Ozone levels in the Spanish Sierra de Guadarrama mountain range are above the thresholds for plant protection: analysis at 2262, 1850, and 995 m a.s.l.

  • S. Elvira
  • I. González-Fernández
  • R. Alonso
  • J. Sanz
  • V. Bermejo-Bermejo
Article

DOI: 10.1007/s10661-016-5581-z

Cite this article as:
Elvira, S., González-Fernández, I., Alonso, R. et al. Environ Monit Assess (2016) 188: 593. doi:10.1007/s10661-016-5581-z

Abstract

The Sierra de Guadarrama mountain range, located at 60 km from Madrid City (Spain), includes high valuable ecosystems following an altitude gradient, some of them protected under the Sierra de Guadarrama National Park. The characteristic Mediterranean climatic conditions and the precursors emitted from Madrid favor a high photochemical production of ozone (O3) in the region. However, very little information is available about the patterns and levels of O3 and other air pollutants in the high elevation areas and their potential effects on vegetation. Ozone levels were monitored at three altitudes (2262, 1850, and 995 m a.s.l.) for at least 3 years within the 2005–2011 period. NOx and SO2 were also recorded at the highest and lowest altitude sites. Despite the inter-annual and seasonal variations detected in the O3 concentrations, the study revealed that SG is exposed to a chronic O3 pollution. The two high elevation sites showed high O3 levels even in winter and at nighttime, having low correlation with local meteorological variables. At the lower elevation site, O3 levels were more related with local meteorological and pollution conditions. Ozone concentrations at the three sites exceeded the thresholds for the protection of human health and vegetation according to the European Air Quality Directive (EU/50/2008) and the thresholds for vegetation protection of the CLRTAP. Ozone should be considered as a stress factor for the health of the Sierra de Guadarrama mountain ecosystems. Furthermore, since O3 levels at foothills differ from concentration in high elevation, monitoring stations in mountain ranges should be incorporated in regional air quality monitoring networks.

Keywords

Ozone critical levels Ozone risk assessment Mediterranean mountain range Surface ozone Iberian peninsula Sierra de Guadarrama Mountains National Park 

Introduction

Tropospheric ozone (O3) levels have been increasing over the last decades linked to emissions of its precursors from industrial and agricultural processes, such as non-methanic volatile organic compounds (NMVOCs), carbon monoxide (CO), and nitrogen oxides (NOx). These compounds react photochemically to form O3 and can be transported long distances in the atmosphere, enhancing O3 background concentration at rural areas far from pollution sources (Finlayson-Pitts and Pitts 1997). Natural processes can also contribute to tropospheric O3 levels, such as stratospheric intrusions (Cristofanelli et al. 2015) or emissions of VOCs compounds (VOCs) by vegetation (Vingarzan 2004; Oltmans et al. 2006).

In spite of significant reductions in the emission of O3 precursors due to the emission control policies implemented in Europe, limited improvements have been observed in the O3 levels, and even slight increments have been detected in some monitoring stations, particularly in urban areas (Balzani et al. 2008; Sicard et al. 2013; Querol et al. 2014; EEA 2014). The annual increment rate of O3 in some European mountains has been estimated in 0.2–0.4 nl l−1 between 1992 and 2002 (Ordóñez et al. 2005). Tropospheric O3 represents a more important challenge in southern Europe, where climate conditions, including intense solar radiation, elevated temperatures, and frequent atmospheric stability, foster the photochemical production of ozone. Furthermore, polluted air mass recirculation processes in the Mediterranean Basin also favor the accumulation of even higher concentrations of photochemical species (Millán et al. 2000; Chevalier et al. 2007; Cristofanelli and Bonasoni 2009). Future O3 projections in the context of global change predict increments of O3 levels in the Mediterranean region associated with greater frequency of dry summers and higher temperatures (Meleux et al. 2007).

It the Iberian Peninsula, tropospheric O3 levels continuously exceed the threshold values defined for human and plant health protection. These values are also exceeded in special environmental protection zones such as National Parks or protected areas belonging to the Natura 2000 Network (Ribas and Peñuelas 2006; Sanz et al. 2007a, b; González et al. 2010; Saavedra et al. 2012; Monteiro et al. 2012; Adame and Sole 2013; Villanueva et al. 2014).

Little information is available on O3 levels in mountainous areas due to the difficulty of establishing monitoring stations at high altitudes, the complex topography, and the variability of O3 levels with elevation. Interestingly, the spatial variation of O3 distribution in mountain regions is even greater than temporal variations (Chevalier et al. 2007). While rising O3 concentrations with elevation have been reported in different mountain locations, other studies have not detected this pattern (Alonso and Bytnerowicz 2003 and references therein; Ribas and Peñuelas 2006; Díaz-de-Quijano et al. 2009; Sicard et al. 2009). As a result, the spatial distribution pattern of O3 in mountains cannot be generalized and extrapolating pollutant levels from stations at different altitudes presents a high uncertainty. In the Iberian Peninsula, O3 studies only contemplate middle altitude monitoring sites (Ribas and Peñuelas 2006; Díaz-de-Quijano et al. 2009) or cover short temporal ranges (Sanz et al. 2001; Sánchez et al. 2005). Thus, further studies are urgently needed to characterize O3 pollution in high elevation areas of Mediterranean mountains.

Tropospheric O3 pollution has received great attention not only because it is the fourth in the importance of greenhouse gas contributing to global warming after CO2, methane, and N2O, but also due to its toxicity for living organisms. Ozone is considered by the Intergovernmental Panel on Climate Change as the most harmful air pollutant affecting crops and forests (IPCC 2007). The high oxidative capacity of O3 alters plant biochemistry and physiology, especially affecting gas exchange and carbon allocation, with consequences for plant growth and reproductive ability (Elvira et al. 1998; Alonso et al. 2001; Heath 2008; Wittig et al. 2009; Ainsworth et al. 2012). Moreover, O3 can also cause important indirect effects, increasing the vegetation sensitivity to other environmental stress factors such as water and nutrition deficits, low temperatures, pathogens, or fires (Weinstein et al. 1991; Foot et al. 1997; Grulke et al. 2008). In this sense, O3 alters different ecosystem services, such as carbon fixation (Wittig et al. 2009; Ainsworth et al. 2012), biodiversity (Mills et al. 2011), or water balance (McLaughlin et al. 2007). Ozone threshold values (critical levels) have been established within the framework of the Convention on Long-Range Transboundary Air Pollution (CLRTAP, UNECE) for the protection of different vegetation types (CLRTAP 2011). These critical levels have been the basis for defining the objective values for plant protection in the European Air Quality Directive (2008/50/EC) and are the main tools for O3 risk assessment studies.

