stedman, 1984 - which one of the two below:
Cantrell, C.A., and D.H. Stedman, A possible technique for the measurement of atmospheric peroxy-radicals, Geophys. Res. Lett., 9, 846-849, 1982.
Cantrell, C.A., D.H. Stedman, G.J. Wendel, Measurement of atmospheric peroxy-radicals by chemical amplification, Anal. Chem., 56, 1496-1502, 1984
Cantrell CA, Shetter RE, Gilpin TM, and J.G. Calvert, Peroxy radicals measured during Mauna Loa observatory photochemistry experiment 2: The data and first analysis, J. Geophys. Res., 101, 14643-14652, 1996.
Cantrell CA, Shetter RE, Gilpin TM, Calvert JG, Eisele FL, Tanner DJ, Peroxy radical concentrations measured and calculated from trace gas measurements in the Mauna Loa observatory photochemistry experiment 2, J. Geophys. Res., 101, 14653-14664, 1996.
Cantrell CA, Shetter RE, Calvert JG, Dual-inlet chemical amplifier for atmospheric peroxy radical measurements, ANAL CHEM 68: (23) 4194- 4199 DEC 1 1996
Cantrell CA, Shetter RE, Calvert J., Peroxy radical chemistry during FIELDVOC 1993 in Brittany, France, ATMOS ENVIRON 30: (23) 3947- 3957 DEC 1996
Carpenter, L.J., P.S. Monks, I.E. Galbally, C.P. Meyer, B.J. Bandy and S.A. Penkett, A Study of Peroxy Radicals and Ozone Photochemistry at Coastal Sites in the Northern and Southern Hemispheres, J. Geophys. Res. 102, 25,417-25,427, 1997.
Carpenter, L.J., K.C. Clemitshaw, R.A. Burgess, S.A. Penkett, J.N. Cape, and G.C. McFadyen, Investigation and evaluation of the NOx/O3 photochemical steady state, Atm. Env., 32, 3353-3365, 1998. - Carpenter only mentioned in Table 3.3.2
Carpenter, L.J., W.T. Sturges, S.A. Penkett, P.S. Liss, B. Alicke, K. Hebestreit, and U. Platt, Short-lived alkyl iodides and bromides at Mace Head, Ireland: Links to biogenic sources and halogen oxide production, J. Geophys. Res., 104, (D1), 1679-1689, 1999.
Carsey, T. P., et al., Nitrogen oxides and ozone production in the North Atlantic marine boundary layer, J. Geophys. Res., 102, 10653-10665, 1997.
Carslaw, N., L.J. Carpenter, J.M.C. Plane, B.J. Allan, R.A. Burgess, K.C. Clemitshaw, H. Coe and S.A. Penkett, Simultaneous measurements of nitrate and peroxy radicals in the marine boundary layer, J. Geophys. Res., 102, 18,917-18,933, 1997a.
Carslaw, N., J.M.C. Plane, H. Coe and E. Cuevas, Observations of the Nitrate Radical in the Free Troposphere at Iza�a de Tenerife, J. Geophys. Res., 102, 10,613-10,622, 1997b.
Cars law , N., D .J. Creasey, D.E. Heard, A.C. Lew is , J .B. M cQ uaid, M.J. Pilling, P.S. Monks, B.J. Bandy, and S .A. P enkett, Modeling O H, HO₂, and RO₂ radicals in the mar ine boundary layer : 1. Model cons tr uction and comparison w ith f ield measurements , J. G eophys. Res., 104, 30,241- 30,255, 1999a.
Carslaw N, Jacobs PJ, Pilling MJ, Modeling OH, HO₂, and RO₂ radicals in the marine boundary layer: 2. Mechanism reduction and uncertainty analysis, J. Geophys. Res., 104, 30257-30273, 1999b.
Chameides, W. and J.C.G. Walker, A photochemical theory for tropospheric ozone, J. Geophys. Res., 78, 8751-8760, 1973.
Chameides, W.L., R.W. Lindsay, J. Richardson, and C.S. Kiang, The role of biogenic hydrocarbons in urban photochemical smog - Atlanta as a case-study, Science 241, 1473-1475, 1988. P56, L31
Chameides, W.L., and E.B.Cowling, The State of the Southern Oxidants Study (SOS): Policy-relevant findings in ozone pollution research 1988-1994, College of Forest Resources, North Carolina State University, Raleigh, North Carolina, 94p, 1995.
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.
Clemitshaw, K.C., L.J. Carpenter, S.A. Penkett and M.E. Jenkin, A calibrated peroxy radical chemical amplifier (PERCA) for atmospheric measurement, J. Geophys. Res., 102, 25405-25416, 1997. Only mentioned in table 3.3.2.
Collins, W.J., D.S.Stevenson, C.E.Johnson and R.G.Derwent, Tropospheric ozone in a global-scale three-dimensional lagrangian model and its response to NOx emission controls. J.Atmos.Chem., 26, 223-274, 1997.
Comes, F.J., O. Forberich, J. Walter, OH field measurements: A critical input into model calculations on atmospheric chemistry, J. Atmos. Sci., 54,1886-1894,1997 Only mentioned in Table 3.3.1
Cox, R.A., Ozone and peroxy radical budgets in the marine boundary layer: Modelling the effect of NO_x, J. Geophys. Res., 104, 8047-8056, 1999.
Crawford, J., et al., Photostationary state analysis of the NO2-NO system based on airborne observations from the western and central North Pacific, J. Geophys. Res., 101, 2053-2072, 1996.
Creasey D.J., P.A. Halford-Maw, D.E. Heard, M.J. Pilling, and B.J. Whitaker, Implementation and initial deployment of a field instrument for measurement of OH and HO2 in the troposphere by laser-induced, fluorescence, J. Chem. Soc.- Faraday trans., 93, 2907-2913, 1997. Only mentioned in Table 3.3.2
Crutzen, P.J., Photochemical Reactions initiated by and influencing ozone in unpolluted tropospheric air, Tellus, 26, 47-57, 1973.
Crutzen and co-workers, in 1977 Is it below
Crutzen, P.J., and J. Fishman, Average concentrations of Oh in the troposphere and the budgets of CH₄, CO, H₂, and CH₃CCl₃, Geophys. Res. Lett., 4, 321-324, 1977
Crutzen, P. J., and P. H. Zimmermann, The changing photochemistry of the troposphere, Tellus, 43AB, 137-151, 1991.
