Air Quality in Brno City Parks

: Parks embody an important element of urban infrastructure and a basic type of public space that shapes the overall character of a city. They form a counterweight to built-up areas and public spaces with paved surfaces. In this context, parks compensate for the lack of natural, open landscapes in cities and thus have a fundamental impact on the quality of life of their inhabitants. For this reason, it is important to consider the quality of the environment in urban parks, air quality in particular. Concentrations of gaseous pollutants, namely, nitric oxide (NO), nitrogen dioxide (NO 2 ), and ozone (O 3 ), were measured in parks of Brno, the second-largest city in the Czech Republic. Relevant concentration values of PM 10 solids were determined continuously via the nephelometric method, followed by gravimetric method-based validation. The results obtained through the measurement of wind direction, wind speed, temperature, and relative humidity were used to identify potential sources of air pollution in parks. The “openair” and “openairmaps” packages from the OpenSource software R v. 3.6.2 were employed to analyze the effect of meteorological conditions on air pollution. Local polar concentration maps found use in localizing the most serious sources of air pollution within urban parks. The outcomes of the analyses show that the prevailing amount of the pollution determined at the measuring point most likely originates from the crossroads near the sampled localities. At the monitored spots, the maximum concentrations of pollutants are reached especially during the morning rush hour. The detailed time and spatial course of air pollution in the urban parks were indicated in the respective concentration maps capturing individual pollutants. Significantly increased concentrations of nitrogen oxides were established in a locality situated near a busy road (with the traffic intensity of 33,000 vehicles/d); this scenario generally applied to colder weather. The highest PM 10 concentrations were measured at the same location and at an average temperature that proved to be the lowest within the entire set of measurements. In the main city park, unlike other localities, higher concentrations of PM 10 were measured in warmer weather; such an effect was probably caused by the park being used to host barbecue parties.


Introduction
Urban green spaces, namely, city parks, are very often considered localities providing the best air quality in a city, and thus they become frequently targeted by citizens seeking relaxation and active recreation. However, there are very few studies supporting this generally accepted claim.
Xing et al. [1][2][3] noticed improved air quality in small urban parks within a distance from surrounding streets due to the dispersion of air pollutants within park areas. Importantly in this context, trees can reduce wind speed and potentially trap pollutants. Most available studies point to stress the choice and variability of greenery. Hewitt et al. [17] discussed several options of how to improve air quality by using different types of green infrastructure, introducing a novel conceptual framework as policy guidance; the authors' interpretation of the problem includes a flow chart to aid decision-making as regards the "green infrastructure to improve urban air quality".
Air pollution poses a major risk to human health, causing premature deaths and potentially reducing the quality of life. Quantifying the role of vegetation in curbing air pollution concentrations is an important step. Most current methods to calculate pollution cutback procedures are static and thus represent neither atmospheric transport of pollutants nor pollutants and meteorology interaction. The focus on urban parks as a tool to facilitate air purification and climate regulation embodies the basis of articles by Vieira et al. [18] and Mexia et al. [19]. These authors concluded that ecosystem service strongly depends on the vegetation type; thus, for example, air purification is more pronounced in mixed forest, and carbon reduction is influenced by tree density. Further, Jones et al. [20] developed a method to calculate health benefits directly from changes in pollutant (including PM2.5, NO2, SO2, and O3) concentrations, exploiting an atmospheric chemistry transport model.
In our paper, the concentrations of PM10 solids were determined continuously, by utilizing the nephelometric method followed by gravimetric method-based validation. To identify potential sources of air pollution in parks, we evaluated the air quality within the local environment via correlation with measurements of wind direction, wind speed, temperature, and relative humidity. The "openair" and "openairmaps" packages from the OpenSource software R were employed to analyze the effects of wind on air pollution. Local polar concentration maps found application in locating the directions of wind coming from the most serious air pollution sources. Sampling and analyses were performed to confirm the assumption that the main sources of the pollution at the measuring point are most likely the roads and/or crossroads near the sampled localities.
Due to the information gap concerning air quality in city parks, the goal of our study was to obtain data on air pollution in urban parks and associated details relevant to the relationship between this pollution and meteorological parameters, prominently including temperature, wind speed, and wind direction; in this context, our efforts also involved comparing these data with pollution around the parks. Based on the findings, we then aimed to estimate the sources of air pollution in the monitored parks.

