SEARCH

SEARCH BY CITATION

Keywords:

  • background ozone;
  • photochemical ozone;
  • ozonesonde;
  • residence time;
  • long-term trend;
  • anthropogenic emissions

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[1] Long-term ozonesonde data (1970–1999) from Hohenpeissenberg, Germany (48°N, 11°E) and Payerne, Switzerland (47°N, 7°E) have been studied for the boundary layer (from 900 to 750 hPa) and lower troposphere (from 750 to 550 hPa) using back air trajectory analysis. The residence times of air masses over central Europe have been estimated for each day utilizing the trajectory data and are tagged with the corresponding mixing ratios of ozone. It is shown that mixing ratios increase with increasing residence time in summer at a rate of about 2 ppbv/day, but stabilize after about 6 days. Mixing ratios corresponding to central European residence times of 4–6 days are defined as “photochemically aged” ozone. Correlation and slope analysis is made between ozone and residence time (1–6 days) using a statistical regression model, and ozone mixing ratios extrapolated to zero day are defined as “background” ozone. The maximum in “photochemically aged” ozone occurs in summer. “Background” ozone shows lower values and generally a broad maximum extending from spring to summer. Absolute values and variations of ozone in “photochemically aged” and “background” air are quite similar at Hohenpeissenberg and Payerne, indicating their good regional representativeness for central Europe. Regional monthly averaged ozone build-up (above the “background” value) over central Europe is about 10 ppbv in summer. Central Europe is found to be a net source of ozone throughout the year, except in winter. However, an increasing trend of ozone in “background” and “photochemically aged” air has been observed in winter during the 1980s. Only the boundary layer “photochemically aged” ozone showed a decreasing trend (−1.6 ± 0.2 ppbv/year), consistent with the NOx emissions trend over central Europe during the 1990s. Long-term changes in “photochemically aged” and “background” ozone suggest that intercontinental transport has a significant effect on ozone mixing ratios in the lower troposphere as well as in “background” ozone in the boundary layer over central Europe. The present analysis ignores transport of ozone from the stratosphere, which might also contribute to interannual variations and changes.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[2] Long-term tropospheric ozone measurements from Europe show a significant increase in ozone concentrations from the 1950s to the middle of the 1980s [e.g., Staehelin et al., 1994; Oltmans et al., 1998; Logan et al., 1999]. These changes in ozone concentrations have generally been considered in a regional context, but some recent numerical simulations suggest that changes in ozone over one continent could have a significant contribution from other continents [e.g., Berntsen et al., 1996; Jacob et al., 1999]. Knowledge of background ozone concentrations plays an important role in such studies. However, definition of the background ozone concentration is also a topic of discussion. Background ozone has been estimated to be 20–45 ppbv in the boundary layer over the US by different methods [Altshuller and Lefohn, 1996; Hirsch et al., 1996; Lin et al., 2000] and it has been shown that there is a contribution in this background ozone from other continents, mainly from Asia [Lin et al., 2000; Fiore et al., 2002]. Such studies, particularly long-term changes in background ozone are limited for European sites. It has also been shown that the background ozone concentration over a region is an important factor for pollution control strategies over that region [Altshuller and Lefohn, 1996]. Policy makers in Europe [Simpson et al., 1999] are proposing to establish critical ozone levels, which involve integration of hourly average mixing ratios of ozone with a threshold level in the range of 40–60 ppbv, sufficiently low for the contribution from background ozone to be very significant.

[3] Long-range transport of primary and secondary pollutants from the North American continent across the Atlantic Ocean has been documented in a few case studies. It was shown during the NARE campaign in 1993 that North American pollution influences the air pollutant concentrations over the western part of the Atlantic Ocean [e.g., Moody et al., 1996]. Other studies, including investigations considering the role of warm conveyor belts (WCB), show the signature of ozone plumes from North America in the middle and upper troposphere over Europe [e.g., Wild et al., 1996; Stohl and Trickl, 1999; Wild and Akimoto, 2001]. Recently, Forster et al. [2001] have shown that emissions from a Canadian forest fire were transported to Europe at lower altitude.

[4] Surface ozone measurements from European sites (including those from high altitude sites) have been reported in many publications [e.g., Laurila, 1999; Bronnimann et al., 2000; Monks, 2000; Schuepbach et al., 2001]. In some studies, backward trajectory analysis was used to classify the origin of ozone in the ambient air [e.g., Derwent et al., 1998; Pochanart et al., 2001]. These studies were usually based on the direction of the transport of the air mass, except the recent study of Pochanart et al. [2001], in which surface ozone data at Arosa were successfully classified using trajectory analysis based on the residence time of air masses over the European continent. To our knowledge, long-term ozonesonde data have not been studied in much detail, except for analysis of trends using statistical approaches. In the present study, we have used two of the most extensive data set of ozonesondes (Hohenpeissenberg, Germany and Payerne, Swiss plateau), which started in the late 1960s and which have been analyzed by statistical methods for long-term trends in the stratosphere and the troposphere in several studies [e.g., Logan, 1985; Tiao et al., 1986; Staehelin and Schmid, 1991; Harris et al., 1997; Logan et al., 1999]. These studies show a large increase (7.3%–13.9% per decade) in tropospheric ozone over Europe since the beginning of the 1970s [see, e.g., Weiss et al., 2001]. In the present paper, we have analyzed the ozonesonde data using a residence time based method, first used by Pochanart et al. [2001], that allows discrimination between European photochemical ozone formation and intercontinental transport.