Danielsen, E. F., Stratospheric-tropospheric exchange based on radioactivity, ozone and potential vorticity, J. Atmos. Sci., 25, 502-518, 1968.
Davies, T.D., and E. Schuepbach, Episodes of high ozone concentrations at the earth s surface resulting from transport down from the upper troposphere/lower stratosphere: A review and case studies, Atmos. Environ., 28, 53-68, 1994.
Davis, D.D., et al., Assessment of ozone photochemistry in the western North Pacific as inferred from PEM-west A observations during the fall 1991, J Geophys.Res., 101, 2111-2134, 1996.
Davis, D.D., J. Crawford, S. Lui, S. McKeen, A. Bandy, D. Thornton, F.S. Rowalnd and D. Blake, J. Geophys. Res, 101, 2135-2147, 1997.
Derwent, R.G., and Curtis, A.R., Two dimensional model studies of some trace gases and free radicals in the troposphere, AERE Report R8853, London HMSO, 1977
Derwent, R.G. et al. 1977? - is it the reference above
Derwent, R.G., 1994 - which of the two options below.
Derwent, R.G., P.G. Simmonds, and W.J.Collins, Ozone and carbon-monoxide measurements at a remote maritime location, Mace Head, Ireland, from 1990 to 1992, Atmos. Environ., 28, 623-2637, 1994
Derwent, R.G., and T.J. Davies, Modeling the impact of NOx or hydrocarbon control on photochemical ozone in Europe, Atmos Environ., 28, 2039- 2052, 1994.
Diab, R.D., et al., Vertical ozone distribution over southern Africa and adjacent oceans during SAFARI-92, J. Geophys. Res., 101, 23,823- 23,833, 1996.
Dickerson, R.R., G.J.Huffman, W.T.Luke, L.J.Nunnermacker, K.E.Pickering, A.C.D.Leslie, C.G.Lindsey, W.G.N.Slinn, T.J.Kelly, P.H.Daum, A.C.Delany, J.P.Greenberg, P.R.Zimmerman, J.F.Boatman, J.D.Ray and D.H.Stedman, Thunderstorms: An important mechanism in the transport of air pollutants. Science, 235, 460-465, 1987.
Dlugokencky, E. J., K. A. Masarie, P. M Lang, and P. P. Tans, Continuing decline in the growth rate of the atmospheric methane burden, Nature, 393, 447-450, 1998.
Dorn H.P., U. Brandenburger, T. Brauers, M. Hausmann, and D.H. Ehhalt, In-situ detection of tropospheric OH radicals by folded long-path laser absorption. Results from the POPCORN field campaign in August 1994, Geophys. Res. Lett., 23, 2537-2540 1996.
Draxler, R.R., Boundary layer isentropic and kinematic trajectories during the August 1993 North Atlantic regional experiment intensive, J. Geophys. Res., 101, 29255-29268, 1996.
ECMWF
Ehhalt, D.H., in Zellner book 1999.
Eisele, F.L., D.J. Tanner, C.A. Cantrell, and J.G. Calvert, Measurements and steady state calculations of OH concentrations at Mauna Loa observatory, J. Geophys. Res., 101, 14665-14679, 1996.
Eisele, F.L., G.H. Mount, D. Tanner, A. Jefferson, R. Shetter, J.W. Harder, and E.J. Williams, Understanding the production and interconversion of the hydroxyl radical during the Tropospheric OH Photochemistry Experiment, J. Geophys. Res., 102, 6457-6465, 1997.
Elbern, H. and H. Schmidt, A four-dimensional variational chemistry data assimilation scheme for Eulerian chemistry transport modeling, J. Geophys.Res., 104, 18583-18598, 1999.
Emmons, L. K., et al., Data composites of airborne observations of tropospheric ozone and its precursors, submitted to J. Geophys. Res., 1999.
Fehsenfeld, F.C., P. Daum, W.R. Leaitch, M. Trainer, D.D. Parrish, and G. Hubler, Transport and processing of O₃ and O₃ precursors over the North Atlantic: An overview of the 1993 North Atlantic Regional Experiment (NARE) summer intensive, J. Geophys. Res., 101, 28,877-28,891, 1996. not in text
Fishman, J., C.E.Watson, J.C.Larsen and J.A.Logan, The distribution of tropospheric ozone determined from satellite data, J.Geophys.Res., 95, 3599-3629, 1990.
Fishman's diagrams and maps pages 52 and 53
Flat�y, F., and �. Hov (1995). p44 - cannot find - is it 1996?
Flat�y, F., �.Hov and H. Smit, 3D model studies of exchange processes of ozone in the troposphere over Europe, J. Geophys. Res., 100, 11,465- 11,481, 1995.
Flat�y, F., �. Hov, C. Gerbig and S.J. Oltmans, Model studies of the meteorology and chemical compiosition of the troposphere over the North Atlantic during August 18-30, 1993, J. Geophys. Res., 101, 29317-29334, 1996.
Flatoy F, and �. Hov, Three-dimensional model studies of the effect of NO_x emissions from aircraft on ozone in the upper troposphere over Europe and the North Atlantic, J. Geophys. Res, 101, 1401-1422, 1996.
Flat�y, F., �.Hov and H.Schlager, Chemical forecasts used for measurement flight planning during POLINAT 2, Geophys. Res. Lett., 2000.
Frost, G.J. et al., Photochemical ozone production in the rural southeastern United States during the 1990 rural oxidants in the southern environment (ROSE) program. J.Geophys.Res., 103, 22491-22508, 1998.
Galbally et al (no year given) page 25, line 19
Garmisch ref (p52, ln24)
Garstang, M., P.D. Tyson, R. Swap, M. Edwards, P. Kallberg, and J.A. Lindesay, Horizontal and vertical transport of air over southern Africa, J. Geophys. Res., 101, 23,721-23,736, 1996.
Gates, W.L., AMIP: The atmospheric model intercomparison project. Bull. Amer. Meteor. Soc., 73, 1962-1970, 1992
GEIA refs
George et al., 1999 In Cox Table
Hard and O Brien
JGR 1999 Hard et al., 1992
Hastemrath, S., Climate and Circulation in the Tropics, D. Reidel, Norwell, Mass., 1985. (not in text)
Hauglustaine DA, Madronich S, Ridley BA, Flocke SJ, Cantrell CA, Eisele FL, Shetter RE, Tanner DJ, Ginoux P, Atlas EL, Photochemistry and budget of ozone during the Mauna Loa Observatory Photochemistry Experiment (MLOPEX 2), J. Geophys. Res., 104, 30275-30307, 1999.
