Size distribution of particulate matter in runoff from different leaf surfaces during controlled rainfall processes

Environmental Pollution 255 (2019) 113234

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Size distribution of particulate matter in runoff from different leaf surfaces during controlled rainfall processes* Xiaowu Xu a, b, Xinxiao Yu a, *, Le Bao a, Ankur R. Desai b a b

Beijing Laboratory of Urban and Rural Ecological Environment, Beijing Forestry University, Beijing, China Department of Atmospheric and Oceanic Sciences, University of Wisconsin, Madison, WI, United States

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 April 2019 Received in revised form 9 September 2019 Accepted 10 September 2019 Available online 11 September 2019

The presence of plant leaves has been shown to lower the risks of health problems by reducing atmospheric particulate matter (PM). Leaf PM accumulation capacity will saturate in the absence of runoff. Rainfall is an effective way for PM to “wash off” into the soil and renew leaf PM accumulation. However, little is known about how PM wash-off varies with PM size and health problems caused by particulate pollution vary with PM size. This study thus used artificial rainfall with six plant species to find out how size-fractioned PM are washed off during rain processes. Total wash-off masses in fine, coarse and large fractions were 0.6e10.3 mg/cm2, 1.0e18.8 mg/cm2 and 4.5e60.1 mg/cm2 respectively. P. orientalis (cypress) and E. japonicus (evergreen broadleaved shrub) had the largest wash-off masses in each fraction during rainfall. P. cerasifera (deciduous broadleaved shrub) had the largest cumulative wash-off rates in each fraction. Rainfall intensity had more influence on wash-off masses and rates of large particles for six species and for small particles in evergreen species, but limited effect on wash-off proportions. Wash-off proportions decreased in large particles and increased in small particles along with rainfall. The results provide information for PM accumulation renewal of plants used for urban greening. © 2019 Elsevier Ltd. All rights reserved.

Keywords: Water-insoluble particles Rain Plants Size fractions

1. Introduction Although air quality has improved in many countries through policies and technologies that reduce harmful emissions, air pollution is still a major risk factor to human health (WHO, 2013). Airborne particulate matter (PM), known as one of the common ‘criteria pollutants’, is a heterogeneous solideliquid mixture, containing toxic substances and transported in the atmosphere, sometimes over long distances (WHO, 2003). Health problems caused by exposure to particulate pollution are related to the sizes of atmospheric particles (e.g., total suspended particles (TSP), PM10, and PM2.5) (Kim et al., 2015). Compared to large particles, small particles are not easy to achieve natural sedimentation on land surfaces (Petroff et al., 2008). Vegetated ecosystems, through leaf accumulation, can filter atmospheric PM out of the air and reduce their potential harm to human health. The role of increased urban forests in PM removal has been confirmed by some studies in cities around the world (e.g.

*
so Targino. This paper has been recommended for acceptance by Dr. Admir Cre * Corresponding author. E-mail address: [email protected] (X. Yu).

https://doi.org/10.1016/j.envpol.2019.113234 0269-7491/© 2019 Elsevier Ltd. All rights reserved.

Yin et al., 2011; Tallis et al., 2011; Nowak et al., 2013). Studies have also demonstrated that PM accumulation on various leaf surfaces differed among size fractions, providing substantial insights into selecting effective plant species according to specific environment conditions (Sæbø et al., 2012; Dzierzanowski et al., 2011; Popek et al., 2013). Since most particles are not taken up inside leaves through their stomata, plants serve mainly as an intermediary sink to store PM which could be washed off by rain, resuspended by wind or fall back to ground via leaf senescence in autumn (Beckett et al., 2000a; Ould-Dada and Baghini, 2001; Pullman, 2009). Meanwhile, rain may carry water-insoluble particles to be intercepted on leaf surfaces and wind may increase airborne PM deposition rates when windspeed is less than 10 m s
1 (Schaubroeck et al., 2014). Washing off accumulation from plant surfaces into more permeable soil by rainfall is typically considered as a net-removal of PM out of atmosphere (Beckett et al., 2000b; Schaubroeck et al., 2014). Moreover, rainfall could infiltrate water-soluble fractions of PM and take them away with raindrops, especially for small size particles where water-soluble fractions are abundant (Dao et al., 2014; Levia et al., 2011; Velali et al., 2015). This runoff is also essential to recover plant PM accumulation ability and reduce phytotoxic effects after

