Surface ozone concentration trends and its relationship with weather types in Spain (2001–2010)

Atmospheric Environment 101 (2015) 10e22

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Surface ozone concentration trends and its relationship with weather types in Spain (2001e2010)
Carlos Gonza
lez-Hidalgo b, Arturo Sanchez-Lorenzo c, d, Ana Santurtún a, *, Jose María Teresa Zarrabeitia a a

Unit of Legal Medicine, Department of Physiology and Pharmacology, University of Cantabria, Av Cardenal Herrera Oria s/n, 39011 Santander, Spain Department of Geography, IUCA, University of Zaragoza, Pedro Cerbuna 12, 50009 Zaragoza, Spain Department of Physics, University of Girona, Girona, Spain d Pyrenean Institute of Ecology, Spanish National Research Council (CSIC), Zaragoza, Spain b c

h i g h l i g h t s
Tropospheric ozone concentration shows an upward trend throughout all seasons.
Ozone upward trend is in line with a reported decrease of NOX emissions and with an increase in surface solar radiation.
Synoptic meteorology is associated with ozone levels.
Median concentrations were significantly lower in days with Anticyclonic weather.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 August 2014 Received in revised form 2 November 2014 Accepted 3 November 2014 Available online 4 November 2014

This paper assesses the temporal variations of surface ozone concentrations during the period 2001 e2010 in 3 regions of Spain with different geographical and socioeconomic features (northern coastland, central inland and northeast inland), as well as its link with atmospheric circulation. Specifically, daily surface atmospheric patterns over the aforementioned regions are characterized using NCEP/NCAR reanalysis data and an objective classification scheme in order to study the relationship between synoptic weather types and daily ozone levels. The results show that tropospheric ozone concentration has a tendency towards an increase during the study period, both during daytime and nighttime. Moreover, in general, this upward trend is seen throughout all of the seasons. The observed trends are in line with a reported decrease of NOX emissions and increase in surface solar radiation during the 2000s in Spain. On the other hand, interestingly, median concentrations were statistically significantly lower in days with anticyclonic weather conditions than in the rest of meteorological situations, while days with a directional weather type showed higher median levels of ozone concentration, with maximum values in days with northern and eastern component. Due to the detrimental effect that ozone has on human health, the relationship between synoptic weather patterns and daily ozone levels shown in this work could potentially be used for implementing pollution level alert protocols depending on forecast weather types. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Tropospheric ozone Weather types Air pollution trends

1. Introduction Since the discovery of ozone and its first measurements in Europe at the end of the nineteenth century, background ozone concentrations have more than doubled, showing a significant increase during the last few decades, both in rural and urban areas (e.g., Volz and Kley, 1988; Parrish et al., 2012; Querol et al.,

* Corresponding author. E-mail address: [email protected] (A. Santurtún). http://dx.doi.org/10.1016/j.atmosenv.2014.11.005 1352-2310/© 2014 Elsevier Ltd. All rights reserved.

2014; Paoletti et al., 2014). This increment in ozone concentrations is alarming given its effects on living beings and their ecosystems, but also because, even though different hypotheses have been formulated to explain its cause, the concrete reasons behind this upward trend are still debated, making it difficult to implement control measures for ozone levels. The analysis of tropospheric ozone concentrations is highly important since, apart from being a greenhouse gas, its strong oxidant properties, at certain levels, can affect animals, vegetation, materials and have an effect on human health not only for predisposed patients, such as asthma sufferers and children, but also

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

for previously healthy individuals (Yanga et al., 2012; Halonen et al., 2010). Ozone is a secondary pollutant which follows clear seasonal and daily cycles, presenting higher values in the summer and during daytime and lower concentrations in the winter and at night, which is determined by its photochemical generation processes (Sebald ~ as et al., 2004; Gerasopoulos et al., 2006; et al., 2000; Duen Zvyagintsev et al., 2008). Its formation process is highly dependent on air mass exchange between the stratosphere and the troposphere, surface dry deposition, temperature, solar radiation, NOX emissions and environmental concentrations of volatile organic compounds (e.g., Trainer et al., 2000). Thus, most studies have shown a strong relationship between ozone concentration levels and solar radiation, air temperature, relative humidity and wind speed and direction (e.g., Adame et al., 2010; Thompson et al., ~ as et al., 2002; Demuzere et al., 2009; Sekiya and Sudo, 2001; Duen 2012). Synoptic scale meteorological patterns determine the conditions for the long-range transport of ozone, while also affecting the interaction among ozone precursors, facilitating its formation and destruction. Specifically, several approaches have been used with the intention of describing their relationships to ozone concentrations and the meteorological conditions that affect its formation, destruction and transport processes, such as categorizing a wide variety of complex meteorological variables in different atmospheric circulation (e.g., García et al., 2005; Tang et al., 2009). For decades, climatologists have dealt with the topic of systematize atmospheric circulation conditions by means of a catalog of weather types (WTs), which has led to the existence of several classification methods (for additional details see, for example, Philipp et al., 2010). Traditional manual subjective methods, such as those proposed by Lamb and Britain (1972) for the British Islands, or by Hess and Brezowsky (1952) for Central Europe, have been combined with or even replaced by objective or semi-automatized techniques which allow for the analysis of large amounts of data in less time and effort (e.g., Esteban et al., 2006; Beck and Philipp, 2010; Philipp et al., 2010). The method devised by Jenkinson and Collison (1977) can be considered an automatized version of Lamb's classification, and it is based on a group of indices related to the direction and vorticity of the geostrophic flux calculated on sea level pressure (SLP) data. Several authors have used this classification method in recent years with the aim of finding a relationship between WTs and different environmental phenomena of natural and anthropogenic origin such as heavy snowfalls, droughts, landslides, soil erosion, and pollution (e.g., Cortesi et al., 2013; Nadal-Romero et al., 2013; Andrade et al., 2011). This paper assesses the temporal variations (i.e., cycles and trends) of surface ozone concentrations for the period 2001e2010 in 3 regions of Spain with different geographical and socioeconomic features and suggests a new hypothesis to explain the increase in ozone concentrations in Spain. Daily surface synoptic circulation patterns over the aforementioned regions are also characterized using NCEP/NCAR reanalysis data in order to study the relationship between synoptic weather patterns and daily ozone levels. Our results could potentially be used for implementing pollution level alert protocols depending on forecast weather types. The datasets and methods used in this study are described in Section 2. The results are presented and discussed in Section 3, both the ozone variability and trends (Section 3.1) and its relationship with atmospheric circulation patterns (Section 3.2). Finally, conclusions of this study are presented in Section 4.