Hausamann et al., 1996 (p32, ln 12)
Haynes, P.H., C.J. Marks, M.E. McIntyre, T.G. Shepherd, and K.P. Shine, On the downward control of extratropical diabatic circulations by eddy-induced mean zonal forces, J. Atmos. Sci., 48, 651-687, 1991.
Hebestreit, K., J. Stutz, D. Rosen, V. Matveev, M. Peleg. M. Luria and U. Platt, DOAS measurements of tropospheric bromine oxide in mid-latitudes, Science, 283, 55-57, 1999.
Heintz, F., U. Platt, H. Flentje and R. Dubois, Long-term observation of nitrate radicals at the TOR Station, Kap Arkona (R�gen), J. Geophys. Res., 101, 22,891-22,910, 1996.
Hofzumahaus, A., U.Aschmutat, M. He�ling, F. Holland and D.H. Ehhalt, The measurement of tropospheric OH radicals by laser-induced fluorescence spectroscopy during the POPCORN field campaign, Geophys. Res. Lett., 23, 2541-2544, 1996.
Hofzumahaus A, Aschmutat U, Brandenburger U, Brauers T, Dorn HP, Hausmann M, Hessling M, Holland F, Plass-Dulmer C, Ehhalt DH, Intercomparison of tropospheric OH measurements by different laser techniques during the POPCORN campaign 1994, J. Atm. Chem., 31, 227-246, 1998.
Holton, J. R., P.H. Haynes, M.E. McIntyre, A.R. Douglass, R.B. Rood, and L. Pfister, Stratosphere-troposphere exchange, Rev. Geophys., 33, 403-439, 1995.
Horowitz, L.W., Liang, J.Y., Gardner, G.M., and Jacob, D.J., Export of reactive nitrogen from North America during summertime: Sensitivity to hydrocarbon chemistry, J. Geophys. Res, 103, 13451-13476, 1998.
Houghton, J.T., L.G.Meira Filho, B.A.Callender, N.Harris, A.Kattenberg and K.Maskell (1996) Climate Change 1995. The science of climate change. Cambridge University Press, Cambridge, United Kingdom, 572 p.
Hov, �. (ed.) (1997) Tropospheric ozone research. Vol. 6 in Transport and chemical transformation of pollutants in the troposphere. Springer, Berlin, 499 p.
Hov, �. and F.Flat�y, Convective redistribution of ozone and oxides of nitrogen in the troposphere over Europe in summer and fall. J. Atmos. Chem., 28, 319-337,1997.
Hov, �., F. Flat�y, J.S. Foot, K. Dewey, H. Richer, A. Kaye, J. Kent, D. Tiddeman, I. Hawke, D. Kley, C. Gerbig, S. Schmitgen, S.A. Penkett, G. Mills, S. Bauguitte, J.M. Baker, B.J. Bandy, R.A. Burgess, T.J. Green, P.S. Monks, H.M. McIntyre, C.E. Reeves, F. Stordal, N. Schmidbauer, S. Solberg, M. Beekmann, E. Dugault, M. Lattuati, K.S. Law, J.A. Pyle, M.J. Evans, M. Cahill, P.-H. Plantevin and D. Shallcross, Testing atmospheric chemistry in anticyclones (TACIA). Final report to the European Commission on contract ENV4-CT95-0038, NILU, NO-2027 Kjeller, Norway, 164 p, 1998.
Hu, J., and D.H. Stedman, Atmospheric RO(x) radicals at an urban site -comparison to a simple theoretical-model, Environ. Sci. Technol. 29, 1655-1659, 1995.
Imhoff, R.E., R. Valente, J.F. Meagher, and M. Luria, The production of O-3 in an urban plume - airborne sampling of the Atlanta urban plume, Atmos. Environ., 29, 2349-2358, 1995.
Jacob, D.J., J.A. Logan, and P.P. Murti, Effect of rising Asian emissions on surface ozone in the United States, Geophys. Res. Lett, 26, 2175-2178, 1999.
Jaegl� L., D.J. Jacob, P.O. Wennberg, C.M. Spivakovsky, T.F. Hanisco, E.J. Lanzendorf, E.J. Hintsa, D.W. Fahey, E.R. Keim, M.H. Proffitt, E.L. Atlas, F. Flocke, S. Schauffler, C.T. McElroy, C. Midwinter, L. Pfister and J.C. Wilson, Observed OH and HO₂ in the upper troposphere suggest a major source from convective injection of peroxides, Geophys. Res. Lett., 24, 3181-3184, 1997.
Jaegle L, D.J. Jacob, W.H. Brune, D. Tan, I.C. Faloona, A.J. Weinheimer, B.A. Ridley, T.L. Campos, G.W. Sachse, Sources of HOx and production of ozone in the upper troposphere over the United States, Geophys. Res. Lett, 25, 1709-1712, 1998. (not in text)
Jaegle L, D.J. Jacob, W.H. Brune, I.C. Faloona, D. Tan, Y. Kondo, G.W. Sachse, B. Anderson, G.L. Gregory, S. Vay, H.B. Singh, D.R. Blake, and R. Shetter, 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, 1999.
Jaffe, D.A., R.E. Honrath, L. Zhang, H. Akimoto, A. Shimizu, H. Mukai, K. Murano, S. Hatakeyama and J. Merrill, Measurements of NO, NO_y , CO and O₃ and estimation of the ozone production rate at Oki Island, Japan, during PEM-West, J. Geophys. Res., 101, 2037-2048, 1996.
Jonquieres, I., A. Marenco, A. Maalej, and F. Rohrer, Study of ozone formation and transatlantic transport from biomass burning emissions over West Africa during the airborne Tropospheric Ozone Campaigns TROPOZ I and TROPOZ II, J. Geophys. Res., 103, 19,059-19,073, 1998.
Kanaya Y, Sadanaga Y, Matsumoto J, Sharma UK, Hirokawa J, Kajii Y, Akimoto H, Nighttime observation of the HO2 radical by an LIF instrument at Oki island, Japan, and its possible origins Geophys. Res. Lett.,26, 2179-2182, 1999.