2

X. Xu et al. / Environmental Pollution 255 (2019) 113234

plant surfaces accumulated PM for a long period (Tomasevi
c et al., 2005; Przybysz et al., 2014; Cai et al., 2017). Rainfall also facilitated PM accumulation on leaf surfaces when PM concentration was high (Wang et al., 2015). However, runoff from leaf surfaces may cause environmental problems, such as leaching limited bioavailable contaminants into soil and underground water, disturbing native microflora, limiting plant growth and posing a risk to human health (Ali et al., 2013). Given its association with threats to human health, it is critical to enhance our understanding of how size-fractionated PM interacts with plants as well as how rainfall influences sizefractionated PM removal (Song et al., 2015). Earlier studies showed that particle accumulation relates to leaf micromorphology and wettability, which may be affected by rain erosion, but it remains qualitative in terms of size distribution (Neinhuis and Barthlott, 1998). Although some studies investigated wash-off processes of dry or wet deposition from plant surfaces by rainfall, they focused mainly on ionic chemical elements in soluble ultrafine particles (Lovett and Lindberg, 1984; Van Stan et al., 2011; Germer et al., 2007). Freer-Smith et al. (2005) observed that there was no major reduction of fine and coarse water-insoluble particles on leaf surfaces of five species after rainfall. Przybysz et al. (2014) indicated that rainfall washed off most large and coarse water-insoluble particles, but not so much for water-insoluble fine particles from pine shoots. However, Liu et al. (2018) found size distribution differed by plant species after rain events. These results highlight the importance of considering PM accumulation size distribution during rainfall processes. The analysis of deposited PM after rainfall in previous studies gives only a partial picture. Recently, Schaubroeck et al. (2014) developed a multi-layered PM removal model for forest canopies and performed a case study on a Scots pine stand to show PM2.5 mass balance within 6 h on tree surfaces. Xu et al. (2017) applied this framework into rainfall control experiments for leaf surfaces of broadleaf species, analyzing the total water-insoluble PM (0.2e100 mm) balance of leaf accumulation before rainfall, wash-off in runoff during and leaf retention after rainfall processes. Another study used a particle number concentration method to discuss PM reduction rates in terms of size fractions (i.e. PM10, PM2.5 and PM1) in green wall plants during controlled rainfall processes (Weerakkody et al., 2018a). Yet, few studies have examined sizespecific distribution of water-insoluble PM dynamics in runoff from leaf surfaces of common urban greenery during rainfall processes. Here, we seek to investigate the effect of three policy-relevant PM size fractions (0.2~3 mm, 3e10 mm, 10e100 mm) on post rainfall wash-off removal from both evergreen and deciduous trees and shrubs commonly found in urban areas. Our objectives were to: (1) quantify the dynamics of PM wash-off masses in size fractions, (2) evaluate cumulative PM wash-off rates among size fractions, and (3) characterize PM wash-off proportions of size fractions during rain processes. Our hypothesis is that large particles would tend to be washed off independent of rainfall intensity, while increasing rainfall intensity would improve removal efficiency of small particles. 2. Materials and methods 2.1. Plant materials Six plant species were selected for this study, including common trees and shrubs in urban greenery. Sophora japonica L. and Populus tomentosa Carri ere are common deciduous broadleaf street trees in Beijing (Yang et al., 2005). Pinus tabulaeformis Carri ere and Platycladus orientalis L. Franco are two typical evergreen needle leaf pine

