Assessment of the Air Pollution Level in the City of Rome (Italy) - MDPI

sustainability Article

Assessment of the Air Pollution Level in the City of Rome (Italy) Gabriele Battista *, Tiziano Pagliaroli, Luca Mauri, Carmine Basilicata and Roberto De Lieto Vollaro Department of Engineering, University of Roma TRE, via della Vasca Navale 79, Rome 00146, Italy; [email protected] (T.P.); [email protected] (L.M.); [email protected] (C.B.); [email protected] (R.D.L.V.) * Correspondence: [email protected]; Tel.: +39-06-5733-3289 Academic Editor: Pietro Buzzini Received: 20 May 2016; Accepted: 19 August 2016; Published: 23 August 2016

Abstract: Exposure to pollutants is usually higher in cities than in the countryside. Generally, in the urban areas pollution sources as traffic, power generator and domestic heating system are more intense and spatially distributed. The pollutants can be classified as a function of long-term toxicological effects due to an exposure and inhalation. In the present work, several kinds of pollutants concentration generated in Rome during 2015 have been analyzed applying different advanced post-processing technique. In particular, statistic and cross-statistic have been computed in time and phase space domain. As main result, it is observed, as expected, that all the pollutant concentrations increase during the winter season into a couple of time ranges despite of [O3 ] that has high values in summer. It can be clearly concluded that Rome has a strongly unsteady behaviour in terms of a family of pollutant concentration, which fluctuate significantly. It is worth noticing that there is a strong linear dependence between [C6 H6 ] and [NO] and a more complex interdependence of [O3 ] and [C6 H6 ]. Qualitatively is provided that, to a reduction of [C6 H6 ] under a certain threshold level corresponds an increase of [O3 ]. Keywords: air quality; emissions; pollutant; statistic analysis; domestic heating; urban traffic; human exposure

1. Introduction Outdoor air pollution has myriad sources, both natural and anthropogenic. It is a mixture of mixtures, and the mix of contaminants in outdoor air varies widely in space and time, reflecting variation in its sources, weather, atmospheric transformations and other factors. In any particular place, the pollution in outdoor air comes not only from local sources, but also from sources that affect air quality regionally and even globally [1]. The main reason for the air quality problems is urban population growth, because people are constantly moving from rural to urban areas [2]. The air pollutants result higher in urban area because of a combination of many elements such as industrial activities, energy production plants and domestic heating [3]. They are noticeable by the toxicological effects resulting from long-term exposure via inhalation [4]. The major pollutants produced are related to human activities, especially those produced by combustion and industrial processes. The International Agency for Research on Cancer (IARC) focused the attention on the human exposures to PM (Particulate Matter). The risk of being exposed to a mixture of pollutants depends on particulate matter smaller than 2.5 µm, which is a common useful indicator [4]. Therefore, the PM2.5 can be taken as indicator of population exposure to outdoor air pollution. In 2010, 3.2 million of people died because of cardiovascular disease, caused by the exposure

Sustainability 2016, 8, 838; doi:10.3390/su8090838

www.mdpi.com/journal/sustainability

Sustainability 2016, 8, 838

2 of 15

to ambient fine particles (PM2.5 ) and 22,300 people died because of lung cancer. China and East Asia show the largest number of people who lost their life [5,6]. With reference to NO2 , SO2 and (PMs) there is general agreement in the scientific literature that they are the main agents responsible for the damage encountered on monuments and historical buildings in urban areas [7]. Atmospheric composition is of unquestionable importance in the study of the damage produced on building materials of artistic interest, since it directly influences the species characteristics and entity of the degradation mechanism occurring on the cultural heritage. The urban areas modified the environmental features that contributed to the increase of pollution. As a matter of fact, the large concentration of the built environment, road pavement and the high building materials capacitance changed the local micrometeorological conditions. Air temperature, humidity and wind velocity and direction are altered in the urban environment compared to rural areas. Furthermore, road traffic, domestic heating, industrial activities and lack buildings energy performance involves high discomfort for users [8–20]. Besides the increase of pollutions, urbanization has led to an increase of the urban heat island intensity (spatially-averaged surface or air-temperature difference between an urban and surrounding rural area(s) [21]). Several studies are focused to reduce the urban heat island effect with different mitigation techniques [22–28]. The people exposed to air pollutants are even more evident considering the weak ventilation, because of the presence of high buildings with a consequent reduction of the dispersion of air masses. For this reason, the contaminants formed below the building height remain at the pedestrian level and increase the health damages especially during thermal inversion episodes. Several studies were conducted to analyse the correlation between the street canyon features and the pollution dispersion [29]. If the ratio between the average height of the buildings (H) and the width of the canyon (W) is high enough to establish skimming flow conditions (at least higher than 0.65), the retention of pollutants within the urban canopy layer will be amplified [30]. The major street canyons in the cities have high value of the ratio H/W with a consequent established helical vortex with an axis parallel to the canyon direction. In this case, the pollutants that go out of the canyon are reduced [31]. The identification of analysis tools and methods, pollutant concentrations measurement, comparison with the threshold values prescribed by law, are the activities foreseen by the legislations in order to monitor the air quality and predict rehabilitation through the definition of plans of interventions. As a first step, in order to plain a control strategy of the pollution concentration in medium and high scale cities, proper measurement and data processing are required to highlight the achievement of dangerous concentration levels of pollutants and formulate a prediction model. Actually, the main active control strategy is based on the introduction of some limits of the urban traffic (e.g., number of the vehicles and vehicle categories that are authorized to transit). Other interventions include the increase of efficiency of the heating systems in buildings such as the replacement of traditional boilers with condensation ones integrated with more performing regulation systems based on energy load tracking. In the present work, several kinds of pollutants concentration such as [CO], [SO2 ], [NOx], [NO], [NO2 ], [C6 H6 ], [PM10 ], [PM2.5 ] and [O3 ] generated in Rome during 2015 have been analysed. These pollutants were taken from the Directive 2008/50/EC, the main legislation about ambient air quality. In these analyses we applied different advanced post-processing techniques. Statistic and cross-statistic have been computed in time and Fourier domain. In particular, probability distribution, Kurtosis, Skewness, Poincaré sections and cross-correlation of the different pollutants were analysed in order to assess the air pollution level in the city of Rome and the correlation of anthropogenic sources with the pollutant emission. The extreme value theory was applied to the experimental data. Especially using the generalized extreme value (GEV) distribution, several fittings of the experiment probability density functions were calculated. GEV distribution introduced by Fisher and Tippett [32] is commonly applied in environmental science to model a wide variety of natural extremes,