Sampling
The sampling was carried out in three pre-selected city parks in Brno, the Czech Republic; two of the parks are located in areas with a high traffic impact (near main roads), while one is found in a low traffic load environment (a small park inside a courtyard). The main city park of Lužánky exhibits the largest surface area of all the monitored parks, and it is located near the city center, surrounded by roads with heavy traffic. Two air quality monitoring spots were positioned in the park: one place in the middle of the area, and the other on the edge of the park, near a playground and the trafficladen roads. This park is frequently visited and used for sports and leisure activities, including picnics.
The Koliště park is adjacent to a road with heavy traffic (33,000 cars/d). It occupies a large walking-friendly area, and there is a very popular restaurant in the middle of the park. However, due to the traffic-laden road, the location is not a popular target for sports, children's activities, or picnics. The air quality was measured near a junction of two main roads.
Tyršův sad is a very small park in the city center, situated inside a courtyard. This park is mostly used only for short walks, especially with dogs. The air quality measurement was performed in the middle of the area.
The devices were installed together at the place of interest.

Instrumentation
The NO, NO2, O3, and PM10 concentrations were determined by using two Airpointer units (Recordum Messtechnik GmbH, Austria). These devices measure pollutant concentrations via separate modules utilizing type-approved reference methods (NO2/NOX, O3) classified as relevant by the EU, WHO, US-EPA, and other competent responsible organizations worldwide.
The measurement principle to define the levels of NO2/NOx is chemiluminescence (EN14211). The Airpointer NOX module was equipped with a delay loop to measure NO and NO2 from the same sample. An external calibration gas with a concentration of 425 ppb NO in N2 (SIAD, Italy) was employed to periodically check the span point.
The O3 measurement principally exploits UV absorption (EN 14625); for the given purpose, an internal ozone generator to allow regular span point checking was applied.
The parameters are calibrated annually by the Slovak Hydrometeorological Institute. The Airpointer PM10 module utilizes nephelometry for measuring solid particles' concentrations. Gravimetric measurements of PM10 concentrations executed within 24 h intervals were carried out to calibrate the nephelometric method. Sequential samplers SVEN LECKEL SEQ 47/50-CD (Sven Leckel Ingenieurbüro GmbH, Germany) were employed for the calibration. The particles were collected on cellulose nitrate filters with the porosity of 1.2 µm (Merck, Germany) and weighed on a Mettler Toledo MX/A microbalance.
The meteorological parameters (air temperature, relative humidity, air pressure, wind speed, and wind direction) were measured by using a compact meteorological station integrated with the Airpointer. These parameters are regularly calibrated by the Czech Metrology Institute.
The data from the Airpointer were downloaded as CSV files and saved in the form of Microsoft Excel files (XLSX). The concentrations measured in ppb were converted to concentrations in µg m −3 . The medians, upper and lower quartiles, and other percentiles for the monitored pollutants, temperature, relative humidity, and wind speed were calculated in MS Excel. The results were then processed by the Origin program (OriginLab, USA) to yield graphs. The dependencies of and relationships between the pollutant concentrations on the wind speed and direction were processed via the "openair" and "openairmaps" packages of OpenSource program R [21,22]. The package "openairmaps" supports "openair" for plotting on various maps. The maps include those available via the "ggmap" package, e.g., Google Maps, and leaflet ones to facilitate plotting bivariate polar plots. Our research utilized the "Esri.WorldImagery" map source and the "Non-parametric Wind Regression" (NWR) technique to display the concentration maps as bivariate polar plots.

Measurement Conditions and Positioning of Instruments
The concentrations of PM10 and also those of the gaseous pollutants NO, NO2, and O3 in three parks within the city of Brno, the Czech Republic, were measured in one-minute intervals. The same scenario was applied to the meteorological conditions, namely, air temperature (T), relative humidity (RH), air pressure (p), wind speed (WS), and wind direction (WD). The NO2 and PM10 measurements at automated air pollution monitoring stations operated by the Czech Hydrometeorological Institute were employed for comparing the measurement results with those acquired at a heavy traffic locality (Údolní, the Hot Spot) and background localities (Arboretum-the natural city background station, and Dětská nemocnice-the commercial city background station). Tables 1 and 2 show the geographic coordinates of the localities and display the time intervals of the measurement.  3 Site alongside an adjacent roadway.