11

2. Methods 2.1. Sites description The present work analyzes ozone concentration series from 3 Spanish regions (Fig. 1): central inland (Madrid), northeast inland (Saragossa) and northern coastland (Santander). The first area of study is situated at the center of the Iberian Peninsula under Mediterranean climate conditions, affected by its inland location. This area is characterized by different stations located at the metropolitan area of Madrid, the national capital and the most populated city in Spain. The surrounding area (the Autonomous Community of Madrid) ranks first in regional contribution to the national Gross Domestic Product (GDP). The industrial and manufacturing sector accounts for 13% of its economy. The second studied area is located in the Ebro valley and is characterized by stations located in Saragossa city. In this area (northeast inland Iberian Peninsula) climate characteristics are that of a semi-desert Continental Mediterranean climate. Saragossa is the fifth most populated city in Spain, and ranks fourth in the Spanish Economy
mica, IAE). The third and Activity Index (
Indice de Actividad Econo last region included in this work is Cantabria, situated in the northern region of Spain, which has mountainous and maritime characteristics. Its capital and most populated city is Santander. Cantabria has a humid temperate oceanic climate, and it is much less populated than the other two regions in this study, its population amounting to 1.26% of that of the country. Hourly data of tropospheric ozone concentration for a period of 10 years between 1 January 2001 and 31 December 2010 were obtained from the Consejería de Medio Ambiente of Madrid, the City Council of Saragossa and the Consejería de Medio Ambiente of Cantabria. These government agencies measure ozone concentrations using ultraviolet absorption-based instruments. Two different filtering criteria were applied when selecting the measuring stations: 1) Only stations which were free from the direct influence of local point sources and were therefore representative of background boundary layer air would be included; 2) Once data was validated and daily average values were calculated, stations having fewer than 85% of daily data points were discarded. In addition, the homogeneity of the different series was confirmed by plotting and comparing their respective histograms. The resulting 9 measuring stations were (Fig. 1):
Cantabria: Castro Urdiales (east of Santander), Reinosa (south of
n (in the urban center of Santander) and Santander), Tetua
n (southwest of Santander). Zapato
Madrid: Aranjuez (south of Madrid city), Majadahonda (north
stoles (southwest of Madrid city). east of Madrid city) and Mo
Saragossa: Roger the Flor and Picarral (both located in the urban center of Saragossa city). The number of data points and completeness of the validated hourly series (once anomalous data were removed) are shown in Table 1. Daily concentration values were calculated by averaging the 24 h period between 8 a.m. and 7 a.m. of the following day. 2.2. Weather types classification The objective method to classify the original Lamb's WT catalog developed by Jenkinson and Collison (1977) was used to classify WTs, using daily mean SLP data (2.5  2.5 latitudeelongitude) from the NCEP/NCAR Reanalysis Project (Kistler et al., 2001). Daily circulation WTs were determined using physical or geometrical approaches, such as the direction and strength of airflow and the degree of vorticity. The following indices were

12

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

Fig. 1. Location of the 3 regions and of the 9 measuring stations with ozone concentration data used in this study over Spain.

used: southerly flow (SF), westerly flow (WF), total flow (F), southerly shear vorticity (ZS), westerly shear vorticity (ZW), and total shear vorticity (Z). These 6 indices were computed using SLP values obtained from 16 grid points over Western Europe, with a central point centered close to the corresponding region (i.e., 40 N and 5 W for Madrid; 42 300 N and 5 W for Santander; and 42 ; 42 300 N and 0 E for Saragossa). Fig. 2 shows an example of the 16 grid points (p1ep16) used for the WT classification in Madrid. Specifically, following the methodology described by Trigo and DaCamara (2000), the following equations were used in order to calculate these indices:

SF ¼ 1:305½0:25ðp5 þ 2p9 þ p13 Þ  0:25ðp4 þ 2p8 þ p12Þ WF ¼ 0:5ðp12 þ p13 Þ  0:5ðp4 þ p5 Þ ZS ¼ 0:85½0:25ðp6 þ 2p10 þ p14 Þ  0:25ðp5 þ 2p9 þ p13 Þ  0:25ðp4 þ 2p8 þ p12 Þ þ 0:25ðp3 þ 2p7 þ p11 Þ ZW ¼ 1:12½0:5ðp15 þ p16 Þ  0:5ðp8 þ p9 Þ  0:91½0:5ðp8 þ p9 Þ  0:5ðp1 þ p2 Þ 1=2  F ¼ SF2 þ WF2 Z ¼ ZS þ ZW

Also according to Trigo and DaCamara (2000), the following conditions were established: 1. Direction of flow was given by tan1 (WF/SF), 180 being added if WF was positive. The appropriate direction was computed using an eight-point compass, allowing 45 per sector. 2. If jZj, the flow is essentially straight and was considered to be of a pure directional type (eight different cases, according to the directions of the compass). 3. If jZj > 2F, the pattern was considered to be of a pure cyclonic type if Z > 0, or of a pure anticyclonic type if Z < 0. 4. If F
A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

13

Table 1 Number of data points of the validated hourly series during the period 2001e2010 for the 3 regions and respective stations. Coastland stations

Inland stations

Cantabria

Daily Data Percentage

Madrid

Castro Urdiales

Reinosa

Tetu
an

3635 99.5%

3630 99.4%

3382 92.6%


n Zapato

Aranjuez

3598 98.5%

3371 92.3%

Fig. 2. Location of 16 grid points centered over the Madrid region used to compute the vorticity and flow indices.

original 26 WTs were regrouped in 10 distinct types, by classifying hybrid types as purely directional (Russo et al., 2014). The motivation behind this reclassification lies in the necessity of giving the appropriate relevance to the direction of air masses, which are highly influential in the transport of ozone precursors. An example of average maps of SLP for all 10 WTs of the classification centered over Madrid for the period 2001e2010 can be seen in Fig. 3. In addition, Fig. 4 shows a selection of the average maps of SLP for 5 WTs in summer (JJA) and winter (DJF), which were selected as they are the WTs best associated with the median value of tropospheric ozone and frequency of occurrence (Section 3.2). 2.3. Statistical analysis A monthly data series was created for each region with the intention of comparing cyclic behavior between them, by averaging measured values in every station of each region. A monthly calendar was also computed for each measuring station. Daytime and nighttime ozone concentration trends were calculated from hourly data. Daytime was defined as the 12 h average between 8 a.m. and 7 p.m., and nighttime as the average between 7 p.m. and 8 a.m. of the following day. Since our data does not follow a parametric statistical distribution, the Kendall rank correlation coefficient (Kendall, 1938) with a 95% CI was used for trend calculations. A mean monthly anomaly time series was also created from the average of all monthly anomaly regional series, the trend of which was calculated using least squares linear fitting and its significance estimated by the Kendall correlation coefficient. Moreover, given that ozone is a

Saragossa Majadahonda


stoles Mo

Picarral

Roger de F.