Kasibhatla, P., H. Levy II, A. Klonecki and W.L. Chameides, Three-dimensional view of the large-scale tropospheric ozone distribution over the North Atlantic Ocean during the summer. J. Geophys. Res., 101, 29305-29316, 1996.
Kaye, J. and S.A. Penkett (co-chairmen), NASA Report No. 1339 on Concentrations, Lifetimes and Trends of CFCs, Halons and Related Species, January 1994.
Keene, W.C., A.A.P. Pszenny, D.J. Jacob, R.A. Duce, J.N. Galloway, J.J. Schultz-Tokos, H. Sievering,H. and J.F. Boatman, Global Biogeochem. Cycles, 4, 407,1990.
Kirchoff, V.W.J.H., J.R. Alves, F.R. Dasilva, and J. Fishman, Observations of O₃ concentrations in the Brazilian cerrado during the TRACE-A field expedition, J. Geophys. Res., 101, 24,029-24,042, 1996. Note in text
Kley, D., P.J. Crutzen, H.G.J. Smit, H. Vomel, S.J. Oltmans, H. Grassl and V. Ramanathan, Science, 274, 230-233, 1996.
Law, K.S., P.-H. Plantevin, D.E. Shallcross, H. Rogers, C. Grouhel, V. Thouret, A. Marenco, and J.A. Pyle, Evaluation of modeled O₃ using MOZAIC data, J. Geophys. Res., 103, 25,721-25,740, 1998.
Law, K.S., P-H. Plantevin, W.A.H. Asman, M. Lawrence, V. Thouret, C. Grouhel, D. Hauglustine, J-F. Muller, M. Kanakidou and A. Marenco, Comparison between global chemistry transport model results and measurement of ozone and water vapor by Airbus in-service aircraft (MOZAIC) data, in press, J. Geophys. Res., 105, 1503-1525, 2000.
Leighton's book
Lelieveld, J., A.M. Thompson, R.D. Diab, �. Hov, D. Kley, J.A. Logan, O.J. Nielsen, W.R. Stockwell, and X. Zhou, Tropospheric ozone and related processes, in Scientific Assessment of Ozone Depletion: 1998, World Meteorological Organization, Report No. 44, pp. 8.1-8.42, Geneva, 1999.
Lelieveld, J., and F.J. Dentener, What controls tropospheric ozone? J. Geophys. Res., 105, 3531-3551, 2000.
Levy, H., Normal atmosphere: large radical and formaldehyde concentrations predicted, Science 173, 141-143, 1971
Levy, H., Photochemistry of the lower troposphere, Planet. and Space Sci., 20, 919-935, 1972.
Levy, H., II, W. J. Moxim, A. A. Klonecki, and P. S. Kasibhatla, Simulated tropospheric NOx: Its evaluation, global distribution, and individual source contributions, J. Geophys. Res., 104, 26279-26306, 1999.
[Levy Ref].page 27
Liang, J., L.W.Horowitz, D.J.Jacob, Y.Wang, A.M.Fiore, J.A.Logan, G.M.Gardner and J.W.Munger, Seasonal variations of reactive nitrogen species and ozone over the United States, and export fluxes to the global atmosphere. J. Geophys. Res., 103, 13435-13450, 1998.
Lin, X., B.A.Ridley, J.Walega, G.F.H�bler, S.A.McKeen, E.-Y.Hsie, M.Trainer, F.C.Fehsenfeld and S.C.Liu, Parameterization of subgrid scale convective cloud transport in a mesoscale regional chemistry model, J. Geophys. Res., 99, 25615-25630, 1994.
Lin, X., M.Trainer and E.-Y.Hsie, A modeling study of tropospheric species during the North Atlantic Regional Experiment (NARE), J. Geophys. Res., 103, 13593-13614, 1998.
Liu, S.C., M.Trainer, F.C.Fehsenfeld, D.D.Parrish, E.J.Williams, D.W.Fahey, G.H�bler and P.C.Murphy (1987) Ozone production in the rural atmosphere and the implications for regional and global ozone distribution. J.Geophys. Res., 92, 4191-4207, 1987.
Mak, J.E., C.A.M. Brenninkmeijer, and M.R. Manning, Eevidence for a missing carbon-monoxide sink based on tropospheric measurements of (CO)-C-14, Geophys. Res. Lett., 19, 1467-1470, 1992.
Mak, J.E., C.A.M. Brenninkmeijer, and J. Tamaresis, Atmospheric (CO)-C-14 observations and their use for estimating carbon monoxide removal rates, J. Geophys. Res., 99, 22915-22922, 1994.
Marenco, A., H. Gouget, P. Nedelec, J.P. Pages, F. Karcher, Evidence of a long-term increase in tropospheric ozone from Pic du Midi data series-consequences - positive radiative forcing, J. Geophys. Res., 99, 16617- 16632, 1994.
Marenco, A., et al., Measurement of O3 and water vapor by Airbus in-service aircraft: The MOZAIC airborne program, an overview, J. Geophys. Res., 103, 25,631-25,642, 1998.
Mather et al., 1997 Cannot find on web - too many names with Mather in them.
Mauldin, R.L., S. Madronich, S.J. Flocke, F.L. Eisele, G.J. Frost, and A.S.H. Prevot, New insights on OH: Measurements around and in clouds, Geophys. Res. Lett., 24, 3033-3036, 1997.
Mauldin, R.L., G.J. Frost, G. Chen, D.J. Tanner, A.S.H. Prevot, D.D. Davis, and F.L. Eisele, OH measurements during the First Aerosol Characterization Experiment (ACE 1): Observations and model comparisons, J. Geophys. Res., 103, 16713-16729, 1998.
Mauldin, RL, D.J. Tanner DJ, Eisele FL, Measurements of OH during PEM-Tropics A, J. Geophys. Res., 104, 5817-5827, 1999.
McKeen, S.A., G. Mount, F. Eisele, E. Williams, J. Harder, P. Goldan, W. Kuster, S.C. Liu, K. Baumann, D. Tanner, A. Fried, S. Sewell, C. Cantrell, R. Shetter, Photochemical modeling of hydroxyl and its relationship to other species during the Tropospheric OH Photochemistry Experiment, J. Geophys. Res., 102, 6467-6493, 1997.
Meagher, J.F., E.B. Cowling, F.C. Fehsenfeld, and W.J. Parkhurst, Ozone formation and transport in southeastern United States: Overview of the SOS Nashville Middle Tennessee Ozone Study, J. Geophys. Res., 103, 22213-22223, 1998.