2. Database and Analytical Methodology

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

2.1. Ozonesonde Data

[5] Ozonesonde data (1970–1999) from Hohenpeissenberg, Germany (48°N, 11°E, 975 m asl) and Payerne, Switzerland (47°N, 7°E, 490 m asl), which are separated by approximately 300 km, have been analyzed for the boundary layer and lower troposphere. Hohenpeissenberg is approximately 75 km south west of Munich and Payerne is approximately 45 km south west of Bern. These two ozonesonde records, along with the record from Uccle (Belgium) form the three longest series of records in Europe. Brewer-Mast sensors are used for the measurements of vertical distribution of ozone at both stations. Total ozone measurements from Dobson spectrophotometers are used from Hohenpeissenberg and from Arosa (Switzerland) for the Payerne series to derive the correction factors to linearly scale the ozone profiles, as recommended in the World Meteorological Organization (WMO) standard procedure. The ozonesonde record of Payerne was recently edited by Stubi et al. [1998]. Both series are available from the WOUDC, Toronto in Canada. Ozone mixing ratios (reported in pressure intervals) for each day are averaged for two vertical layers, representing the boundary layer (BL) (from 900 hPa to 750 hPa), and the lower troposphere (LT) (from 750 hPa to 550 hPa). Measurements between the surface and 900 hPa were not considered in present analysis to minimize the possible direct influence of surface emissions, effect of deposition and to reduce the influence due to change in balloon launch time in the Payerne series [Logan, 1985; Staehelin and Schmid, 1991]. Correction factors have been applied for measurements between 900 and 800 hPa at Payerne as proposed by Staehelin and Schmid [1991] in order to minimize the influence of the change in balloon launch time (during 1970–1981). More details on these ozonesonde data, measurement procedures, and observation sites are given by Claude et al. [1987], Staehelin and Schmid [1991], Claude et al. [1998], Kohler and Claude [1998], and Logan et al. [1999].

[6] The vertical ozone profiles are available between 2 and 3 days per week to as much as seven days over some periods. We have used 2857 ozone profiles from Hohenpeissenberg and 3800 ozone profiles from Payerne in the present analysis. Figure 1 shows the seasonal average vertical distributions of ozone at Hohenpeissenberg and Payerne. Both the sites show quite similar seasonal variations with maximum ozone values during summer, lowest values during winter and some variability in upper troposphere. The ozone values are approximately 20 ppbv higher in summer than in winter.

image

Figure 1. Average seasonal vertical distributions of ozone in the troposphere at Hohenpeissenberg and Payerne for the period 1970–1999.

Download figure to PowerPoint

2.2. Trajectory Simulations

[7] A Global Meteorological system (FRSGC-GMET) has been employed for the isentropic trajectory simulations. The FRSGC-GMET is similar to the trajectory calculation package of NIES [Hayashida-Amano et al., 1991] except that it uses NCEP reanalysis wind data (2.5° resolution), which are archived every 6 hours. A fourth-order Runge-Kutta method is used for the time integration of the ordinary differential equations of particle motion. Space interpolation is made linearly both in horizontal and vertical directions. Time interpolation is also made linearly as in most other trajectory models [e.g., Knudsen and Carver, 1994; Knudsen et al., 1996]. The calculation time step is 30 min and time step of the output is 60 min.

[8] Clusters of four trajectories, with latitudinal and longitudinal displacements of 0.5 degrees with respect to the locations of Hohenpeissenberg and Payerne, are generated each day. There are some cases when the four trajectories are in-coherent. A filtering criterion is adapted to discard such cases. If the estimated residence times from the four trajectories show large variabilities (i.e., if three times standard deviation, 3 sigma, is greater than the average residence time of the four trajectories) the analysis for that day is rejected and it is not included in further study. Trajectories from about 13% of days are rejected by this criterion. However, a sensitivity analysis (not shown here) also included the in-coherent trajectories and demonstrated that rejection of this class of trajectories did not alter the conclusion of the present study. Therefore, the limitations and uncertainties that apply to individual isentropic trajectory calculations may to some extent be overcome by simulating clusters of trajectories. Further, we did not consider the path of the air mass or its origin or direction in our analysis, but only considered the residence times of air masses over Europe (defined here as 35°N–60°N; 10°W–30°E).

[9] A measure of the errors in trajectory calculations is the absolute spherical distance (ASD) or relative spherical distance (RSD) between two positions of particles [e.g., Knudsen and Carver, 1994; Knudsen et al., 1996; Stohl and Koffi, 1998]. A comparison of trajectories calculated for a few days during different months with the trajectory results from NIES [Pochanart et al., 2001] and also with results obtained from KNMI [Scheele et al., 1996] showed that the RSD is only about 2–5%. The correlation coefficients are estimated to be better than 0.99 for latitudinal and longitudinal comparisons on days with stable conditions. Integration errors are estimated by computing the backward trajectory from the end-point position of its forward counterpart. These errors are smaller than 2%, which is similar to many other trajectory simulations. It has been shown that trajectory accuracy mainly depends on the spatial and temporal resolution of the meteorological data used [e.g., Knudsen and Carver, 1994; Morris et al., 1995; Stohl and Koffi, 1998]. Errors and differences in different trajectory calculations due to their different integration schemes and interpolation are smaller. More details on isentropic trajectory calculations can be found elsewhere [e.g., Harris and Kahl, 1994; Moody et al., 1995; Knudsen et al., 1996; Merrill, 1996].