The current study aims to characterize air pollution and the possible effects in a representative high mountain area of the Mediterranean Basin. The study was located in the Sierra de Guadarrama (SG) mountain range, in the central Iberian Peninsula, around 60 km North of Madrid City. The city constitutes an important source of air pollution, mainly due to its heavy road traffic emitting high amounts of O3 precursors (Pujadas et al. 2000). These precursors, together with the topographic and climatic conditions of the Madrid watershed, favor the high O3 levels registered in the SG (Sánchez et al. 2005; Alonso et al. 2009). The Sierra de Guadarrama National Park has recently been declared, with the aim to protect different high ecologically valuable ecosystems representing the high Mediterranean mountain. However, no information is available on the possible threat that air pollution could be posing for the protection of these habitats.

The specific objectives of this study were: (1) determining through a long-term monitoring study the levels, patterns, and seasonality of surface O3 and other air pollutants at the slopes and summits of the SG Mountains; (2) defining the local meteorological parameters most related to the observed O3 levels; and (3) developing an O3 risk assessment for the most characteristic vegetation types at different altitudes. Given the complete absence of atmospheric pollution monitoring stations in elevated areas, two new monitoring sites were established at 2262 and 1840 m a.s.l.; additionally, data from the only rural station located at the SG (995 m a.s.l.) from the Madrid regional Air Quality Network were also incorporated to the analysis. The length of the monitoring period considered (2005–2011) allowed a risk assessment according with the EU Directive time frame (3 years for vegetation, 5 for human health). This study also extends the scarce information available on O3 levels on Mediterranean Basin Mountains above 2000 m a.s.l.

Materials and methods

Study area

The SG Mountains belongs to the Spanish Central Range, which crosses the Iberian Peninsula from east to west. The area is characterized by a continental Mediterranean climate, with accentuated cold and wet winters and hot and dry summers. The SG extends from the foothills at 900 m.a.s.l. to the highest elevation peak at 2428 m (Peñalara peak). The altitudinal range produces a climatic gradient sustaining different natural vegetation belts (Fig. 1), also subjected to traditional forestry management.
Fig. 1.

Vegetation zones along the elevation gradient in the Guadarrama Mountains and location of the monitoring stations for atmospheric pollutants

Atmospheric pollutant monitoring stations

Station at 2262 m a.S.L. (Alto de las Guarramillas)

The monitoring station at the highest altitude was established in 2009 at Alto de las Guarramillas summit, inside the Bola del Mundo TV-tower facility at the edge of the SG National Park (2262 m a.s.l., 40° 47′ 10″ N, 3° 58′ 35″ W). The area is surrounded by a characteristic psicroxerophyll pasture, endemic of the summits of Mediterranean mountains. The concentrations of O3, SO2, and NOx were continuously recorded from June 2009, according to standard measurement methods accomplishing USA and EU regulations (SIR monitors 5014, S-5001, and S-5012 models, respectively, SIR S.A., Spain). Monitors were calibrated in situ every year by a certified company, which also performed the equipment maintenance every 6 months. Air was sampled through a 3-m Teflon tube placed at 3 m height above ground. Due to the low winter temperatures (minimum temperatures could reach −15 °C), the air sampling tube was insulated and heated for maintaining an air flux through the monitors at a constant temperature rounding 18 °C. Moreover, monitors were placed inside a thermally insulated cabinet. Ambient air temperature and relative humidity (RH) were registered using HOBO gauges (Pro-V2 Model, Onset Computer Corporation, USA) placed at 3 m above ground.

Station at 1850 m a.s.l. (Cotos)

The monitoring station at 1850 m a.s.l. was established close to the Cotos visitor center of the SG National Park (40°49′ 31″ N, 3° 57′ 40″ W) and located in a Scots Pine (Pinus sylvestris) forest clearing. Ozone levels were continuously recorded since June 2005 (SIR monitor 5014 model, SIRSA, Spain). Monitor calibration and maintenance followed the same protocols explained before. Air was sampled using a 1-m Teflon tube placed at a 3m height above ground. Air temperature and RH were recorded since 2008 using a HOBO sensor (Pro-V2 Model, Onset Computer Corporation, USA). Air temperature and RH for the 2005–2007 period were obtained from the Zabala station (2079 m a.s.l.) belonging to the SG National Park Meteorological Network (Environmental and Territorial Management of Madrid Regional Government). Temperature and relative humidity data for the period 2008–2010 showed a very good correlation between Zabala and Cotos stations, allowing the use of Zabala meteorological data as a good proxy for the Cotos site (temperature, R2 = 0.96, slope = 1.03, p < 0.05; HR, R2 = 0.93, slope = 0.99, p < 0.05). The Zabala station is sheltered by a slope; thus, despite its higher altitude, meteorological conditions are closer to Cotos than to the Alto de las Guarramillas. Wind speed and velocity were obtained from the Cotos meteorological station of the SG National Park Meteorological Network.

Station at 995 m a.s.l. (El Atazar)

The El Atazar station (995 m a.s.l., 40° 54′ 37″ N, 3° 27′ 60″ W) belongs to the Air Quality Network of the Regional Government of Madrid and it represents the only rural background station of the network located at the SG foothills. The area is located in the transition zone between the lower elevation distribution limit of Pyrenean oak (Quercus pyrenacica) forests and the upper limit of holm oak (Quercus ilex) forests. Concentrations of O3, SO2, and NOx and meteorological parameters including solar radiation were continuously monitored meeting the requirements of the European Air Quality Directive. More information about the parameters and methods can be obtained from the website (http://gestiona.madrid.org/aireinternet).

Statistical analyses

The relationships among O3 levels at the three different altitudes were tested with Pearson’s correlations using hourly mean concentrations. Correlations were analyzed considering: (1) yearly periods and (2) two contrasted seasonal periods: the warmest (spring–summer) and the coldest (winter–autumn). In order to define the most influential local meteorological parameters determining the O3 levels in the SG, multiple regression analyses were performed for each altitude including the following variables (hourly means): O3, temperature, RH, photosynthetic active radiation, wind velocity, and direction. The analyses were also performed considering: (1) yearly periods and (2) two seasonal periods. Alpha level was set at 0.05. The SPSS V14 statistical package was used.