3370 92.3%

3634 99.5%

3624 99.2%

3240 88.7%

photochemical reaction product of air oxygen, seasonal trends were also computed from daily data, in order to evaluate the contribution of different intra-year periods to the global series trend. The relative overall and seasonal frequencies of each WT were calculated in the 3 analyzed regions, in order to assess spatial differences. The relationship between ozone concentration levels and WTs was calculated using the non-parametric Kruskal Wallis test (Kruskal and Wallis, 1952). Since only one WT series exists for each region, and with the intention of enhancing the statistical validity of the analysis, a new ozone concentration series was created for each region by concatenating the measured values in each station in the region and assigning them a WT value according to the measurement date. This resulted in three single regional ozone concentration series with varying lengths: 14,245 data points for Cantabria, 10,375 data points for Madrid and 6864 data points for Saragossa. The ManneWhitney test (Mann and Whitney, 1947) was also used for comparing the relative prevalence of each WT against the A type. The latter was chosen for being the most frequent WT in two out of the three studied regions. The Bonferroni correction (Abdi, 2007) was also applied to counteract the problem of multiple comparisons. This same analysis was performed both station and regionwise. 3. Results and discussion 3.1. Seasonal ozone variability and trends Monthly ozone mean series for each region are plotted in Fig. 5, which shows the cyclical behavior of the compound in the three studied provinces. A comparison between the values in the 3 regions shows that the maximum average level is found in Madrid (with an average daily concentration of 51.9 mg/m3), followed by Cantabria (49.8 mg/m3), while Saragossa recorded the lowest average level (30.6 mg/m3). It is noteworthy that, in the last 2 years of the study (2009e2010), the measurements of the 3 provinces tend to have similar values. Monthly mean values at the nine ozone stations are presented in Fig. 6. During the first months of the year there is a continual increase in the mean value of ozone in the four stations in Cantabria (coast stations) due to solar radiation and temperature increase, amongst other factors; in April and May, the compound reaches its peak levels; between June and December, the pollutant follows a downward trend although a new local maximum can be found in the month of August. However, the monthly mean values of ozone in the 5 inland stations (Madrid and Saragossa) show a progressive increase until April; during May the values are maintained; and they grow again in June, peaking in July, at which time concentration levels gradually decrease to a minimum in December. These cycles are dependent on a number of processes including photochemistry, deposition and transport, acting on local, regional and global scales (e.g., Derwent et al., 2004). The differences between the minimum (recorded in December in all stations) and maximum monthly average values is close to 30 mg/m3 for stations in Cantabria and Saragossa; however, for stations in Madrid, the average

14

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

Fig. 3. Average maps of the SLP (period 2001e2010) for all 10 WTs of the modified Jenkinson and Collinson classification centered over Madrid.

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

15

Fig. 4. Average maps of the SLP (period 2001e2010) for summer (JJA) and winter (DJF) for 5 WTs of the modified Jenkinson and Collinson classification centered over Madrid.

16

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

Fig. 5. Mean monthly ozone series January 2001 to December 2010 (by region).

maximumeminimum monthly difference is much bigger, reaching around 50 mg/m3. It is worth mentioning that Madrid is the city with the largest industrial activity in the study, which could be the main cause of the increased difference. The differences in ozone annual cycles have been described by other researchers. Thus, some studies have attributed the difference in the seasonal ozone maximum to the latitude of the studied point. For example, Scheel et al. (1997) compared different European cities and reported that, while those located in the north reached the maximum in spring, the southern ones did so in summer. In addition, other studies have also found spring maximums in coastal cities and summer highs in inner cities, coinciding
ndez-Ferna
ndez et al., 2011). Given the with our results (e.g., Ferna similarity in latitude for the cities in this work, the second hypothesis seems to align better with the results of this study. Thus, although the cause for the difference in ozone cycles by proximity to the sea is unknown, the spring maximum could be attributed to increased photochemical activity as well as stratospheric intrusions at this time of the year (Atlas et al., 2003), while higher values in Madrid and Saragossa in June and July could be explained by the large temperature increase that happens in these regions in the summer. In general, inland regions in the Iberian Peninsula present higher solar radiation than coastal zones and clear temperature maximum values during the summer, causing the peak in ozone concentration levels to appear later in the year. However, in coastal cities, the influence of the sea results in smoother temperature changes, and, thus, in smaller temperature ranges. Moreover, in polluted areas, a summer maximum is more likely, due to the influence of local photochemical ozone production from precursor emissions (Kalabokas et al., 2000), and Saragossa and Madrid both have a much higher industrial activity than Cantabria. Apart from the aforementioned seasonal cycle, the level of surface ozone exhibits pronounced diurnal variations (not shown), which present different patterns depending on the geographical

characteristics of the studied site (Monks, 2000; Adame et al., 2008). In the studied regions, diurnal contaminant levels are, in general, and as expected, higher than nighttime concentrations every day of the week. However, there is a contrast in concentration differences by time of day; dayenight differences are particularly marked in Madrid, where the average difference (calculated from hourly data over the 10-year period and the 3 measurement stations) is 25.1 mg/m3, a value which is much smaller in Saragossa (4.6 mg/m3) and Cantabria (10.8 mg/m3). Additionally, monthly calendars for diurnal and nocturnal values show that in two of the
n and Cantabria stations, both located closest to the sea (Tetua Castro Urdiales), in some winter months (January and December in Tetu
an, and November, December and January in Castro Urdiales) nighttime ozone levels were higher than during daytime. The calculated Kendall rank correlation coefficients and their corresponding p-value are represented in Table 2, which summarizes the trend results obtained from the nine monitoring sites during the study period. Tropospheric ozone concentration in 6 of the 9 studied stations shows an upward trend both during daytime and nighttime. In fact, the mean monthly anomaly time series derived from the average values in all regions, which presents a significant increase (p < 0.05) of 0.068 mg/m3 per month (equivalent to an increase 8.16 mg/m3 per decade, or 18.49% per decade relative to the mean value), corroborates the existence of an upward trend (Fig. 7). This significant increase has been found by other authors in urban and industrial sites in Spain, and has been attributed to a decrease of NOX emissions, particularly NO (Querol et al., 2014). Under lower NO concentrations in an urban environment, less ozone is consumed to oxidize this pollutant into NO2; this is also the most accepted explanation for the “weekend effect” (Heuss
nez et al., 2005; Sanchezet al., 2003; Murphy et al., 2007; Jime Lorenzo et al., 2012). This could also explain why in 3 of the sta
stoles) no upward trend for tions (Reinosa, Majadahonda and Mo ozone concentrations could be found as these 3 measurement points are located in low-traffic, less densely populated areas, in which a potential decrease in nitrogen oxides may not have occurred or may not be noticeable, keeping ozone oxidation unaltered. Another factor that may explain the aforementioned upward trend of tropospheric ozone is the increase in surface solar radiation during the 2000s, also known as brightening period (Wild et al., 2005), due to a decrease in cloudiness and atmospheric aerosols over Spain (Sanchez-Lorenzo et al., 2013; Mateos et al., 2014). Given that surface solar radiation shows a varying tendency depending on the moment of the year (Sanchez-Lorenzo

Fig. 6. Monthly mean values at the nine ozone stations grouped by proximity to the sea.

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

17

Table 2 Kendall rank correlation coefficients and their p-values for the nine stations relative to the time of day and season. Statistically significant values are marked with an asterisk, and bold when positive. Ozone stations

Stat. param.