2.3. Air Mass Categorization

[10] Ten day backward trajectory simulations are made for the mean heights of the two layers of ozone (see section 2.1), one at 1500 m asl representing the BL and the other at 3000 m asl representing LT. Trajectories are calculated for the time and day of each individual ozonesonde measurement at Hohenpeissenberg and Payerne. These trajectory data are used to estimate the residence times of the air masses over the polluted European continent (central Europe, southern Europe, England, part of eastern Europe, and southern part of Norway and Sweden) (see Figure 2) covering the latitude region from 35°N to 60°N and longitude region from 10°W to 30°E. This region is selected on the basis of emissions of NOx, and VOCs (ozone precursors) over Europe [EMEP, 1996, 2001] and will be described as central Europe hereinafter. In this analysis, we do not distinguish regions of different strengths of emission over central Europe. However, an attempt was also made to analyze the data over the regions of relatively higher NOx and VOCs emissions (see section 3.1). Air masses are assigned as 1 day residence time if they have stayed over central Europe for 0.5 to 1.5 days (12 hours to 36 hours); 1.5 to 2.5 days period is assigned as 2 days residence time and similarly for the other days. Figure 2 shows examples of 10-day backward trajectories reaching Payerne at 1500 m altitude and displaying different residence times (1–2 days, 5–6 days, and 10 days) over central Europe. Variations in the altitude of the air during the 10 days prior to arrival are also shown.

image

Figure 2. 10 days backward isentropic trajectories at Payerne in the boundary layer (1500m). These are examples of different residence times [(a) 1–2 days; (b) 5–6 days; (c) 10 days] of the air masses over central Europe (35°N-60°N and 10°W-30°E; region shown as rectangle). Six days trajectory data are used in the present analysis (see section 3.2). Altitude variations of the respective air masses are also shown.

Download figure to PowerPoint

[11] Average ozone values each day at Hohenpeissenberg and Payerne are tagged with the corresponding residence times of the air masses. The tagged ozone values each day for the two layers (BL and LT) are averaged over months or years for further analysis.

3. Results and Discussions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[12] Figure 3a shows the seasonal variations in the average residence times of trajectories reaching Payerne in the boundary layer and lower troposphere. The calculated trajectories show slightly longer residence times over central Europe in spring, summer and autumn than in winter. The average residence time in the boundary layer is 3–4 days from spring to autumn and 3 days in winter. The average residence time is slightly shorter in the lower troposphere than in the boundary layer, due to higher wind speed at the higher altitude. Similar results in average residence times were also found for the trajectories reaching Hohenpeissenberg (not shown). Figure 3b shows the probability distribution of residence times for air masses arriving at Hohenpeissenberg and Payerne in both layers during 1970–1999. Variations in all the four groups are similar. Maximum numbers of trajectories show residence times of 1–2 days. The fraction of trajectories with residence times of 7 days and longer is less than 3%.

image

Figure 3. (a) Seasonal variations in average (1970–1999) residence time (in days) for the trajectories reaching to Payerne in the boundary layer (BL) and lower troposphere (LT). (b) Variations in percentage contribution of trajectories having different residence times at Hohenpeissenberg and Payerne in BL and LT for the period 1970–1999.

Download figure to PowerPoint

3.1. Correlation Between Ozone and Residence Time

[13] Figure 4 shows the correlations between the residence times of air masses and corresponding ozone mixing ratios at Hohenpeissenberg and Payerne in BL and LT layers by seasons for the entire period 1970–1999. Average ozone mixing ratios increase with an increase in residence times in summer and spring. A regression analysis of ozone mixing ratio as function of the residence time shows that the values of 95% confidence intervals are less than 20% of slope value in summer and spring. The average increasing tendency for ozone at both sites and layers is 2.08 ± 0.36 ppbv/day in summer and 1.72 ± 0.34 ppbv/day in spring. There is no significant change in ozone during winter. In fact, ozone shows a slight decrease (−0.18 ± 0.25 ppbv/day) in LT layer during winter. Average ozone increase rate is 0.87 ± 0.29 ppbv/day in autumn. These estimates of ozone rates are made using the data corresponding to residence time of up to six days (see next section).

image

Figure 4. Changes in ozone mixing ratios in the boundary layer and lower troposphere at Hohenpeissenberg and Payerne with respect to changes in residence times of air masses over the central Europe in four seasons during 1970–1999. Vertical bars are 95% confidence intervals, which are “mean ± standard error × zc”, where zc = 1.96.

Download figure to PowerPoint

[14] The high increase in ozone with increasing residence time in summer shows that the accumulation of ozone precursor gases and their photoxidation lead to increases in ozone mixing ratios. The intensity of solar radiation is much smaller in winter and therefore photochemical ozone formation is much slower in this season. The decrease in ozone in winter is possibly due to titration of ozone by NO. An attempt was also made to discriminate the ozone value when air mass passed over the regions of relatively higher NOx and VOCs emissions (France, Germany, Italy, and UK). Mixing ratios of ozone are found to be somewhat higher for these air masses when compared to those for entire central Europe (not shown). However, mixing ratio of ozone shows large variability, partially due to the much smaller numbers of events and therefore we did not further consider such a refined approach.

3.2. Estimates of Ozone in Photochemically Aged and Background Air

[15] Figure 4 shows that average ozone mixing ratios increase in summer and spring until residence times of 5 or 6 days are reached, after which they are approximately stable. This feature has also been observed in the analysis of surface ozone data from Arosa [Pochanart et al., 2001]. The approximately stable ozone values after about 6 days may be caused by unfavorable weather conditions for photochemical ozone formation met during such long stay of the air mass over central Europe. Loss of ozone due to titration by NO during this longer period of residence time under such condition may also contribute. The fraction of days with residence times longer than 6 days over central Europe is rather small (see Figure 3b). Variability in ozone is also much greater for residence times longer than 6 days (see Figure 4). Hence, ozone values corresponding to 4–6 days residence time over Europe, which could have been sufficiently oxidized regionally, are described here as “photochemically aged” ozone.

[16] In order to obtain the ozone value before the air enters the European continental region, ozone values from six to one days are extrapolated linearly to the zero day condition as in the study of Pochanart et al. [2001]. These estimated ozone values of zero day residence time would be described as “background” ozone hereinafter. Here, “background” ozone is basically representative of Atlantic air masses not influenced by European emissions. We feel that estimated “background” ozone in the boundary layer may be affected by local photochemistry, however in the lower troposphere it should be more representative of background levels.