Ozone risk assessment for vegetation and human health

The European Air Quality Directive (AQD) establishes a target and a long-term objective values for plant protection. Both values are expressed using the AOT40 O3-exposure accumulated index calculated as the sum of the differences between hourly O3 means and 40 nl l−1 when the hourly mean exceeds 40 nl l−1, using only the 1-h values measured between 08:00 and 20:00 hours of the Central European Time (CET) from May to July. Additionally, the AOT40 can be estimated for a 6-month period from April to September for forest protection. Thus, AOT40 indexes accumulated during the periods May–July (3 months) and April–September (6 months) were calculated for the three monitoring stations at different elevation. These values were compared with the target and long-term objective values of the AQD for plant protection presented in Table 1. Concentration-based O3 critical levels (CLec) proposed within the CLRTAP for the protection of different vegetation types (Table 1) were used as well for risk assessment. In the last years, new O3 critical levels based on the stomatal O3 flux uptake have been proposed for different species by the CLRTAP (Phytotoxic O3 Dose, POD; CLRTAP 2011). These critical levels have more biological meaning, but they require complex calculations for determining O3 uptake using species-specific parameterization to model stomatal conductance depending on meteorological variables. Unfortunately, these models have not been parameterized yet for many representative species of the Mediterranean mountain ecosystems present in the SG. Therefore, the O3 risk analysis conducted in the present study is O3 concentration based. Moreover, taking into account the high O3 concentrations during nighttime at the high-altitude stations, the nocturnal 3-month AOT40 index was also calculated (from 20:00 to 08:00 hours CET).
Table 1

Ozone threshold values for plant protection according to CLRTAP (2011) and the EU Air Quality Directive (2008/50/CE), and threshold values for health protection according to the EU Directive 2008/50/CE

Concentration-based ozone critical levels according to CLRTAP (Clec)

Receptor

Parameter

CLec (μl l−1 h)

Period

Effect

Tree species

6-month AOT40

5

April–September

Growth reduction (10 %)

Annual pastures

3-month AOT40

3

May–July

Growth reduction/seed production (10 %)

Perennial grasslands

6-month AOT40

5

April–September

Growth reduction (10 %)

Ozone objective values according to the EU Directive 2008/50/CE

Vegetation

Target value: 3-month AOT40 (May–July) 18,000 μg m−3 h (9 μl l−1 h; mean 5 years)

Objective value: 3-month AOT40 (May–July) 6000 μg m−3 h (3 μl l−1 h; mean 5 years)

Human health

Target value: maximum daily 8-h mean, 120 μg m−3 (60 nl l−1) not to be exceeded on more than 25 days per year over 3 years

Objective value: maximum daily 8-h mean, 120 μg m−3 (60 nl l−1) not to be exceeded

CLec concentration-based critical levels

EU Directive guidelines were followed when missing values of the database were up to 10 % in order to avoid indices biases. Then, the AOT40 value was corrected according to the formula: AOT40estimate = AOT40measured × total possible number of hours/number of measured hourly values (the number of hours within the time period of AOT40 definition 08:00 to 20:00 hours CET during the 3- or 6-month period considered).

The risk of O3 effects for human health was assessed according to the EU Air Quality Directive using the maximum daily 8-h mean concentration. All the critical levels, and target and objective values for the protection of vegetation and human health are presented in Table 1.

Results

Meteorological conditions

The meteorological conditions recorded at the three sites during the study period (Fig. 2) reflect the altitudinal climatic gradient existing in the SG Mountains. At the highest elevation, 2262 m a.s.l., climatic conditions were characterized by low air temperatures and small day–night oscillations compared to lower altitudes. Temperature remained almost constantly below 0 °C during winter while maximum monthly averages were in the range of 15–16 °C during July and August. At this high altitude, the RH was over 80 % during most of the year but decreased in summer to a minimum daylight value below 50 %. At mid-elevation (1850 m a.s.l.), mean temperatures were higher compared with the summit, with maximum temperatures 4–5 °C higher during the warmer months, while the RH in summer was similar in both sites. The station at the lowest altitude (995 m) was the warmest during the whole year, with summer monthly means close to 25 °C, and maximum hourly values around 35 °C. The RH was the lowest at this elevation, especially in summer, with minimum hourly averages around 30 %. Over the study period, inter-annual variations of meteorological parameters were detected (data not shown). The warmest period was recorded in July and August 2009 and 2010 at the three altitudes.
Fig. 2

Average seasonal daily profiles of air temperature (°C) (a) and relative humidity (b) throughout the studied period at the different monitoring stations

Wind velocity and direction records showed that predominant winds changed with altitude. At Cotos station (1850 m a.s.l., Fig. 3a, b), the dominant wind direction during daylight hours came from the West. During nighttime, the wind direction was more variable and more frequently coming from NE, although it was also frequent the W direction (Fig. 3b). However, at 995 m of altitude, the wind direction presented a higher frequency of the components S and SE during daylight hours, whereas at night the predominant winds had a W component (Fig. 3c, d).
Fig. 3

Mean values for the period 2005–2011 of wind velocity and direction at Cotos—1850 m a.s.l. (a, b) and El Atazar—995 m a.s.l. (c, d) during day time (09:00–19:00 hours GMT; a, c) and night time (20: 00–08:00 hours GMT; b, d)

Atmospheric pollutants

Ozone levels

The average O3 concentrations presented an increasing gradient with altitude (Figs. 4 and 5; Table 2). Mean O3 levels for the period 2009–2010 decreased from 46.5 to 44.3 and 41.8 nl l−1 at 2262, 1950, and 995 m a.s.l. respectively, but this gradient slightly varied among different years. This elevation gradient was more evident during summer when 2009–2010 summer averages were 53.4, 51, and 49 nl l−1 at 2262, 1850, and 995 m a.s.l., respectively (data not shown). However, this elevation gradient was not observed when accumulated exposures expressed as AOT40 were considered (Table 2). Similarly, maximum O3 hourly values in the high elevation sites were up to 90 nl l−1 while maximum hourly concentrations reached up to 102 nl l−1 in El Atazar station.
Fig. 4

Monthly averages for O3 (a), NOx (b), and SO2 (c) throughout the study period at the three monitoring sites: 2262 m a.s.l- Alto de las Guarramillas, 1850 m a.s.l.- Cotos, and 995 m a.s.l.- El Atazar. Mean values ± SE

Fig. 5

Hourly mean daily profiles per season of O3 (a), NOx (b), and SO2 (c) concentrations measured at the different monitoring stations throughout the period of study

Table 2

Ozone indices for vegetation and health risk assessment

Altitude site

Year

Mean O3 year (nl l−1)

Risk vegetation indices

Risk health indices (EU)

Available data (%)a

Diurnal 3-month AOT40 (μl l−1 h)b

Nocturnal 3-month AOT40 (μl l−1 h)c

Available data (%)

Diurnal 6-month AOT40(μl l−1 h)d

>60 nl l−1 (days)e

% > 60 nl l−1 (% days)f

Alto de las Guarramillas—2262 m a.s.l.

