Key factors determining indoor air PM10 concentrations in naturally ventilated primary schools in Belgrade, Serbia
Article Category: Original scientific article
Published Online: Oct 09, 2017
Page range: 227 - 235
Received: May 31, 2016
Accepted: Jul 27, 2017
DOI: https://doi.org/10.1515/sjph-2017-0031
Keywords
© 2017 National Institute of Public Health, Slovenia
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
The quality of indoor air of homes, offices, or other public or private dwellings could be accounted as one of the essential determinants of a healthy life and wellbeing of each individual (1). School indoor air quality (IAQ) is expected to have a key role in the assessment of children’s personal exposure to air pollution, concerning the fact that they spend at least a third of their time inside school buildings, approximately 7 hours a day (2-5). Children are particularly vulnerable to all types of pollutants, because their breathing and metabolic rates are high. In a school, they have much less floor space than adults working in a typical office. Their breathing zone tends to be closer to pollutant sources, such as a new carpet, and less likely to be well ventilated, as it is below the window level. The immune system of young children is yet immature, and exposure to pollutants can mean allergic reactions or ill health (6).
Poor indoor environment in schools may be attributed to three primary causes: inexistence or inadequate operation and maintenance of ventilation systems, infrequent and unthoroughly cleaned indoor surfaces, and a large number of students in relation to room area and volume, with constant re-suspension of particles from the surface (1). Furthermore, IAQ can be attributed to various phases of the building process, including poor site selection, choice of materials, roof design, poor construction quality, improper installation, or any number or combination of these factors (7). Therefore, it is of utmost importance to provide good IAQ in classrooms, to help minimize these effects (8, 9). Sources of IA (indoor air) pollution could be: furnishings, IT equipment, bio-effluents, and external pollutants, such as nitrogen-dioxide and carbon-monoxide (3, 10-12). In the indoor environment, in which people spend most of their time, both indoor and outdoor sources contribute to PM levels. Indoor PM is affected by ambient concentrations, air exchange rates, penetration factors, as well as deposition and re-suspension mechanisms. In this complex microenvironment, activities such as cleaning, walking, playing, and, particularly, smoking cause the formation of PM in indoor air (13,14).
The main objective of this paper was to study the difference of PM10 concentration values, sampled both inside classrooms of ten Belgrade primary schools and in ambient air in front of them, and relate it to different classroom and school characteristics, using SEARCH Project methodology. A specific objective was to study the difference of concentration of the PM10 measurements inside and outside the chosen classrooms. In addition, we aimed to correlate a specific quantitative indicator of the thermal comfort zone, as given by the ASHRAE Standards (4), that is, space occupancy (<2m2 of indoor space per child, not suitable), to the measured values of indoor PM10.
The cross-sectional SEARCH (School Environment and Respiratory Health in Children) study has involved 6 European countries (Albania, Bosnia and Herzegovina, Hungary, Serbia, Slovakia, and Italy). In Serbia, the research was undertaken in the capital city of Belgrade. The project has defined that a sample of 10 schools per country shall be enough to get a clear picture on the level of indoor exposure of children in primary schools. They were chosen (sampled) to be with heterogeneous characteristics. Primarily, schools were grouped by their location, i.e. the level of their exposure to potential sources of air pollution in their vicinity (be it traffic or industrial facility), as suburb schools, schools in broader urban area, and downtown ones, shown in Table 1 (15). As for the vicinity of busy traffic, it can be easily comprehended by consulting the GIS map of their address, together with traffic characterisation (Figure 1). The criterion for the choice of a classroom was its orientation, either towards the street or to the school yard.