3 of 15

including floods, rainfall, wind speeds, wave height, snow depths, earthquake, and other maxima. For the present research activity, probability density functions fitting of the pollutant concentration were calculated. The GEV distribution turned out to be very attractive mathematical tool since its inverse has an analytical form, and its parameters are can be easily estimated [33–38]. This last feature allows us to compute the return period: the likelihood of an event to occur. This post-processing strategy is common applied to pollution database to develop model. Shen et al. developed a statistical model using extreme value theory to estimate changes in ozone episodes [39]. 2. Materials and Methods 2.1. Characteristics of the Study Area Rome is the capital of Italy, and the largest city situated in the west-central part of the country. It is also one of the most populous cities in the European Union with 2.9 million residents in 1285 km2 . Considering the metropolitan area of Rome, the population reaches up to 4.3 million residents in 5352 km2 . With a history of more than two and half thousand years, Rome is called the Eternal City because of the number of open-air museums. It is a mixture of a modern city and a plethora of monuments, piazzas, villas, museums, churches, Egyptian obelisks, the Forum and Vatican City. Its climate is typically Mediterranean: winters are cool and humid and summers are hot and dry. In the coldest month (January) the temperature can reach about 0 ◦ C, and in the warmest months (July and August) the temperature can reach 36 ◦ C. The problem has two faces: there are a lot of pollutant activities and the public transport is not efficient enough to reduce the number of cars. Pedestrian and public transport are only 20% of the total mobility, while 60% of journeys are made by private means of transport; in the historical centre this modal share changes into 34% of pedestrians, 29% of public transport and 37% of private transport [40]. In order to partially solve the problem of pollution, the Government has developed a lot of policies to improve the net of the means of transport [41]. Studying the air pollutant in the city of Rome is important because there is a high density of population and for the effects of pollutant on the variety of open-air museum. 2.2. Pollutant Legislation The European Union policy on air quality has the goal of implementing appropriate instruments aimed at improving air quality. The control of emissions from mobile sources, improving fuel quality and promoting and integrating environmental protection requirements into the transport and energy sector are the main goals of this policy [42]. The Directive 2008/50/EC of the European Parliament [43] is the main legislation about ambient air quality. It fixes health based standards and objectives for a number of pollutants in air. These standards are applied in different time spans because the observed health impacts associated with the various pollutants occur over different exposure times. Table 1 shows the standard adopted in the Directive 2008/50/EC. Table 1. Directive 2008/50/EC standards of main pollutant. Pollutant PM10 PM10 PM2.5 NO2 NO2 SO2 SO2 O3 CO C6 H6

Concentration µg/m3

50 40 µg/m3 25 µg/m3 200 µg/m3 40 µg/m3 350 µg/m3 125 µg/m3 120 µg/m3 10 mg/m3 5 µg/m3

Averaging Period

Permitted Excess Each Year

24 h 1 year 1 year 1h 1 year 1h 24 h Maximum daily 8 h mean Maximum daily 8 h mean 1 year

35 18 24 3 25 days averaged over 3 years -


4 of 15

2.3. Post Processing Techniques In this paper, different statistics and cross-statistics have been computed in time and Fourier domain. These correspond to time-frequency analysis that study a signal in both the time and the frequency domain simultaneously. In particular, the different techniques adopted are:

• •

• •

• • •

Probability distribution describes the possible values that a random variable can take within a given range. Kurtosis is a measure of whether the data have a flattening or elongation from the normal distribution. High kurtosis indicates a flattering distribution, while low values indicate an elongation distribution. Skewness is a measure of the asymmetry of the distribution. A data set is symmetric if it looks the same to the left and right of the center point. Poincaré sections are a way to represent a dynamical system. The surface of section presents a trajectory in n-dimensional phase space in an (n-1)-dimensional space. By picking one phase element constant and plotting the values of the other elements each time the selected element has the desired value, an intersection surface is obtained. The phase space is a surface that describes all the possible states of a system. Cross-correlation is a measure of similarity of two data series as a function of the lag of one relative to the other. Coefficient of variation normalizes the standard deviation with the mean of a data. This index gives information about the variability of a data set. Generalized Extreme Value distribution is often applied to analyse a large set of data characterized by small or large value. In this approach three simpler distributions into a single form are combined, allowing a continuous range of possible shapes.

2.4. Monitoring Station Network ARPA Lazio [44], the regional environmental agency, operates several air quality monitoring sites in the Lazio Region, including Rome. The Rome monitoring network consists of 10 monitoring stations of [CO], [SO2 ], [NOx], [NO], [NO2 ], [C6 H6 ], [PM10 ], [PM2.5 ] and [O3 ] that are shown in Figure 1. The monitoring network acquires concentration data every hour and every stations is set to monitor different type of concentrations. All the stations taken into account are placed inside high density urban areas that are characterized by traffic and domestic heating system pollutant sources. The monitoring stations are constituted by fixed structures in which the entire measurement instrument, the acquisition systems and local storage are placed inside. The procedures and the control of the quality of the measures are assured by ARPA Lazio [45]. A zero value was considered when an instruments errors were found. All the measurement instruments are made by Project Automation. Table 2 shows the sensors used for each concentration. Table 2. Sensors used for each concentration measurement. Pollutant

PMs

NOx

MP101MC SWAMDC FAI Sensors

SWAM5a FAI SWAM DC FAI TE SHARP 5030

SO2

O3

TE 43i M200 A-API

CO

C6 H6

TE 48i

AIR Toxic

M300E API

CP 7001

M400E API M100E API

Sustainability 2016, 8, 838 Sustainability 2016, 8, 838

5 of 15 5 of 15

Figure 1. Google Earth map Rome.Red Red circles circles are stations taken intointo account. Figure 1. Google Earth map ofofRome. arethe themonitoring monitoring stations taken account.

3. Results and Discussion

3. Results and Discussion

In the present study, the data recorded in 2015 were considered. It is useful to consider the

