5.5 Characterisation of the ozone phenomenology

5.5 Characterisation of the ozone phenomenology from a meteorological perspective

Surface ozone concentrations depend on complex combinations of chemical processes and meteorological influences. Characterising the ozone phenomenology in the European Union is therefore essential to better assess the impact of man-made emissions. The following sections give an overview of the main results of JRC-ERLAP's study concerning the interpretation of widespread ozone episodes observed in the European Union in 1994, 1995 and 1996 (Koffi et al., 1998). The first two sections focus on the characterisation of the Europe scale meteorology associated with exceedances of the population information threshold (i.e. concentrations above 180 mg.m^-3). The contribution to observed ozone levels of the photochemical formation, horizontal advection and vertical transport processes are then discussed for eight selected episodes, on the basis of additional chemical and meteorological measurements and backward trajectory studies. The tables and figures mentioned in this section are given in Appendix 1.

5.5.1 Ozone exceedances and weather type over Middle Europe

The Grosswetterlagen (GWL) daily classification from Deutscher Wetterdienst in Offenbach (Hess and Brezowsky, 1969) was firstly used as a convenient description of the synoptic meteorology that prevailed during 1994, 1995 and 1996 growing seasons (i.e. April to August periods). It defines 29 different synoptic weather situations over Middle Europe, in terms of cyclonic/anticyclonic circulations and predominant wind direction (Table I.1):

• More than 89% of the exceedances reported in Austria, Belgium, Denmark, Finland, France, Germany, Luxembourg, the Netherlands, Sweden and United Kingdom were observed during anticyclonic conditions over middle Europe (Table I.2).
• The situation for which more ozone exceedances were reported at a European scale correspond to a ridge of high pressure over Middle Europe. Previous studies focusing on Germany and the Netherlands (e.g. Davies et al., 1992; Guicherit and van Dop, 1977) also showed high surface ozone levels to be mostly associated in summer time with a ridge of high pressure over Middle Europe. Nevertheless, this result is partly due to the fact that theses situations (BM situations) are the most frequently occurring.
• Calculating the number of exceedances per day of weather situation (Table I.3) shows that the differences in the exceedances occurrence as a function of the predominant wind observed between the above-mentioned Member States can be related, in a large part, to the influence of long-range transport of ozone and/or its precursors, i.e. whether regions of high precursor emissions are encountered in the catchment area of the air mass trajectories.
• United-Kingdom, Ireland, Greece, Spain, Portugal and, to a lesser extent, Italy show different trends compared to Middle and North Europe countries. This is firstly due to the fact that the GWL classification is not well adapted to define the meteorology systems that prevail in those outlying countries. Moreover, particular meso-scale meteorological conditions (e.g. longer periods of hot weather, higher incidence of sunlight, complex orography and daily cycles of local winds and sea/land breeze, for South Europe countries; strong oceanic influence for Portugal, United Kingdom and Ireland) also play a major role on the ozone concentrations as discussed in the following sections.

5.5.2 Ozone conducive meteorological parameters

Eight characteristic wide-spread ozone episodes were selected during the 1994, 1995 and 1996 growing seasons, in order to characterise the meteorological situations which are conducive to high ozone levels (Table I.4). The selection was made on the basis of the previous analysis and the temporal evolution of the exceedances of the population information threshold reported in each of the 15 Member States. Daily European scale thematic maps were prepared for the 62 days of the selected episodes for the following parameters (see examples in Figure I.1).

The main conclusions are:

• 850 mb pressure surface:

The beginning of the occurrence was always associated with the centre of an anticyclone, or with less well defined high pressure systems. The spatial and temporal evolution of the ozone exceedances area is shown to be usually associated to a corresponding displacement, in time and space, of the high pressure system. The end or decrease of the ozone exceedances were linked either to the arrival of low pressure systems over Europe or to a slackening of the anticyclone, or to a transition between two different but anticyclonic weather types over Middle Europe. Particular differing scenarios are also observed in the spatio-temporal distribution of the ozone exceedances that cannot be explained by the associated spatio-temporal evolution of the anticyclone. Moreover, horizontal advection can also be suspected during the eight episodes.