GIS map of 10 schools’ geographic distribution on the Belgrade City mapFigure 1
Characteristics of sampling sitesSchool name School code Spatial characteristics of the surroundings GIS coordinates “Aca Milosavljevic” 1 Suburb Rušanj village, in the valley downhill the regional highway; 44°41’01,10” N 20°26’20,86” E “Kosta Abrašević” 2 Suburb Resnik, residential area, trees in between the street and school 44°52’20,13” N 20°27’18,7” E “Nikola Tesla” 3 Rakovica, suburban municipality, ex-industrial zone, ordinary urban traffic mode 44°42’33,95” N 20°27’06,52” E “I. G. Kovačić” 4 Broad city centre, isolated from dense traffic, residential area; 44°49’18,19” N 20°24’15,75” E “Skadarlija” 5 Downtown, high density traffic, backyard towards pedestrian zone, bus stop in front 44°44’30,27” N 20°25’44,89” E “Stevan Sremac” 6 Borča III, urbanized suburb, with no heavy traffic 44°48’17,27” N 20°29’01,08” E “Drinka Pavlović” 7 Downtown, close to two busy streets, 1 tunnel 44°48’16,20” N 20°27’43,91” E “Petar Petrović Njegoš” 8 Downtown, high traffic density 44°48’21,72” N 20°27’21,81” E “Radojka Lakić” 9 Squeezed between two streets with high density traffic (lots of heavy traffic), downtown, next to the Central Rail Station 44°49’06,12” N 20°27’49,41” E “Ivan Gundulić” 10 New Belgrade, broader urban zone, frequent traffic 44°48’48,32” N 20°27’54,66” E
For both indoor and ambient air PM10 sampling, a portable HAZ-DUST EPAM-5000 particulate monitor (Ambient Portable Direct Reading Aerosol Monitor for Measuring Lung Damaging Airborne Particles) was used. It uses light scattering to measure particle concentration and provide real-time determinations and data recordings of airborne particle concentration in mg/m3. By the model specification, its accuracy is ±10% to filter gravimetric SAE fine test dust; sensing range: .001-20.0 mg/m3 or optional .01-200.0 mg/m3 or 0.1-2000.0 mg/m3; particle size range is .1-100 μm, and precision ±.003 mg/m3 (3 μg/m3) (16). For indoor air monitoring, it was positioned in classrooms, at 1.5 m height, away from the walls, to prevent influence of chipping. The monitor was positioned outside of the school building, in front of the indicated classroom, for ambient air PM10 sampling. Monitoring lasted for one whole working week, in both school shifts, while the children were present indoors, only. Monitoring lasted for 4 days during heating season in February 2008, simultaneously with the procedures undertaken in Albania, Bosnia and Herzegovina, and Slovakia. During each of the 4 measuring days, authors would fill in the checklist of the questionnaire for the classrooms, with details on the presence of pupils, and conditions concerning natural ventilation through the windows. Measuring instruments were looked after by the teachers (indoor) and authors who were mostly present in the school to check upon the equipment. The following measurements took place in the chosen classrooms: Combination of diffuse sampling during a 4-day exposure period for formaldehyde (HCHO), nitrogen dioxide (NO2), BTX, and continuous 24h measuring for carbon monoxide (CO), carbon dioxide (CO2) and PM10, during school hours. Air temperature and relative humidity were measured as well. Parallel to these IAQ monitoring activities, outdoor air quality was followed for the same specific pollutants, close to school building (selected classroom). Besides air sampling procedures, the study protocol included two standardized questionnaires, namely: for school characteristics (filled in by the school administrator); for classroom characteristics (filled in by the teacher holding classes in it).
Simple descriptive statistics, such as mean ± standard deviation, was used for continuous variables, IAQ and OAQ PM10, the number and % of IAQ interval distributions, by schools and schools’ position, while numbers (percentages) were used for categorical variables. The Kolmogorov-Smirnov test was used to check if IAQ and OAQ PM10 had a normal distribution. Quantitative variables were compared using ANOVA F test, and categorical variables were compared using contingency tables and Chi-Square or Kruskal Wallis test. Chi-square test was used to compare IAQ PM10 between groups – schools or schools’ position. For correlations between variables, we used Pearson Correlation for the linear relationship between two variables, by schools. A P-value less than 0.05 were considered statistically significant.
Table 2 shows comparative values of IAQ PM10 and OAQ PM10 by groups of schools. Both indoor and outdoor PM10 measured values are significantly higher in suburban schools than in those located in the broader urban zone: (for PM10 IAQ: K-W test=107.86, p<0.0001; PM10 OAQ: K-W test=39.43, p<0.0001). A similar level of significance appears when correlating values measured in suburban schools, with the values in schools located in the strictly urban zone: (K-W test=93.01; p<0.0001), and for PM10 OAQ, (K-W test=27.74; p<0.0001).