In the present study, the data recorded 2015 were considered. It is useful to consider the mean mean values of contaminants of each in station in order to represent the average pollutant values of contaminants of each station in order to represent the average pollutant concentrations concentrations of Rome. of Rome. Before calculating the mean values concentrations of Rome, Figures 2 and 3 show the mean, standardcalculating deviation, Kurtosis andvalues Skewness values of eachof pollutant the2 year These aremean, Before the mean concentrations Rome, during Figures and2015. 3 show the analysed in order to find the behaviour of each pollutant in the different stations. In the case of are standard deviation, Kurtosis and Skewness values of each pollutant during the year 2015. These [PM10], there is only one station that indicates a different Kurtosis value than the others. Despite this analysed in order to find the behaviour of each pollutant in the different stations. In the case of consideration, in all the concentrations the Kurtosis and Skewness values of the different stations are [PM10 ], there is only one station that indicates a different Kurtosis value than the others. Despite this similar and it suggest that the mean of the network is possible. consideration, in all the concentrations Kurtosis and Skewness values of theduring different From Figures 4–11 is shown how the the mean concentration of all stations work the stations year in are similar and it suggest that the mean of the network is possible. each hour. It can been seen that the major pollutant concentrations happen in the winter seasons From 4–11 shown howthere the mean concentration during from 8Figures to 13, and fromis17 to 2, when is a greater flow of carsof inall thestations city. Thework highest values the are year in each hour. from It can19been seen that majorsystems pollutant concentrations happen in there the winter seasons recorded to 23, when thethe heating are turned on. During the night are lower values due to the turned off heating systems that is usually in Italy. In the summer, the high values from 8 to 13, and from 17 to 2, when there is a greater flow of cars in the city. The highest values are in these hours duewhen to the the urban traffic. systems are turned on. During the night there are lower recorded from 19 are to 23, heating It is worth noticing that the [CO], [NO], [NO ] and [C6in H6Italy. ] haveInsimilar trend during the year values due to the turned off heating systems that is 2usually the summer, the high values in suggesting the presence of pollutant source simultaneously. these hours are due to the urban traffic. The [O3] is recorded to be higher in the summer and during the hottest hours, when the solar It is worth noticing that the [CO], [NO], [NO2 ] and [C6 H6 ] have similar trend during the year radiation is highest. As a matter of fact, the presence of solar radiation allows the reaction of nitrogen suggesting presence of pollutant source simultaneously. dioxidethe (NO 2) in the formation of ozone (O3). The The [O3 ][SO is recorded to be higher in the summer and during the hottest hours, when the solar 2] have values up to 8 μg/m3 compared with the limit of 125 μg/m3 reported in Table 1. radiation is highest. As a matter fact, presence radiation allows thepower reaction of and nitrogen That is foreseeable because theof [SO 2] isthe mainly causedofbysolar the combustion of fuel at plants dioxide (NO formation ozone other industrial facilities. As aof matter of(O fact, 2 ) in the 3 ). in the center of Rome there is not this kind of system. 3 compared 3 reported The 2concentration and PM 10 has highwith values the of winter season and reachinthe The [SO ] have valuesof upPM to 2.58 µg/m theinlimit 125 µg/m Table 1. 3 values in the last of 2015, with by values close to 50 μg/m of PM 2.5 and values That maximum is foreseeable because the two [SO2months ] is mainly caused the combustion of fuel at power plants and 3 of PM10. to 70 μg/m otherclose industrial facilities. As a matter of fact, in the center of Rome there is not this kind of system. The concentration of PM2.5 and PM10 has high values in the winter season and reach the maximum values in the last two months of 2015, with values close to 50 µg/m3 of PM2.5 and values close to 70 µg/m3 of PM10 .

Sustainability 2016, 8, 838 Sustainability 2016, 8, 838 Sustainability 2016, 8, 838

6 of 15 6 of 15 6 of 15

Figure 2. Mean (blue) and standarddeviation deviation (red) values allall pollutant species in the Figure 2. Mean (blue) and standard valuescalculated calculatedfor pollutant species in the Figure 2. Mean (blue) and standard deviation (red) (red) values calculated forfor all pollutant species in the different stations. Blue line is the mean of the mean values of the different stations and the red line is different stations. Blue line is the mean of the mean values of the different stations and the red different stations. Blue line is the mean of the mean values of the different stations and the red line is line the mean of the standard deviation values of the different stations. (a) [CO]; (b) [NO]; (c) [NO2]; is the valuesofofthe thedifferent differentstations. stations. [CO]; [NO]; (c) [NO 2 ]; 2]; themean meanofofthe thestandard standard deviation deviation values (a) (a) [CO]; (b) (b) [NO]; (c) [NO (d) [C6H6]; (e) [O3]; (f) [SO2]; (g) [PM2.5]; (h) [PM10]. (d) [C H ]; (e) [O ]; (f) [SO ]; (g) [PM ]; (h) [PM ]. 2.5]; (h) [PM10]. (d)6 [C66H6]; (e) [O 3 3]; (f) [SO 2 2]; (g) [PM2.5 10

Figure 3. Kurtosis (blue) and Skewness (red) values calculated for all pollutant species in the

Figure 3. Kurtosis (blue) and Skewness (red) values calculated all forpollutant all pollutant species in the Figure 3. Kurtosis (blue) Skewness calculated species different stations. Blue and line is the mean(red) of thevalues Kurtosis values of for the different stations and in thethe reddifferent line different stations. Blue line is the mean of the Kurtosis values of the different stations and the red line the stations. is Skewness the meanvalues of theof Kurtosis values of the(a) different stations and the line is the Blue meanline of the the different stations. [CO]; (b) [NO]; (c) [NO 2]; red (d) [C 6H6is ]; is the mean of the Skewness values of the different stations. (a) [CO]; (b) [NO]; (c) [NO2]; (d) [C6H6]; mean Skewness the[PM different stations. (a) [CO]; (b) [NO]; (c) [NO2 ]; (d) [C6 H6 ]; (e) [O3 ]; (e)of[Othe 3]; (f) [SO2]; (g)values [PM2.5];of(h) 10]. (e) [O3]; (f) [SO2]; (g) [PM2.5]; (h) [PM10]. (f) [SO2 ]; (g) [PM2.5 ]; (h) [PM10 ].

Sustainability 2016, 8, 838 Sustainability 2016, 838 Sustainability 2016, 2016, 8, 8, 838 838 Sustainability Sustainability 2016,8, 8, 838

hour hour hour hour

24 24 24 22 24 22 22 20 22 20 20 18 20 18 18 16 18 16 16 14 16 14 14 12 14 12 12 10 12 10 10 8 10 8 868 6 646 4 424 2 22

7 of 15 77 of 15 of 15 15 77of of 15

30 30 30 30

60 60 60 60

90 90 90 90

120 120 120 120

150 150 150 150

180 180 day 180 180 day day day

210 210 210 210

240 240 240 240

270 270 270 270

300 300 300 300

330 330 330 330

360 360 360 360

12 12 12 12 10 10 10 10 8 8 88 6 6 66 4 4 44 2 2 22 0 0 00

Figure µg/m333.. . Figure 4. Hourly annual concentration variation of C during in Figure 4. 4. Hourly Hourly annual annual concentration concentration variation variationof ofC C666H H6666 during during 2015 2015 in in μg/m μg/m Figure Figure 4. 4. Hourly Hourly annual annual concentration concentration variation variation of of CC66H H6 during 2015 2015 in μg/m μg/m33..

hour hour hour hour

24 24 24 24 22 22 22 20 22 20 20 18 20 18 18 18 16 16 16 14 16 14 14 12 14 12 12 10 12 10 10 10 8 8 868 6 646 4 424 2 22

30 30 30 30

60 60 60 60

90 90 90 90

120 120 120 120

150 150 150 150

180 180 day 180 180 day day day

210 210 210 210

240 240 240 240

270 270 270 270

300 300 300 300

330 330 330 330

360 360 360 360

3.5 3.5 3.5 3.5 3 3 33 2.5 2.5 2.5 2.5 2 2 22 1.5 1.5 1.5 1.5 1 1 11 0.5 0.5 0.5 0.5 0 0 00

3 Figure 5. annual variation CO 2015 in mg/m Figure 5. 5. Hourly Hourly annual annual concentration concentration variation variation of of CO CO during during 2015 2015 in in mg/m mg/m3333.... Figure Figure 5.Hourly Hourly annualconcentration concentration variationof of COduring during 2015in in mg/m mg/m .