2009

47

37

19.1

14.5

52

36.3

48

27

2010

46

100

13.9

11.1

96

26.7

50

18

2011

53

80

16.6

13.3

89

36.3

85

31

Mean

49

 

16.5

13.3

 

33.1

61

25.3

Mean 2009–2010

46.5

 

16.5

12.8

 

31.5

  

Cotos—1850 m a.s.l.

2005

55

98

27.3

16.8

92

52,1

139

47

2006

51

64

30.3

20.2

62

48.6

152

42

2007

39

63

8.8

3.5

71

20,3

30

9

2008

37

58

6.1

2.4

73

16,2

22

6

2009

43

98

7.4

3.3

79

21,3

40

12

2010

46

82

11.0

6.5

45

22,9

20

13

Mean

45

 

15.2

8.8

 

26.6

67.2

21.5

Mean 2009–2010

44.3

 

9.2

4.9

 

22.1

  

El Atazar—995 m a.s.l.

2007

43

100

16,2

6.3

99

33,3

76

21

2008

43

100

17,6

7.1

99

33,7

75

21

2009

39

100

10,7

3.4

99

23,3

54

15

2010

45

100

18,0

7.6

99

33,5

78

21

2011

48

100

16,0

4.6

99

32,8

73

38

Mean

43.6

 

15.7

5.8

 

31.3

71.2

23.2

Mean 2009–2010

41.8

 

14.3

5.5

 

28.4

  

Significance in bold are the annual ozone values

aPercentage of available data

bDiurnal 3-month AOT40 from May to July

cNocturnal 3-month AOT40 from May to July

dDiurnal 6-month AOT40 from April to September

eNumber of days with the maximum daily running 8-h mean > 60 nl l−1 according to the EU Directive

fPercentage of days with the maximum daily running 8-h mean > 60 nl l−1 according to the EU Directive

Marked inter-annual variations were recorded during the study (Fig. 4a). At 2262 m, the maximum O3 levels were measured during the summer 2009 and in August 2010. At the 1850 m station, with the longest data series (2005–2010), the highest concentrations were recorded at the end of spring and beginning of summer in 2005 and 2006, with monthly means almost constantly above 55 nl l−1. Other periods of high O3 levels were August 2009 and July 2010. At El Atazar station (995 m), the highest O3 levels were recorded in the summers of 2007 and 2008, and in July 2010.

The seasonal pattern was similar in the three stations but the variations were less marked with increasing altitude (Fig.5a). The highest O3 concentrations were always recorded during the summer months at all altitudes. The spring values were also high at the lowest altitude station, but the spring peak was not present every year at the summit and mid altitude stations. During the coldest months, the high altitude stations presented relatively high and constant O3 concentrations above 40 nl l−1 (Fig. 5a). However, at El Atazar station during the coldest months (November–February), the concentrations decreased to monthly means between 28 and 35 nl l−1 (Fig. 4a).

The daily profile of O3 also exhibited a clear seasonal and altitudinal component. During the cold months (October–March), very small daily oscillations were detected especially at the high elevation stations (Fig. 5a). In spring and summer, the daily cycle intensified, with minima values recorded in the early morning hours and maxima afternoon. The amplitude of the daily O3 profile was more intense at lower altitudes. The differences between daylight and nighttime concentrations during summer at El Atazar station were 20–25 nl l−1 while at higher elevation stations were 10–15 nl l−1.

NOx levels

The NOx concentrations measured in the SG summits and foothills were very low throughout the studied period (Fig. 4b). At the highest altitude station, the NOx monthly averages ranged between 0.5 and 2 nl l−1 without showing a clear seasonal pattern. Occasionally, some NOx short-term high values were recorded, most probably related to periodic facility maintenance labors or to helicopter mountain rescue training operations in the area. These values represented less than 1 % of the annual data and were discarded from the final dataset to avoid interference in the comparison with other stations. A significant episode during March 2010 resulted in maximum NOx hourly values around 10 nl l−1. At the summit station, maxima NOx values were found in early morning during the summer and towards the end of the afternoon in winter (Fig. 5b).

At the foothills station, the NOx concentrations were also low (Fig.4b) but higher than at the summit A marked seasonality was observed, with the highest values recorded during the cold months of the autumn and winter, reaching monthly means up to 8 nl l−1. During those months, hourly means up to 50–70 nl l−1 were reached occasionally. The daily cycle was sharper during the coldest months, with maxima between 17:00 and 19:00 hours CET (Fig. 5b); whereas in summer, the daily oscillation was limited.

SO2 levels

The SO2 levels at the SG summit were very low with monthly averages below 2 nl l−1 (Fig. 4c). The only noteworthy discrepancy of this pattern was a slight increase in June–July 2009 values and an episode in March–April 2010, coinciding with the exceptionally high values of NOX and low O3 concentrations. This pollutant showed a mild daily cycle in summer, whereas values were close to the detection limits of the monitor during the rest of the year (Fig. 5c).

At the foothills station, the SO2 concentrations were also very low but slightly higher than in the mountain summit (Figs. 4c and 5c). Monthly values were always below 2.5 nl l−1, exhibiting practically constant levels throughout the day (Fig. 4c).