IAQ and OAQ PM10 concentration related to school’s geographic positions (μg/m3).Type of schools by location No. of exposed children Mean SD 95% Cor Mean Median/Range Min Max lower upper PM10 IAQ Suburb schools 244 109.18 47.66 103.17 115.19 96/164 33 197 Schools in broader urban area 220 66.08 37.36 61.12 71.05 53/126 32 158 Downtown 271 71.09 26.90 67.88 74.31 70/79 32 111 Total 735 82.24 42.43 79.17 85.31 70/165 32 197 PM10 QAQ Suburb schools 244 153.90 130.39 137.46 170.34 116/515 34 549 Schools in broader urban area 220 83.46 55.77 76.05 90.87 55/168 22 190 Downtown 271 77.30 28.53 73.89 80.72 80/89 30 119 Total 735 104.57 89.85 98.07 111.08 82/527 22 549
Indoor PM10 concentrations are significantly lower in schools located within a broad urban zone, when correlated to ones in a strictly urban zone, i.e.downtown (K-W test=12.943, p<0.0001). On the other hand, it does not count in the case of outdoor PM10 values (K-W test=2.228, p=0.135). PM10 IAQ measured values are significantly highest in suburban schools, (K-W test=133.454, p<0.0001), together with PM10 OAQ, (K-W test=69.86, p<0.0001).
Among schools, a statistically significant difference is proved for the distribution of IAQ PM10 concentration (p<0.0001). School 4 has significantly higher frequency of measured values IAQ PM10 in the range lower than 50 g/m3. On the other hand, schools No. 1, 2, 3 and 8 had highest average values, and in all of them each of measured indoor concentrations was in the interval beyond 50 μg/m3 (Table 3).
Interval distribution IAQ PM10 Concentration >50 or ≤49.9 μg/m3 by schoolsPM10 ranges School 1 School 2 School 3 School 4 School 5 School 6 School 7 School 8 School 9 School 10 ≤49.9 μg/m3 0.0 0.0 0.0 67.2 39.7 23.4 32.1 0.0 57.8 37.8 >50 pg/m3 100.0 100.0 100.0 32.8 60.3 76.6 67% 42.2 62.2
The following Table 4 presents indoor PM10 mean concentrations (μg/m3), standard deviation, median and range. The maximum concentration values of PM10 (162.12±41.93 μg/m3) was registered in school No. 2, while a significantly lower concentration value of PM10 (44.72±12.48 μg/m3) was at school 4 (p<0.001), and its value is below 50 μg/m3.
IAQ PM10 (μg/m3) concentration by schools.School Groups School PM10 (μg/m3) IAQ Ci 95% Median Range Mean SD Lower Upper Suburban „A.Milosavljevic“ 109.16 18.440 105.48 112.84 105.00 46.00 „ 41.926 151.97 172.27 183.00 106.00 „S. Sremac“ 62.47 23.592 57.11 67.82 52.00 63.00 Broad urban area „Nikola Tesla“ 99.70 45.620 89.35 110.06 72.00 107.00 „ 12.479 41.53 47.92 38.00 30.00 „I. Gundulic“ 50.40 6.224 49.03 51.77 53.00 16.00 Downtown „Skadarlija“ 58.24 18.276 54.12 62.36 70.00 46.00 „D. Pavlovic“ 65.68 26.899 58.26 73.09 51.00 63.00 „P.P. Njegos“ 103.87 8.523 101.92 105.82 110.00 18.00 „R. Lakic“ 52.31 11.222 49.51 55.12 44.00 26.00
Ambient air mean concentration of PM10 (μg/m3), together with standard deviation, median and range by groups and schools, is given in Table 5. The maximum concentration values of PM10 in ambient air (320.82±137.79 μg/m3) were in front of a suburban school No.2, while significantly lower concentration values of PM10 OAQ (38.33±9.65 μg/m3) were outside school 6, located in the broad urban area, slightly hidden away from frequent traffic flow (p<0.01), with a value below 50 μg/m3. The highest average values of both IAQ PM10 and OAQ PM10 concentration are accounted to school 2 (suburb schools).