400 400 400 400 300 300 300 300

hour hour hour hour

24 24 24 22 24 22 22 20 22 20 20 18 20 18 18 16 18 16 16 14 16 14 14 12 14 12 12 10 12 10 10 8 10 8 868 6 646 4 424 2 22

200 200 200 200 100 100 100 100 30 30 30 30

60 60 60 60

90 90 90 90

120 120 120 120

150 150 150 150

180 180 day 180 180 day day day

210 210 210 210

240 240 240 240

270 270 270 270

300 300 300 300

330 330 330 330

360 360 360 360

0 0 00

3 Figure 6. annual concentration variation of NO during 2015 in μg/m Figure 6. 6. Hourly Hourly annual annual concentration concentration variation variation of of NO NO during during 2015 2015 in in μg/m μg/m333... Figure Figure6.6. Hourly Hourly annual annualconcentration concentrationvariation variationof ofNO NOduring during2015 2015in inµg/m μg/m . Figure Hourly

hour hour hour hour

24 24 24 24 22 22 22 22 20 20 20 20 18 18 18 18 16 16 16 16 14 14 14 14 12 12 12 12 10 10 10 10 8 8688 6466 4244 222

30 30 30 30

60 60 60 60

90 90 90 90

120 120 120 120

150 150 150 150

180 180 day 180 180 day day day

210 210 210 210

240 240 240 240

270 270 270 270

300 300 300 300

330 330 330 330

3 Figure 7. Hourly annual concentration variation of NO during 2015 in μg/m Figure 7. 7. Hourly Hourly annual annual concentration concentration variation variation of of NO NO222 during during 2015 2015 in in μg/m μg/m3333... Figure Figure7.7.Hourly Hourlyannual annualconcentration concentrationvariation variationof ofNO NO22 during during 2015 2015in inµg/m μg/m .. Figure

360 360 360 360

160 160 160 160 140 140 140 140 120 120 120 120 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 000


8 of 15 8 of 15

hour hour

8 of 15 8 of160 15

hour hour

Sustainability 2016, 8, 838 Sustainability 2016, 8, 838 24 22 24 20 22 24 18 20 22 16 18 20 14 16 18 12 14 16 10 12 14 12 8 10 8 6 10 4 68 2 46 24 30 60 2

30 30

60 60

90

90 90

120

120 120

150

150 150

180 day 180

day 180 day

210 210 210

240

240 240

270

270 270

300

300 300

330

330 330

160140 160 140120 140 120100 120 10080 100 80 60 80 60 40 60 40 40 20 20 360 20 360 360

24 24 22 22 24 20 20 22 18 18 20 16 16 18 14 14 16 12 12 14 12 10 10 8 10 8 6 68 4 46 2 24

Figure 8. Hourly annual concentration variation of O3 during 2015 in μg/m3.

8 8 8 6 6 6

hour hour

hour hour

33. 3 Figure Hourlyannual annualconcentration concentration variation of O during 2015 ininμg/m Figure 8. 8. Hourly ofO O3333during during 2015 µg/m . 3. Figure 8. Hourly annual concentrationvariation variation of 2015 in μg/m

2

4 4 4 2 2 2

30 30

30

6060

60

9090

90

120 120

120

150 150

150

180 180 day 180 day

day

210 210

240 240 240

270 270 270

300 300 300

330 330 330

Figure 9. Hourly annual concentrationvariation variation of 2015 ininμg/m . 33. 3 . Figure Hourly annual concentration variation of SO SO during 2015 Figure 9. 9. Hourly annual concentration SO22222during during 2015 inμg/m µg/m 3. Figure 9. Hourly annual concentration variation of SO during 2015 in μg/m 3

Figure 10. Hourly annual concentration variation of PM2.5 during 2015 in μg/m3.

Figure 10. Hourly annual concentrationvariation variation of PM2.5 during 2015 in μg/m . 3 3. Figure 10.10. Hourly during 2015 in µg/m Figure Hourlyannual annualconcentration concentration variation of 2.5 of PM2.5 2.5 during 2015 in μg/m 3. 3

Figure 11. Hourly annual concentration variation of PM10 during 2015 in μg/m3. Figure 11. Hourly annual concentration variation of PM10 during 2015 in μg/m3. 3 Figure Hourlyannual annualconcentration concentrationvariation variation of of PM 10 during 2015 in μg/m 3. 3 Figure 11.11. Hourly PM10 10 during 2015 in µg/m .

0 0 360 360 0 360


9 of 15 9 of 15

The main pollutant species terms probability density function (pdf) have been analysed The main pollutant species in in terms of of probability density function (pdf ) have been analysed and and represented in normalized form in order to have zero mean and unitary standard deviation represented in normalized form in order to have zero mean and unitary standard deviation defined as defined follows:as follows: x − hxi xnorm == − 〈 〉 (1) (1) σ σx where as as h·i〈∙〉 is denoted the ensemble deviation. where is denoted the ensembleaverage averageand andas as σσisisindicated indicatedthe the standard standard deviation. It It is is shown in Figure 12 that [CO], [NO], [SO ] and [C H ] have a narrow distribution around 2 6 6 shown in Figure 12 that [CO], [NO], [SO2] and [C6H6] have a narrow distribution around the themost mostfrequent frequentvalue valueand andsignificant significant positive tail. The latter aspect can be ascribed to intermittent positive tail. The latter aspect can be ascribed to intermittent energetic events embedded in in concentration time histories. To To better clarify thisthis aspect all pdf are are fitted energetic events embedded concentration time histories. better clarify aspect all pdf using the Generalized Extreme Value (GEV) function. By means of the GEV function, the return level fitted using the Generalized Extreme Value (GEV) function. By means of the GEV function, the and period were in calculated order to evaluate occurrence time of these events as reported return level andcalculated period were in order the to evaluate the occurrence time of these events as in reported in Figure 13. With reference the return period of [NO], every two days a high energetic Figure 13. With reference to the return to period of [NO], every two days a high energetic peak in [NO] peak in [NO]can concentration On the other hand, in [NO 2 ] time history intermittent concentration be detected.can Onbe thedetected. other hand, in [NO ] time history intermittent events are rarely 2 events are rarely observed. This aspect is a footprint of chaotic behavior of the Rome in terms of observed. This aspect is a footprint of chaotic behavior of the Rome in terms of pollutant generation pollutant generation could during be takenthe into account during thethis modelling and could be taken intoand account modelling stage of process.stage of this process. Instead [NO 2],[PM [PM 2.5], ], [PM 10]] and [O 3] ]have a distribution with a more flatflat trend. In In particular, Instead [NO ], [PM and [O have a distribution with a more trend. particular, 2 2.5 10 3 [NO 2] pdf is well fitted by the Gaussian pdf except for a small region over 3σ where a departure of the [NO2 ] pdf is well fitted by the Gaussian pdf except for a small region over 3σ where a departure of In In thethe latter case, wewe can consider, in in a first approximation, thepositive positivetail tailcan canbebeclearly clearlyobserved. observed. latter case, can consider, a first approximation, equally probable concentrations in a certain range. equally probable concentrations in a certain range.