Correlation between O3 levels at the monitoring stations

The two sites at higher altitudes, 2262 and 1850 m a.s.l., presented the highest correlations among their O3 levels (Table 3) considering the total data set (r = 0.80), only the diurnal data (r = 0.86) or the data grouped by seasonal periods (r = 0.75–0.84). On the other hand, the correlation between the sites located in the extremes of the altitudinal gradient, 2262 and 995 m a.s.l., presented the lowest correlation (r = 0.59). Moreover, this correlation varies between seasonal periods being higher during the warmer months of spring–summer (r = 0.53) than in autumn–winter (r = 0.21). The correlations between the stations at different altitudes slightly increased when only daylight data were considered.
Table 3

Correlation among ozone levels measured at the different monitoring sites considering all the data, only diurnal data (08:00–20:00 hours CET) or data seasonally grouped: Pearson coefficient of correlation (r), p value, and n

 

2262 m a.s.l.—Alto Guarramillas

1850 m a.s.l.—Cotos

All data

 2262 m a.s.l.—Alto Guarramillas

 1850 m a.s.l.—Cotos

r = 0.80, p < 0.01, n = 6836

 995 m a.s.l.—El Atazar

r = 0.59, p < 0.01, n = 15,486

r = 0.70, p < 0.01, n = 7584

Diurnal data

 2262 m a.s.l.—Alto Guarramillas

 1850 m a.s.l.—Cotos

r = 0.85, p < 0.01, n = 3710

 995 m a.s.l.—El Atazar

r = 0.61, p < 0.01, n = 8364

r = 0.75, p < 0.01, n = 4112

Spring–summer (diurnal data)

 2262 m a.s.l.—Alto Guarramillas

 1850 m a.s.l.—Cotos

r = 0.75, p < 0.01, n = 2224

 995 m a.s.l.—El Atazar

r = 0.53, p < 0.01, n = 5920

r = 0.66, p < 0.01, n = 2539

Autumm–winter (diurnal data)

 2262 m a.s.l.—Alto Guarramillas

 1850 m a.s.l.—Cotos

r = 0.84, p < 0.01, n = 1486

 995 m a.s.l.—El Atazar

r = 0.21, p < 0.01, n = 2445

r = 0.55, p < 0.01, n = 1573

Meteorological parameters determining ozone levels

Table 4 shows the results of the multiple regression analyses for each altitude, indicating those local meteorological variables that were significantly selected by the models (p < 0.05) to explain the variability in O3 levels. The regression models for the high elevation sites, although significant, only explained around 28–27 % of the O3 variability. At these high altitudes, the variables with the highest relative importance (higher β coefficients) selected by the models were temperature and RH, while the variables related to wind were selected with a lower β value. This pattern was also observed when the data were analyzed grouped for the different seasons of the year.
Table 4

Local meteorological variables determining ozone levels at the different monitoring sites

Altitude/site

Year

Total

Spring

Summer

Autumn

Winter

2262 m a.s.l.—Alto de las Guarramillas

2009–2011

n = 5146, R2 = 0.28

0.5, 0.1, 0.07, 0.07

T, RH, WD, PAR

n = 1144, R2 = 0.38

0.6, 0.09, 0.06, 0.05

T, RH, WD, WV

n = 2075, R2 = 0.17

0.4, 0.11, 0.11, 0.09

T, RH, WD, PAR

n = 1927,R2 = 0.27

0.36, 0.37

RH, T

1850 m a.s.l.—Cotos

2007–20101

n = 10581, R2 = 0.27

0.5, 0.14, 0.06, 0.02

T, RH, PAR, WD

n = 3399, R2 = 012

0.3, 0.1, 0.1

T, PAR, WV

n = 4406, R2 = 0.16

0.4, 0.18, 0.08, 0.07

T, WV, WD, PAR

n = 1120, R2 = 0.38

0.6, 0.13, 0.13, 0.06

RH, T, PAR, WD

n = 840, R2 = 0.41

0.6, 0.1, 0.09, 0.06, 0.12

RH, WV, T, WD, PAR

995 m a.s.l.—El Atazar

2007–2011

n = 29629, R2 = 0.49

0.7, 0.2, 0.13, 0.08, 0.04

RH, T, WV, WD, PAR

n = 8110, R2 = 0.34

0.6, 0.14, 0.1

RH, WD, WV

n = 8710, R2 = 0.53

0.7, 0.34, 0.05, 0.05

RH, T, PAR, WD

n = 7720, R2 = 0.39

0.6, 0.2, 0.12, 0.07

RH, WV, T, PAR

n = 5089, R2 = 0.47

0.6, 0.3, 0.04

RH, WV, PAR

Multiple regression parameters: number of data (n), R2 adjusted (R2), β standardized coefficients of partial regression, and significant variables selected at p < 0.05

T air temperature, RH relative humidity, WD wind direction, WV wind velocity, PAR photosynthetic active radiation

1There is not available PAR data from 2005 to 2006 at Cotos site

For the site at the lowest altitude, regression models could explain 49 % of the O3 variability, a percentage that increased until 53 % during the summer months. At this lower altitude, the RH was the first meteorological variable selected by the models and wind related variables acquired a relative greater relevance compared with higher altitude sites.

Ozone exposure indices for risk assessment analysis

Table 5 presents the values of the cumulative 3 and 6-month AOT40 indexes for O3 risk assessment considering herbaceous and forest species, respectively, and the index for health risk analysis according to the EU Air Quality Directive. The mean values of the 3-month AOT40 index for the three altitudes were high with 2009–2010 average values between 9.2 and 16.5 μl l−1 h. Similarly, the 2009–2010 mean values of the 6-month AOT40 index were also high, between 22.1 and 31.5 μl l−1 h. The differences between altitudes became more apparent when only the night O3 concentrations were considered: the highest elevation at 2262 m a.s.l., presented the highest 3-month nocturnal AOT40 index (12.8 μl l−1 h), which decreased to 4.9 and 5.5 μl l−1 h at 1850 and 995 m a.s.l., respectively. There was a significant inter-annual variation in all the monitoring sites, although it was observed more clearly when analyzing the longer dataset at Cotos (1850 m a.s.l.). At this site, the 3-month AOT40 index varied from 6.1 to 30.3 μl l−1 h depending on the year.
Table 5

Ozone risk analysis for the different plant communities along an elevation gradient in the SG, based on the exceedance of the concentration-based AOT40 critical levels defined according to the CLRTAP (2011), EU Air Quality Directive, and recent peer-review literature

Ozone monitoring site

Bioclimatic zone (altitudinal range)

Plant community (representative species)

Mean AOT40 (μl l−1 h)

Concentration-based critical level—AOT40 (μl-l h)

Risk damage

AOT40

Exposure period

Reference

Alto de las Guarramillas—2262 m a.s.l.

Oro- and crioromediterranean (1800–2500 m a.s.l.)