Outdoor PM10 (μg/m3) concentration by schools.School Groups School PM10 (μg/m3) OAQ Ci 95% Median Range Mean SD Lower Upper Suburban „A.Milosavljevic“ 105.84 26.945 100.46 111.21 106.00 78.00 „ 137.797 287.47 354.18 309.00 380.00 „S. Sremac“ 68.27 36.277 60.04 76.51 41.00 85.00 Broad urban area „Nikola Tesla“ 134.00 55.399 121.43 146.57 141.00 135.00 „I. G. Kovacic“ 80.33 35.626 71.20 89.45 75.00 84.00 „ 9.648 36.21 40.45 42.00 26.00 Downtown „Skadarlija“ 83.64 30.003 76.88 90.41 96.00 87.00 „D. Pavlovic“ 69.85 31.935 61.05 78.65 52.00 83.00 „P.P. Njegos“ 98.00 13.862 94.83 101.17 101.00 46.00 „R. Lakic“ 51.17 3.444 50.31 52.03 51.00 9.00
Descriptive statistical analysis of the data for particulate matter mass concentrations (PM10) measured outdoors and in the classrooms, by schools is given in Table 6. None of the 10 schools satisfies the World Health Organization (WHO) standard for PM10 annual average of 20 μg/m3 (17, 18). However, they meet the National Ambient Air Quality Standards (12) and WHO standards for PM10 24-hour average, which have been set at 150 μg/m3 and 50 μg/m3, respectively (19).
Correlations between IAQ and OAQ by school (signif.)School PM10 outdoor median (μg/m3) PM10 indoor median (μg/m3) R P N „A.Mlosavljevic“ 106.00 105.00 0.799 0.000 99 „K.Abrasevic” 309.00 183.00 0.756 0.000 68 „N.TesIa” 141.00 72.00 0.457 0.000 77 „I.G.Kovacic” 75.00 38.00 0.956 0.000 61 „Skadarlja” 96.00 70.00 0.598 0.000 78 „S.Sremac” 41.00 52.00 0.744 0.000 77 „D.PavIovic” 52.00 51.00 0.937 0.000 53 „P.P.Njegos” 101.00 110.00 0.725 0.000 76 „R.Lakic” 51.00 44.00 0.160 0.207 65 „I.Gundulc” 42.00 53.00 0.453 0.000 82
Table 7 presents the distribution of the occupancy rate (according to ASHRAE) for each school, in m2 per present child in the indicated classroom. Values are given as the percentage of children exposed to overcrowdedness (<2m2/per child), or to convenient spatial conditions (>2m2/per child). Indicators of crowdedness are the number of children in the classroom (less/more than 20), and space available in the classroom, per one child, of less/more than 2m2 (4). Statistically significance is proven for the distribution of occupancy rate (m2/per child), for each school, χ2=340.70, p<0.0001.
Child occupancy rate per school distribution (m2/per child) vs. PM10 values (μg/m3), by schools.School No. School name <2 m2/child >2 m2/child IAQ PM10 mean 1 „A.Mlosavljevic“ 778 22.2 109.16 2 „K.Abrasevic” / 100 162.12 3 „N.TesIa” / 100 99.70 4 „I.G.Kovacic” 16.4 83.6 44.72 5 „Skadarlja” 38.6 61.4 58.24 „? „S.Sremac” / 100 62.47 7 „D.PavIovic” / 100 65.68 8 „P.P.Njegos” 47.4 52.6 103.87 9 „R.Lakic” / 100 52.31 „0 „I.Gundulc” / 100 50.40
As an outcome of our study, when differentiating between OAQ PM10 and IAQ PM10 concentration values in relation to some classroom and school characteristics, we singled out the following moments: the highest average values of IAQ PM10 and OAQ PM10 concentration were measured in the school No. 2 (suburb school), while only in one school measured values IAQ PM10 were below 50 μg/m3, that is in the school No. 4. The school No. 10, located in New Belgrade, in a broader urban zone with frequent traffic, had PM10 outdoor mean value below 50 μg/m3, which could be explained with the terrain’s topography. New Belgrade is, in geographical means, a flat terrain, with broad boulevards and widely spread buildings, belonging to the geographical entity of the Pannonian plain, enabling the build-up of high ambient air concentrations of traffic-related pollutants. This particular school is located in a residential block, built in the 1960s. School building is encircled by greenery and residential buildings, acting as a physical barrier to three streets with very busy traffic, of which one is the E-75 highway. It was, moreover, reported by the City of Belgrade Institute of Public Health, that in 2008, a series of days in a row were characterized with SE (south-east) wind (‘Košava’), with episodes of wind speed reaching 12m/s, which caused a decrease of concentration of all ambient air pollutants measured by this institution (20).