101

101

100

100

10

-1

10-1

10

-2

10-2

10-3

10-3

10-4 -2

0

2

4

6

8

10

10-4 -2

101

101

0

100

10-1

10-1

10-2

10-2

10

-3

10-3

10

-4

10

-2

0

2

4

6

8

10

10-4 -2

0

2

4

6

8

10

0

2

4

6

8

10

Figure12. 12. Probability (red(red triangles) and Generalized Extreme Value fitting Figure Probabilitydensity densityfunction function triangles) and Generalized Extreme Value(blue fitting circles) of pollutant concentrations calculated for samples acquired in a year. (blue circles) of pollutant concentrations calculated for samples acquired in a year.


10 of 15


10 of 15 Analysing the probability distributions in comparing with the standards reported in Table 1, it could be notice that [NO ], [PM ], [PM ], [C H ] and [O ] exceeds the limit of legislation 2 2.5 10 in comparing 6 6 3 the standards reported in Table 1,during Analysing the probability distributions with it 3 , 23% of time the [PM ] the year. In particular, 64% of time the [NO ] was over the limit of 40 µg/m 2 could be notice that [NO2], [PM2.5], [PM10], [C6H6] and [O3] exceeds the limit of legislation during the 2.5 3 , and 21% of time [PM ] was over the limit 3 Regarding the was over limit of 2564% µg/m of of 40time µg/m 10 limit of 40 μg/m3, 23% year. the In particular, of time the [NO2] was over the the .[PM 2.5] was 3. Regarding concentration of [Cof6 H and 3[O time in which concentration the limit of legislation over the limit 256 ]μg/m , and 21% of time [PM10]the was over the limitisofover 40 μg/m the is 3 ] the 3 concentration [C6H6] and4% [Oof 3] the time the concentration is of over limit and of legislation is the less than previousofpollutant: time thein[Cwhich over the limit 40the µg/m 3% of time 6 H6 ] was pollutant: 4% of3time the [C6H6] was over the limit of 40 μg/m3 and 3% of time the [O3 ] less wasthan overprevious the limit of 120 µg/m . 3 [O ] was over the limit 120 μg/m In32015, regarding theoflimit in the. short averaging period (from 1 to 24 h), [NO2 ], [SO2 ] and [CO] In 2015, regarding the limit in the shortFor averaging (from 1 toµg/m 24 h),3[NO ], [SO 2] and [CO] haven’t exceeded the legislation thresholds. [PM10 ]period the limit of 50 for 2an averaging period haven’t exceeded the legislation thresholds. For [PM10] the limit of 50 μg/m3 for an averaging period of 24 h was exceeded 39 times that is more than 35 permitted exceedances each year by the legislation. of 24 h was exceeded 39 times that is more than 35 permitted exceedances each year by the These exceedances are concentrated in December when there was a combination of exogenous state legislation. These exceedances are concentrated in December when there was a combination of variable (i.e., temperature, humidity and wind velocity and traffic flow) and low exogenous state variable (i.e., temperature, humidity and and winddirection, velocity and direction, and traffic 3 for an averaging period of 8 h was exceeded 7 times. For this rainfall. For [O ] the limit of 120 µg/m 3 rainfall. For [O3] the limit of 120 μg/m3 for an averaging period of 8 h was exceeded 7 flow) and low concentration, the concentration, legislation permitted 25 exceedances over 3 years. times. For this the legislation permitted 25days exceedances days over 3 years.

Figure 13. Dimensionless Return level trend for all the pollutant concentrations.

Figure 13. Dimensionless Return level trend for all the pollutant concentrations.

As stated previously and as further confirmed by the values assumed by Kurtosis and As stated (see previously asreference further confirmed by[CO], the values and Skewness Skewness Table 3),and with to the set of [NO], assumed [SO2] and by [C6Kurtosis H6] a high Kurtosis value is evaluated, indicating that these pollutants are present on Rome in constant quantities. On (see Table 3), with reference to the set of [CO], [NO], [SO2 ] and [C6 H6 ] a high Kurtosis value is evaluated, the other thepollutants high Kurtosis values suggests distribution around theother mode.hand Whereas indicating thathand these are present on Romea sharped in constant quantities. On the the high the coefficient of variation of [NO] have large value indicating thatWhereas there are the highcoefficient energetic extreme Kurtosis values suggests a sharped distribution around the mode. of variation value (seelarge Figurevalue 12). indicating that there are high energetic extreme value (see Figure 12). of [NO] have Table 3. Kurtosis and Skewness values calculated for all pollutant species.

Table 3. Kurtosis and Skewness values calculated for all pollutant species. Specie [CO] [NO] [SO2] [C6H6] [NO2] [PM2.5] [PM10] [O3]

Kurtosis 14.3 Kurtosis 17.1 [CO] 14.3 9.6 [NO] 17.1 9.4 [SO2 ] 9.6 3.9 [C6 H6 ] 9.4 4.2 [NO2 ] 3.9 3.7 [PM2.5 ] 2.8 4.2

Specie

[PM10 ] [O3 ]