Shrubs

6 months

5

6 months

CLRTAP 2011

Yes

Citysus oromediterraneus

33.1

9–3

3 months

EU Directive

Yes

Perennial pastures

3 months

5

6 months

CLRTAP 2011

Yes

Festuca curvifolia and Senecio pyrenaica

16.5

9–3

3 months

EU Directive

Yes

Cotos—1850 m a.s.l.

Supramediterranean (1300–1800 m a.s.l.)

Conifer forests

6 months

5

6 months

CLRTAP 2011

Yes

Pinus sylvestris

26.6

9–3

3/6 months1

EU Directive

Yes

 

3 months

4.7

6 months

Karlsson et al. (2004)2

Yes

 

15.2

    

El Atazar—995 m a.s.l.

Supramediterranean (1000–1400 m a.s.l.)

Deciduous forests

6-months

5

6 months

CLRTAP 2011

Yes

Quercus pyrenaica and Fagus sylvatica)

31.3

9–3

3/6 months

EU Directive

Yes

  

2.4–18

6 months

Karlsson et al. (2004)3

Yes

  

18

6 months

Calatayud et al. (2011)4

Yes

Mesomediterranean

(650–1000 m a.s.l.)

Broadleaf evergreen forests

6 months

5

6/12 months

CLRTAP (2011)

Yes

Quercus ilex

31.3

9–3

3/6 months

EU Directive

Yes

  

21.8

1 year

Alonso et al. (2014)5

Yes

  

46

6 months

Calatayud et al. (2011)6

No

Annual pastures

3 months

3

3 months

CLRTAP (2011)

Yes

Trifolium sps.

15.7

9–3

3 months

EU Directive

Yes

  

3.1

1.5 months

Sanz et al. (2016)7

Yes

The analysis based on the mean AOT40 values for the years considered for each site

1EU Air Quality Directive considers 3-month AOT40 values for the protection of vegetation or 6-month ATO40 values for forest trees

2CLec for coniferous sensitive species (Pinus sylvestris + Picea abies) according to Karlsson et al. (2004)

3CLec for broadleaf deciduous species sensitive (beech and birch, 2.4) and less sensitive (oak, 18) according to Karlsson et al. (2004)

4CLec for Mediterranean deciduous species (Querus robur + Q. pyrenaica + Q. faginea) according to Calatayud et al. (2011)

5CLec for Mediterranean broadleaf evergreen species according to Alonso et al. (2014)

6CLec for Quercus ilex according to Calatayud et al. (2011)

7CLec for biomass production of annual species according to Sanz et al. (2016)

The target value for the protection of human health was exceeded every year in the three stations except the years 2008 and 2010 in Cotos (1850 m a.s.l.). Although the percentage of days with the maximum daily running 8-h mean O3 concentrations above 60 nl l−1 was similar or bigger at the highest elevation site, the number of days with exceedances was smaller than at the foothills.

Discussion

Air pollution levels in the Sierra de Guadarrama

The O3 concentrations recorded in the SG were constantly high through the entire study period at the three altitudes analyzed. Ozone levels were permanently above 40 nl l−1 during the whole summer period, resulting in high AOT40 values accumulated over 3 and 6 months. Moreover, hourly concentrations during summer frequently rounded 80 nl l−1 reaching up to 90 nl l−1, the threshold for information for the protection of sensitive population in the EU Air Quality Directive. Thus, O3 levels in SG exceeded the standards for the protection of vegetation and human health. These high O3 levels confirm previous short-term studies performed in the SG during summer (Sánchez et al. 2005). The O3 levels recorded in SG are within the range found in other mountains located in the Mediterranean basin and southern Europe (Mechergui et al. 2009; Cristofanelli et al. 2015; Dalstein and Vas 2005; Sicard et al. 2011; Díaz-de-Quijano et al. 2009).

A different pattern was found for the NOx and SO2 concentrations, which generally remained very low or only increased sporadically. The NOx values at 995 m a.s.l. exhibited a seasonal pattern with higher concentrations during autumn and winter linked to the variation in emissions. Those values were similar to concentrations recorded in other mountainous areas at similar altitudes in the Pyrenees (Ribas and Peñuelas 2006). At the highest altitude (2262 m a.s.l.), the monthly means of NOx did not exceed 1.2 nl l−1 and did not present seasonal variations. These low levels and seasonal behavior are characteristic of high mountain areas and they are representative of atmospheric background regional levels (Pandey Deolal et al. 2012). An episode of high NOx concentrations lasting several days was recorded between the end of March and the beginning of April 2010. NOx and SO2 levels increased significantly while O3 levels remained below 30 nl l−1. This event might be related to long distance transport of pollutants from an important forest fire that occurred in central Spain. Increases in NOx concentration related with forest fires have been reported in other natural areas (Di Carlo et al. 2015). Concentrations of both NOx and SO2 in SG were always lower than the standards for the protection of vegetation and human health according to the EU Air Quality Directive. Thus, the pollution climate of the SG was primarily characterized by a chronic tropospheric O3 pollution.

Seasonal and inter-annual variations of ozone levels

Ozone concentration in the SG at the three altitudes followed the typical seasonal pattern with higher levels during late spring and summer, and lower in autumn and winter. These results agree with studies carried out in other rural areas of the Iberian Peninsula consistently highlighting the elevated O3 levels in the region and their marked seasonal pattern (González et al. 2010; Fernández-Fernández et al. 2011; Kulkarni et al. 2011; Notario et al. 2012; Adame and Sole 2013). In other background monitoring sites located in mountainous areas, high O3 values recorded in spring have been related with stratospheric O3 intrusion episodes and with long-distance transport of background pollution in the northern hemisphere (Vingarzan 2004; Balzani et al. 2008; Cristofanelli et al. 2015). On the other hand, the O3 maxima during the summer have been more related with local or regional scale pollution episodes linked to the increase in photochemical processes favored by heatwaves and climate conditions (Bonasoni et al. 2000; Chevalier et al. 2007; Cristofanelli et al. 2015). The seasonal variations of O3 levels were also revealed by the changes in the O3 daily cycle amplitude. While autumn and winter exhibited mild daily oscillations, the daily pattern was amplified in spring, reaching the maximum amplitude during summer. This variation of the daily O3 profile coincided with the daily oscillation pattern found for air temperature. Accordingly, temperature was the main meteorological factor related to the variability in O3 concentrations.