A statistically significant correlation exists between PM10 indoor and outdoor concentration for each school (p<0.0001), except for the school No. 9, ‘Radojka Lakic’ (p=0.207), although it is located in a strictly urban zone, close to the juncture of two streets with very busy traffic. The increase of outdoor PM10 concentration is significantly correlated to the increase of indoor PM10 values (except for the school No. 9).
Considering this school, located close to the heavy traffic and Central rail station, with no statistical significance between indoor and outdoor measurement, we can conclude that, in this case, indoor concentration could be influenced by activities and movements of occupants, allowing re-suspension of previously deposited particles or their delayed deposition or settling. Fromme et al. found an average indoor particulate matter concentration higher than corresponding outdoor level in a German primary school (21). Similar was confirmed in a Belgian survey (22). Oeder et al. detected indoor PM10 concentration even 5-fold higher than outdoor ones in six schools in Munich (23). In addition, the presence of children, together with their movements, could affect indoor PM levels through the interception of personal clouds (primarily comprising of course particles), recorded by sampling devices (1).
In the school No. 1, which is both in the suburb and has all its measured values of IAQ PM10 beyond 50 μg/m3, with an average PM10 IAQ of 109.16±18.44 μg/m3, a significantly higher number of children is exposed to classroom indoor environment in a space with less than 2 m2 per child. Concerning the fact that there is no busy traffic close to the school, high occupancy rate, together with bad ventilation habits and cleaning practice could be the reason for such results. This may be due to large number of students in relation to the room area and volume, whose movements cause re-suspension of settled particles (24).
In the school No. 8, 47.4% of its pupils have less than 2 m2 per child, while all measured IAQ PM10 were above 50 μg/m3, in which case the high occupancy rate could be accounted for one of the reasons for it. Some authors have determined that re-suspension is a significant factor affecting indoor particle concentration with suspension rates increasing with the particle size (25).
The school No. 4 is worth mentioning, with only its 16.4% pupils residing in the space with <2 m2 per child, located in a broad urban zone of the city. IAQ PM10 values were below 50 μg/m3, with the average PM10 IAQ concentration being the lowest compared to other schools, 44.72±12.48 μg/m3.
On the contrary, the school No. 3 has 100% PM10 IAQ measurements in space with >2 m2 per child, located in a broader urban zone, with all measured values of IAQ PM10 beyond 50 μg/m3. The average indoor PM10 concentration in it is 99.70±45.62 μg/m3. This school is cleaned mostly with the broom, and combined with the use of chemicals. This cleaning practice, being predominant as a cleaning pattern in this school, obviously facilitates the process of particle re-suspension (24).
As one of the defined elements of the thermal comfort zone, occupancy rate was chosen to be one of the key indicators in this study, together with CO2 indoor concentration, relative humidity, and classroom air temperature. Statistically significant smaller chances exist for formaldehyde indoor concentrations to be below 1.01 mg/m3 in classrooms with more than 2 m2 of space per child. For indoor CO2 concentration to increase above 1000 ppm, the number of children in the classroom above 20 (N>20) is a statistically significant predictor. Namely, chances for CO2 concentrations to be higher than 1000 ppm are 3.6-fold bigger in classrooms hosting more than 20 children. For indoor (classroom) air temperature within a comfort zone, a statistically significant predictor is classroom crowdedness, i.e. space in square meters per child. In fact, chances for indoor temperature to be within the comfort zone are less likely to occur in classrooms with less than 2 m2 of space per child (4). For indoor values of relative air humidity, a statistically significant predictor is classroom crowdedness, i.e. space in square meters per child. In other words,, chances for this parameter to be within the comfort zone are close to 18-fold higher in classrooms hosting less than 20 children. Chances for PM10 to be below 50μg/m3 are smaller, with high statistical significance, in classrooms hosting more than 20 children, having a blackboard, with chalk to write on.
