The PM2.5 capture capability of 25 species in Beijing and Chongqing were tested by a chamber device. We obtained PM2.5 capture capability per unit leaf area and per tree, which can be helpful for selecting proper vegetation in urban settings. We tested the same species in Chongqing and Beijing at the same environment (same pollution level, temperature, and humidity). Surprisingly, there was a difference between the two sites. It can be seen that the APM2.5 in Beijing were larger, possibly due to the rich leaf morphology.
Leaf features of coniferous species can cause high air turbulence inside the tree crowns, leading to an increase in the interception capacity of contaminants (Bunzl et al. 1989). In this study, conifers did not show a significant advantage to capture APM2.5 compared with broadleaf species. However, PM2.5 accumulation capacity of conifers was superior to most broadleaf species for larger leaf areas per tree.
Leaf morphology appears to be a dominant factor in particle deposition (Mitchell et al. 2010). In previous studies, leaf morphology was qualitatively analyzed, while few quantified results were made (Chai et al. 2002). More detailed classifications and quantification of leaf morphology need to be further refined. In our study, groove proportions were quantified to evaluate the roughness. Meanwhile, leaf hair and stomatal density were quantified by counting them in a fixed leaf area. In addition, some studies revealed that particulate matter can get into leaves through the stomata, where fine particulates often crowded (Song et al. 2015; Lehndorff et al. 2006). Therefore, stomata size was also quantified. Our main results reveal that grooves are the main parts of a blade that capture PM2.5. A strong correlation between the PM2.5 accumulation and groove proportion proves that leaf surface roughness is a facilitator for PM2.5 capture (Fig. 6a). In addition, stomata size is an important influence factor for PM2.5 capture capability.
In this research, NaCl was used as a PM2.5 source, thus the mass of NaCl is lower than the natural PM2.5 source. The mass of captured PM2.5 per tree in our study was lower than the results in Song et al. (2015). Overall, the method used in this study is suitable for comparing the difference between species qualitatively merely.
Leaf characteristic and leaf morphology
Though all sampled species can capture PM2.5 and leaf surfaces have a considerable capacity (Wang et al. 2015), the amount differ significantly depending on leaf morphology. B. papyrifera is the most hairy for both lower and upper sides of the leaf and has the high groove proportion. The stomata size and stomatal density of B. papyrifera are not counted because leaf hair covered the stoma on leaf surface. Koelreuteria bipinnat has the highest groove proportion and stomata size, while the leaf hair and stomatal density of K. bipinnat are low. All of them do not show the highest PM2.5 capture capability. G. robusta is the most efficient broadleaf species and C. lanceolate is the most efficient conifer with high groove proportion and low stomata size. Species with high groove proportion and low stomata size is most effective at capturing PM2.5. Earlier studies also reported that mounts of PM2.5 captured on rough tree leaves with low stomatal density were high (Hwang et al. 2011; Räsänen et al. 2013).
Nguyen et al. (2015) found that trees with leaf hairs have high PM2.5 capture capability. Species with densely haired leaves were most effective at capturing PM (Dzierżanowski et al., 2011; Weber et al. 2014). However, we found no significant correlation between PM2.5 capture capability and leaf hair (Fig. 6d). This may attribute to differed methodological approaches and limited tree species in this research. A large number of tree species need to be studied in the future.
The role of stoma activity in particle deposition is ambiguous. On one hand, transpiration of water through stomata cools the surface which is conducive to attracting PM2.5; on the other hand, transpired water repels PM2.5 due to diffusiophoresis (Hinds 1999). No statistically significant correlations are found between PM2.5 capture capability and stomata size (Fig. 6b), stomatal density (Fig. 6c) in Chongqing and Beijing. For the stomata size, significant correlation with PM2.5 accumulation exists when tree species are classified into two groups according to stoma size:(1) The bigger one: PO, MS, CC, FM, KB, SS, PT, FP, which average stomata size are 111.19 μm. (2) The smaller one: CJ, BP, GB, FS, GR, LE, CL, PM, PA, EV, SJ, PLO, which average stomata size are 44.41 μm (Fig. 7). It can be seen that the correlation of the smaller one is higher than the bigger one, which can be explained by the restrain effect due to small stomata size. When the stomata size grow, the restrain effect become smaller.
Difference between coniferous and broadleaf species
Conifers shows the highest particle capture efficiency of tested tree species, which coincides with previous studies. Among the conifers, pines captured significantly more PM2.5 than cypresses (Beckett et al., 2000a, b). In this study, C. lanceolata is the most effective species in PM2.5 accumulation. Nevertheless, P. orientalis belongs to cypresses and has the least efficiency of PM2.5 accumulation, which coincides with the results of Song et al. (2015). It may be due to that pine trees deposited more PM2.5 than cypress ones.
More complex structure of the foliage of the conifers explained their greater effectiveness at capturing particles (Beckett et al., 2000a, b). However, in this study, conifers did not show a significant advantage to capture APM2.5 comparing with broadleaf species (Fig. 8a), which may attribute to that the structure of the conifer crowns were not considered. The total leaf area per tree of the conifer is higher than broadleaf species. Therefore, PM2.5 accumulation capacity per tree of conifers are superior to most of broadleaf species (Fig. 8b).
In addition, some trees are better able to survive in smoky and polluted conditions due to differences in physiological mechanisms of varied species. All in all, the best choices for pollution-control plantings are coniferous and broadleaved species with rough leaf surfaces and high adaptability (Beckett et al. 1998, 2000a, b; Silli et al. 2015).
Difference between two sites
A number of studies have demonstrated the effects of pollution on tree leaves. Deposition of PM was responsible for the change on leaf surface morphology (Gupta et al. 2015). Furthermore, it was found that the effects of PM2.5 on leaves relate to their acidity, salinity, and trace metal content properties (Grantz et al. 2003). Leaf density and thickness are altered when exposed to pollution environment and higher levels of NOX (Jochner et al. 2015). Pääkkönen et al. (1997) found that higher stomatal density and thicker leaves result in a greater tolerance to pollution. In addition, it was likely that PM2.5 might have an indirect effect via altering soil chemistry, which is also believed to be the major effect of PM on trees (Grantz et al. 2003). Trees strengthened the characteristics of their leaf structures under polluted conditions, which are regarded as adaptive and compensative to the adverse effects of air pollution (Chaturvedi et al. 2013). Studies also showed that trees develop different morphologies under polluted conditions (Karenlampi 1986; Veselkin 2004).
In addition, the chemical composition and wax structure may also be different in Beijing and Chongqing, which are significant for PM2.5 capture (Burkhardt 2010). Therefore, it is worthy of further study regarding the effect of PM2.5 pollutions on leaf morphology, including chemical composition, wax structure, groove proportion, leaf hair, stomatal density, and stomata size.