PM10 Calibration
As nephelometric measurements are performed in one-minute intervals, the conversion factor was calculated for each 24 h measurement interval according to the formula where is the gravimetric PM10 concentration over 24 h (µg/m 3 ), and is the average nephelometric PM10 concentration over 24 h (µg/m 3 ) The calculated emission factor is discontinuous, and was thus smoothed by the function where Factor is the smoothed conversion factor in time t fi is the conversion factor for the ith day The PM10 concentration was calculated for every minute by the function An example of the factors' calculation for the site Tyršův sad is shown in Figure 5.

Results and Discussion
Each measurement at a park is represented by a dataset with 20,160 observations, and the measurement at the Koliště road locality is represented by a dataset with 25,920 observations. Therefore, the results were summarized as percentiles and mean values to be calculated in MS Excel. Table 3 shows the intervals in which 90% of the measured values are considered for each parameter. The results of the measurements at the automatic air pollution monitoring stations were used to compare the air pollution concentrations in the parks and their vicinity. The hourly averages of the NO2 and PM10 concentrations were compared, as the data are measured in hourly intervals. The results are shown in Table 4.   Figure 6 compares the individual mean values (medians) of the measured pollutant concentrations and meteorological parameters. The dispersions of these values are represented through the upper and lower quartiles, an interpretation that is more plausible than that rendered via the mean and standard deviations because the data have an asymmetric statistical distribution. This is also clearly seen in Figure 6: the vertical lines, whose length represents the size of the first and the third quartiles, are not identically long.
The highest NO concentrations were measured in the immediate vicinity of the road adjacent to the Koliště park and then directly in the park; in both cases, the measurement was performed during a cold season (January, November). Similarly, the highest NO2 concentrations were determined next to the road adjacent to the Koliště park and directly in the park (but also in Tyršův sad); in all of the cases, the measurement was carried out during a cold season (January, November, February). The highest total concentrations of nitrogen oxides (NOx) were acquired, as in the NO, in the immediate vicinity of the road adjacent to the Koliště park and directly in the park, during a cold season (January, November). The highest O3 concentrations were measured in springtime, the lowest one in winter. The solid particles detected at Lužánky SVC and Lužánky SS exhibited a higher concentration in August than in the colder months (March, September), which is not a normal effect. This deviation arises from the fact that, in these localities, people often gather for barbecue parties and use the parks' public cooking facilities during the summer months, whereas the other parks are not frequented for this purpose. Figure 6f,g shows also the differences between the speeds and variations between the temperatures at the sampling sites, respectively. The March, February, and January temperatures reached significantly lower than the September, August, and June ones. Figure 7a compares graphically the NO2 air pollution in the parks, with the pollution measured at the reference stations, while Figure 7b displays, in the same manner, the air pollution caused by PM10. The individual measurement campaigns are separated by the red lines. As can be seen, the air pollution in the parks, with the exception of the Koliště park for PM10, was lower than that at the traffic locality, and the pollution at the background localities approached the value. The exception concerning the Koliště park was probably due to the fact that this area is relatively narrow compared to Lužánky; thus, in wintertime, when the vegetation is leafless, it provides less from the dust generated on the nearby busy roads. Moreover, it is obvious from the representation that the parks ensure better air protection from nitrogen oxides than against dust. The average concentrations of the measured pollutants were compared with the legal limits [23,24], Table 5. The excess values are marked in pink. It is possible to claim that in most localities the NOx limit for ecosystems and vegetation The analysis of the relationship between the individual pollutants' concentrations, wind speed, and wind direction was utilized to identify the places from which the highest pollutant concentrations reached the sampling site. The concentration scale of the measured pollutants is shown in Figure 12.