and cypress species that are widely planted in North China. Prunus cerasifera Ehrhar f. is a deciduous broadleaf shrub, which were brought into Beijing during the Olympic Games in 2008 because of its beautiful flowers. Euonymus japonicus Thunb is an evergreen broadleaf shrub with thrifty foliage and branches and is broadly used in road greenbelts in Beijing. Leaf traits data were derived from literature (Table 1). Coarseness was estimated by atomic force microscopy (Multimode Nanoscope IIIa, Bruker, GER) and analysed by Nanonavi II (SII Nano Technology Inc., Japan). (Zhang et al., 2016; Zhao, 2018). The scanning range is 5 mm * 5 mm on coniferous leaves and on the upper surfaces of broadleaves, because most of PM accumulated on the upper surfaces (Wang et al., 2015). Stomatal densities were scanned by scanning electron microscope (FESEM, Quanta 200, FEG, USA) and analysed by Image J (Yang et al., 2015). Contact angle were measured by angle instrument (Kono SL200A, Kono Industrial Co., Ltd.) with 3 mL droplets of distilled water (Wang, 2012; Yang et al., 2015). Leaf samples for stomata density and contact angle are 5 mm  5 mm. Wax contents were extracted by trichloromethane and weighed by a balance (Wang, 2012). Our experiments were conducted in the laboratory of artificial rainfall experiments (40 040 N, 116 060 E, 145 m a.s.l.), located in Jiufeng Mountains, Beijing. All saplings of the six species were in good condition and had been growing outdoors for six months in 2015 near the laboratory. These saplings experienced several heavy natural rains among six months and exposed to the similar PM concentrations after rainfall. It has been found that leaf surfaces would reach maximum accumulation without any rain events (Liu et al., 2013; Cai et al., 2017), and thus, our saplings accumulated PM for 20 days without natural rain events prior to the controlled rainfall experiments. Saplings were about 1.5 m and 0.8 m high for trees and shrubs respectively. Before rainfall, twigs (20e30 cm length) and leaves were cut off from saplings to use in rainfall experiments and in PM measurements for initial accumulation. 2.2. Sample collections To analyse PM initial accumulation before rainfall, leaf samples without rainfall were cut from the same saplings in similar positions of twigs as for the rainfall experiments. Samples for each species were cut from different pots of saplings for replicates. Leaf samples were taken back laboratory in sealed plastic bags carefully to avoid PM loss and kept at 4  C in a refrigerator, in advance of wash procedures. Leaf samples were put in glass beakers with 250 mL of ultrapure water and washed for 60 s with an ultrasonic cleaner (KQ2200, Kun Shan Ultrasonic Instruments Co., Ltd., Jiangsu, China). Finally, we used a tweezer to pick leaves and spray ultrapure water on leaf surfaces to wash off residual PM into suspensions. Suspensions for initial accumulation were collected with sampling bottles and kept in a refrigerator. Washed leaf samples for initial accumulation were kept in a refrigerator with plastic bags after they were naturally dried and prepared for leaf area calculation. Rainfall experiments were controlled by an artificial rainfall simulation system (QYJY-503C, Xi 'an Qingyuan measurement and control technology Corp., Xi'an, China). The raindrops from height of 12 m was controlled by computer to set rainfall areas and intensities. Rainfall intensity in this system ranges from 0 to 300 mm/ h and their coefficients of uniformity are more than 85% (Sun et al., 2016). Our sampling intervals were divided into each 2.5 mm of rainfall until maximum continuous rainfall reached 17.5 mm. Time intervals for application were every 10, 5, and 3 min under 15, 30 and 50 mm/h rainfall intensities, respectively, until their total rainfall durations reached 70, 35 and 21 min for 17.5 mm rainfall. Wash-off suspensions from leaf surfaces were collected from a

X. Xu et al. / Environmental Pollution 255 (2019) 113234

3

Table 1 Leaf traits of tested plant species. Species

Morphology

Coarseness (nm)

Contact angle ( )

Wax content (g/m2)

Stomata density (N/mm2)

P. tomentosa S. japonica P. tabulaeformis P. orientalis P. cerasifera E. japonicus

smooth pilose resin canal scale-like shallow groove deep groove

85.68 ± 11.15 122.11 ± 34.32 227.54 ± 45.56 183.45 ± 31.21 55.53 ± 11.23a e

70.39 ± 4.20 85.43 ± 1.36 66.5 ± 4.2 e 60.65 ± 3.18 64.96 ± 1.56

0.48 ± 0.13a 0.35 ± 0.07 1.25 ± 0.12 e 0.46 ± 0.05a e

168.00 ± 3.50 109.00 ± 2.70 68.86 ± 12.17 151.83 ± 24.88 277.00 ± 5.70 247.00 ± 2.80

a

Data were derived from species in the same genus with similar leaf traits.