Mihelic, D., D. Klemp, P.M�sgen, H.W. P�tz and A. Volz-Thomas, Simultaneous measurements of peroxy and nitrate radicals at Schauinsland, J. Atmos. Chem., 16, 313-335, 1993.
Mill�n, M.M., R.Salvador, E.Mantilla and B.Art��ano, Meteorology and photochemical air pollution in southern Europe: Experimental results from EC research projects, Atmos. Environ., 30, 1909-1924, 1996.
Mill�n, M.M., R.Salvador, E.Mantilla and G.Kallos, Photooxidant dynamics in the Mediterranean basin in summer: Results from European research projects, J. Geophys. Res., 102, 8811-8823, 1997.
Miyoshi, A., S. Hatakeyama, and N. Washida, OH radical-initiated photooxidation of isoprene: An estimate of global CO production, J. Geophys. Res., 99, 18779-18787, 1994.
Monks, P.S., L.J. Carpenter, S.A Penkett and G.P. Ayers, Night-time peroxy radical chemistry in the remote marine boundary layer over the Southern Ocean, Geophys. Res. Lett. 23, 535-538, 1996.
Monks, P.S., L.J. Carpenter, S.A. Penkett, G.P. Ayers, R.W. Gillett, I.E. Galbally and C.P. Meyer, Fundamental ozone photochemistry in the remote marine boundary layer: the SOAPEX experiment, measurement and theory, Atmos. Environ., 32, 3647-3664, 1998.
Monks, P.S., G .P. A yer s, G. H olland, S.A. Penkett, and G. S alisbury, need title, Atmos.Env., 34, 2547- 2561, 2000.
Murphy and Fahey (199..) Is it the one below?
MURPHY DM, FAHEY DW, MATHEMATICAL TREATMENT OF THE WALL LOSS OF A TRACE SPECIES IN DENUDER AND CATALYTIC-CONVERTER TUBES, ANAL CHEM 59: (23) 2753- 2759 DEC 1 1987
Novelli, P. C., K. A. Masarie, and P. M. Lang, Distributions and recent changes of carbon monoxide in the lower troposphere, J. Geophys. Res., 103, 19015-19033, 1998.
Noxon, J.F., R.B. Norton and E. Marovich, NO 3 in the troposphere, Geophys. Res. Lett., 7, 125-128, 1980.
Noxon, J.F., NO₃ and NO₂ in the Mid-Pacific Troposphere, J. Geophys. Res., 88, 11017-11021, 1983.
Oltmans, S.J., and H. Levy II, Surface O₃ measurements from a global network, Atmos. Environ., 28, 9-24, 1994.
Parrish, D.D., M. Trainer, J.S. Holloway, J.E. Yee, M.S. Warshawsky, F.C. Fehsenfeld, G.L. Forbes, and J.L. Moody, Relationships between ozone and carbon monoxide at surface sites in the North Atlantic region, J. Geophys. Res., 103, 13357-13376, 1998.
Penkett, S.A., B.J. Bandy, C.E. Reeves, D. McKenna, and P. Hignett, Measurements of Peroxides in the Atmosphere and their Relevance to the Understanding of Global Tropospheric Chemistry, Faraday Discussions, 100, 155-174, 1995.
Penkett, S.A., P.S. Monks. L.J. Carpenter, K.C. Clemitshaw, G.P. Ayers, R.W. Gillet, I.E. Galbally, and C.P. Meyer, Relationships between ozone photolysis rates and peroxy radical concentrations in clean marine air over the Southern Ocean, J. Geophys. Res., 102, 12,805-12,817, 1997.
Penkett, S.A., C.E. Reeves, B.J. Bandy, J.M. Kent, H.R. Richer, Comparison of calculated and measured peroxide data collected in marine air to investigate prominent features of the annual cycle of ozone in the troposphere, J. Geophys. Res., 104, (D11), 13377-13388, 1998.
Perner on page 13, line 8
Perner, D., Ehhalt, D.H., Paetz, H.W., Platt U., Roth, E.P., and Voltz, A., OH radicals in the lower troposphere, Geophys. Res. Lett., 3, 466-468, 1976.
Perner, D., Platt U., M.Trainor, G.Hubler, J.Drummond, W.Junkermann, J.Rudolph, B.Schubert, A.Volz and D.H. Ehhalt, Measurements of tropopheric OH concentrations; A comparison with Field data with model predictions, J.Atmos.Chem. 5, 185-216, 1987.
Pickering, K.E., R.R. Dickerson, G.J. Huffman, J.F. Boatman and A. Schanot, Trace gas transport in the vicinity of frontal convective clouds, J. Geophys. Res., 93, 759-773, 1988.
Pickering, K.E., A.M. Thompson, J.R. Scala, W.-K. Tao, R.R. Dickerson and J. Simpson, Free tropospheric ozone production following entrainment of urban plumes into deep convection, J.Geophys.Res., 97, 17985-18000, 1992.
Plane, J.M.C., and C.-F. Nien, A Study of Nighttime NO₃ Chemistry by Differential Optical Absorption Spectroscopy. In: H.I. Schiff, Measurement of Atmospheric Gases (The International Society for Optical Engineering, Bellingham, 1991), 8-20.
Plass-Dulmer C., T. Brauers T, J. Rudolph, POPCORN: A field study of photochemistry in North-Eastern Germany, J. Atm. Chem., 31, 5-31, 1998. not mentioned in text
Platt, U., D. Perner, G.W. Harris, A.M. Winer and J.N. Pitts, Detection of NO₃ in the polluted troposphere by differential optical absorption, Geophys. Res. Lett., 7, 89-92, 1980.
Platt, U., D. Perner, J. Schroder, C. Kessler and A. Toennissen, The diurnal variation of NO3 J. Geophys. Res., 86, 11965-11970, 1981.
Platt, U., A.M. Winer, H.W. Biermann, R. Atkinson, and J.N. Pitts Jr., Measurement of nitrate radical concentrations in continental air, Environ. Sci. Technol., 18, 365-369, 1984.
Platt, U., and F. Heintz, Nitrate radicals in tropospheric chemistry, Israel J. Chem., 34, 289-300, 1994.
Prinn et al 1995
Prinn et al.