• Surface temperature and solar radiation:

A good correlation between the highest daily maximum surface temperature and the location of the ozone exceedances is observed for five out of the eight episodes, whereas the main location of exceedances coincided with high surface temperatures also during the three other. On the other hand, no clear connection is observed between the pollution area and the average diurnal surface solar radiation (12-18 hours UTC). The relative sunshine duration, or the daily maximum surface radiation might be more appropriate predictors. Unfortunately, no data is available for these parameters at the required spatial and temporal scales.

• Horizontal wind field:

The stations which reported ozone exceedances at the beginning of the episodes (i.e. the first and second days) were mainly located within areas with relatively low wind speeds in the lower tropospheric layer, whereas no obvious relationship was observed during the following days. The previous assumption about horizontal advection processes that would be involved during most of the selected episodes is supported in all cases by the wind speed and direction in the low and middle troposphere. The wind fields also provided useful information for the understanding of the above-mentioned special spatio-temporal evolution of the pollution area that were related to wind direction changes, from continental to oceanic, and vice versa. The persistence phenomenon observed during one of the selected episodes was shown to be associated with persistent low winds in the boundary layer, which induced stagnation/accumulation of the pollution.

Whilst the continental thematic maps provide a useful description of the European scale meteorological conditions associated with ozone episodes, they do not allow a detailed interpretation. However, local to synoptic meteorological effects are discernible from additional chemical and meteorological data and backward-trajectory studies, as illustrated in the three following sections.

5.5.3 Photochemical formation of ozone

As shown in the above section for Middle and North European countries, major episodes of high concentrations of ozone are associated with slow-moving, high-pressure systems. The often cloudless and warm conditions, associated with these systems are favourable for the photochemical production of ozone. These systems are also characterised by widespread sinking of air through most of the troposphere, which is warmed adiabatically and thus tends to produce a pronounced high elevation (1-3 km) inversion of the normal temperature profile serving as a strong lid to contain pollutants in a shallow layer in the troposphere. Moreover, because associated winds are generally light, there is greater chance for pollutants to accumulate in the atmospheric boundary layer (National Research Council, 1992). All these conditions lead to accumulation of ozone and other chemical pollutants from one day to the next. As an example, the wide-spread anticyclone observed during the episode of June 1996 ( during a period with a ridge of high pressure over Middle Europe) and the location of the ozone exceedances are shown in Figure I.2.

Previous studies focusing on South European countries showed that the diffusion, transport and chemical cycles of atmospheric pollutants in Mediterranean areas are significantly different from those observed in North Europe. The complex processes that cannot be highlighted from the sparse information collected in these countries in the framework of Ozone Directive, are summarised below:

• The MEDCAPHOT-TRACE experimental field campaign (Ziomas, 1998), which took place in summer 1994 clearly demonstrated the impact of land-sea-breeze-circulations on the occurrence of ozone episodes in the Athens Basin. During the warm period of the year, a special synoptic situation favours northerly winds. The weakening of these winds, called Etesians, allows the development of the land-sea-breeze-circulations, with a potential increase of air pollution levels: first the sea breeze tends to stratify the atmosphere above Athens thus trapping air pollutants at a relatively small height above ground. In addition, a recirculation of air pollutants takes place consisting of pollutant transport by the land breeze onto the sea and re-advection back to the basin by the sea breeze, resulting in a abrupt increase of pollutant concentrations in the Athens basin during the day.
• The results from the MECAPIP project indicate that meso-scale circulations generated in the Mediterranean area, play an important role in the formation of photo-oxidants over Spain and the Western Mediterranean (Millan et al., 1996). They include surface wind convergence over the Iberian peninsula, large-scale compensatory subsidence over the surrounding coastal areas and the formation of re-circulatory cells as a result of the sea breezes combining with up-slope winds and their return and compensatory flows. These processes combined with related specific air pollution dynamics form a large 'photochemical reactor' which operates almost every day from spring to fall and which can generate ozone levels 2 to 3 times higher than the Ozone Directive threshold for protection of vegetation (i.e. 65 mg.m^-3 for 24h average).