3.7 2.8

Skewness

Coefficient of Variation

2.7 Coefficient of Variation 66% Skewness

2.7 3.3 2.2 2.1 0.7 1.3 1.0 0.7

3.3 2.2 2.1 0.7 1.3 1.0 0.7

66% 155% 91% 72% 48% 60% 46% 80%

155% 91% 72% 48% 60% 46% 80%


11 of 15

In the case of [NO2 ] the values assumed by the Kurtosis are close to those of a Gaussian Sustainability 2016, 8, 838 11 of 15 Sustainability 8, 838 of 15 distribution as2016, confirmed by the comparison in Figure 12, indicating the presence of a 11 species of pollutant In that significantly around an average thefluctuate case of [NO 2] the values assumed by the value. Kurtosis are close to those of a Gaussian In the case of [NO2] the values assumed by the Kurtosis are close to those of a Gaussian Given that the most important benzene source is the12,urban traffic, concentration of benzene distribution as confirmed by the comparison in Figure indicating thethe presence of a species of distribution as confirmed by the comparison in Figure 12, indicating the presence of a species of that anits average value. can be taken into account as reference. may pollutant be used as anfluctuate index ofsignificantly city traffic around flux and time history pollutant that fluctuate significantly around an average value. Given that the most important benzeneby source is the urban traffic, the concentration benzene The production of benzene is characterized small oscillations in Rome during theof window Given that the most important benzene source is the urban traffic, the concentration oftime benzene may be used as an index of city traffic flux and its time history can be taken into account as reference. investigated. Furthermore, [C H ] are strongly related to the concentration of pollutants species 6 traffic 6 may be used as an index of city flux and its time history can be taken into account as reference. The production of benzene is characterized by small oscillations in Rome during the time window such The as [CO], [NO] of and [NO2 ], confirmedby bysmall the cross-correlation values in Figure 14. production benzene is as characterized oscillations in Rome duringreported the time window investigated. Furthermore, [C6H6] are strongly related to the concentration of pollutants species such investigated.isFurthermore, [C6H 6] are stronglyoxide relatedand to the concentration of pollutants species such and This correlation due to the fact that nitrogen carbon oxides are combustion products as [CO], [NO] and [NO2], as confirmed by the cross-correlation values reported in Figure 14. This as [CO], [NO] and are [NOdue 2], asto confirmed by the cross-correlation values reported in Figure 14. This that their fluctuations the oscillations of the city traffic flux. It is noteworthy that [O ] is correlation is due to the fact that nitrogen oxide and carbon oxides are combustion products and that 3 correlation is due to the fact that nitrogen oxide and carbon oxides are combustion products and that weakly related to theare production of benzeneofbut is traffic indeed in It phase opposition To further their fluctuations due to the oscillations the it city flux. is noteworthy thatwith [O3] isit.weakly their fluctuations are due to the the oscillations of section the city traffic flux. It is noteworthy that [O3]] upon is weakly investigate this interesting aspect, Poincaré of [C H ] upon [NO] and [C H [O3 ] are 6 related to the production of benzene but it is indeed in6 phase opposition with 6it. 6To further related to the production of benzene but it is indeed in phase opposition with it. To further represented in Figure 15. As noted above there is a strong linear dependence between [C H ] and investigate this interesting aspect, the Poincaré section of [C6H6] upon [NO] and [C6H6] upon 6 [O 6 3] are [NO] investigate this interesting aspect, the Poincaré section of [C6H6] upon [NO] and [C6H6] upon [O3] are in Figure 15. As noted of above strong linear dependence 6H6] and and arepresented more complex interdependence [O ]there and is [Ca H specifically anbetween inverse [C proportionality 6 ]. More represented in Figure 15. As noted above3 there is a6 strong linear dependence between [C6H6] and [NO]observed and a more complex interdependence of [O3] and [C6H6]. More specifically an inverse is clearly as formalized in the following: [NO] and a more complex interdependence of [O3] and [C6H6]. More specifically an inverse proportionality is clearly observed as formalized in the following: proportionality is clearly observed as formalized in the following: [O d [C[C d [O 3 ] 6 HH6 ]] ] ∼−− [O ] [C H ] ~ dt dt ~−

(2) (2)

Figure 14. Cross-correlation betweenpollutant pollutant concentrations; concentrations; benzene concentration is taken into into Figure 14. Cross-correlation between benzene concentration is taken Figure 14. Cross-correlation between pollutant concentrations; benzene concentration is taken into account as a reference signal; the timeaxis axisisisnormalized normalized respect totoa a reference time: t* = t* 86,400 s. account as a reference signal; the time respect reference time: = 86,400 account as a reference signal; the time axis is normalized respect to a reference time: t* = 86,400 s. s.

Figure 15. Poincaré sections[C[CH 6H6] over [NO] (a) and [C6H6] over [O3] (b). Figure 15. Poincaré sections [NO](a) (a)and and[C[C 6 6H66]] over 66H 6 ] over 3 ] (b). Figure 15. Poincaré sections [C over [NO] 6H ] over [O3][O (b).

(2)


12 of 15

Qualitatively is provided that to a reduction of [C6 H6 ] under a certain threshold level corresponds an increase of [O3 ]. Such behaviour could be attributed to exogenous state variable, which were not taken into account in the present work. 4. Conclusions Observance of air quality standards represents a great challenge in cities, especially in the ones in which traffic and other additional sources are combined with bad weather conditions. With reference to this issue, the results herein provided shown that the major pollutant concentrations are observed in the winter seasons during the intense traffic flow and the threshold values are often exceeded. Furthermore, according to the coefficient of variation values, Rome has a strongly unsteady behaviour in terms of a family of pollutant concentration which fluctuate significantly. [C6 H6 ] are strongly related to the concentration of pollutants species such as [CO], [NO] and [NO2 ], as confirmed by the cross-correlation analysis. This correlation is due to the fact that nitrogen oxide prompts and carbon oxides are combustion products and that their fluctuations are caused by the oscillations of the city traffic flux. [O3 ] is weakly related to the production of benzene and it is also in phase opposition with it. For this reason, the Poincaré section of [C6 H6 ] upon [NO] and [C6 H6 ] upon [O3 ] was investigated. It is worth noticing that there is a strong linear dependence between [C6 H6 ] and [NO] and a more complex interdependence of [O3 ] and [C6 H6 ]. Qualitatively is provided that, to a reduction of [C6 H6 ] under a certain threshold level which corresponds to an increase of [O3 ]. Such behaviour could be attributed to exogenous state variable. The main objective for the reduction of air pollutant is to reach, thanks to the implementation of suitable mobility policies, an urban sustainable development, i.e., to improve traffic mobility conditions, to increase road safety and to decrease traffic caused by pollution and to re-qualify urban spaces. It includes rationalizing public space, safeguarding citizens’ health and life quality, and preserving historical and architectural heritage. The government is promoting both long and short-term activities to reach these goals. From the analysis of the current air situation, it is clear that traffic is the reason why the pollutant concentrations are so high. Traffic is the main source of [CO], [C6 H6 ] and [PM10 ] concentrations. In order to reduce the pollution in Italian cities and in particular in Rome, measures are needed to decrease the level of the various substances dispersed into the air. One of the main actions that can be performed is the reduction or elimination of the use of the most polluting cars, i.e., cars Euro 0, 1 and 2. As a matter of fact, Roma Capitale has imposed the circulation reduction of these cars permanently from 15 December 2015 [46]. Regarding the reduction of the pollutants due to buildings heating systems, it is necessary to replace traditional boilers with more efficient systems, such as condensing boilers. As a matter of fact, the condensing boilers allow reducing the utilization of combustion and a consequence decrease of emission. However, the boilers are only one element of the heating systems. Its efficiency depends on other elements such as distribution, emission and regulation. Using condensing boilers coupled with other high heating systems elements, can improve the total efficiency and reduce the environmental emission. The Poincaré sections in Figure 15 gives a significant contribution to the modeling purpose: such as dynamical model (ODE) or regression approach (LUR) (see among many [47,48]). More specifically, the experimental analysis shows a correlation between [C6 H6 ] over [NO] and [C6 H6 ] over [O3 ]. These aspects are well known for small scale zero-dimensional reactor, but for very large scale problems as the city domain is not commonly investigated. Therefore, a reduction of a three dimensional large scale process to zero-dimensional phenomenon as in homogeneous volume can be consider the first step for a mathematical model development by means of an ODE system as follows: x = f [ x (t)] where x (t) = {[NO] , [O3 ] , [C6 H6 ] , . . .}