Trend Time of day

Cantabria

Castro Reinosa
n Tetua
n Zapato

Madrid

Aranjuez Majadahonda
stoles Mo

Saragossa

El Picarral Roger de Flor

tau p tau p tau p tau p tau p tau p tau p tau p tau p

MK MK MK MK MK MK MK MK MK

Season

Day

Night

Winter

Spring

Summer

Autumn

0.09* <0.0001 0.07* <0.0001 0.12* <0.0001 0.08* <0.0001 0.17* <0.0001 0.01 0.28 0.04* <0.0001 0.22* <0.0001 0.17* <0.0001

0.09* <0.0001 0.10* <0.0001 0.07* <0.0001 0.05* <0.0001 0.10* <0.0001 0.01 0.26 0.05* <0.0001 0.17* <0.0001 0.13* <0.0001

0.07* 0.002 0.09* <0.0001 0.08* 0.001 0.10* <0.0001 0.10* <0.0001 0.04 0.06 0.08* <0.0001 0.10* <0.0001 0.13* <0.0001

0.15* <0.0001 0.15* <0.0001 0.17* <0.0001 0.12* <0.0001 0.36* <0.0001 0.08* <0.0001 0.14* <0.0001 0.32* <0.0001 0.19* <0.0001

0.09* <0.0001 0.15* <0.0001 0.17* <0.0001 0.02 0.34 0.32* <0.0001 0.11* <0.0001 0.08* <0.0001 0.38* <0.0001 0.33* <0.0001

0.18* <0.0001 0.03 0.24 0.17* <0.0001 0.14* <0.0001 0.2 <0.0001 0.01 0.58 0.02 0.31 0.19* <0.0001 0.11* <0.0001

et al., 2009, 2013), seasonal (winter, DJF; spring, MMA; summer, JJA; autumn, SON) ozone trends were also calculated (Table 2). In general, the aforementioned upward trend continues throughout all of the seasons, also in line with the widespread increase observed in surface solar radiation in all seasons over Spain
n (Cantabria) and (Sanchez-Lorenzo et al., 2013), although in Zapato
stoles (Madrid) the trend was not statistically significant in in Mo summer and autumn, respectively. Equally, in Aranjuez (Madrid), the pollutant presented a downward trend during the entire evaluated period. Ozone in Reinosa (Cantabria) also showed a downward tendency during the studied period except in autumn, in which no statistically significant trend was found. In Majadahonda (Madrid), the downward trends were statistically significant (with a 95% CI) only in spring and summer. 3.2. Weather types and ozone concentration The 3652 days of the studied period were classified according to the Jenkinson and Collinson WT classification, after grouping each

Fig. 7. Monthly anomaly time series of average ozone values from all regions (solid line). The linear regression equation is indicated (dash line).

hybrid type with its corresponding purely directional type, and leaving the A and C types as pure WT. Table 3 shows the recorded frequencies for each weather type. In Cantabria and Madrid, the most frequent type was A, with an average frequency of 27.5% and 22.6% respectively. It was followed by the E type, with an annual average of around 49 days per year (13.5%) in Cantabria, and of 59 days per year (16.1%) in Madrid. However, the NE type was the most frequent in Saragossa during the study period with an annual average of around 73 days per year (20.0%), followed by the A type (18.6%). The S type was the less frequent synoptic atmospheric pattern in the three locations of our study. On seasonal basis (Fig. 8), type A was found to be more prevalent in Cantabria and Madrid throughout all seasons, except for summer in Madrid, when type E was more frequent. In Saragossa, NE was the most common type throughout all the year, with the exception of winter, when there were more days with type A. For all regions, types NE and E are more frequent during the summer, while types SW and W are less common in this season. The median ozone concentration for every WT and in each region is shown in Fig. 9. The results of the KruskalleWallis test show that there are highly significant differences (p < 0.001 in all cases) in median ozone concentration across WTs. It is noteworthy that, even if ozone concentration levels are different in the three regions, their median values under the A and C weather types are very similar and higher in the latter case. The detailed results of the ManneWhitney pairwise comparisons between type A days and days with other WTs for the 9 studied measuring stations are shown in Table 5, which confirm that the A WT category had significantly lower median ozone concentration than the C WT category in Cantabria and Madrid; however in Saragossa no significantly differences were found between those two WTs. In addition, comparison of O3 concentrations between type A days and days with directional WTs showed some statistically significant differences depending on the region (Table 4). Thus, in Cantabria, the difference in ozone by WT was found for all directional WTs except S and SE. It is also noteworthy that whenever a statistically significant difference was found between the A and a directional WT, the latter showed higher concentration values than the former. The highest levels were recorded in type N days, showing a median concentration of 59.6 mg/m3, 17.4 6 mg/m3

18

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

Table 3 Frequency expressed in number of days and percentage for each weather type per region. WT

A C N NE NW S SE SW E W

Cantabria

Madrid

Saragossa

Frequency (no. Days)

Percentage (%)

Frequency (no. Days)

Percentage (%)

Frequency (no. Days)

Percentage (%)

1005 282 243 317 302 103 243 246 494 417

27.5% 7.7% 6.7% 8.7% 8.3% 2.8% 6.7% 6.7% 13.5% 11.4%

825 258 360 410 336 110 280 171 588 314

22.6% 7.1% 9.9% 11.2% 9.2% 3.0% 7.7% 4.7% 16.1% 8.6%

680 286 365 729 241 148 217 225 445 316

18.6% 7.8% 10.0% 20.0% 6.6% 4.1% 5.9% 6.2% 12.2% 8.7%

more than the type A days median. On N days, air intake originates from northernmost and oceanic regions, which is associated with an increase in precipitation and a decrease in temperature. Rainfall contributes to NO deposition, therefore NO air concentration

Fig. 8. Seasonal Frequency of each weather type in Cantabria (A), Madrid (B), and Saragossa (C).

decreases, which would stop O3eNO interactions, causing an increase in ozone levels. Moreover, in days with this type of circulation, air intake from the British Isles could transport ozone precursors into Cantabria, the interaction of which could increase ozone levels. Nevertheless, it is worth noting that days with type N are more frequent in the summertime, the period of the year that presents the highest ozone concentration values due to elevated photochemical activity that may be favored as the Iberian Peninsula is dominated by the Azores high pressure system during the N WT in summer (Fig. 4). In Madrid, the results also show that days in the A WT category had a significantly lower median ozone concentration than those in directional categories, except for the S and SW types, for which no statistically significant difference was found. Moreover, in this region, the highest ozone levels were also recorded in type N days, which had a median concentration of 59.6 mg/ m3, 17.9 mg/m3 more than the median for type A days. In Saragossa, days with A weather type showed statistically significant differences in concentration with days of types N, NE, SW and E, with type A days presenting lower ozone levels. E WT days registered the highest ozone concentration, with a median of 55.28 mg/m3, 12.48 mg/m3 higher than that of type A days. High ozone values under eastern component situations in the Iberian Peninsula have already been described by other authors. The presence of an anticyclone located between the British Isles promote the transport of hot and dry air masses from Central Europe (Saavedra et al., 2012), which foments the formation of this secondary pollutant (Russo et al., 2014). Additionally, it is well known that pressure change rates in the air columns from the surface to any altitude differ with the temperature of air masses, being faster under cold conditions and lower under warm ones. This suggests that the A WT during the winter could be related to low surface temperatures (the opposite holding for the summer) due to low level temperature inversions (Fig. 4), and consequently A conditions in the winter may not be maintained along the air column but they may in the summer. Thus, we have analyzed the relationship between the A weather type and the levels of ozone concentration in the winter (DJF) and

Fig. 9. Median of ozone concentration per weather type in each region.