Ozone levels in the SG also presented important inter-annual variations detected more clearly in the Cotos station (1850 m a.s.l.) with the longer temporal data series. During the years 2005–2006, the recorded O3 levels were very high with respect to the following years. These years have been classified as very warm and dry in the Iberian Peninsula, standing out for their historical minimum levels of precipitation and for the great number of forest fires (MAGRAMA 2012). These conditions would have favored photochemical O3 formation. Other authors have described the relationship between high O3 levels and the frequency and intensity of forest fires far from the monitoring sites in the Iberian Peninsula at that time (Martins et al. 2012). In contrast, the period 2007–2008, which exhibited lower O3 values, has been classified as a cold period with low forest fire incidence. Thus, lower photochemical O3 formation and concentrations were expected.

Ozone elevation gradient

Average concentrations in SG generally showed increasing values with elevation from 995 to 2262 m a.s.l. during the study period. However, this pattern was not observed when daylight AOT40 values were considered. Previous studies using passive samplers described an O3 gradient from the city of Madrid (600 m a.s.l.) towards the SG related to increasing elevation and distance from the city (Sanz et al. 2001; Alonso et al. 2009). Interestingly, the elevation gradient between 1000 and 2000 m a.s.l. (similar to the range of the current study) was only apparent during the summer, while it faded during the rest of the year (Alonso et al. 2009). Passive samplers collect O3 levels during the whole day including also nighttime concentrations. This nocturnal gradient would be then responsible for the elevation gradient manifested when whole day average concentrations are considered.

The elevation gradient was also manifested in the differences observed in the diurnal profiles of O3 concentrations. The daily cycle was consistent throughout the year at the lower elevation site while a clear daily profile was only detected at the high elevation sites during summer. This reduction in the daily oscillation related with altitude has been described in several studies (Brodin et al. 2010; Burley and Bytnerowicz 2011). In rural areas, the decrease in O3 levels during the night can be attributed to deposition or destruction processes under contact with ground and vegetation surfaces (Brodin et al. 2010). This type of behavior has also been described in other Spanish rural areas with low NOx concentrations (Ribas and Peñuelas 2004; MAGRAMA 2009; Santurtún et al. 2015). By contrast, locations at higher altitudes would be above the thermal inversion layer, within an air mass uncoupled from lower altitudes, showing O3 concentration relatively stable due to the lack of O3 deposition and destruction.

Correlation between ozone levels and meteorological parameters

The variations in O3 levels at the two high altitudes (2262 and 1850 m a.s.l.) were highly correlated, but the relationship was weaker with the lower elevation station at 995 m a.s.l. This behavior indicates that, in general, the processes of O3 formation and destruction at high altitudes are different from those operating at low elevation sites. Indeed, the O3 concentration at 995 m a.s.l. were highly correlated with local meteorological variables (R2 = 0.49), while O3 levels at the high elevation stations were less related with local meteorological conditions (R2 = 0.28). In this sense, O3 monitor stations in SG above 1800 m a.s.l. could represent O3 background regional concentrations. This scheme has been described in other mountainous areas studies, although the elevation of the mixing layer varies depending on the area (Chevalier et al. 2007; Brodin et al. 2010). Previous studies on pollutant dispersion carried out in the Madrid basin situated the mixing layer at 1000–1500 m a.s.l. (Plaza et al. 1997; Pujadas et al. 2000). The current study is the first long-term monitoring dataset showing that monitoring stations located at the foothills are not representative of the O3 levels present in the SG mountain range.

Ozone levels at the lowest elevation station (995 m a.s.l.) were highly correlated with local air temperature and RH, but also with wind related variables. Dominant winds in the Madrid basin come predominantly from S to SE direction, dragging O3 precursors from the city, resulting in the high O3 concentrations recorded in the SG. Previous pollutant dispersion studies in the Madrid area also indicated that the transport of air pollutants generated by the city was predominantly towards the SG mountain range (Pujadas et al. 2000; Palacios et al. 2005).

Ozone risk assessment for human health

Ozone concentrations in the SG exceeded the target values for the protection of human health established in the EU Air Quality Directive at the three monitoring sites. At the highest elevation site (2262 m a.s.l.), the maximum daily 8-h mean index exceeded the target value 61 days per year on average, compared with the 25 days limit value established by the Air Quality Directive. At 1850 m a.s.l. (Cotos), the index for health protection was surpassed all the years except in 2008. The worst scenario was registered in 2006 with 152 days of exceedances. These results highlight the importance to include air quality monitoring stations in mountain ranges, particularly where high value areas, such as national parks, are frequently visited and used for practicing sport activities.

Ozone risk assessment for plant communities

Ozone risk assessment has been performed for the communities representative of the different vegetation zones along the elevation gradient in the SG. The summits of this mountain range are above tree line with perennial psychro-xerophilous pastures dominated by the gramineae Festuca curvifolia. This ecosystem presents a high number of valuable species due to their endemic character or because they represent the latitudinal distribution limit of the species. High altitude slopes are also occupied by endemic species forming a dense scrubland of legume-shrubs and junipers (Citysus oromediterraneus and Juniperus communis subsp. Alpina). These communities of high Mediterranean mountains are adapted to endure extreme winter conditions and summer droughts. The summits of the SG remain snow-covered since October/November until April/May, depending on the year. Therefore, plant physiological activity starts in May, becomes most active between June–July, but strongly decreases with the dry season in August (Elvira et al. 2011). To analyze the O3 risk for this vegetation, the AOT40 values for the station located at 2262 m a.s.l. (Alto de las Guarramillas) were considered. Both the 3-month and the 6-month AOT40 values always exceeded the threshold values defined for the protection of these vegetation types by EU Air Quality Directive and the CLRTAP (Table 5). The O3 sensitivity of these communities is currently unknown, but they are sensitive to other disturbances related to global change such as the increment of temperature or summer droughts (Körner 2003; Pauli et al. 2012; Gutiérrez-Girón and Gavilán 2013). Thus, O3 could represent an additional stress factor for them. Furthermore, some species of these communities such as Senecio pyrenaica, one of the valuable species of the mountain, have shown stomata partially open at night (Elvira et al. 2011). Under such physiological behavior, the high O3 levels recorded at night at medium–high altitudes could be relevant for intensifying the O3 risk damage (Musselman and Minnick 2000).