Rao, S.T., I.G.Zurbenko, R.Neagu, P.S.Porter, J.Y.Ku and R.F.Henry, Space and time scales in ambient ozone data. Bull. Am. Met. Soc. 78, 2153- 2166, 1997.
Reiner T, Hanke M, Arnold F, Atmospheric peroxy radical measurements by ion molecule reaction mass spectrometry: A novel analytical method using amplifying, chemical conversion to sulfuric acid, J. Geophys. Res., 102, 1311-1326, 1997.
Reiner T, M. Hanke, F. Arnold, H. Ziereis, H. Schlager, and W. Junkerman W., Aircraft-borne measurements of peroxy radicals by chemical conversion/ion molecule reaction mass spectrometry: Calibration, diagnostics, and results, J. Geophys. Res., 104, 18,647-18,659, 1999.
Ridley et al., 1987; WHICH OF THE 1987 ONES BELOW?
Ridley, B.A., M. Mcfarland, A.L. Schmeltekopf, M.H. Proffitt, D.L. Albritton, R.H. T.L. Thompson, Seasonal differences in the vertical distributions of NO, NO₂, and O₃ in the stratosphere near 50�N, J. Geophys. Res., 92, 11,919-11,929, 1987.
Ridley, B.A., M.A. Carroll, G.L. Gregory, Measurements of nitric oxide in the boundary layer and free troposphere over the Pacific Ocean, J. Geophys. Res., 92, 2025-2047, 1987.
Ridley, B.A., and E. Robinson, The Mauna-Loa Observatory photochemistry experiment, J. Geophys. Res., 97, 10285-10290, 1992.
Ridley et al., page 58
Roberts et al., 1998 - Too many to find on the web.
Roelofs, G.-J., and J. Lelieveld, Model study of the influence of cross-tropopause O3 transports on tropospheric O3 levels, Tellus, 49B, 38-55, 1997.
Rossow, W.B. and R.A. Schiffer, ISCCP cloud data sets, Bull. Amer. Meteor. Soc., 72, 2-20, 1991.
Rudolph, J., and D. H. Ehhalt, Measurements of C2-C5 hydrocarbons over the north Atlantic, J. Geophys. Res., 86, 11959-11964, 1981.
Rudolph, J., and F. J. Johnen, Measurements of light atmospheric hydrocarbons over the Atlantic in regions of low biological activity, J. Geophys. Res., 95, 20583-20591, 1990.
Schall, C. and K.G. Heumann, GC determination of volatile organoiodine and organobromine compounds in Arctic seawater and air samples, Fres. J. Anal. Chem., 346, 717-722, 1993.
Schaller, E., and A.Wenzel, Evaluierung regionaler atmosph�rischer Chemie-Transport- Modelle. Fall 1: TRACT, 16.9.1992, 13-17 Uhr MESZ, Lehrstuhl f�r Umweltmeteorologie, Brandenburgische Technische Universit�t, Cottbus, Germany, 89p, 1999.
Schiffer, R.A. and W.B.Rossow, The International Satellite Cloud Climatology Project (ISCCP): The first project of the World Climate Research Program. Bull., Amer. Meteor. Soc., 64, 779-784, 1983.
Schlager, H., P. Schulte, F . Flatoy, F. Slemr, P. van Velthoven, H. Zier eis, U. Schumann, Regional nitric oxide enhancements in the North A tlantic flight corr idor obs er ved and modeled during PO LIN AT 2 - A cas e s tudy, G eophys Res. Lett., 26, 3061-3064, 1999.
Schultz et al., 1999 - Cannot find - need more information
Schumann, U., et al., Pollution fr om aircr af t emissions in the N orth Atlantic flight corridor: Overview on the P OLI NA T projects , J ournal Geophys. Res., in pres s, 2000.
Sillman S., D.Y. He, M.R. Pippin, P.H. Daum, D.G. Imre, L.I. Kleinman, J.H. Lee, J. Weinstein-Lloyd, Model correlations for ozone, reactive nitrogen, and peroxides for Nashville in comparison with measurements: Implications for O3-NOx-hydrocarbon chemistry, J. Geophys. Res., 103, 22,629-22,644, 1998.
Simmonds, P.G., S. Seuring, G. Nickless, and R.G. Derwent, Segregation and interpretation of ozone and carbon monoxide measurements by air mass origin at the TOR Station Mace Head, Ireland from 1987 to 1995, J. Atmos. Chem., 28(1-3), 45-59, 1997.
Simpson, D. (1992) Long-period modelling of photochemical oxidants in Europe. Model calculations for July 1985. Atmos. Envir., 26A, 1609-1634. - not mentioned in text
Singh et al. and others
Singh and others 1975
Singh, H.B., Atmospheric halocarbons; evidence in favour of a reduced average hydroxyl radical concentrations in the troposphere, Geophys. Res. Lett., 4, 101-104, 1977
Singh, H. B., and L. J. Salas, Measurement of selected light hydrocarbons over the Pacific Ocean: latitudinal and seasonal variations, Geophys. Res. Lett., 9, 842-845, 1982.
Singh, H.B., D. Herlth, R. Koyler, R. Chatfield, W. Viezee, L.J. Salas, Y. Chen, J.D. Bradshaw, S.T. Sandholm, R. Talbot, G.L. Gregory, B. Anderson, G.W. Sachse, E. Browell, A. S. Bachmeier, D.R. Blake, B. Heikes, D. Jacob, and H. Fuelberg, Impact of biomass burning emissions on the composition of the South Atlantic troposphere: Reactive nitrogen and ozone, J. Geophys. Res., 101, D19, 24,203-24,219, 1996.
Singh, H.B., A.M. Thompson, H. Schlager, 1997 SONEX airborne mission and coordinated POLINAT-2 activity: Overview and accomplishments, Geophys. Res. Lett., 26, 3053-3056, 1999.
Smith, N., J.M.C. Plane, C.-F. Nien, and P.A. Solomon, Nighttime radical chemistry in the San Joaquin Valley, Atmos. Environ., 29, 2887-2897, 1995.
Spivakovsky, C.M., S.C. Wofsy, M.J. Prather, A numerical method for parameterization of atmospheric chemistry - computation of tropospheric OH, J. Geophys. Res., 95, 18433-18439, 1990.
Stevens et al., 1997 Need more information
Suhre, K., J.P. Cammas, P. Nedelec, R. Rosset, A. Marenco, and H.G.J. Smit, Ozone-rich transients in the upper equatorial Atlantic troposphere, Nature, 388, 661-663, 1997.