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

found near the Iberian Peninsula, while in summer low pressure areas are much less defined (Fig. 4). Moreover, throughout the winter, N conditions present significantly higher ozone values than those found in type A days, while in summer this difference is not statistically significant in Madrid and Zaragoza; this could be explained by the fact that, in the winter, under this configuration, strong northern winds are present, potentially transporting polluted air masses from the UK, while the most prominent feature in the summer is the Azores anticyclone located west of the Iberian Peninsula, which favors cloudless skies (Fig. 4). While most of the previous literature describes an association between low O3 levels and cool and cloudy cyclonic conditions (e.g., Comrie and Yarnal, 1992), and between higher ozone concentrations and stationary or slowly migrating anticyclonic or high-pressure systems that reduce pollution, dispersion, diffusion and deposition (Schichtel and Husar, 2001; Rao et al., 2003), the low ozone concentrations found in this work in type A days coincide with other recent results obtained in Portugal (Russo et al., 2014). The explanation for these findings could be related with the considerable increase of nitrogen oxides (NOX) under anticyclonic

Table 4 Results of the ManneWhitney pairwise comparison between each Weather type and A (Anticyclonic). Statistically significant differences found against the A type are marked with a point. Weather type

Cantabria Madrid Saragossa

C

N

NE

NW

C C

C C C

C C C

C C C

S

SE

SW

E

W

C

C C C

C C

C C

19

summer (JJA) separately. Analysis of the association between tropospheric ozone and weather types in winter and summer (Table 6) shows that in both seasons the lowest ozone levels are found in type A days, although during the summer the differences in concentration between WTs were lower, especially in Madrid and Saragossa. Moreover, in winter, ozone levels in type C days were found to be lower than in days with a directional WT. The differences in ozone levels in type C days (which are smaller in the winter) could be associated to the disparity in configuration depending on the season: in winter, an area of low pressure can be

Table 5 Results of ManneWhitney pairwise comparisons between type A days and days with other WTs by station (ManneWhitney U coefficients and p-values). Statistically significant values with a confidence level of 95% are bold and marked with an asterisk. Stat. param.

Cantabria

Castro Reinosa Tetu
an
n Zapato

Madrid

Aranjuez Majada-honda
stoles Mo

Saragossa

El Picarral Roger de Flor

M.W. p M.W. p M.W. p M.W. p M.W. p M.W. p M.W. p M.W. p M.W. p

U U U U U U U U U

Weather type C

N

NE

NW

S

SE

SW

E

W

6.66* <0.0001 9.12* <0.0001 6.71* <0.0001 7.85* <0.0001 6.46* <0.0001 3.60* 0.0003 5.55* <0.0001 2.44 0.015 2.38 0.017

13.38* <0.0001 13.30* <0.0001 9.05* <0.0001 12.43* <0.0001 8.76* <0.0001 9.57* <0.0001 11.38* <0.0001 11.83* <0.0001 7.45* <0.0001

12.92* <0.0001 14.84* <0.0001 8.59* <0.0001 8.23* <0.0001 9.16* <0.0001 8.43* <0.0001 8.28* <0.0001 16.07* <0.0001 10.96* <0.0001

13.72* <0.0001 6.43* <0.0001 8.52* <0.0001 13.82* <0.0001 5.70* <0.0001 7.36* <0.0001 9.91* <0.0001 4.29* <0.0001 1.6 0.11

1.1 0.271 1.02 0.304 0.995 0.32 0.255 0.798 3.62* <0.0001 1.59 0.112 1.13 0.258 2.14 0.033 1.42 0.155

1.3 0.193 0.092 0.927 2.4 0.016 0.049 0.961 6.78* <0.0001 3.80* 0.0001 3.13 0.002 0.268 0.789 3.24 0.001

6.93* <0.0001 8.6 <0.0001 6.74* <0.0001 6.95* <0.0001 1.52 0.129 2.8 0.005 0.512 0.609 3.61* 0.0003 1.29 0.198

7.49* <0.0001 10.68* <0.0001 6.34* <0.0001 2.57 0.01 9.32* <0.0001 9.96* <0.0001 7.46* <0.0001 4.47* <0.0001 5.27* <0.0001

7.65* <0.0001 4.92* <0.0001 4.79* <0.0001 6.30* <0.0001 2.88 0.004 3.89* <0.0001 6.80* <0.0001 0.877 0.38 1.9 0.057

Table 6 Median and 10th and 90th percentiles (P10 and P90) under each WT, by region. Statistically significant results of ManneWhitney pairwise comparison between type A days and days with other WTs are marked in bold and with an asterisk.

Cantabria

Winter

Summer

Madrid

Winter

Summer

Saragossa

Winter

Summer

Median P10 P90 Median P10 P90 Median P10 P90 Median P10 P90 Median P10 P90 Median P10 P90

A

C

N

NE

NW

S

SE

SW

E

W

31.35 12.10 60.82 46.04 31.03 66.39 21.25 8.13 47.15 68.46 46.82 92.95 8.14 2.13 21.79 41.32 20.57 63.87

40.45* 17.82 71.54 57.26* 36.89 87.72 37.44* 15.13 66.37 66.96 40.65 100.46 14.18* 4.56 38.06 38.05 14.24 68.61

57.84* 35.80 82.15 51.56* 33.49 73.45 44.17* 18.29 69.88 71.49 52.03 95.89 25.95* 12.39 49.85 43.05 22.27 63.14

54.41* 26.72 78.90 53.41* 31.82 81.60 31.36* 15.22 53.33 75.71* 53.34 99.56 25.66* 6.67 56.01 43.52 22.84 63.14

62.08 35.00 81.83 49.56 31.63 67.10 50.63* 27.17 72.38 63.17 45.52 85.56 18.91* 5.95 44.33 39.69 19.73 63.03

36.04 13.60 65.96 57.3* 35.73 84.54 33.63* 9.42 66.33 70.73 33.51 117.49 9.78 4.08 23.68 38.66 13.33 61.37

36.04 13.60 65.96 53.45* 32.88 79.56 33.93* 10.41 63.14 74.6* 52.79 116.99 10.00 3.70 31.82 40.03 14.61 59.32

35.71* 10.24 61.94 54.88* 39.26 76.45 29.75* 9.81 57.62 70.46 47.03 93.04 10.40 4.07 27.60 32.38 8.69 53.44

52.79* 23.93 77.42 52.62* 33.59 77.29 33.68* 14.50 62.51 74.10* 48.20 101.74 11.31 2.09 35.14 47.68* 24.44 67.24

42.94* 20.23 71.28 48.51 32.17 69.32 45.65* 22.00 68.42 58.37* 40.78 82.80 15.53* 3.04 35.35 28.09* 13.23 44.51