3.3. Seasonal Variations of Ozone in Photochemically Aged and Background Air

[17] Average seasonal variations of ozone in “photochemically aged” and “background” air at Hohenpeissenberg and Payerne during 1970–1999 are shown in Figures 5 and 6, respectively. Ozone in “photochemically aged” air shows a distinct maximum during summer at both sites in the boundary layer (Figures 5b and 6b). In the lower troposphere, it shows a similar but slightly broader maximum that extends to late spring (Figures 5a and 6a). “Background” ozone shows a broad maximum during late spring and summer at both sites in both layers. Minimum mixing ratios of both categories of ozone occur in winter at both sites in both layers. Table 1 shows mixing ratios of ozone in “photochemically aged” air and “background” air at Hohenpeissenberg and Payerne in BL and LT layers during the four seasons. The ozone mixing ratios at both sites agree well within statistical uncertainties for both types of air and both layers.

image

Figure 5. Seasonal variations in ozone in “photochemically aged” and “background” air at Hohenpeissenberg for the period 1970–1999. Seasonal variations in background ozone at other European sites (Jungfraujoch (46.5°N, 8°E, 3580 m; 1988–1997) [Schuepbach et al., 2001], Arosa (46.8°N, 9.7°E, 1840 m; 1996–1997) [Pochanart et al., 2001] and Mace Head (53.3°N, 9.9°W, 10 m; 1990–1994) [Derwent et al., 1998]) are also shown. Upper graph is for the lower troposphere and lower graph is for the boundary layer. Vertical bars are 95% confidence intervals.

Download figure to PowerPoint

image

Figure 6. Same as Figure 5, but for Payerne.

Download figure to PowerPoint

Table 1. Average Ozone in “Photochemically Aged” (PC) and “Background” (BG) Air at Hohenpeissenberg and Payerne in the Boundary Layer and Lower Troposphere for the Period 1970–1999a
 HohenpeissenbergPayerne
PCBGPCBG
  • a

    Values with ± ranges are 95% confidence limits. All values are in ppbv.

Boundary Layer
Winter (DJF)36.2 ± 1.832.9 ± 2.632.6 ± 2.633.1 ± 2.4
Spring (MAM)46.2 ± 3.142.7 ± 4.050.6 ± 3.043.9 ± 3.0
Summer (JJA)55.5 ± 2.344.9 ± 3.656.3 ± 2.944.3 ± 3.1
Autumn (SON)39.3 ± 2.634.6 ± 2.640.9 ± 2.335.5 ± 3.5
Annual44.5 ± 1.838.8 ± 1.945.1 ± 2.139.2 ± 1.7
 
Lower Troposphere
Winter (DJF)41.2 ± 2.243.0 ± 2.540.0 ± 2.140.6 ± 2.8
Spring (MAM)55.8 ± 2.148.0 ± 3.254.9 ± 3.050.8 ± 2.9
Summer (JJA)61.4 ± 2.651.7 ± 3.959.5 ± 2.750.7 ± 2.8
Autumn (SON)47.9 ± 2.842.0 ± 3.046.1 ± 2.242.2 ± 2.7
Annual51.7 ± 1.846.2 ± 1.750.0 ± 1.946.1 ± 1.6

[18] Surface ozone at different European sites shows different seasonal variations with either a broad summer maximum or spring maximum and summer minimum [e.g., Scheel et al., 1997]. Central Europe shows a broad summer maximum [e.g., Monks, 2000]. The seasonal cycle in background ozone (derived by a criteria that is based on concentrations of CO and NOx) at different sites over Switzerland shows a spring maximum and an autumn/winter minimum [Bronnimann et al., 2000], which is similar to the variations in “background” ozone at Hohenpeissenberg and Payerne. Also, the seasonal variation in surface ozone for cleaner air masses at Jungfraujoch, Switzerland [Schuepbach et al., 2001] is similar to those at Hohenpeissenberg and Payerne (Figures 5a and 6a). However, the analysis of surface ozone at Arosa, Switzerland [Pochanart et al., 2001], using a similar technique of trajectory analysis, shows a distinct minimum in background ozone data during summer and a maximum during spring (Figures 5b and 6b) that is different from the present study. The study at Arosa shows a broad maximum for polluted ozone (not shown) during spring-summer that is similar to the variations in “photochemically aged” ozone at Hohenpeissenberg and Payerne.

[19] Another study at a coastal site in northwestern Europe (Mace Head, Ireland) [Derwent et al., 1998] also shows a clear summer minimum in background ozone (Figures 5b and 6b), similar to that of Arosa. These studies at Arosa and Mace Head are based on surface ozone measurements in which dry deposition could be more important than at the High Alpine station at Jungfraujoch and in the ozonesonde data. In addition, ozonesonde measurement in the boundary layer represents a larger layer (900–750 hPa), which experiences extensive mixing and might have been more influenced by photochemistry compared to the air near the surface.

3.4. Build-Up in Ozone

[20] “Photochemically aged” air has experienced European emissions and has been photochemically processed for 4–6 days, whereas “background” air is just entering the European region. Therefore the difference of ozone levels in these two air masses is expected to provide information on ozone formation due to European precursors. Figures 7a and 7b show the seasonal variations in ozone build-up at Hohenpeissenberg and Payerne in LT and BL layers respectively. Both the sites show largest European ozone formation in summer in both BL and LT layers. Maximum ozone build-up occurs rather at the end of summer (August) at Hohenpeissenberg in the boundary layer, whereas the maximum build-up is slightly earlier (June) at Payerne. Seasonal variations in ozone build-up are very similar at both sites in the lower troposphere, except for some differences during early winter. Average ozone build-up is observed to be about 10 ppbv in summer at both sites and in both layers. Wintertime ozone formation is very low and sometimes negative (i.e., net ozone destruction occurs). Similar seasonal variations and a similar amount of ozone build-up at higher altitudes (LT layer) at both sites indicate that local processes over this region do not affect this build-up significantly in the LT layer. Therefore, we believe that this ozone formation in the LT layer is representative of central Europe.

image

Figure 7. Seasonal variations in ozone build-up at Hohenpeissenberg and Payerne during the period 1970–1999. These variations are obtained by the difference of ozone levels in “photochemically aged” air and “background” air. Upper graph is for the lower troposphere and lower graph is for the boundary layer. Net ozone build-up at other European sites is also shown.