Examples of the impact of such meso-scale meteorological processes on the surface ozone levels observed in Spain and Greece during the years 1994, 1995 and 1996 are given below:

• In Table I.5 are reported the annual and seasonal average values of the daily maximum ozone concentrations for some selected stations in the EU, over the 1994-1995 period. Because O₃ concentrations in urban stations are lowered by the chemical destruction of ozone from NO traffic emissions, O_x (O₃ + NO₂) concentrations in urban stations are also reported. Statistics show significantly higher levels throughout the year in Athens and Barcelona compared to Berlin, whereas they are significantly higher throughout the year in Athens and during summer in Barcelona compared to London (Figure I.3).
• The effect of orography is shown in Figure I.4 for a rural site located in a sheltered zone of South of Pyrenees in Spain, where stagnation and vortex circulation of the air masses in the low troposphere induced particular daily evolution of the surface ozone concentration, with high concentrations also reported during the night.
• The interaction between the European and local scales weather situations on the photochemistry in the Athens basin is illustrated in Figure I.5 (Suppan, et al., 1998). The meteorological and photochemical situation in this region can be divided into three categories: (i) days with strong synoptic northerly Etesians winds which suppress the land/sea breeze and lead to low pollution level inside the basin, (ii) days with moderate northerly winds and development of weak sea breeze cells near the coast line, called normal days, and (iii) days with weak northerly winds and fully developed sea breeze cells affecting the whole basin of Athens, called sea breeze days. During sea breeze days, the ozone concentrations can be up to 57% higher than during normal days.

5.5.4 Long-range transport

A back-trajectory model was used to study the advection processes involved during the eight selected ozone episodes (Koffi, 1997). The backward trajectories calculated for twenty selected reception sites in nine out of the 15 Member States clearly demonstrate:

• the cross-border character of the ozone pollution, with long-range transport processes observed during all the selected episodes;
• that continental air masses usually induce an increase in ozone concentrations in favourable conditions, whereas oceanic air masses have a cleaning effect (see example for United Kingdom in Figure I.6);
• the high dependence of the ozone surface concentrations on the meeting of the air mass trajectories with areas of high emissions of ozone and/or its precursors (as shown for Germany in Figure I.7).

5.5.5 Downward transport from the upper troposphere and lower stratosphere

Accumulated air pollution in the higher troposphere and even stratospheric ozone intrusion can intrude the lower atmospheric layers, where by vertical turbulence they are mixed down to ambient air masses. There has been a continuing debate since the 1960s/1970s about the role of vertical transport vs. local photochemical formation with regard to the annual tropospheric ozone budget. The current consensus view is that in situ chemical production is the major contributor to the observed ozone levels in the ambient tropospheric air. Nevertheless, the contribution of ozone transport down from the free troposphere may not be negligible (Davies and Schuepbach, 1994).

The ozone episode of May 1995 is in fact a characteristic example of such vertical exchanges: The wind vertical velocity over Europe shows a subsidence of the air masses during the three days of high ozone level concentrations, with its centre over Benelux (Figure I.8). The concentrations of Beryllium 7 (a tracer of air coming from the upper troposphere) measured in Luxembourg in 1995 show a significant peak during the corresponding days (Figure I.9). Moreover, peaks of ⁷Be concentrations are also observed during the 1995 summer ozone episodes, which suggest a contribution of free tropospheric ozone in addition to photochemical formation, during summertime.