constructed experimental device with sampling bottles, PVC pipes, plastic funnels, foam board, and an upholder as shown in Xu et al. (2017). The following describes sampling needed to collect the PM. Further details on calculations are provided in Part 2.3. Wash-off from leaf surfaces flowed from subjacent plastic funnels through PVC pipes to sampling bottles. At the end of each rainfall interval, funnels were washed by spraying distilled water to remove PM residue. The plastic bottles with wash-off and control (no leaves above the funnel) samples were then replaced with new empty bottles to collect wash-off in the next interval. After rainfall experiments, all bottles with suspensions were labelled and kept at 4  C in a refrigerator, in preparation for laboratory analysis. Sampling events were repeated three times under each rainfall intensity for three replicate samples in every rainfall interval. Total number of collected resuspension samples was 378 (6 species * 7 intervals * 3 intensities * 3 replicates). Leaf samples after rainfall were dried and kept in similar fashion to those for initial accumulation. 2.3. Quantitative analysis of PM Water-insoluble surface PM was analysed as Dzierzanowski et al. (2011) and all measurements were performed in three biological replicates. Size fractions were divided according to the standard of atmospheric criteria pollutants (US EPA, 2013). Suspensions were first filtered with a 100-mm metal sieve and then filtered through a glass filter connecting to a vacuum pump (HPD25A, Tianjin Heng'ao Development Corp., China) with hydrophilic membrane filters of 10-mm, then 3-mm, and finally 0.2-mm (EMD Millipore Corp., USA). The membrane filters were weighed using a balance sensitive to 10-mg (BT125D, Sartorius Corp., China) before filtration. After filtration, samples were weighed again to calculate their weight differences. Before weighing, all membrane filters were dried in an oven for 30 min at 50  C and stored advance for 24 h in an 80 cm  80 cm  80 cm polytetrafluoroethylene (PTFE) balance box at 25  C and 40% relative humidity, which was controlled by a balance and a humidity controller (WHD48-11; ACREL Co., Ltd., Jiangsu, China). As the rainwater from rainfall system is neither natural rainwater nor distilled water, we collected blank samples with only artificial rainwater under each rainfall intensity. Final PM wash-off masses eliminated PM masses in blank samples (20 ± 4 mg/mL) from wash-off samples to exclude the potential influence of PM in raindrops. Sampled broadleaves were recorded using a scanner (HP Scanjet 4850, Hewlett-Packard Co., Ltd., Beijing, China), whose areas were calculated using Photoshop CS6 software (Adobe Corp., USA). Sampled needle-leaf areas were calculated from average length and the volume following published allometric equations (Chen et al., 2016; Xu et al., 2018). Three fractions of PM were thus collected and designated as large particles (10e100 mm), coarse particles (3e10 mm), and fine particles (0.2e3 mm). The sum of three fractions was designated as total PM. The sum of fine and coarse particles was designated as small particles (0.2e10 mm). PM wash-off masses, rates and

proportions during rainfall processes were expressed by:

PðN; nÞ ¼ ðM2
M1 Þ=S

(1)

. RPðN; nÞ ¼ PðN; nÞ P0  100%

(2)

. PPðN;nÞ ¼ PðN;nÞ PðN;tÞ

(3)