20

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

circulations (Buchhold et al., 2010; Le
sniok et al., 2010), resulting in an increase in ozone oxidation due to NO and therefore a reduction in the total concentration of surface ozone. Moreover, airflow along the flanks of cyclonic systems can transport ozone precursors, creating the conditions for an ozone event (Lennartson and Schwartz, 1999; Tanner and Law, 2002). This, coupled with our reduced WT classification (in which hybrid WTs have been classified according to their directional component), could explain the results presented in this paper. For this reason, another set of ManneWhitney tests was performed using the original classification of 26 WTs, with the intention of assessing the relative prevalence of each WT against the A type. These results show that while in many cases ozone concentrations under hybrid A WTs (higher-pressure days with a directional component that could favor precursor transport) are significantly higher than those under purely A WT (particularly when a North component is present), the highest concentration levels are still found in nonanticyclonic WT (N for Cantabria, CE for Madrid and NE for Zaragoza). Most authors have predicted an increase in tropospheric ozone concentration levels linked to climate change and due to an increased frequency of stable, anticyclonic conditions with little boundary layer ventilation and associated high temperatures, cloud-free conditions and large solar radiation inputs (Ebi and McGregor, 2008). Similarly, our analysis finds an ascending trend in this secondary pollutant but with maximum values occurring under non-anticyclonic synoptic conditions, which contrasts to the findings from most of the previous studies. In order to explain these results, as mentioned above, we hypothesize that since ozone and nitrogen monoxide have a mutually antagonistic behavior, the reduction in nitrogen monoxide during the last decade may explain the ozone increment, and given that the highest levels of nitrogen oxides are reached under anticyclonic circulations, it is expectable for ozone levels to decrease under this circulation pattern; another possible cause for our results are that the positive ozone trend and its association with directional weather types could be related to stratospheric ozone intrusions. The exchange between the stratosphere and troposphere mainly occurs in tropopause folds, cut-off lows and streamer-type systems (Rao et al., 2008). Stratospheric intrusions in tropopause folds are mainly associated to midlatitude cyclones (Browning, 1997), where enhanced ozone mixing ratios can subsequently mix with tropospheric background air (Ancellet et al., 1994; Stohl et al., 2003); however, this study shows that maximum concentration values happen under directional weather types, which could be attributed to the absence of tropopause folds directly over the Iberian Peninsula, making the effects of other folds only noticeable when their intrusions are transported. The proportion of tropospheric ozone attributed to stratospheric intrusions is close to 9% (Seinfeld and Pandis, 2006), but modeling studies (Roelofs and Lelieveld, 1997) indicate that the stratospheric contribution to ozone in the troposphere could be as large as that from net photochemical production, which was also confirmed in a multi-model ensemble simulation (Stevenson et al., 2006; Skerlak et al., 2013). 4. Conclusions In this study we have described the surface ozone concentration variability and trends in Spain, as well as its relationship with daily atmospheric patters, using records from 9 stations of different regions during the period 2001e2010. Annual monthly series show that ozone follows a cyclic pattern with differences depending on proximity to the sea; coastal and inland stations show higher values in April and June, respectively. Additionally, from the results of this study we can conclude that

those cities with higher industrial activity and a bigger population, and thus a higher concentration of ozone precursors, do not necessarily present higher values of ozone concentrations (since Saragossa shows lower concentrations than Cantabria). Moreover, ozone concentrations show an upward trend, even as more restrictive legislation on emissions is implemented. These two facts should motivate future research into the causes of the increase in ozone levels beyond anthropogenic precursor emissions. Specifically, tropospheric ozone concentration in 6 of the 9 studied stations shows an upward trend. Moreover, in general, this upward trend is seen throughout all of the seasons. These results are in line with the reported decrease of the concentration levels of ozone precursors over the studied period in Spain due to more restrictive European laws (Querol et al., 2014), which reduce the ozone loss by the NOX-catalyzed mechanism. Equally, the registered increase in ozone concentrations could be also in part related to the strong increase in surface solar radiation observed in Spain since the 2000s due to a decrease of clouds and atmospheric aerosol load (Sanchez-Lorenzo et al., 2013; Mateos et al., 2014). The study of weather types (WTs) following the Jenkinson and Collinson classification and reducing the 26 original types to 10 categories show that the Anticyclonic synoptic pattern is the most frequent in Cantabria and Madrid, followed by the East type, while the Northeast type is the most common in Saragossa. In addition, the findings of this study highlight that, interestingly, the concentration of ozone levels were statistically significantly lower in days with Anticyclonic weather than in the rest of meteorological situations. Maximum ozone concentration values in Spain are found in days with directional WTs (i.e., northern and eastern component). In any case, it is worth noting that the WT classification used here only relies on surface pressure conditions, i.e., it does not consider the upper levels of the atmosphere (particularly 500 hPa), where divergence (high pressure) or convergence (low pressure) could be detected and indicate the nature of surface pressure (reinforced or not). This is currently a drawback of the WT classification, and further research is needed in order to consider the middle levels of the troposphere. Finally, due to the detrimental effect that ozone has on human health, the relationship between synoptic weather patterns and daily ozone levels shown could potentially be used for formulating pollution mitigation strategies depending on forecast weather types.

Acknowledgments We would like to thank Dr Cortesi (researcher for the Centre
en de Recherche et de Formation Avance
e en Calcul ScienEurope tifique), the environmental departments (Consejería de Medioambiente) in Cantabria and in Madrid, and the environmental and sustainability agency (Agencia de Medio Ambiente y Sostenibilidad) of City of Zaragoza for their collaboration and contributions.
n Mapfre and This work has also been partially funded by Fundacio by project HIDROCAES (CGL2011-27574-C02-01). A. SanchezLorenzo was supported by the Postdoctoral Fellowships 2011-BPB-00078 and JCI-2012-12508, and the Spanish Ministry of Science and Innovation Project CGL2010-18546.

References  a
k Corrections for Multiple Comparisons. In: N. J. Abdi, H., 2007. Bonferroni and Sid Encyclopedia of Measurement and Statistics. Sage, Thousand Oaks, CA. Adame, J.A., Bolívar, J.P., de la Morena, B.A., 2010. Surface ozone measurements in the southwest of the Iberian Peninsula (Huelva, Spain). Environ. Sci. Pollut. Res. 17, 355e368.