Tan D, Faloona I, Brune WH, Weinheimer A, Campos T, Ridley B, Vay S, Collins J, Sachse G, In situ measurements of HOx in aircraft exhaust plumes and contrails during SUCCESS, Geophys. Res. Lett., 25, 1721- 1724, 1998.
Thompson, A.M., K.E. Pickering, D.P. McNamara, M.R. Schoeberl, R.D. Hudson, J.H. Kim, E.V. Browell, V.W.J.H. Kirchoff, and D. Nganga, Where did tropospheric ozone over southern Africa and the tropical Atlantic come from in October 1992? Insights from TOMS, GTE TRACE-A and SAFARI 1992, J. Geophys. Res., 101, 24,251-24,278, 1996.
Thompson, A.M., K.E.Pickering, R.R.Dickerson, W.G.Ellis Jr., D.J.Jacob, J.R.Scala, E.-K.Tao, D.P.McNamara and J.Simpson, Convective transport over the central United States and its role in regional CO and ozone budgets, J.Geophys.Res., 99, 18703-18711, 1994.
Thouret, V., A. Marenco, P. Nedelec, and C. Grouhel, Ozone climatologies at 9-12~km altitude as seen by the MOZAIC airborne program between September 1994 and August 1996, J. Geophys. Res., 103, 25,653-25,679, 1998a.
Thouret, V., A. Marenco, P. Nedelec, C. Grouhel, and J. Logan, Comparisons of O₃ measurements from the MOZAIC airborne program and the O₃ sounding network at eight locations, J. Geophys. Res., 103, 25,695- 25,720, 1998b.
Tilmes, S. and J.Zimmermann, Investigation on the spatial scales of the variability in measured near-ground ozone mixing ratios, Geophys.Res.Lett. 25, 3827-3830, 1998.
Torres, A. L., and A. M. Thompson, Nitric oxide in the equatorial Pacific boundary layer: SAGA 3 measurements, J. Geophys. Res., 98, 16949- 16954, 1993.
Trainer on page 63.
Trainer et al., 1987 - which of the two below.
Trainer, M., E.Y. Hsie, S.A. McKeen, R. Tallamraju, D.D. Parrish, F.C. Fehsenfeld, S.C. Liu, Impact of natural hydrocarbons on hydroxyl and peroxy radicals at a remote site, J. Geophys. Res., 92, 11879-11894, 1987.
Trainer, M., E.J. Williams, D.D. Parrish, M.P. Buhr, E.J. Allwine, H.H. Westberg, F.C. Fehsenfeld, S.C. Liu, Models and observations of the impact of natural hydrocarbons on rural ozone, Nature, 329, 705-707, 1987.
Tuck, A. F. et al., Strat-trop exchange, in Atmospheric ozone 1985, World Meteorological Organization, Report No. 16, pp. 151-240, Geneva, 1985.
Vogt, R. and F.P. Finlayson-Pitts, A diffuse reflectance infrared Fourier-transform spectroscopic (drifts) study of the surface reaction of NaCl with gaseous NO₂ and HNO₃, J. Phys. Chem., 98, 3747-3755, 1994.
Vogt, R., et al., 1994 (is this correct)
Vogt, R., P.J. Crutzen, and R. Sander, A mechanism for halogen release from sea-salt aerosol in the remote marine boundary layer, Nature, 383, 327- 330, 1996.
Vogt, R., R. Sander, R. Von Glasow and P.J. Crutzen, Iodine chemistry and its role in halogen activation and ozone loss in the marine boundary layer: A model study, J. Atmos. Chem., 32, 375-395, 1999.
Volpe, C., M. Wahlen, A.A.P. Pszenny and A.J. Spivack, Chlorine isotopic composition of marine aerosols: Implications for the release of reactive chlorine and HCl cycling rates, Geophys. Res. Lett., 25, 3831-3834, 1998.
Wang, C.C., Davis L.I., C.H.Wu, S.Japar, H.Niki and Weinstock, B, Hydroxyl radical concentration measured in ambient air, Science 189, 798-800, 1975. (NOT in TEXT)
Wang, Y., D. J Jacob, and J. A. Logan, Global simulation of tropospheric O₃-NO_x- hydrocarbon chemistry, 1, Model formulation, J. Geophys. Res., 103, 10713-10725, 1998a.
Wang, Y., J. A. Logan, and D. J Jacob, Global simulation of tropospheric O₃-NO_x- hydrocarbon chemistry, 2, Model evaluation and global ozone budget, J. Geophys. Res., 103, 10727-10755, 1998b.
Weaver et al., 1996; Not possible to find without more information
Weinstock, B., Carbon Monoxide, Residence time in the atmosphere, Science, 166, 224-225, 1969.
Weinstock and Niki 1968 Need more information
Weinstock, B., and Niki, H., Carbon Monoxide balance in nature, Science, 176, 290-292, 1972.
Wild, O., K.S.Law, D.S.McKenna, B.J.Bandy, S.A.Penkett and J.A.Pyle, Photochemical trajectory modeling studies of the North Atlantic region during August 1993, J.Geophys.Res., 101, 29269-29288, 1996.
WMO 1994
WMO (1999) Scientific assessment of ozone depletion: 1998. World Meteorological Organization, Global ozone research and monitoring project Report No. 44, Geneva, Switzerland.
Yienger, J., and H. Levy II, Emperical model of global soil-biogenic NOx emissions, J. Geophys, Res., 100, 11447-11464, 1995.
Yvon, S.A., J.M.C. Plane, C.-F. Nien, D.J. Cooper and E.S. Saltzman, Interactions between nitrogen and sulfur cycles in the polluted marine boundary layer, J. Geophys. Res., 101, 1379-1386, 1996.
Zanis, P., P.S. Monks, E. Schuepbach and S.A. Penkett, On the relationship of HO2+RO2 during the Free Tropospheric Experiment (FREETEX 96) at the Jungfraujoch Observatory (3580 m above sea level) in the Swiss Alps, J. Geophys., 104, 26,913-26,926, 1999.
Zanis, P., P.S. Monks, E. Schuepbach, L.J. Carpenter, T.J. Green, G.P. Mills, S. Bauguitte, and S.A. Penkett, In-situ Ozone Production under Free Tropospheric conditions during FREETEX 98 in the Swiss Alps, submitted to J. Geophys. Res. 2000.
