Introduction

Air pollution is a major global health problem. As the world’s largest developing country, China’s economic growth has been accompanied by a rapid and large increase in NO2 concentration1. In the past, China was committed to reducing atmospheric NO2 emissions at the national level2. For example, the NO2 emission reduction strategy formulated in the 13th Five-Year Plan (2016–2020) has markedly contributed to reducing NO2 emissions3. However, the current study shows that spatially clustered population exposure to NO2 still exists in some regions and provincial capitals of large urban agglomerations in China4. As an important tropospheric trace gas and precursor of photochemical smog, NO2 accumulation causes serious air pollution5 and has a significant ecological impact on the surrounding environment6. In addition, long-term exposure to NO2 is deleterious to human health7 or even causes death8. For example, Hu et al.9 showed that NO2 increases the mortality of individuals with cardiovascular and respiratory diseases.

The present study paid more attention to the tropospheric NO2 column than the surface. However, ground NO2 concentration is more closely associated with anthropogenic emissions, is considerably related to transportation10, and directly affects human health2.

Although the key to reducing NO2 concentrations in the air is to reduce emissions11, it is necessary to explore alternative solutions because of the difficulty of solving air pollution by completely controlling pollution sources. As an important part of urban ecosystems, plants have received widespread attention because of their ability to evaluate air pollution status and as an effective indicator of air pollution12 and reduce NO2 concentrations through multiple pathways13. Researchers tend to conduct such research in three ways. One is to select locally common road-grown plants and screen plant species that effectively alleviate NO2 concentrations through laboratory fumigation13,14. Second, ENVI-Met, FLUENT, MISKAM, OSPM, and other air pollutant diffusion models have been used to conduct numerical simulations15,16. The third is to explore the purification effect of vegetation by measuring the NO2 concentration in plant communities in the field17,18. Using these methods, we believe that measuring the change in NO2 concentration in the real environment is more conducive to exploring the purification effect of plants under various complex conditions. Interestingly, field studies have shown that the purifying effect of plants on NO2 concentrations depends on, and often contradicts, the climatic conditions at the time of research and the vegetation type and structure19,20,21. These studies mainly focused on urban green spaces, such as forests and parks22,23, and were less concerned with whether street-scale green spaces can reduce NO2 concentrations. However, this green space is in direct contact with NO2 from traffic emissions, and we feel its importance may be overlooked.

Based on the subject matter discussed above, our current study focused on common street plant species, such as street trees, shrubs, and herbs. By studying different plant communities in summer (vigorous plant growth) and winter (poor plant growth) in Nanjing, we analyzed the influence of different plant communities on the NO2 concentration and its variation in the horizontal and vertical directions under careful consideration of street meteorological factors.

Materials and methods

Overview of the study area

Nanjing is the capital of Jiangsu Province. It is located in the southwest of Jiangsu Province and the lower reaches of the Yangtze River, with latitude 31°14–32°37 N and longitude 118°22–119°14′ E. In recent years, road construction in Nanjing has developed rapidly. There are 445 roads in the urban area, including 75 main roads, 77 secondary roads, and 303 branch roads. The total length of roads in the city has reached 734, and 670.10 km are green, with a greening rate of 91.29%. By 2016, the road density had reached 5.39 km/km2, and the road area ratio (road area/built-up area) had reached 19.8%. Based on the investigation of all the roads in the urban area of Nanjing, we selected five roads in four main urban areas of Nanjing (Jianye District, Xuanwu District, Gulou District, and Jiangning District) as the research sites (Fig. 1). On the basis, we conducted a 5% random sampling survey on the plants of these 5 roads and selected widely represented plant communities (with trees, shrubs, and grasslands as the main vegetation types of the study).

Figure 1
figure 1

(a) Nanjing, Jiangsu Province, China, (b) location of five roads and five sampling sites in Nanjing, (c) aerial photo of sampling point of Mengdu Street, (d) sampler set at 6 m height of sidewalk in Mengdu Street, (e) sampler set at 6 m height of green space in Mengdu Street. (The base images for (a) and (b) are from Google Maps. Version number: 11.60.0703. URL link: https://www.google.com.hk/maps/place/%E4%B8%AD%E5%9B%BD%E6%B1%9F%E8%8B%8F%E7%9C%81%E5%8D%97%E4%BA%AC%E5%B8%82/@32.0554234,118.67779,59275m/data=!3m1!1e3!4m5!3m4!1s0x35b58c9b668dcd83:0x8ffbb60b79df1b06!8m2!3d32.0583799!4d118.79647!5m1!1e4?hl=zh-CN).