Download figure to PowerPoint

[21] Figure 7 also shows a comparison of the estimates of ozone build-up at Jungfraujoch, Arosa, and Mace Head. There are differences in the magnitudes of ozone build-up at the different sites but the seasonal variations are similar. Mace Head shows mainly a net sink (about 2 ppbv) of ozone, with some build-up of ozone during summer (2–6 ppbv) [Simmonds et al., 1997; Derwent et al., 1998]. It has been shown that Mace Head is exposed mainly to cleaner air masses from the Atlantic. The seasonal variation at Arosa is similar to the variations at Hohenpeissenberg and Payerne, but maximum ozone build-up at Arosa is considerably higher (by 8–10 ppbv) than the maximum build-up at Hohenpeissenberg and Payerne. The difference in maximum ozone build-up is mainly due to difference in “background” ozone values in summer at these sites. The amount of ozone build-up in the boundary layer at Hohenpeissenberg and Payerne thus derived here may be underestimated due to the contribution of local photochemistry to the “background” level as noted before. Ozone build-up at Jungfraujoch is smaller than that at Hohenpeissenberg and Payerne.

3.5. Trends of Ozone in Photochemically Aged and Background Air

[22] Long-term changes of ozone in “photochemically aged” and “background” air at Hohenpeissenberg and Payerne are shown in Figures 8 and 9 for the boundary layer and lower troposphere respectively. Changes in the two categories of ozone and their mixing ratios are quite similar at the two sites in the boundary layer and also in the lower troposphere. It has also been mentioned and shown in Table 1 that mixing ratios of these two categories of ozone are similar at the two sites. Therefore we conclude that the ozone sounding data at Hohenpeissenberg and Payerne have good regional representativeness for central Europe. Amounts and variations in “background” ozone mixing ratios are different than those in “photochemically aged” ozone in summer at both sites (Figures 8b and 9b). However, they are similar in winter (Figures 8a and 9a) due to the fact that ozone in winter does not depend on the residence time of the air over central Europe (see Figure 4).

image

Figure 8. A comparison between long-term changes in two categories of ozone at Hohenpeissenberg and Payerne in the boundary layer during winter and summer. Vertical bars are 95% confidence intervals.

Download figure to PowerPoint

image

Figure 9. Same as Figure 8, but for the lower troposphere.

Download figure to PowerPoint

[23] The temporal evolution of ozone in “photochemically aged” and “background” air for the combined data set of two sites are shown in Figure 10. Each point is estimated using two years data at Hohenpeissenberg and Payerne. The “photochemically aged” ozone in summer shows most obvious change in the boundary layer with a peak around 1990 and a clear decrease (−1.6 ± 0.2 ppbv/year) thereafter (Figure 10b). This seems to be consistent with the temporal evolution in central European NOx emissions (Figure 11). However, “photochemically aged” ozone in the lower troposphere is almost stable and shows much less decrease (−0.4 ± 0.1 ppbv/year) after the year of 1990 (Figure 10a). The summertime “background” ozone and wintertime averaged “background” and “photochemically aged” ozone do not show decreases during 1990s both in BL and LT layers. “Background” ozone in summer seems to have peaked around 1985 and stabilized during 1990s in the boundary layer, whereas it leveled off after 1985 in the lower troposphere.

image

Figure 10. Long-term changes of ozone in “photochemically aged” (black circle) and “background” (gray triangle) air during summer in the lower troposphere (upper graph) and boundary layer (lower graph). Ozone mixing ratios are averaged for “photochemically aged” and “background” air (white square) in winter, as they are quite similar (see Figures 8a and 9a and text for details). Estimates are with data of Hohenpeissenberg as well as Payerne.

Download figure to PowerPoint

image

Figure 11. Changes in anthropogenic NOx emissions over the US [EPA, 1999], central Europe (countries within previously define region) [EMEP, 2001] and East Asia (China, Japan, Korea, Taiwan) (black square, [Kato and Akimoto, 1992]; white square, [Streets et al., 2001]; gray square [Aardenne et al., 1999]). Extrapolated central European values during the 1970s are on the bases of OECD [1993]. Decrease in emissions over central Europe during the 1990s is due to decrease in emissions mainly over Germany and UK.

Download figure to PowerPoint

[24] Model simulation has shown that Europe could be significantly affected by intercontinental transport of pollution from the US [Wild and Akimoto, 2001]. There have been other recent studies, showing transport of emissions from Canada and the US to Europe [Stohl and Trickl, 1999; Forster et al., 2001; Fiore et al., 2002]. Therefore, one would expect that change in emissions from the US during 1990s (Figure 11) could be playing a role in ozone variations over central Europe at least in the lower troposphere as well as in “background” ozone in the boundary layer. In the present analysis, “background” ozone is expected to be representative of the Atlantic air mass, which receives a stronger influence from the US. Wotawa et al. [2000] have shown that higher net ozone production over central Europe could not be explained due to European emissions and the importance of background pollution was suggested. The summertime ozone formation rate in the boundary layer (Figure 12) is rather similar to NOx emissions over central Europe during the three decades with no increase from 1980s to 1990s, but in the lower troposphere it shows some similarity with NOx emissions over the US with slightly increasing tendency from 1980s to 1990s. The present analysis ignores the input of ozone from the stratosphere, which is also a significant part of the tropospheric ozone budget and which might be the subject of large interannual variations and changes. This could impact “background” ozone. Further, any change in the apparent latitude of the air trajectories entering the European boundaries could also have some influence on ozone in the sonde measurements, which is, however, ignored in this analysis.

image

Figure 12. Summertime ozone formation rate during the three decades in the boundary layer (BL) and lower troposphere (LT). These estimates are made using the approach shown in Figure 4. Vertical bars are 95% confidence intervals.