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22 Adame, J.A., Lozano, A., Bolívar, J.P., de la Morena, B.A., Contreras, J., Godoy, F., 2008. Behavior, distribution and variability of surface ozone at an arid region in the south of Iberian Peninsula (Seville, Spain). Chemosphere 70, 841e849. Ancellet, G., Beekmann, M., Papayannis, A., 1994. Impact of cutoff low development on downward transport of ozone in the troposphere. J. Geophys. Res. 99, 3451e3468. Andrade, C., Santos, J.A., Pinto, J.G., Corte-Real, J., 2011. Large-scale atmospheric dynamics of the wet winter 2009e2010 and its impact on hydrology in Portugal. Clim. Res. 46, 29e41. Atlas, E.L., Ridley, B.A., Cantrell, C.A., 2003. The tropospheric ozone production about the spring equinox (TOPSE) experiment: introduction. J. Geophys. Res. Atmos. 108 http://dx.doi.org/10.1029/2002JD003172. Beck, C., Philipp, A., 2010. Evaluation and comparison of circulation type classifications for the European domain. Phys. Chem. Earth Parts A B C 37 (9e12), 374e387. http://dx.doi.org/10.1016/j.pce.2010.01.001. Browning, K.A., 1997. The dry intrusion perspective of extra-tropical cyclone development. Meteorol. Appl. 4, 317e324. Buchholza, S., Junka, J., Kreina, A., Heinemannb, G., 2010. Lucien Hoffmanna air pollution characteristics associated with mesoscale atmospheric patterns in northwest continental Europe. Atmos. Environ. 44 (39), 5183e5190. http:// dx.doi.org/10.1016/j.atmosenv.2010.08.053. Comrie, A.C., Yarnal, B., 1992. Relationships between synoptic-scale atmospheric circulation and ozone concentrations in metropolitan Pittsburgh, Pennsylvania. Atmos. Environ. Part B. Urban Atmos. 26, 301e312. http://dx.doi.org/10.1016/ 0957-1272(92)90006-E. Cortesi, N., Trigo, R.M., Gonzalez-Hidalgo, J.C., Ramos, A.M., 2013. Modelling monthly precipitation with circulation weather types for a dense network of stations over Iberia. Hydrol. Earth Syst. Sci. 17, 665e678. Demuzere, M., Werner, M., van Lipzig, N.P.M., Roeckner, E., 2009. An analysis of present and future ECHAM5 pressure fields using a classification of circulation patterns. Int. J. Climatol. 29, 1796e1810. ~ as, C., Ferna
ndez, M.C., Can ~ ete, S., Carretero, L., Liger, E., 2002. Assessment of Duen ozone variations and meteorological effects in an urban area in the Mediterranean coast. Sci. Total Environ. 299 (1e3), 97e113. http://dx.doi.org/10.1016/ S0048-9697(02)00251-6. Derwent, R.G., Stevenson, D.S., Collins, W.J., Johnson, C.E., 2004. Intercontinental transport and the origins of the ozone observed at surface sites in Europe. Atmos. Environ. 38, 1891e1901. ~ as, C., Ferna
ndez, M.C., Carretero, J., Liger, E., 2004. Analyses of ozone in urban Duen and rural sites in Malaga. Chemosphere 56, 631e639. Ebi, K.L., McGregor, G., 2008. Climate change, tropospheric ozone and particulate matter, and health impacts. Environ. Health Perspect. 116 (11), 1449e1455. http://dx.doi.org/10.1289/ehp.11463. Esteban, P., Martin-Vide, J., Mases, M., 2006. Daily atmospheric circulation catalogue for Western Europe using multivariate techniques. Int. J. Climatol. 26, 1501e1515.
ndez-Fern
Ferna andez, M.I., Gallego, M.C., García, J.A., Acero, F.J., 2011. A study of surface ozone variability over the Iberian Peninsula during the last fifty years. Atmos. Environ. 45, 1946e1959.
nchez, M.L., Pe
rez, I.A., de la Torre, B., 2005. Ground level ozone García, M.A., Sa concentrations at a rural location in northern Spain. Sci. Total Environ. 348, 135e150. Gerasopoulos, E., Kouvarakis, G., Vrekoussis, M., Donoussis, C., Mihalopoulos, N., Kanakidou, M., 2006. Photochemical ozone production in the eastern Mediterranean. Atmos. Environ. 40, 3057e3069. Halonen, J.I., Lanki, T., Tiittanen, P., Niemi, J.V., Loh, M., 2010. Ozone and causespecific cardiorespiratory morbidity and mortality. J. Epidemiol. Community Health 64, 814e820. Hess, P., Brezowsky, H., 1952. Katalog der großwetterlagen europas (catalog of the european large scale weather types). Ber. Dt. Wetterd. in der US-Zone 33. Heuss, J.M., Kahlbaum, D.F., Wolff, G.T., 2003. Weekday/weekend ozone differences: what can we learn from them? J. Air Waste Manage. 53, 772e788. Hoinka, K.P., Castro, M., 2003. The iberian Peninsula thermal low. Q. J. R. Meteorol. Soc. 129, 1491e1511. http://dx.doi.org/10.1256/qj.01.189. Jenkinson, A.F., Collison, F.P., 1977. An Initial Climatology of Galesover the North Sea. Synoptic Climatology Branch Memorandum, No. 62. Meteorological Office, Bracknell.
nez, P., Parra, R., Gasso
, S., Baldasano, J.M., 2005. Modeling the ozone weekend Jime effect in very complex terrains: a case study in the Northeastern Iberian Peninsula. Atmos. Environ. 39 (3), 429e444. http://dx.doi.org/10.1016/ j.atmosenv.2004.09.065. Jones, P.D., Hulme, M., Briffa, K.R., 1993. A comparison of lamb circulation types with an objective classification scheme. Int. J. Climatol. 13, 655e663. Kalabokas, P.D., Viras, L.G., Bartzis, J.G., Repapis, C.C., 2000. Mediterranean rural ozone characteristics around the urban area of Athens. Atmos. Environ. 34, 5199e5208. Kendall, M., 1938. A new measure of rank correlation. Biometrika 30 (1e2), 81e89. http://dx.doi.org/10.1093/biomet/30.1-2.81. Kistler, R., Kalnay, E., Collins, W., Saha, S., White, G., Woollen, J., Chelliah, M., Ebisuzaki, W., Kanamitsu, M., Kousky, V., van den Dool, H., Jenne, R., Fiorino, M., 2001. The NCEP-NCAR 50-year reanalysis: monthly means CD-ROM and documentation. Bull. Am. Meteorol. Soc. 82, 247e267. Kruskal, W.H., Wallis, W.A., 1952. Use of ranks in one-criterion variance analysis. J. Am. Stat. Assoc. 47 (260), 583e621.