(3)


13 of 15

This aspect is the fundamental task to predict the time evolution of the pollutant species generated within an urban domain. Furthermore, as a result, it can be clearly concluded that Rome has a strongly unsteady behaviour in terms of pollutant concentration. All-time series are positive skewed, indicating that some short time rare events, of pollutant concentration, have an order of magnitude bigger than the expected values. Such as intermittent behaviour must be taken into account for modelling purpose. Acknowledgments: The authors are grateful to ARPA Lazio which provided the experimental data. Author Contributions: The work was designed by Gabriele Battista, Tiziano Pagliaroli and Roberto De Lieto Vollaro. Gabriele Battista and Tiziano Pagliaroli wrote the paper and perform the statistic and cross-statistic analysis. English corrections were revised by Luca Mauri and Carmine Basilicata. Finally, Roberto De Lieto Vollaro supervised the work related to the paper. All authors read and approved the final manuscript. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2. 3.

4.

5.

6. 7. 8. 9.

10.

11.

12. 13. 14. 15.

International Agency for Research on Cancer (IARC). Outdoor air pollution. In IARC Monograph on the Evaluation of Carcinogenic Risks to Humans; International Agency for Research on Cancer: Lyon, France, 2015; Volume 109. Mayer, H. Air pollution in cities. Atmos. Environ. 1999, 33, 4029–4037. [CrossRef] Guerrieri, M.; Corriere, F.; Rizzo, G.; Casto, B.L.; Scaccianoce, G. Improving the Sustainability of Transportation: Environmental and Functional Benefits of Right Turn by-Pass Lanes at Roundabouts. Sustainability 2015, 7, 5838–5856. [CrossRef] Schiavon, M.; Redivo, M.; Antonacci, G.; Rada, E.C.; Ragazzi, M.; Zardi, D.; Giovannini, L. Assessing the air quality impact of nitrogen oxides and benzene from road traffic and domestic heating and the associated cancer risk in an urban area of Verona (Italy). Atmos. Environ. 2015, 120, 234–243. [CrossRef] Lim, S.S.; Vos, T.; Flaxman, A.D.; Danaei, G.; Shibuya, K.; Adair-Rohani, H.; Adair-Rohani, H.; Amann, M.; Anderson, H.R.; Andrews, K.G.; et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet 2013, 380, 2224–2260. [CrossRef] Straif, K.; Cohen, A.; Samet, J. Air Pollution and Cancer. IARC Scientific Publication No. 161; International Agency for Research on Cancer: Lyon, France, 2015. Ghedini, N.; Gobbi, G.; Sabbioni, C.; Zappia, G. Determination of elemental and organic carbon on damaged stone monuments. Atmos. Environ. 2000, 34, 4383–4391. [CrossRef] Coppi, M.; Quintino, A.; Salata, F. Numerical study of a vertical channel heated from below to enhance natural ventilation in a residential building. Int. J. Vent. 2013, 12, 41–49. [CrossRef] Salata, F.; Golasi, I.; de Lieto Vollaro, E.; Bisegna, F.; Nardecchia, F.; Coppi, M.; Gugliermetti, F.; de Lieto Vollaro, A. Evaluation of different urban microclimate mitigation strategies through a PMV analysis. Sustainability 2015, 7, 9012–9030. [CrossRef] De Lieto Vollaro, R.; Evangelisti, L.; Carnielo, E.; Battista, G.; Gori, P.; Guattari, C.; Fanchiotti, A. An Integrated Approach for an Historical Buildings Energy Analysis in a Smart Cities Perspective. Energy Procedia 2014, 45, 372–378. [CrossRef] Evangelisti, L.; Battista, G.; Guattari, C.; Basilicata, C.; de Lieto Vollaro, R. Influence of the Thermal Inertia in the European Simplified Procedures for the Assessment of Buildings’ Energy Performance. Sustainability 2014, 6, 4514–4524. [CrossRef] Salata, F.; de Lieto Vollaro, A.; de Lieto Vollaro, R.; Davoli, M. Plant reliability in hospital facilities. Energy Procedia 2014, 45, 1195–1204. [CrossRef] Evangelisti, L.; Battista, G.; Guattari, C.; Basilicata, C.; de Lieto Vollaro, R. Analysis of Two Models for Evaluating the Energy Performance of Different Buildings. Sustainability 2014, 6, 5311–5321. [CrossRef] De Lieto Vollaro, R.; Vallati, A.; Bottillo, S. Differents Methods to Estimate the Mean Radiant Temperature in an Urban Canyon. Adv. Mater. Res. 2013, 650, 647–651. [CrossRef] D’Orazio, A.; Fontana, L.; Salata, F. Experimental study of a semi-passive ventilation grille with a feedback control system. Rev. Sci. Instrum. 2011, 82, 085107. [CrossRef] [PubMed]


16. 17. 18.

19. 20.

21. 22. 23.

24. 25. 26. 27.