Full size image

Sample site selection and plant community characteristics

We selected five sampling sites in five plant communities along five roads in Nanjing. Each site was represented by its unique plant species. As shown in Table 1, according to the different plant community types, the study divided the plant configuration patterns into five types: a control point (hard pavement square), lawn, arbor-grass, arbor-shrub, and arbor-shrub-grass. The control point CK was located in Beijing East Road, lawn A1 in Mufu South Road, arbor-grass community B1 in Shuanglong Avenue, arbor-shrub community C1 in Jiyin Avenue, and arbor-shrub community D1 in Mengdu Street.

Table 1 Characteristics of the plant community structure in the street green spaces.
Full size table

Data collection

The data used in this study were collected from the field. The study was conducted in the winter of 2021 and the summer of 2022, when the concentration of NO2 reaches its maximum and minimum in winter and summer, respectively24,25. The data were as follows: (1) location information, including geographic coordinates, altitude, longitude, and latitude. (2) Environmental factors including temperature, humidity, wind speed, wind direction, light, radiation, air pressure, noise and traffic flow. (3) Plant information included plant species, tree height (H), diameter at breast height (DBH), crown size, canopy area (CA) and canopy density (CD). (4) The concentration of NO2. Location information, environmental factors, and pollutant concentrations were recorded using a self-developed multifunctional lifting environmental detector (Table 2, Fig. 2) on weekdays without precipitation in February 2022 and July 2022. The concentration of NO2 was considerably lower on the rest day than on the working days26. The detector was mainly composed of sensors, circuit boards and related accessories. To ensure the reliability of the measurement, we went to the national automatic air quality monitoring point for ground monitoring and data comparison before the experiment every month. Taking temperature, humidity, wind speed, light, radiation, air pressure, noise, traffic flow and NO2 concentration as a set of data, we got a total of 1080 sets of effective data. Plant information was obtained using field surveys and measurements.

Table 2 The measuring range and accuracy of multifunctional lifting environmental detector.
Full size table
Figure 2
figure 2

The analytical diagram of self-developed multifunctional lifting environmental detector.

Full size image

As shown in Fig. 3, a 20 × 20 m sampling site was set in the street green space, and three types of sampling points in the horizontal direction were set: the control point, sidewalk, and green space. At each sampling site, NO2 concentrations were recorded from 7:00 to 9:00 (morning peak), 12:00–14:00 (off-peak), and 17:00–19:00 (evening peak) on weekdays, with no precipitation and wind speeds lower than 2 m/s at heights of 0, 0.5, 1.5, 3, and 6 m above the ground.

Figure 3
figure 3

Diagram of horizontal and vertical position of multifunctional lifting environmental detector in sampling sites.

Full size image

Data analysis

The formula for calculating the percentage of pollutant purification in the green belts of different plants is as follows:

$${\text{P}}_{{\text{n}}} = ({\text{C}}_{{\text{c}}} - {\text{C}}_{0} )/{\text{C}}_{{\text{c}}} \times {1}00\%$$

where Pn is the purification percentage of various pollutants by the green belt, Cc is the pollutant concentration on the side of the motor vehicle lane close to the green belt, and C0 is the green belt far from the edge of the motor vehicle. The concentration of pollutants (control concentration). * P < 0.05 and ** P < 0.01 were considered statistically significant and highly significant, respectively.

Result

Seasonal variation characteristics of meteorological factors and NO2 concentration in different plant communities

Comparing the meteorological factors of the plant communities in the different seasons (Table 3), it was observed that temperature, radiation, light, and noise increased, whereas air pressure decreased from winter to summer. One-way analysis of variance showed that seasonal changes significantly (P < 0.05) affected temperature, air pressure, and noise. It is speculated that the limitations of weather conditions on the day are more likely to affect changes in humidity, wind speed, wind direction, radiation, and light than seasonal changes.