21

Lamb, H.H., Britain, G., 1972. British Isles Weather Types and a Register of the Daily Sequence of Circulation Patterns 1861e1971. HM Stationery Office. Le
sniok, M., Małarzewski, Ł., Nied
zwied
z, T., 2010. Classification of circulation types for southern Poland with an application to air pollution concentration in upper Silesia. Phys. Chem. Earth Parts A B C 35 (9e12), 516e522. http://dx.doi.org/ 10.1016/j.pce.2009.11.006. Lennartson, G.J., Schwartz, M.D., 1999. A synoptic climatology of surface-level ozone in eastern Wisconsin, USA. Clim. Res. 13, 207e220. Mann, H.B., Whitney, D.R., 1947. On a test of whether one of two random variables is stochastically larger than the other. Ann. Math. Stat. 18 (1), 50e60. http:// dx.doi.org/10.1214/aoms/1177730491. Martin-Vide, J., 2001. Limitations of an objective weather-typing system for the Iberian Peninsula. Weather 56, 248e250.
n, M., Cachorro, V.E., Calbo
, J., Costa, M.J., Mateos, D., Sanchez-Lorenzo, A., Anto Torres, B., Wild, M., 2014. Quantifying the role of aerosols and clouds to the strong brightening since the early 2000s in the Iberian Peninsula. J. Geophys. Res. Atmos. 119 (17), 10382e10393. Monks, P.S., 2000. A review of the observations and origins of the spring ozone maximum. Atmos. Environ. 34, 3545e3561. Murphy, J.G., Day, D.A., Cleary, P.A., Wooldridge, P.J., Millet, D.B., Goldstein, A.H., Cohen, R.C., 2007. The weekend effect within and downwind of Sacramento e part 1: observations of ozone, nitrogen oxides, and VOC reactivity. Atmos. Chem. Phys. 7, 5327e5339. http://dx.doi.org/10.5194/acp-7-5327-2007.
lez-Hidalgo, J.C., 2013. Weather types, runoff Nadal-Romero, E., Cortesi, N., Gonza and sediment yield in a Mediterranean mountain landscape. Earth Surf. Process. Landforms 39, 427e438. http://dx.doi.org/10.1002/esp.3451. Parrish, D.D., Law, K.S., Staehelin, J., Derwent, R., Cooper, O.R., Tanimoto, H., VolzThomas, A., Gilge, S., Scheel, H.E., Steinbacher, M., Chan, E., 2012. Long-term changes in lower tropospheric baseline ozone concentrations at northern midlatitudes. Atmos. Chem. Phys. 12, 13881e13931. http://dx.doi.org/10.5194/acpd12-13881-2012. Paoletti, E., Marco, A., Beddows, D., Harrison, R., Manning, W., 2014. Ozone levels in European and USA cities are increasing more than at rural sites, while peak values are decreasing. Environ. Pollut. 192, 295e299. http://dx.doi.org/10.1016/ j.envpol.2014.04.040. Philipp, A., et al., 2010. Cost733cat e a database of weather and circulation type classifications. Phys. Chem. Earth Parts A B C 35 (9e12), 360e373. http:// dx.doi.org/10.1016/j.pce.2009.12.010. Querol, X., et al., 2014. 2001e2012 trends on air quality in Spain. Sci. Total Environ. 490, 957e969. http://dx.doi.org/10.1016/j.scitotenv.2014.05.074. Rao, S.T., Ku, J.Y., Berman, S., Zhang, D., Mao, H., 2003. Summertime characteristics of the atmospheric boundary layer and relationships to ozone levels over the eastern United States. Pure Appl. Geophys. 160, 21e55. Rao, N.T., Arvelius, J., Kirkwood, S., 2008. Climatology of tropopause folds over a European Arctic station (esrange). J. Geophys. Res. Atmos. (1984e2012) 113. http://dx.doi.org/10.1029/2007JD009638. Roelofs, G.J., Lelieveld, J., 1997. Model study of the influence of cross-tropopause O3 transports on tropospheric O3 levels. Tellus, Ser. B 49, 38e55. Russo, A., Trigo, R.M., Martins, H., Mendes, M.T., 2014. NO2, PM10 and O3 urban concentrations and its association with circulation weather types in Portugal. Atmos. Environ. 89, 768e785. http://dx.doi.org/10.1016/j.atmosenv.2014.02.010. Saavedra, S., Rodríguez, A., Taboada, J.J., Souto, J.A., Casares, J.J., 2012. Synoptic patterns and air mass transport during ozone episodes in northwestern Iberia. Sci. Total Environ. 441, 97e110. Sanchez-Lorenzo, A., Calbo, J., Brunetti, M., Deser, C., 2009. Dimming/brightening over the Iberian Peninsula: trends in sunshine duration and cloud cover, and their relations with atmospheric circulation. J. Geophys. Res. 114, D00D09. http://dx.doi.org/10.1029/2008JD011394.
, J., Vogl, S., Sanchez-Lorenzo, A., Laux, P., Hendricks-Franssen, H.-J., Calbo Georgoulias, A., Quaas, J., 2012. Assessing large-scale weekly cycles in meteorological variables: a review. Atmos. Chem. Phys. 12, 5755e5771.
, J., Wild, M., 2013. Global and diffuse solar radiation in Sanchez-Lorenzo, A., Calbo Spain: building a homogeneous dataset and assessing their trends. Glob. Planet. Change 100, 343e352. Scheel, H.E., Areskoug, H., Geib, H., Gomiscek, B., Granby, K., Haszpra, L., Klasinc, L., Kley, D., Laurila, T., Lindskrog, A., Roemer, M., Schmitt, R., Simmonds, P., Solberg, S., Toupance, G., 1997. On the spatial distribution and seasonal variation of lower-troposphere ozone over Europe. J. Atmos. Chem. 28, 11e28. Schichtel, B.A., Husar, R.B., 2001. Eastern North American transport climatology during high- and low-ozone days. Atmos. Environ. 35, 1029e1038. Sebald, L., Treffeisen, R., Reimer, E., Hies, T., 2000. Spectral analysis of air pollutants. Part 2: ozone time series. Atmos. Environ. 34, 3503e3509. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics: from Air Pollution to Climate Change, second ed. John Wiley, Hoboken, NJ. Sekiya, T., Sudo, K., 2012. Role of meteorological variability in global tropospheric ozone during 1970e2008. J. Geophys. Res. Atmos. 117, D18. http://dx.doi.org/ 10.1029/2012JD018054. Skerlak, B., Sprenger, M., Wernli, H., 2013. A global climatology of stratosphereetroposphere exchange using the ERA-interim data set from 1979 to 2011. Atmos. Chem. Phys. Discuss. 13 (5), 11537e11595. Stevenson, D.S., et al., 2006. Multimodel ensemble simulations of present-day and near-future tropospheric ozone. J. Geophys. Res. 111 http://dx.doi.org/10.1029/ 2005JD006338.

22

A. Santurtún et al. / Atmospheric Environment 101 (2015) 10e22

Stohl, A., Wernli, H., James, P., Bourqui, M., Forster, C., Liniger, M.A., Seibert, P., Sprenger, M., 2003. A new perspective of stratosphere-troposphere exchange. Bull. Am. Meteorol. Soc. 84 (11), 1565e1574. Trigo, R.M., DaCamara, C.C., 2000. Circulation weather types and their influence on the precipitation regime in Portugal. Int. J. Climatol. 20, 1559e1581. Tang, L., Karisson, P.E., Gu, Y., Chen, D., 2009. Synoptic weather types and long-range transport patterns for ozone precursos during high-ozone events in southern Sweden. Ambio 38, 459e464. Tanner, P.A., Law, P.T., 2002. Effects of synoptic weather systems upon the air quality in an Asian megacity. Water Air Soil Pollut. 136, 105e124. Thompson, M.L., Reynolds, J., Cox, L.H., Guttorp, P., Sampson, P.D., 2001. A review of statistical methods for the meteorological adjustment of tropospheric ozone. Atmos. Environ. 35 (3), 617e630. http://dx.doi.org/10.1016/S1352-2310(00) 00261-2.

Trainer, M., Parrish, D., Goldan, P., Roberts, J., Fehsenfeld, F., 2000. Review of observation-based analysis of the regional factors influencing ozone concentrations. Atmos. Environ. 34, 2045e2061. Volz, A., Kley, D., 1988. Evolution of the Montsouris series of ozone measurements made in the nineteenth century. Nature 332, 240e242. Wild, M., Gilgen, H., Roesch, A., Ohmura, A., Long, C.N., Dutton, E.G., Forgan, B., Kallis, A., Russak, V., Tsvetkov, A., 2005. From dimming to brightening: decadal changes in surface solar radiation. Science 308, 847e850. Yanga, C., Yang, H., Guo, S., Wang, Z., Xu, X., Duan, X., Kan, H., 2012. Alternative ozone metrics and daily mortality in Suzhou: the China air pollution and health effects study (CAPES). Sci. Total Environ. 426, 83e89. Zvyagintsev, A.M., Tarasova, O.A., Kuznetsov, G.I., 2008. Seasonal and daily cycles of surface ozone in the extratropical latitudes. Izvestiya Atmos. Ocean. Phys. 44 (4), 474e485.