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

14 of 15

Coppi, M.; Quintino, A.; Salata, F. Fluid dynamic feasibility study of solar chimney in residential buildings. Int. J. Heat Technol. 2011, 29, 1–5. Salata, F.; Golasi, I.; de Lieto Vollaro, R.; de Lieto Vollaro, A. Outdoor thermal comfort in the Mediterranean area. A transversal study in Rome, Italy. Build. Environ. 2016, 96, 46–61. [CrossRef] Salata, F.; Alippi, C.; Tarsitano, A.; Golasi, I.; Coppi, M. A first approach to natural thermoventilation of residential buildings through ventilation chimneys supplied by solar ponds. Sustainability 2015, 7, 9649–9663. [CrossRef] Vallati, A.; De Lieto Vollaro, A.; Golasi, I.; Barchiesi, E.; Caranese, C. On the impact of urban micro climate on the energy consumption of buildings. Energy Procedia 2015, 82, 506–511. [CrossRef] Battista, G.; Carnielo, E.; Evangelisti, L.; Frascarolo, M.; de Lieto Vollaro, R. Energy Performance and Thermal Comfort of a High Efficiency House: RhOME for denCity, Winner of Solar Decathlon Europe 2014. Sustainability 2015, 7, 9681–9695. [CrossRef] Memon, R.A.; Leung, D.Y.C.; Liu, C. An investigation of urban heat island intensity (UHII) as an indicator of urban heating. Atmos. Res. 2009, 94, 491–500. [CrossRef] Goia, F.; Zinzi, M.; Carnielo, E.; Serra, V. Characterization of the optical properties of a PCM glazing system. Energy Procedia 2012, 30, 428–437. [CrossRef] Paolini, R.; Zinzi, M.; Poli, T.; Carnielo, E.; Mainini, A.G. Effect of ageing on solar spectral reflectance of roofing membranes: Natural exposure in Roma and Milano and the impact on the energy needs of commercial buildings. Energy Build. 2014, 84, 333–343. [CrossRef] Goia, F.; Zinzi, M.; Carnielo, E.; Serra, V. Spectral and angular solar properties of a PCM-filled double glazing unit. Energy Build. 2015, 87, 302–312. [CrossRef] Zinzi, M.; Carnielo, E.; Federici, A. Preliminary studies of a cool roofs’ energy-rating system in Italy. Adv. Build. Energy Res. 2014, 8, 84–96. [CrossRef] Carnielo, E.; Zinzi, M. Optical and thermal characterisation of cool asphalts to mitigate urban temperatures and building cooling demand. Build. Environ. 2013, 60, 56–65. [CrossRef] De Lieto Vollaro, A.; Galli, G.; Vallati, A.; Romagnoli, R. Analysis of thermal field within an urban canyon with variable thermophysical characteristics of the building’s walls. J. Phys. Conf. Ser. 2015, 655, 012056. [CrossRef] De Lieto Vollaro, A.; Galli, G.; Vallati, A. CFD Analysis of Convective Heat Transfer Coefficient on External Surfaces of Buildings. Sustainability 2015, 7, 9088–9099. [CrossRef] Chan, A.T.; So, E.S.P.; Samad, S.C. Strategic guidelines for street canyon geometry to achieve sustainable street air quality. Atmos. Environ. 2001, 35, 4089–4098. [CrossRef] Oke, T.R. Street design and urban canopy layer climate. Energy Build. 1988, 11, 103–111. [CrossRef] Britter, R.E.; Hanna, S.R. Flow and dispersion in urban areas. Annu. Rev. Fluid Mech. 2003, 35, 469–496. [CrossRef] Fisher, R.A.; Tippett, L.H.C. Limiting forms of the frequency distribution of the largest or smallest member of a sample. Math. Proc. Camb. Philos. Soc. 1928, 24, 180–190. [CrossRef] Martins, E.S.; Stedinger, J.R. Generalized maximum-likelihood generalized extreme-value quantile estimators for hydrologic data. Water Resour. Res. 2000, 36, 737–744. [CrossRef] Kütchenhoff, H.; Thamerus, M. Extreme value analysis of Munich air pollution data. Environ. Ecol. Stat. 1996, 3, 127–141. [CrossRef] Smith, R.L. Extreme Value Analysis of Environmental Time Series: An Application to Trend Detection in Ground-Level Ozone. Stat. Sci. 1989, 4, 367–377. [CrossRef] Rao, S.T.; Sistla, G.; Henry, R. Statistical analysis of trends in urban ozone air quality. J. Air Waste Manag. Assoc. 1992, 42, 1204–1211. [CrossRef] Nor, A.M.A.; Mohd, B.A.; Ahmad, Z.A. Bayesian Extreme for Modeling High PM10 Concentration in Johor. Procedia Environ. Sci. 2015, 30, 309–314. Shi, L.; Stuart, B.; Feng-Chiao, S.; Bhramar, M. Addressing extrema and censoring in pollutant and exposure data using mixture of normal distributions. Atmos. Environ. 2013, 77, 464–473. Shen, L.; Mickley, L.J.; Gilleland, E. Impact of increasing heat waves on US ozone episodes in the 2050s: Results from a multimodel analysis using extreme value theory. Geophys. Res. Lett. 2016, 43, 4017–4025. [CrossRef] [PubMed]


40.

41. 42. 43.

44. 45. 46.

47.

48.

15 of 15

Roma Capitale, Piano Generale del Traffico Urbano (in English: City Urban Traffic General Plan). April 2015. Available online: https://www.comune.roma.it/pcr/it/dip_mob_delibere.page (accessed on 22 August 2016). United States Environmental Protection Agency (EPA). Air Quality Planning and Standards. Available online: https://www3.epa.gov/airquality/ (accessed on 22 August 2016). European Commission. Available online: http://ec.europa.eu/ (accessed on 22 August 2016). European Union. Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on Ambient Air Quality and Cleaner Air for Europe. Available online: http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=OJ:L:2008:152:0001:0044:en:PDF (accessed on 22 August 2016). ARPA Lazio—Centro Regionale della Qualità dell’Aria (in English: Regional Center of Air Quality). Available online: http://www.arpalazio.net/ (accessed on 22 August 2016). ARPA Lazio. Valutazione della Qualità dell’aria—2015 (In English: Air Quality Assessment—2015). RT/DAI/16/01; ARPA Lazio: Rieti, Italy, 2015. Roma Capitale, Fascia Verde e Anello Ferroviario (in English: Green Band and Ring Rail). Available online: http://www.agenziamobilita.roma.it/it/servizi/orari-ztl/fascia-verde-anello-ferroviario.html (accessed on 22 August 2016). Beelen, R.; Hoek, G.; Vienneau, D.; Eeftens, M.; Dimakopoulou, K.; Pedeli, X.; Tsai, M.; Künzli, N.; Schikowski, T.; Marcon, A.; et al. Development of NO2 and NOx land use regression models for estimating air pollution exposure in 36 study areas in Europe—The ESCAPE project. Atmos. Environ. 2013, 72, 10–23. [CrossRef] ˙ ˙ A.; Eeftens, M.; Beelen, R.; de Hoogh, K.; Bellander, T.; Cesaroni, G.; Cirach, M.; Declercq, C.; Dedel e, Dons, E.; de Nazelle, A.; et al. Development of land use regression models for PM2.5, PM2.5 absorbance, PM10 and PMcoarse in 20 European study areas; results of the ESCAPE project. Environ. Sci. Technol. 2012, 46, 11195–11205. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).