The average indoor PM10 concentrations are lower in the classrooms where indicators of thermal comfort zone are satisfactory. In the school No. 4, the average indoor PM10 concentration is lower in classrooms with achieved standards for indoor comfort zone indicators, with high statistical significance Z -test=6.540, p<0.0001, while in the schools No. 5 (Z=0.105, p=0.916), No. 8 (Z=1.614, p=0.107) and No. 10 (Z=0.948, p=0.343) it is lower, but with no such high significance. Significantly highest mean values of IAQ PM10 concentration was measured in the schools 1, 2, 3 and 8, where comfort zone was not achieved (p<0.001).
Values of classroom indoor PM10 concentrations, from measurements in similar studies in different countries, during the heating season, are close to the values reported in this study in Serbia: in three schools in Portugal, PM10 average concentrations ranged from 30 to 146 μg/m3; in a German study implemented in 64 urban and rural schools, it was 105 μg/m3 (16.3-313.2 μg/m3), while in HESE study, IAQ was monitored in 21 schools, both urban and rural, with the average PM10 concentration 112 μg/m3 (91-133 μg/m3) (1, 10, 26).
In most of the SEARCH1 countries, ambient PM10 concentrations were significantly increased in school zones close to frequent traffic streets, compared to those located further from such sources of air pollution. On the other hand, this difference in the traffic frequency of streets surrounding schools, has not significantly influence IAQ PM10 concentrations measured in classrooms, pointing that key sources of this pollutant are mainly in classrooms themselves (15). In the case of Belgrade study, as a part of the mentioned research, 36.2% of pupils study in classrooms with I/O PM10 ratio beyond 1.0, where the key source of PM10 is within the room; 3.7% is in classrooms with the ratio equal to 1.0., while 60.1% attend classes in rooms with the ratio below 1.0, where the source of a particulate matter is in ambient air, mostly from traffic in the vicinity of school buildings.
Although the city is located close to two coal burning power plants (Obrenovac, Kolubara), traffic is seen as the most powerful source of air pollution. Air back trajectories analysis showed that the prevalence of stagnant or week flow regimes (calm conditions) favours the suspension and accumulation of particles produced locally, resulting at the elevation of suspended particles levels (20, 27).
If we compare the location of Belgrade, on the banks of Sava River, with another region on its banks in Zasavje Region in Slovenia, with PM10 ambient air concentration, we spot some differences. Firstly, it is an issue of topography, as Zasavje is surrounded by steep hilly terrain, while Belgrade spreads in parts of the Pannonian Plain. Secondly, sources of air pollution are different in these two cases. In Zasavje, it is the case of industrial PM10 emissions (36.3μg/m3), while in SEARCH1 Belgrade study, traffic was attributed as the key source for high PM10 levels (104.7 μg/m3) (15, 28).
The majority of surveyed children is exposed to high indoor PM10 concentrations (560/735; 76.2%). Maximum PM10 values were measured in suburban schools, away from busy traffic. The increase of outdoor PM10 concentration significantly affects the increase of indoor PM10 values.
Concerning an insufficient achievement of standards for indicators of indoor thermal comfort zone, dominant factors for the increase of PM10 are: high occupancy rate in the classrooms (<2 m2 of space per child), high relative humidity (>75%) and indoor temperature beyond 23°C; bad ventilation habits (keeping windows shut most of the time).
As the authors suggest, measures for the improvement of conditions in classrooms are as follows: schools should be built in places not directly affected by heavy traffic or industry, or any other polluting establishments in the neighbourhood; crowdedness should be avoided in the classrooms; appropriate ventilation regime of the classrooms should be introduced in order to provide good indoor air quality during the whole period of teaching hours, especially in classrooms directed towards crowded streets with presumably high ambient air pollution; in schools being ventilated only through natural means, but located close to traffic-induced air pollution, installing air-conditioning units should be taken into consideration. Cleaning practices should be standardized.