Table 3 Road meteorological factor changes at the different sites.
Full size table

By comparing the change in NO2 concentration (Fig. 4), it was found that the average NO2 concentration was generally higher in the summer than in the winter. The average concentration of NO2 in the five plant communities was higher than 10 μg/m3 in both summer and winter (World Health Organization average annual concentration guidelines). The average NO2 concentrations of CK, A1, B1 and D1 in summer were 98.93, 135.48, 105.78, 100.49 μg/m3, respectively. It exceeds the 24-h average level 2 concentration limit (80 μg/m3) set in the ambient air quality standard (GB 3095-2012). In addition, the concentrations of NO2 at the other sampling sites were lower than the level 2 standard concentration limit. However, by comparing the monthly data of national automatic air quality monitoring points (24-h continuous monitoring) (Fig. 5), it can be found that the average value was significantly lower than our monitoring data (6-h monitoring), and the range of values overlapped to some extent.

Figure 4
figure 4

Seasonal distribution of NO2 concentration.

Full size image
Figure 5
figure 5

(a) NO2 of national control monitoring sites in Nanjing in February 2022 (b) NO2 of national control monitoring sites in Nanjing in July 2022 (The base images for (a) and (b) are from Google Maps. Version number: 11.60.0703. URL link: https://www.google.com.hk/maps/place/%E4%B8%AD%E5%9B%BD%E6%B1%9F%E8%8B%8F%E7%9C%81%E5%8D%97%E4%BA%AC%E5%B8%82/@32.0554234,118.67779,59275m/data=!3m1!1e3!4m5!3m4!1s0x35b58c9b668dcd83:0x8ffbb60b79df1b06!8m2!3d32.0583799!4d118.79647!5m1!1e4?hl=zh-CN).

Full size image

Influence of environmental factors on NO2 concentration

The correlation analysis of meteorological factors and NO2 concentration found that temperature and humidity significantly (P < 0.01) influenced NO2 concentrations in winter and summer, among which temperature was significantly positively correlated while the humidity was significantly negatively correlated. Through analysis and comparison of the different seasons, it can be seen that environmental factors had a greater impact on NO2 concentrations in summer than in winter (Fig. 6).

Figure 6
figure 6

Correlation analysis between road meteorological factors and NO2.

Full size image

As shown in Table 4, the traffic flows in winter and summer significantly (P < 0.05) differed between the peak and off-peak periods. However, there was no significant correlation between NO2 concentration and traffic flow at the five sampling sites. This indicates that the change in NO2 concentration on the road may be jointly affected by traffic flow and other meteorological factors.

Table 4 The daily traffic flow in the 5 plant communities in Nanjing.
Full size table

In light of the above, we conducted multiple linear regression analysis of temperature, humidity, air pressure and NO2 concentration. The results showed that in winter the fitting equation was y = 1543.36 + 0.44x1 − 0.81x2 − 0.01x3 R2 = 0.09, and in summer was y = 11330.36 + 10.25x1 − 1.37x2 − 0.11x3 R2 = 0.64, where x1 was temperature, x2 was humidity, and x3 was atmospheric pressure.

Effects of different plant community structures on NO2 concentration

The average height through which humans breathe is 1.5 m. Considering the NO2 concentration at 1.5 m height, we compared and analyzed the reduction rate of NO2 at the five sampling sites. The reduction rate of NO2 in winter was D1 > C1 > A1 > CK > B1. The reduction rate of NO2 in the summer was D1 > C1 > B1 > CK > A1 (Table 5).

Table 5 Reduction rates of different plant communities.
Full size table

As shown in Fig. 7, we analyzed the changes in the NO2 reduction rate at the five sampling sites and found that the reduction rate of B1 varied greatly in winter and summer (from 10.78% in winter to 22.33% in summer), indicating that the seasonal changes in plants impacted the reduction rate. The reduction rates of CK (15.28% in winter, 16.60% in summer) and A1 (16.98% in winter, 15.07% in summer) were similar and smaller than those of C1 (21.77% in winter, 23.70% in summer) and D1 (28.24% in winter, 24.60% in summer). These results indicate that road green spaces positively affect NO2 concentration reduction.