Download figure to PowerPoint

[25] The wintertime averaged “background” ozone and “photochemically aged” ozone (Figures 10a and 10b) show increasing values during the 1980s when emissions were not increasing in Europe or in the US. The increasing trend of ozone [Akimoto et al., 1994] and NOx emissions (see Figure 11) in East Asia in the 1980s might have some influence but further modeling studies are necessary to quantify the effect. The change in decadal average ozone in “background” and “photochemically aged” air show large increases from the 1970s to 1980s, but not from the 1980s to 1990s (see Table 2). Therefore, it appears that the 1980s are a transition period.

Table 2. Increase of Ozone in “Photochemically Aged” (PC) and “Background” (BG) Air (10-Year Average) From the 1970s to the 1980s and From the 1980s to the 1990s Over Central Europe in the Boundary Layer and Lower Tropospherea
 Increase From the 1970s to the 1980sIncrease From the 1980s to the 1990s
PCBGPCBG
  • a

    Values with ± ranges are 95% confidence limits. All values are in ppbv.

Boundary Layer
Winter (DJF)7.6 ± 2.47.0 ± 3.23.0 ± 1.81.8 ± 2.6
Spring (MAM)12.2 ± 3.011.5 ± 4.7−0.9 ± 3.0−2.8 ± 3.9
Summer (JJA)6.7 ± 2.614.3 ± 3.44.0 ± 3.1−2.4 ± 3.0
Autumn (SON)6.0 ± 3.310.1 ± 4.1−0.5 ± 2.80.2 ± 2.2
Annual7.8 ± 1.210.8 ± 1.21.4 ± 1.2−0.8 ± 0.9
 
Lower Troposphere
Winter (DJF)5.5 ± 2.48.7 ± 3.03.4 ± 2.6−0.2 ± 2.0
Spring (MAM)9.1 ± 2.511.5 ± 2.61.1 ± 2.9−0.6 ± 2.8
Summer (JJA)7.8 ± 2.88.7 ± 3.21.9 ± 2.5−0.5 ± 3.4
Autumn (SON)8.2 ± 2.810.3 ± 3.50.7 ± 2.8−2.0 ± 2.7
Annual7.7 ± 1.29.8 ± 0.91.8 ± 1.1−0.9 ± 0.9

4. Summary and Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[26] Ozonesonde data from Hohenpeissenberg and Payerne for the period 1970–1999 have been analyzed using 10 days back air trajectory calculations. The residence time of air masses over central Europe have been estimated utilizing the trajectory data only of 6 days backward. It is shown that mixing ratios of ozone over central Europe in the boundary layer (900–750 hPa) and lower troposphere (750–550 hPa) are sensitive to regional emissions during summer and spring. Ozone mixing ratios are observed to increase (about 2 ppbv/day) with increasing residence times in summer, however they stabilize after about 6 days residence time. Ozone mixing ratios show no increase with residence time in winter.

[27] Ozone in “photochemically aged” and “background” air are calculated using the estimated residence time statistics. This analysis shows similar ozone mixing ratios and variations at Hohenpeissenberg and Payerne in “background” air as well as in “photochemically aged” air. The maximum mixing ratio of “photochemically aged” ozone occurs in summer (about 55 ppbv in BL and about 60 ppbv in LT). However, “background” ozone generally shows a broad maximum extending from late spring to summer with relatively lower values (43–45 ppbv in BL and 48–52 ppbv in LT). Mixing ratio of “photochemically aged” as well as “background” ozone is minimum in winter (33–36 ppbv in BL and 40–43 ppbv in LT). Estimated “background” ozone mixing ratios in the boundary layer (33–45 ppbv) are in good agreement with those of Bronnimann et al. [2000] (33–50 ppbv for different sites in Switzerland). Background ozone over Mace Head, Arosa, and at other sites in the higher latitudes of Europe has been estimated to be 30–35 ppbv [e.g., Derwent et al., 1998; Laurila, 1999; Pochanart et al., 2001]. These “background” ozone levels over central Europe seem to be toward the upper limit of the background ozone levels (20–45 ppbv) over the US [Altshuller and Lefohn, 1996; Hirsch et al., 1996; Lin et al., 2000].