where P(N, n) are PM masses per unit leaf area for size fraction n (n represents 10e100 mm, 3e10 mm or 0.2e3 mm) within sampling interval N; M2 and M1 are each sampled membrane filter's weight before and after filtration, respectively; S are sampled leaf areas; RP(N, n) are PM wash-off rates for size fraction n within sampling interval N; P0 are PM accumulation masses on leaf surfaces before rainfall; PPn are PM wash-off proportions of size fraction n in total PM within sampling interval N; and P(N, t) are total PM masses within sampling interval N. Wash-off masses could be negative after eliminating raindrop masses at the end of rainfall due to inevitable uncertainties in these measured values (Xu et al., 2017). When calculating proportions, negative values were considered as a sign that there wash-off were equivalent to zero and the whole group with three fractions was excluded. 2.4. Statistical analyses Values are presented in charts by Graphpad Prism 7 (Systat Software Inc, USA) and analysed by Statistical Package for the Social Sciences software (SPSS Inc., USA). Multivariate analysis of large, coarse and fine fractions was performed with a three factors analysis of a) rainfall intensities, b) rainfall intervals, c) plant species using a type 3 method. If there was a 3-way interaction effect, additional multivariate analysis on a) rainfall intensities and b) rainfall intervals was performed in each plant species separately. If there were 2-way interaction effects, one-way analysis of variance (ANOVA) for one interacted factor was analysed in each level of another interacted factor, followed by the Bonferroni (homoscedasticity) or Games-Howell (heteroscedasticity) post hoc comparisons. 3. Results 3.1. PM wash-off masses of size-fractionated particles during rainfall processes For wash-off mass in each interval, rainfall intervals had interaction effects with rainfall intensities or plant species (P < 0.01, N ¼ 18 or 9, Table SI 1). Therefore, we further analysed the influence of rainfall intensities and plant species in three fractions along with rainfall intervals separately. PM wash-off masses decreased with rainfall and approached zero after 12.5 mm for large, fine, or coarse particles (Fig. 1). PM

4

X. Xu et al. / Environmental Pollution 255 (2019) 113234

similar to those of coarse particles. However, one differences was that P. tomentosa approached zero simultaneously with P. tabulaeformis in 10e12.5 mm. 3.2. PM wash-off rates of size-fractionated particles during rainfall processes For cumulative wash-off rates, effects of plant species and rainfall intensities showed interactions in three fractions (P < 0.01, N ¼ 21, Table SI 2). Rainfall intervals had no interaction effects with plant species and rainfall intensities, but a main effect on cumulative wash-off rates (P < 0.01, N ¼ 54, Table SI 2). PM wash-off rates of all three fractions did not grow after 10 mm rainfall (P < 0.05, N ¼ 9). Cumulative PM wash-off rates only differed among rainfall intensities in some plant species (Fig. 3). For deciduous trees, 15 mm/ h had less cumulative wash-off than 30 mm/h and 50 mm/h for P. tomentosa and S. japonica in large particles. In coarse particles, cumulative wash-off rates increased from 15 mm/h to 50 mm/h for S. japonica. But there were no differences among intensities for P. cerasifera in each fraction. For evergreen trees, all three fractions had cumulative wash-off differences among intensities, except for P. orientalis in fine particles. For E. japonicus, 50 mm/h differed significantly with 15 mm/h. The six plant species also showed diverse efficiencies in PM wash-off rates (Fig. 3). For two deciduous trees, their dynamics were similar in P. tomentosa and S. japonica. These ultimate cumulative PM wash-off rates varied from 42% to 79%. The deciduous shrub, P. cerasifera, had average final rates of 79%. For evergreen species, cumulative PM wash-off rates of P. tabulaeformis were typically the smallest among other five species, averaging 32%. For P. orientalis and E. japonicus, cumulative PM wash-off rates had statistically similar dynamics in each size fraction, averaging 61% for small particles and 49% for large particles. 3.3. PM wash-off proportions of size-fractionated particles during rainfall processes

Fig. 1. Scatter plots of wash-off masses in each rainfall interval for three size fractions (N ¼ 54 in each column, stand for samples of all species). Lines in the dots presents the average value in each column. Lowercases marked significant differences in relation to the three rainfall intensities (15, 30, 50 in order) in each rainfall interval (ANOVA, P < 0.05, N ¼ 18). No lowercases indicated no differences from intensity in the interval.

wash-off masses of large particles averaged 6.5 mg/cm2 under 15 mm/h at 2.5 mm rainfall, while the other two sizes averaged 13.4 mg/cm2. Large particles decreased mostly within 2.5e5 mm rainfall, compared to the first rainfall interval. PM wash-off masses in large fraction differed among rainfall intensities after 12.5 mm rainfall. PM wash-off disappeared late under 15 mm/h. For coarse and fine particles, PM wash-off masses did not differ with increased rainfall intensity. PM wash-off masses differed significantly among plant species until 12.5 mm rainfall intervals (Fig. 2, not shown after 12.5 mm due to no differences). For large particles, P. tabulaeformis had the least PM wash-off in each interval. P. tomentosa were the smallest at 2.5 mm, while P. tabulaeformis had no statistical difference. The remaining four species had no difference at first interval, but after that, some differences emerged. S. japonica dropped more quickly than other three in 10e12.5 mm rainfall. E. japonicus had the largest wash-off mass in each interval, followed by P. orientalis, and then P. cerasifera. For coarse particles, E. japonicus were still the leader in drop-off, followed by P. orientalis, which was significantly smaller than E. japonicus after 5 mm. For fine particles, their patterns were