Zenker T, H. Fischer, C. Nikitas, U. Parchatka, G.W. Harris, D. Mihelcic, P. Musgen, H.W. Patz, M. Schultz, A. Volz-Thomas, R. Schmitt, T. Behmann, M.Weissenmayer, J.P. Burrows, Intercomparison of NO, NO2, NOy, O3, and ROx measurements during the oxidizing capacity of the tropospheric atmosphere(OCTA) campaign 1993 at Izana, J. Geophys. Res., 103, 13615-13634, 1998.

Last modified: Wed Apr 26 09:28:30 CEST 2000

Experiment	Location, Year	Technique(s)	[HO₂] or S[RO₂] pptv	Comments	Reference(s)
	Oregon, 1986, 1987	LIF (HO₂)	8)	One coastal site, one urban site; similar maxima at both sites for clear-sky	Hard et al., [1992]
	Schauinsland, 1990	MIESR	40	Forested location; anti-correlation between RO₂ and NO₃ at night	Mihelcic et al., [1993]
ROSE 1	Alabama, 1990	PERCA	200	Forested location; measured levels compared well with PSS calculations	Cantrell et al., [1992, 1993]
MLOPEX 2	Mauna Loa, 1991-2	PERCA	25	Free troposphere; upslope/downslo pe Effects; measurements low compared with PSS calculations; peroxy radicals observed at night	Cantrell et al., [1996a, 1996b, 1997]
SONTOS	Ontario, 1992	PERCA	23	Rural site	Arias and Hastie, [1996]
OCTA	Iza�a, Tenerife, 1993	3 PERCA instruments and 1 MIESR instrument	80-100	2 PERCA instruments within 25% of MIESR instrument, 1 systematically low	Zenker et al., [1998]
	Denver, Colorado, 1993	PERCA	20-80	Urban site; reactions of O₃ with alkenes Claimed as important source of peroxy radicals at night	Hu and Stedman, [1995]
FIELDVOC	Brittany, 1993	PERCA	60	Measurements much lower than PSS calculations; low night-time levels	Cantrell et al., [1996]
LAFRE	Los Angeles, 1993	LIF (HO₂)	8	Measurements in smog; poor agreement with simple model at midday	George et al., [1999 ]
TOHPE	Idaho Hill, Colorado, 1993	LIF(HO₂), PERCA	6_(HO₂), 60	RO₂/HO₂ ratio 4-15 times larger than predicted; HO₂/OH ratio in range 15-80: agreed well with theory for [NO] > 100 pptv; factor of 3-4 too low under clean conditions	Cantrell et al., [1997]; Stevens et al., [1997]
PRICE I	Schauinsland, 1994	PERCA MIESR i	25 (PERCA), 50 (MIESR)	No systematic difference between PERCA instruments; MIESR instrument measured systematically higher	Carpenter, [1996]
WAO-TIGER	Weybourne, UK, 1993-5	PERCA	12	Positive correlation of peroxy radicals and NO₃ observed at night	Carpenter, [1996], Clemitshaw et al. [1997]; Carslaw et al., [1997]; Carpenter et al., [1998]
BESSE	Bush Estate, Edinburgh, 1994	PERCA	15	Measurements systematically low compared with PSS Calculations	Carpenter, [1996]
SOAPEX I	Cape Grim, Tasmania, 1995	PERCA	11	Relationships of peroxy radicals with j_(O¹D) in clean and polluted air; O₃ production /destruction compensation point ca. 25 pptv NO_x; CH₃O₂ at night; winter / summer comparison; measurements systematically low by factor of 2-3	Monks et al., [1996]; Penkett et al., [1997]; Cox, [1999]; Monks et al., [2000]
ATAPEX	Mace Head, Ireland, 1995	PERCA	5	Compensation point ca. 50 pptv NO	Carpenter et al., [1997]
STRAT	Aircraft, 1995-6	LIF (HO₂)	Measurements in upper troposphere; Acetone photolysis as major HOx source in dry upper troposphere / lower stratosphere	Jaegl� et al., [1997]
SUCCESS	Aircraft, 1996	LIF (HO₂)	3-15	HO₂/OH ratios 30 % higher than modelled	Brune et al., [1998]
	Aircraft, 1996	SICIMS (HO₂)	10-40	4-8 km altitude; good agreement between measurements and steady-state calculations for clear skies	Reiner et al., [1997; 1999]
ACSOE-EASE	Mace Head, Ireland, 1996-7	LIF (HO₂), PERCA	8 _(HO2), 25	Effect of NO_x on radical levels; control of peroxy radicals at night by NO₃ / O₃ reactions	Creasey et al., [1997]; Monks et al., [1999]
SONEX	Aircraft, 1997	LIF (HO₂		Effect of aircraft emissions in troposphere	Brune et al., [1999]; Jaegl� et al., [1999]
FREETEX	Jungfraujoch, Switzerland, 1996, 1998	PERCA	18	Radicals shown to be dependent on V(j(O¹D)) in low-NO_x ozone- production regime; seasonal comparison; photochemical control of diurnal ozone variation	Zanis et al., [1999; 2000]
	Oki Island, Japan, 1998	LIF (HO₂)	14	HO₂ observed at night	Kanaya et al., [1999]
BERLIOZ	Berlin, 1998	LIF (HO₂		No results yet published
PROPHET	Michigan,	LIF (HO₂)		No results yet published
SOAPEX 2	Cape Grim, Tasmania, 1999	LIF (HO₂) PERCA	No results yet published
PUMA	Birmingham, 1999	LIF (HO₂)	No results yet published
PRIME	Ascot, 1999	PERCA Instrument LIF (HO₂)		No results yet published

	Eastern North Atlantic			Continental Europe
	NO_x	NO_y	ozone	NO_x	NO_y	ozone
[c], ppt (dc/dt)chem, ppt/h (dc/dt)dep, ppt/h (dc/dt)v.adv, ppt/h (dc/dt)conv, ppt/h emission, ppt/h (v.adv+conv)/emis	56 -2.6 small -0.0 -0.0 0.7 0.09 (0.32)	474 small -7.4 -2.0 0.1 0.7 2.9 (0.15)	31.4 -36 -29 -57 3.5	1391 -214 small -5.3 -18 290 0.08 (0.07)	3373 small -113.5 -23.4 -47 290 0.24 (0.11)	47.6 1075 -351 -194 -128