[28] Consistent trends are observed for both categories of air masses, seasons, and layers for Hohenpeissenberg and Payerne for the first time. Long-term changes in “background” ozone are different from changes in “photochemically aged” ozone in summer. “Photochemically aged” ozone shows a substantial decrease with −1.6 ± 0.2 ppbv/year in boundary layer after year 1990 that is consistent with temporal variations of central European NOx emissions. However, decreasing rate in lower troposphere is much smaller (−0.4 ± 0.1 ppbv/year). Influences of emissions from the US, particularly in the 1990s, appear to play an important role in ozone variations over central Europe at least in the lower troposphere and in “background” ozone in the boundary layer. A possible influence from East Asia is suggested during the winter in 1980s. Wild and Akimoto [2001] have shown that emissions from East Asia have smaller but significant effects over Europe in the free troposphere. Numerical simulation shows that, Europe may receive larger contribution in surface ozone from the US (2–4 ppbv on average; 5–10 ppbv on transatlantic transport events) than from Asia (less than 2 ppbv on average) [Li et al., 2002]. Whether it is only long-range transport from North America that is contributing to these variations in ozone over central Europe or whether there is an influence from East Asia, needs to be studied further.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[29] We thank Pakpong Pochanart and Oliver Wild for useful discussions and comments into the manuscript. We also thank Shamil Maksyutov and P. F. J. van Velthoven for providing the trajectory data for a comparison. We greatly appreciate the constructive suggestions of the three reviewers.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information
  • Aardenne van, J. A., G. R. Carmichael, H. Levy II, D. Streets, and L. Hordijk, Anthropogenic NOx emissions in Asia in the period 1990–2020, Atmos. Environ., 33, 633646, 1999.
  • Akimoto, H., H. Nakane, and Y. Matsumoto, The chemistry of oxidant generation: Tropospheric ozone increase in Japan, in Chemistry of the Atmosphere: The Impact on Global Change, edited by J. G. Calvert, pp. 261273, Blackwell Sci., Malden, Mass., 1994.
  • Altshuller, A. P., and A. S. Lefohn, Background ozone in the planetary boundary layer over the United States, J. Air Waste Manage. Assoc., 46, 134141, 1996.
  • Berntsen, T., I. S. A. Isaksen, W. Wang, and X. Liang, Impacts of increased anthropogenic emissions in Asia on tropospheric ozone and climate, Tellus, 48B, 1332, 1996.
  • Bronnimann, S., E. Schuepbach, P. Zanis, B. Buchmann, and H. Wanner, A climatology of regional background ozone at different elevations in Switzerland (1992–1998), Atmos. Environ., 34, 51915198, 2000.
  • Claude, H., R. Hartmannsgruber, and U. Kohler, Measurements of atmospheric ozone profiles using the Brewer/Mast sonde, WMO Global Ozone Res. and Monit. Proj., Rep. No. 17, WMO/TD, No. 179, World Meteorol. Org., Geneve, 1987.
  • Claude, H., U. Kohler, and W. Steinbrecht, New trend analyses of the homogenized ozone records at Hohenpeissenberg, in Proceedings of the Quadrennial Ozone Symposium, L'Aquila, Italy, September 1996, R. D. Bojkov, and G. Visconti, pp. 2124, Edigrafital S.p.A.-S. Atto (TE), Italy, 1998.
  • Derwent, R. G., P. G. Simmonds, S. Seuring, and C. Dimmer, Observation and interpretation of the seasonal cycles in the surface concentrations of ozone and carbon monoxide at Mace Head, Ireland from 1990 to 1994, Atmos. Environ., 32, 145157, 1998.
  • EMEP, Estimated dispersion of acidifying agents and of near surface ozone, MSC-W Status Rep. 1996, Part 1, Meteorological Canter-West, Nor. Meteorol. Inst., Oslo, Norway, 1996.
  • EMEP, Emission data reported to UNECP/EMEP: Evaluation of the spatial distribution of emissions, MSC-W Status Rep. 2001, by Vigdis Vestreng, Nor. Meteorol. Inst., Oslo, Norway, 2001.
  • EPA, National air pollutant emission trends, 1900–1998, Rep. EPA-454/R-00-002, Research Triangle Park, N. C., 1999.
  • Fiore, A. M., D. J. Jacob, I. Bey, R. M. Yantosca, B. D. Field, and J. G. Wilkinson, Background ozone over the United States in Summer: Origin, trend and contribution to pollution episodes, J. Geophys. Res., 107, 4279, doi:10.1029/2001JD000982, 2002.
  • Forster, C., et al., Transport of Canadian forest fire emissions to Europe, J. Geophys. Res., 106, 22,88722,906, 2001.
  • Harris, J. M., and J. D. W. Kahl, An analysis of 10-day Isentropic flow patterns for Barrow, Alaska: 1985–1992, J. Geophys. Res., 99, 25,84525,855, 1994.
  • Harris, N. R. P., et al., Trends in stratospheric and free tropospheric ozone, J. Geophys. Res., 102, 15711590, 1997.
  • Hayashida-Amano, S., Y. Sasano, and Y. Iikura, Volcanic disturbance in the stratospheric aerosol layer over Tsukuba, Japan, observed by the National Institute for Environmental Studies Lidar from 1984 through 1986, J. Geophys. Res., 96, 15,46915,487, 1991.
  • Hirsch, A. I., J. W. Munger, D. J. Jacob, L. W. Horowitz, and A. H. Goldstein, Seasonal variation of the ozone production efficiency per unit NOx at Harvard Forest, Massachusetts, J. Geophys. Res., 101, 12,65912,666, 1996.
  • 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, 21752178, 1999.
  • Kato, N., and H. Akimoto, Anthropogenic emissions of SO2 and NOx in Asia: Emission inventories, Atmos. Environ., 26, 29973017, 1992.
  • Kohler, U., and H. Claude, Homogenized ozone records at Hohenpeissenberg, in Proceedings of the Quadrennial Ozone Symposium, L'Aquila, Italy, September 1996, R. D. Bojkov, and G. Visconti, pp. 5760, Edigrafital S.p.A.-S. Atto (TE), Italy, 1998.