Minimum concentration
Maximum concentration Figure 8. The concentration scale for the "openmaps" graphs.
The color scale shown in Figure 8 expresses concentrations depending on wind direction (angle coordinate) and wind speed (radius coordinate). Figures 9-18 introduce the concentration polar maps of the measured pollutants at all of the localities.   (Figure 9d). The lowest PM10 concentration was obtained in north and south winds, meaning that transport embodies the most likely source of the nitrogen oxides; there is a larger amount of PM10 sources; and, probably, the activities pursued within the area contribute to the dust circulation in the park (Figure 9c).  Figure 10a,b shows that the highest concentrations of nitrogen oxides were measured with east and west winds blowing from the adjacent road and the opposite side. The impact of traffic on the road west of the park, which had not manifested itself in January, probably shows here. From the south through the east to the northwest, the lowest ozone concentrations were measured (Figure 10d). The highest PM10 concentrations were acquired in calm weather.   Figure 12a indicates that the highest NO concentrations were measured with a east wind. The highest NO2 concentrations were determined in eastern wind directions, namely, from the south to the north, similarly to PM10 (Figure 12b,c). At low wind speeds, we acquired the lowest O3 concentrations of the (Figure 12d), meaning that both the NO2 and the PM10 had probably been generated by similar sources. The NO had most likely originated from the traffic on the road east of the park.  Figure 13a indicates that the highest NO concentrations were measured with a north wind, similarly to the situation in Figure 14. The highest NO2 concentrations were acquired under eastern wind directions, namely, from the south to the north, similarly to PM10 (Figure 13b,c). This scenario resembles that represented in Figure 16. In eastern wind directions, we measured the lowest O3 concentrations (Figure 13d). The NO had probably originated from the traffic on the road north of the park.  Figure 14b shows that the highest NO2 concentrations were measured under northern wind directions (Figure 14b). In western to northern wind directions, we established the lowest concentrations of O3 (Figure 14d). The nitrogen oxides had probably originated from the traffic on the road north of the park. The PM10 concentrations did not exhibit any significant relationship to the wind direction in this case. Figure 15a-c indicates that the highest NO, NO2, and PM10 concentrations were measured under a northeastern wind direction. In the same wind directions, we acquired the lowest concentrations of O3 (Figure 15d). Both the nitrogen oxides and the PM10 had probably been generated by the traffic on the crossroads to the northeast of the park.  Figure 16 displays a situation similar to that shown in Figure 15. It clearly follows from the images in both of the figures that, at the Lužánky SS locality, the traffic pollution (NO) is contained by the Svojsík srub building. At the Lužánky SVC site (Figures 12-14), conversely, the NO source is blocked by the Leisure Center from the west. There are no significant transport-based air pollution sources near Tyršův sad; the air pollution at this location can be rather generated by long-distance transfer or, especially in wintertime, PM10 from local heating. Figure 17 shows the pollution from eastern directions, and Figure 18 displays the ambiguous situation at the site. As outlined above, the problem of reducing PM concentrations in urban parks has been discussed in diverse papers, e.g., [4][5][6]. Other articles analyzed the impact of urban greenery on NOx, NO2 [14], and O3 [16]. In this study, the outcomes presented within the referenced research reports are followed and developed through such procedural approaches as monitoring the influence of wind and air temperature on pollutant concentrations. The measurements have shown that, in addition to vegetation, seasonal changes of meteorological conditions and human activities in parks embody a substantial aspect modifying the local situation, as observed at Lužánky park in August 2019. The obtained results have confirmed the conclusions proposed by Kumar et al. [12], namely, that progressive steps need to be taken to bring further knowledge in the field. The relationships between O3, NO, and NO2 were studied by Han et al. [25]; interestingly, the outcomes of our research resemble Han et al.'s findings in suggesting that, as regards the study area(s), the daily NO cycle initiated by flue gas emissions from motor vehicles and continued by the related conversion of the pollutant into NO2, had a major impact on the regular ozone process. The daily course of concentrations in these pollutants was similar, too.

Conclusions
In four 14-day campaigns, concentrations of NO, NO2, PM10, and O3 were measured at five diverse locations, of which four were enclosed within Brno parks and one set at a road adjacent to a park. Compared to the average values, significantly higher nitrogen oxide concentrations were determined at the monitored spots of Koliště and Koliště-road in colder weather. Both of the locations are situated near a busy road exhibiting a traffic intensity of 33,000 vehicles/d. In terms of PM10, the highest concentrations were obtained at Koliště park, with an average air temperature that proved to be the lowest among the values adopted for the other measurements. At Lužánky park, the PM10 concentrations measured in warmer weather reached higher than those acquired during colder periods-an effect probably caused by the park being a popular public barbecue place. Using the "openairmaps" software package, we determined the directions pointing to the main sources of pollution at the individual spots. Based on this procedure, it was estimated that the main air pollution sources affecting the parks lie in the adjacent roads and crossroads. In some cases, however, human activities of people in the parks (barbecue) can also be regarded as important or semi-critical. By extension, we established that the overall surface layout, prominently including buildings in the park, can locally shield the impact of traffic on the air quality. Interestingly, the air quality in the parks approached that of the urban background locations, except for Koliště park, which, due to its shape and proximity to a very busy road, showed the characteristics of a regular traffic location.