Figure 7
figure 7

Reduction rate fluctuation of different plant communities.

Full size image

Effects of plant community spatial structure on NO2 concentration

Effects of horizontal direction on NO2 concentration in plant community structure

We compared and analyzed the NO2 concentration in the plant communities in the horizontal direction with the control point, the sidewalk, and green space sampling points and found that the NO2 concentration in summer was generally higher than that in winter, and the NO2 concentrations in CK and A1 were higher than those in other plant community structures (as shown in Fig. 8). In CK and A1, the concentration of NO2 fluctuated greatly. In B1, C1, and D1, the NO2 concentration of the sidewalks was higher than that of the control points in winter. However, in the summer, when plants were growing vigorously, the overall performance was as follows: control point < sidewalk < green space. In general, at the sampling points with green space, the NO2 concentration decreased to a certain extent with an increase in road distance (especially green space < control point), indicating that road green spaces had a positive effect on the reduction of NO2 concentration.

Figure 8
figure 8

Horizontal decline rate in different plant communities.

Full size image

Effect of vertical direction on NO2 concentration in plant community structure

We selected the vertical sampling points in the green space to analyze the vertical variation in NO2 concentration in the five plant communities and found that the vertical variation in NO2 concentration was inconsistent (as shown in Fig. 9). The NO2 concentrations in CK and A1 were higher than those in the other plant community structures. In general, NO2 concentration increased from 0 to 0.5 m and decreased from 0.5 to 1.5 m. C1 and D1 decreased from 0 to 0.5 m, indicating that shrubs (0–0.5 m) had a certain reduction effect on NO2 concentration. Overall, the NO2 concentration increased from 1.5 to 6 m, while B1 and C1 decreased from 1.5 to 6 m. Combined with the vertical structure diagram of the plant community shown in Fig. 7, it was suggested that the tree canopy promoted the reduction of NO2 concentration. However, at the branch points under the tree canopy, where there are no leaves, only the trunk will accumulate and precipitate pollutants, and thus increase the concentration.

Figure 9
figure 9

Vertical reduction rate in different plant communities.

Full size image

Discussion

Relationship between NO2 concentration and environmental factors

This study found that temperature, air pressure, and noise were significantly (P < 0.05) correlated with seasonal changes. We also found that the average concentration of NO2 was generally higher in summer than in winter. This is speculated to be due to the following: 1. As the radiation and light in summer were much higher than in winter, the photochemical reactions in summer led to an increase in the secondary formation of NO227. 2. The temperature was extremely high during the summer sampling period. Goldberg et al.26. showed a higher concentration of atmospheric NO2 at temperatures > 32 °C than at temperatures from 26 to 32 °C. In contrast, the NO2 concentration decreases with increasing temperature at moderate temperatures. However, July was relatively hot in Nanjing, with an average daily temperature of 30.87–34.59 °C at the sampling sites. 3. This may be related to human activity and NO2 emissions in the streets28. Study has shown that NO2 emissions generally peak in the summer29.

The correlation between the meteorological factors and NO2 concentration showed that temperature was mainly positively correlated with NO2 concentration (P < 0.01), whereas humidity was negatively correlated (P < 0.01). Other studies have shown that temperature and humidity are directly related to NO2 concentrations, but the correlation is not the same3,30,31. We believe the positive correlation between temperature and NO2 concentration agrees with our previous conjecture 2. When the temperature reaches a certain level, the concentration of NO2 increases with an increase in temperature.

Normally, the weather factors such as wind direction and wind speed would affect pollutants attenuation to a certain extent. But it did not show significant impact in this study. This may probably because all the sampling sites were heavily affected by windbreaks and leeward, which Yin et al.17 and Irga et al.20 had also found it.

Current study suggests that vehicles are the main source of NO2 in cities and are markedly correlated with NO2 concentrations32. However, our study found that the changes in NO2 concentrations at the five sampling sites were not markedly correlated with traffic flow. Considering the above correlation, we speculated that the changes in street NO2 concentrations might be jointly affected by traffic flow and other meteorological factors which was also proved by our fitting results of multiple linear regression equation.