For wash-off proportions in each interval, plant species and intensities had interaction effects in each size fraction but only for 10 mm rainfall (P < 0.05, N ¼ 3, Table SI 3). PM wash-off proportions fluctuated quite randomly and intensely after 7.5 mm rainfall when total PM wash-off masses were quite small. In later intervals, one or even three components of size fractions failed to be found; thus, only samples within 10 mm rainfall were shown and statistically analysed in Fig. 4. In terms of each plant species, the interaction effects of rainfall intervals and intensities were only found in three size fractions of P. cerasifera (P < 0.05, N ¼ 3, Table SI 4). Specifically, its PM wash-off proportions of large or coarse fractions differed between intensities of 15 mm/h and 50 mm/h in 2.5 mm and 10 mm rainfall. Along with rainfall, large fractions increased and small fractions decreased under 15 mm/h, while showing opposite patterns under 30 mm/h and 50 mm/h. In addition, there was no main effects of rainfall intervals and intensities for E. japonicus. For deciduous trees, P. tomentosa had more evident variations compared to other species. PM wash-off proportions decreased 44% in large particles, while also having a growth of 29% in coarse fractions and of 16% in fine fractions until 10 mm. For S. japonica, PM wash-off proportions varied similarly but with much smaller differences as P. tomentosa. For coniferous trees, PM wash-off proportions of P. tabulaeformis only showed an increase of fine fraction in 2.5e5 mm rainfall interval. For P. orientalis, PM wash-off proportions were relative steady in large fractions and had a gradual decline in coarse fractions and a slight growth in fine fractions.

X. Xu et al. / Environmental Pollution 255 (2019) 113234

5

4. Discussion 4.1. Rainfall effects on size distribution of PM wash-off We hypothesized that rainfall intensity would increase wash-off efficiency of fine and coarse particles. Our results showed that increasing rainfall intensity mainly caused differences in cumulative wash-off rates of fine and coarse particles from evergreen species, consistent with findings for PM1, PM2.5 and PM10 number densities from English ivy (Weerakkody et al., 2018a). However, cumulative wash-off rates of large particles were greatly influenced by rainfall intensities for all species. Wash-off rates of needle-leaves thus are likely to increase in areas experiencing increasing rainfall intensities. Although wash-off masses in each interval only differed in large particles, these results imply a mechanism of increasing force of rain with intensity that reduces the resistance between leaf and particle layers. In contrast, wash-off proportions had no differences among rainfall intensities in this study, implying that rainwater always exerts a strong enough kinetic energy on leaf surfaces to transport mobilizable size-fractionated particles with similar flow (Neinhuis and Barthlott, 1998). PM wash-off masses in three size fractions had the similar rank in each rainfall interval as their accumulation without rainfall or retention after rainfall on leaf surfaces (Popek et al., 2013; Przybysz et al., 2014). PM wash-off masses in three fractions approached zero after 10e15 mm in lab rainfall experiments for PM2.5 on three conifers under 220 mm/h (Pullman, 2009), as well as experiments in terms of number density for PM1, PM2.5 and PM10 on four vertical greenery species under 16 mm/h (Weerakkody et al., 2018a). It appears, regardless of size fraction, leaf accumulations quickly turnover after 10 mm rainfall. However, there are differences with size. The general dynamics of wash-off rates in fine, coarse or large fractions showed an exponential trend, similar with total water-insoluble particles (0.2e100 mm) and water-soluble ions (Van Stan et al., 2011; Xu et al., 2017). Final wash-off rates were larger in particles with larger size in pines (Przybysz et al., 2014), but they had no unified patterns with other plant species of our study and natural light or moderate rains (Liu et al., 2018). Moreover, proportions of large particles decreased and of small particles increased along with rainfall because small particles were often stuck in grooves (Mo et al., 2015) and removed late in post-rainfall. E. japonicus had the greatest leaf water holding capacity, making wash-off proportions more random. 4.2. Differences of PM wash-off among plant species during rainfall PM wash-off masses and rates that differed among species and serve as important indicators to plants selection for phytoremediation. It was found that plant species were generally similar for dynamics of sulfate and potassium wash-off masses within leaching (Potter and Radsdale, 1991), but they differed in water insoluble particles masses of three fractions in this study. Pine needle surfaces had similar smallest cumulative wash-off rates for three size fractions as Przybysz et al. (2014), but much less wash-off masses in this study. They also found larger PM accumulation capacity in small particles and water-soluble ions than other species in field experiments (Beckett et al., 2000b). Since pines were young in our study, their wash-off efficiency may be strengthened as they age. Cypress and an evergreen broadleaved shrub had the largest Fig. 2. Column plots with error bars (Mean ± SD, N ¼ 9) of PM wash-off masses in three size fractions for (A) P. tomentosa, (B) S. japonica, (C) P. tabulaeformis, (D) P. orientalis, (E) P. cerasifera and (F) E. japonicus. Lowercase letters marked significant differences in relation to six plant species in each rainfall interval at 0.05 level.