Scots pine (Pinus sylvestris) is the main species of the forests between 1200/1400 and 2000 m. a.s.l. in SG (Rivas-Martínez et al. 1987). The station located at 1850 m a.s.l. (Cotos) is therefore adequate for evaluating the risk of O3 for this vegetation type. The 6-month AOT40 values exceeded (up to 10 times depending on the year) the values proposed for the protection of these forests by the Air Quality Directive, the CLRTAP and Karlsson et al. (2004). Ozone negative effects on gas exchange, cellular metabolism and growth of P. sylvestris have been experimentally confirmed (Skärby et al. 1987; Huttunen and Manninen 2013). Moreover, alterations in the growth and defoliation in P. sylvestris natural forests of central Europe have been related to O3 levels lower than the concentrations registered in SG (Augustaitis and Bytnerowicz 2008). Thus, further studies are urgently needed to evaluate the possible O3 effects on the P. sylvestris forests in SG where O3-like visible foliar symptoms have already been observed (Elvira et al., pers. comm.).

The deciduous oak forests of Quercus pyrenaica grow below the pine vegetation belt, between 700 and 1200 m a.s.l. Other deciduous species like birch (Betula spss) and beech (Fagus sylvatica) grow also on protected and humid zones at this elevation. Although they occupy small areas, these species are considered relict in Central Spain where they find one of the southernmost limits of their European distribution (Rivas-Martínez et al. 1987). The O3 sensitivity of these two species have been clearly established (Vollenweider et al. 2003; Novak et al. 2008); and they have been selected to define the O3 critical levels for the protection of forests within the CLRTAP framework (CLRTAP 2011). Q. pyrenaica is also considered a relatively O3 sensitive species showing O3 effects on both growth and plant physiology (Calatayud et al. 2011). Below 900 m a.s.l., forests at the foothills of SG are dominated by the broadleaf evergreen holm oak (Quercus ilex). Although more resistant than other deciduous oak species, holm oak has also shown O3-induced effects on physiology and growth in different experimental studies (Alonso et al. 2014; Calatayud et al. 2011). Moreover, O3 represents an additional stress factor that might be impairing tree ability to withstand the drought stress commonly experienced at the foothills of SG (Alonso et al. 2014). Additionally, annual pasture communities dominating holm oak forest understory have been defined as high sensitive to O3 (Calvete-Sogo et al. 2014), especially the annual clovers (Bermejo et al. 2003; Gimeno et al. 2004; Sanz et al. 2007a, b, 2014; Calvete-Sogo et al. 2014). The air quality monitoring station located at 995 m a.s.l. (El Atazar) was used to analyze the risk of O3 effects on all the communities located below the pine belt. All the estimated AOT40 indexes exceeded the threshold values for the protection of the different vegetation types proposed by the Air Quality Directive, CLRTAP, and several experimental studies (Karlsson et al. 2004; Calatayud et al. 2011; Alonso et al. 2014). Recent studies have proposed higher critical levels for the protection of Mediterranean trees, indicating that using current Air Quality Directive and CLRTAP thresholds would overestimate O3 effects in Mediterranean forests (Calatayud et al. 2011; Alonso et al. 2014). Concentrations of O3 in SG foothills also exceeded those higher concentration-based critical levels proposed for Mediterranean forests. Thus, both deciduous and evergreen Mediterranean broadleaf forests and the annual pastures forming the forest understory are threatened by O3 in the foothills of SG.

This study represents an O3 risk assessment based on O3 exposure indices, like the AOT40 index (CLec), that cannot be used for quantifying real effects (CLRTAP 2011). Studies comparing concentration-based risk analyses with those based on O3-absorbed flux indices (PODy) have provided very different results (De Andrés et al. 2012; Büker et al. 2015). Risk analysis based on AOT40 indices does not take into account adverse physiological conditions, especially during summer months, when high air temperature and vapor pressure deficit, and low soil water availability reduce plant stomatal O3 uptake. In this sense, the current risk analysis represents the worst-case scenario and further efforts are needed to fully develop risk analysis approaches based on stomatal O3 uptake for the highly biodiverse Mediterranean vegetation.

Conclusions

The Sierra de Guadarrama mountain range is exposed to high chronic O3 concentrations, while NOx and SO2 levels are usually low except scarce and short-term episodes. The O3 levels registered at the three altitudes, 995, 1850, and 2262 m a.s.l., presented marked seasonal and yearly variations, but the background levels were always high. Although 24 h averages suggested an altitudinal gradient with increasing concentrations with elevation, this trend was not observed when AOT40 indexes were considered. The sites at higher altitudes (1850 and 2262 m a.s.l.) presented a parallel behavior with O3 levels responding to atmospheric processes at synoptic instead of local scale. These results indicate that these monitoring sites, especially at the highest altitude, would be adequate to represent O3 background regional concentrations. At the lower elevation site (995 m a.s.l.), the O3 levels were more determined by local meteorological and pollution conditions. Thus, indirectly, the study indicates the importance of the O3 precursors emitted from the city on the O3 levels registered at low and medium altitude slopes of SG. Ozone levels in the SG mountains and in the SG National Park were chronically over the thresholds for human health and plant protection considering concentration-based indexes.

Thus, surface O3 levels must be considered as a risk factor for the health of plant communities located at the different altitude belts. Moreover, since O3 levels differed at foothills from high elevation, monitoring stations in mountain areas should be incorporated in regional air quality monitoring networks. The development of species-specific O3 flux-based indexes (PODy) is necessary for improving the risk analysis for the Mediterranean vegetation in the SG.

Acknowledgments

This research was funded by ECLAIRE (EU FP7-ENV), AGRISOST (Comunidad de Madrid, S2013/ABI-2717), and NEREA (Spanish Government, AGL2012-37815-C05-03) projects, and by an agreement between the Spanish Ministry of Agriculture, Food, and Environment and CIEMAT on Critical loads and levels. The monitoring station at Alto de las Guarramillas was set up, thanks to an agreement with Abertis-Telecom Company. Authors want to thank the Company and the personnel of the TV-Tower Station Bola del Mundo for their special considerations setting up the monitoring station at the summit and specially for providing access to the station during the hard winter conditions. The monitoring station at Cotos was set up, thanks to an agreement with the Sierra de Guadarrama National Park—Community of Madrid. Special thanks are given to the staff of the Park for their interest and enthusiastic support during the study.

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  • S. Elvira
    • 1
  • I. González-Fernández
    • 1
  • R. Alonso
    • 1
  • J. Sanz
    • 1
  • V. Bermejo-Bermejo
    • 1
  1. 1.Ecotoxicology of Air PollutionCIEMATMadridSpain

Personalised recommendations