  • Knudsen, B. M., and G. D. Carver, Accuracy of the isentropic trajectories calculated for the EASOE campaign, Geophys. Res. Lett., 21, 11991202, 1994.
  • Knudsen, B. M., J. M. Rosen, N. T. Kjome, and A. T. Whitten, Comparison of analyzed stratospheric temperatures and calculated trajectories with long-duration balloon data, J. Geophys. Res., 101, 19,13719,145, 1996.
  • Laurila, T., Observational study of transport and photochemical formation of ozone over northern Europe, J. Geophys. Res., 104, 26,23526,243, 1999.
  • Li, Q., et al., Transatlantic transport of pollution and its effects on surface ozone in Europe and North America, J. Geophys. Res., 107, 10.1029/2001JD001422, in press, 2002.
  • Lin, C.-Y. C., D. J. Jacob, J. W. Munger, and A. M. Fiore, Increasing background ozone in surface air over the United States, Geophys. Res. Lett., 27, 34653468, 2000.
  • Logan, J. A., Tropospheric ozone: Seasonal behavior, trends, and anthropogenic influence, J. Geophys. Res., 90, 10,46310,482, 1985.
  • Logan, J. A., et al., Trends in the vertical distribution of ozone: A comparison of two analyses of ozonesonde data, J. Geophys. Res., 104, 26,37326,399, 1999.
  • Merrill, J. T., Trajectory results and interpretation for PEM-WEST A, J. Geophys. Res., 101, 16791690, 1996.
  • Monks, P. S., A review of the observations and origin of the spring ozone maximum, Atmos. Environ., 34, 35453561, 2000.
  • Moody, J. L., S. J. Oltmans, H. Levy II, and J. T. Merrill, Transport climatology of tropospheric ozone: Bermuda, 1988–1991, J. Geophys. Res., 100, 71797194, 1995.
  • Moody, J. L., J. C. Davenport, J. T. Merrill, S. J. Oltmans, D. D. Parish, J. S. Holloway, H. Levy II, G. L. Forbes, M. Trainer, and M. Buhr, Meteorological mechanism for transporting O3 over the western North Atlantic Ocean: A case study for August 24–29, 1993, J. Geophys. Res., 101, 29,21329,227, 1996.
  • Morris, G. A., et al., Trajectory mapping and applications to data from the upper atmosphere research satellite, J. Geophys. Res., 100, 16,49116,505, 1995.
  • OECD, OECD Environmental Data Compendium 1993, Paris, France, 1993.
  • Oltmans, S. J., et al., Trends of ozone in the troposphere, Geophys. Res. Lett., 25, 139142, 1998.
  • Pochanart, P., H. Akimoto, S. Maksyutov, and J. Staehelin, Surface ozone at the Swiss Alpine site: The hemispheric background and the influence of large-scale anthropogenic emissions, Atmos. Environ., 35, 55535566, 2001.
  • Scheel, H. E., et al., On the spatial distribution and seasonal variation of Lower-Troposphere ozone over Europe, J. Atmos. Chem., 28, 1128, 1997.
  • Scheele, M. P., P. C. Siegmund, and P. F. J. van Velthoven, Sensitivity of trajectories to data resolution and its dependence on the starting point: In or outside a tropopause fold, Meteorol. Appl., 3, 267273, 1996.
  • Schuepbach, E., T. K. Friedli, P. Zanis, P. S. Monks, and S. A. Penkett, State space analysis of changing seasonal ozone cycles (1988–1997) at Jungfraujoch (3580 m above sea abundance) in Switzerland, J. Geophys. Res., 106, 20,41320,428, 2001.
  • 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, 4559, 1997.
  • Simpson, D., L. D. Emberson, M. R. Ashmore, H. M. Cambridge, and J. P. Tuovinen, EMEP modeling of AOT40 (and ozone fluxes?) across Europe: Development and possibilities, in Critical Levels for Ozone Levels, vol. 2, Environ. Doc. No. 115, edited by J. Fuhrer, and B. Achermann, Swiss Agency for the Environ., For. and Landscape, Gerzensee, Switzerland, 1999.
  • Staehelin, J., and W. Schmid, Trend analysis of tropospheric ozone concentrations utilizing the 20-year data set of ozone balloon soundings over Payerne (Switzerland), Atmos. Environ., 25A, 17391749, 1991.
  • Staehelin, J., J. Thudium, R. Buehlfr, and A. Volz-Thomas, Trends in surface ozone concentrations at Arosa (Switzerland), Atmos. Environ., 28, 7587, 1994.
  • Stohl, A., and N. E. Koffi, Evaluation of trajectories calculated from ECMWF data against constant volume balloon flight during ETEX, Atmos. Environ., 24, 41514156, 1998.
  • Stohl, A., and T. Trickl, A text book example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and North America origin in the free troposphere over Europe, J. Geophys. Res., 104, 30,44530,462, 1999.
  • Streets, D. G., N. Y. Tsai, H. Akimoto, and K. Oka, Trends in emissions of acidifying species in Asia, 1985–1997, Water Air Soil Pollut., 130, 187192, 2001.
  • Stubi, R., V. Bugnion, M. Giroud, P. Jeannet, P. Viatte, B. Hoegger, and J. Staehelin, Long term ozone balloon sounding series at Payerne: Homogenization method and problems, in Atmospheric Ozone, Proceedings of the XVIII Quadrennial Ozone Symposium, edited by R. D. Bojkov, and G. Visconti, pp. 179182, Edigrafital S.p.A.-S. Atto (TE), Italy, 1998.
  • Tiao, G. C., G. C. Reinsel, J. H. Pedrick, G. M. Allenby, C. L. Mateer, A. J. Miller, and J. J. Deluisi, A statistical analysis of ozonesonde data, J. Geophys. Res., 91, 13,12113,136, 1986.
  • Weiss, A. K., J. Staehelin, C. Appenzeller, and N. R. P. Harris, Chemical and dynamical contributions to ozone profile trends of the Payerne (Switzerland) balloon soundings, J. Geophys. Res., 106, 22,68522,694, 2001.
  • Wild, O., and H. Akimoto, Intercontinental transport of ozone and its precursors in a 3-D global CTM, J. Geophys. Res., 106, 27,72927,744, 2001.
  • 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, 29,26929,288, 1996.
  • Wotawa, G., H. Kroger, and A. Stohl, Transport of ozone towards the Alps—Result from trajectory analyses and photochemical model studies, Atmos. Environ., 34, 13671377, 2000.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Database and Analytical Methodology
  5. 3. Results and Discussions
  6. 4. Summary and Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.