Influence of different plant community structures on NO2 concentration

Desyana et al.33 found that vegetated areas had lower NO2 concentrations than that at unvegetated sites. The same finding was apparent in our study. Comparing the five sampling sites in winter and summer, we found that the reduction rate of NO2 in winter was as follows: D1 > C1 > A1 > CK > B1, and the reduction rate of NO2 in summer was D1 > C1 > B1 > CK > A1. Deciduous trees accounted for a large proportion of the vegetation at B1. We believe that after plant defoliation in winter, B1 was not open enough to facilitate air flow like A1 and CK, and could not have sufficient evergreen vegetation to purify NO2 like D1 and C134 (as shown in Table 6). In contrast, the plant community structure with trees, shrubs, and grasses can effectively reduce NO2 concentrations, which is consistent with the results of Luo35. Rao et al.36 showed that NO2 purification by Portland trees could provide $7 million per year, which is an interesting topic. In the next phase, we plan to combine the measured pollutant data with the ecological benefits of plant communities to further explore the intrinsic value of these plant communities.

Table 6 Wind speed at different monitoring heights in different plant communities.
Full size table

Internal effects of plant community structure on NO2 concentration

The spatial structure of the plant community was divided into horizontal and vertical directions for comparative analysis. We found that in the horizontal direction, compared with the control point, the NO2 concentration was higher in the sidewalks in winter (B1, C1, and D1), which may be due to the recirculation area of airflow in front of the green space37. However, in summer, when the trees are growing vigorously, NO2 concentration gradually decreases with an increasing distance from the control point, clearly showing a trend that green space < control point. This agrees with results of Fantozzi et al.38, and indicates that road green spaces positively affect reducing NO2 concentration, especially in B1 and D1.

In the vertical direction, the concentration of NO2 in sampling sites with shrubs decreases from 0 to 0.5 m. Considering our current investigation and finding that the height of shrubs is less than or equal to 0.5 m, we believe that the presence of shrubs can reduce the NO2 concentration in green areas to a certain extent39. Overall, the NO2 concentration increased from 1.5 to 6 m, while in B1 and C1 it decreased. Considering the height of the trees under the branches in the community40, we believe that the arbor canopy density could promote the reduction of NO2 concentration, but the height of branch points under the canopy would be affected by the obstruction of air flow by the canopy, thus accumulating and precipitating NO2 and increasing its concentration rise (as shown in Table 6).

In addition, studies have shown that green space could regulate the microclimate, especially the reduction in air temperature and the increase of air humidity in summer41,42. Considering the significant influence of temperature and humidity on NO2 concentration, we analyzed the temperature and humidity variation in different plant communities in summer (as shown in Fig. 10). The results showed that the presence of green space reduced temperature and increased humidity, which also partly explained the variation of NO2 concentrations in horizontal and vertical direction.

Figure 10
figure 10

Temperature and humidity variation in different plant communities in summer.

Full size image

Conclusion

In this study, we investigated the effects of five small plant communities along urban streets on NO2 concentrations in the atmosphere. The results showed that temperature, air pressure, and noise were significantly correlated with seasonal changes, and the average NO2 concentration in summer was generally higher than that in winter. The correlation between environmental factors and NO2 concentration showed that temperature and humidity were significantly (P < 0.01) correlated with NO2 concentration; however, the change in NO2 concentration in the streets may be jointly affected by traffic flow and other meteorological factors. By comparing different plant community structures and the internal space of NO2 concentration changes, we found that the plant community structure with arbors, shrubs, and grass was more effective in reducing pollution, and evergreen plants were more effective. These findings will help predict the impact of plant communities on NO2 concentrations in urban streets and help city managers and planners effectively reduce NO2 pollution in the air.

Since our study focused on winter and summer, we plan to continue the monitoring of pollutants until the length of one year in the future to observe the effects of environmental factors and vegetation changes on pollutant concentrations. In addition, we will import the data of five sites into the i-Tree Eco model to evaluate their ecological benefits and analysis the difference between model calculation of pollutant purification and actual monitoring.