Derwent R.G. et al., 1998, “Observation and interpretation of the seasonal cycles in the surface concentration of ozone and carbon monoxide at Mace Head, Ireland from 1990 to 1994”. Atmospheric Environment 32, Hirsch R.M., 1991, “Selection of methods for the detection and estimation of trends in water quality”. Water Ressources Research 27, Ordóňez C. et al., 2005, “Changes of daily surface ozone maxima in Switzerland in all seasons from 1992 to 2002 and discussion of summer 2003”. Atmospheric Chemistry and Physics 5, Sirois A., 1998, “WMO/EMEP Workshop on Advanced Statistical Methods and their application to Air Quality Data sets”. Helsinki. This work was made possible by the financial support of the “Ecole des Mines de Douai”, the French Ministry of Environment and the environmental agency ADEME. Many thanks to the site operators too. Pierre SICARD *, P. CODDEVILLE, S. SAUVAGE, J.C. GALLOO, Ecole des Mines de Douai, Dpt Chimie et Environnement, 941 rue Charles Bourseul, Douai, France. * Corresponding author – phone number : and address: Annual and seasonal trends surface ozone background levels at rural French monitoring stations over the period Annual and seasonal trends surface ozone background levels at rural French monitoring stations over the period The non-parametric Mann-Kendall test used (Sirois, 1998) for detecting and estimating monotonic trends in the time series of annual values of air concentrations (non-normally distributed, extreme values and missing data, small samples...) and the non-parametric Sen’s method for the magnitude of the trend. E(S i ) = 0 and Var(S i ) = t i : the extent of a given tie in month i and the distribution of S i is normal in the limit as n i → . Xi = f (ti) + εi where f (ti) is a continuous monotonic increasing or decreasing function of time. The residuals εi are assumed to belong to the same distribution with zero mean. Hirsch et al. (1991) have extended the MKT to take into account any seasonality in the data (Seasonal Kendall Test). Even if the test is presented here for 12 monthly values per year, it can also be used 52 weekly values or 4 seasonal values. with x ij and x ik : values for the month i, years j and k (j>k) S i = S i,S j (i j) : functions of independent random variables so Cov(S i,S j ) = 0 References and acknowledgements If f(t) = Qt + B where Q is the slope and B is a constant, Q is the median of these n values A ijk (j>i, t ik t jk ). For each season k (k = 1, 2, …, m) with A (ij)k = x ik = concentration for the year i and season k if Z’ > 0, increasing monotonic trend if Z’ < 0, decreasing monotonic trend The standard normal deviate Z’ = O 3 is a secondary pollutant that is produced during the atmospheric photooxidation of Volatile Organic Compounds under the presence of nitrogen oxides. O 3 is an important air quality issue due to adverse health effects on humans and plants. Furthermore, O 3 acts as a strong greenhouse gas in the free troposphere. In 1990, the French Background Air Pollution Monitoring Network called MERA was implemented by the French Ministry of Environment and the Environmental Agency ADEME (Agence De l’Environnement et de la Maîtrise de l’Énergie), to create a long term high quality data base in atmospheric pollution. Our study examines the spatio-temporal trends of data from 8 sites, during the sampling period from January 1995 to December Moreover, national emissions of NO x and NMVOC (CITEPA data) were examined. There is considerable interest in quantifying surface background ozone concentrations and associated trends, as they serve to define the impacts on ozone of the anthropogenic precursors reduction and to define target values for O Introduction 2. Geographic locations of the MERA stations All sites are located in rural areas, widely distributed over the French territory and are generally not influenced by local anthropogenic emission sources. 7 stations are also included in the European network EMEP. Table 1: Annual and seasonal average ozone air concentration, mean seasonal and annual changes and standard deviations obtained by the Seasonal Kendall test from 8 stations of the MERA network in France over the period a significance levels :  = ***, 0.01 **, 0.05 *, 0.1 +, > 0.1, b ns : non significant trend. 4. Annual and Seasonal Trends 5. Results and Discussion Annual average trends: +1.5 to + 2.0%yr -1 for the high-lying stations, about %yr -1 in the N-E and -1.0 to -1.5 %yr -1 in the other stations. For the whole period : %.year -1 (strongly influenced by the concentrations of the high-lying stations) despite the significant reduction in the precursor emissions in France (- 2.8%yr -1 for NO x and - 3.8%yr -1 for VOCs) and Europe (- 1.1%yr -1 for NO x and - 1.6%yr -1 for VOCs). This situation with respect to near-surface O 3 is comparable to that observed in other countries (Germany, Switzerland, Canada...). 98th Percentiles (127± 10 µg.m -3, ± 2.4 %yr -1 ) : largely attributed to the reduction in precursor emissions within the European Community since Median (66.2± 3.2 µg.m -3, ± 2.8 %yr -1 ) : less sensitive to emission changes than P.98. Spring : %yr -1 and Summer : %yr -1, seasons in which the main driving mechanism is photochemical production, suggests that the negative trend observed is slightly in line with the reduction in emissions of O 3 precursors during the period. Winter : %yr -1 and Autumn : %yr -1 : related to the lower effect of the O 3 loss by titration through NO as a consequence of the decreased emissions of primary during the 90s (Ordóňez et al., 2005). This effect of titration is especially important at the areas under the direct influence of the air masses coming from the NO x emitters countries (Donon and Revin). Possible reasons for observed trends :  The stratospheric–tropospheric ozone exchange mainly in the high-lying stations (Iraty and Le Casset)  Trans-Atlantic transport of North American pollution affects Europe (Derwent et al., 1998)  CH 4 : - 13% between 1990 and 2003 but 0% over the period in France  The bio-geogenic compounds contribute mainly to the O 3 formation in rural station. At Donon ( ), we obtained by the MKT, + 1.1%.year -1 for ethane, + 1.1%.year -1 for propane, + 1.7%.year -1 for isoprene and + 1.0%.year -1 for O 3  At the regional level, exists coherence: when the ratio VOC/NO x increases, the O 3 concentration decreases  The lower effect of the ozone loss by titration through NO  The effect of variations in meteorological conditions can mask the dependence of the long-term trends in O 3 on precursors emissions. 3. Detection and estimation trends with Sgn(x ij -x ik ) = and Figure1 : Geographic locations of the MERA stations in = 0