6

X. Xu et al. / Environmental Pollution 255 (2019) 113234

Fig. 3. Line plots with error bars (Mean ± SD, N ¼ 3) of cumulative PM wash-off rates in three size fractions. Lowercase letters marked significant differences in relation to rainfall intensities in each species during whole rainfall process at 0.05 level. Uppercase letters marked significant differences in relation to plant species at 0.05 level.

X. Xu et al. / Environmental Pollution 255 (2019) 113234

7

angles, but they are also influenced by water-repellency of sticky resin and wax content. Wax features may change by rainfall characteristics. Leaf epicuticular wax tend to capture PM permanently during rainfall because of hydrophobicity (Dzierzanowski et al., 2011), but heavy rain could impose erosion on wax layer and acid rain could modify wax crystal structure (Baker and Hunt, 1986; Percy and Baker, 2010). Leaf trichome density had a significant positive correlation with wettability, where raindrops could form water film on leaf surfaces, pierced by trichomes or water droplets above trichomes (Brewer and Smith, 1997). More fine particles were found adhered to trichomes (Mo et al., 2015; Yang et al., 2015), but wash-off rates in coarse and large particles for S. japonica with trichomes in this study were increased by rainfall intensity. It indicated more raindrops may be pierced in large intensity to increase rates of larger size particle wash-off. Stomatal density has not been found to consistently correlate to wettability in previous studies (Wang, 2012), but it may absorb ions after leaching. 4.3. Future work

Fig. 4. Heat maps of size-fractionated PM wash-off proportions in each rainfall interval. Each small area represents each PM wash-off sample case. A double cross indicates samples were removed. Lowercase letters marked significant differences in relation to rainfall intervals at 0.10 level.

PM wash-off masses in three size fractions among six species, which also had large initial accumulations among 35 plant species (Mo et al., 2015). Deciduous trees had similar PM wash-off rates as in three size fractions (apart from bigger rates of large particles for S. japonica) but smaller wash-off masses. Deciduous shrub had the greatest wash-off rates and small wash-off masses, which should wash effectively in natural rain events. These studies and our results are consistent with recommendations of cypresses and evergreen broadleaved shrubs as choices of plant selections for artificial rainfall strategies to renew PM accumulation in urban greening (Schaubroeck et al., 2014). The findings imply mechanisms behind PM wash-off is intrinsically related to leaf traits. Initially, PM accumulation often varied among plant species with diverse leaf traits (Sæbø et al., 2012; Mo et al., 2015; Weerakkody et al., 2018b), influencing the wash-off masses. Also, as a comprehensive indicator, leaf coarseness changed the raindrops force impacting on leaf surfaces. Rainfall intensities influenced size fractions, especially of the smaller size, in species with higher coarseness in this study. Differences in leaf surface micromorphology were closely related to wettability, resulting in different amounts of accumulated particles by intercepted rain (Neinhuis and Barthlott, 1998). For example, P. cerasifera had the smallest coarseness and contact angle (large wettability) among six species, contributing to the largest reduction rates. But contact angle is not the only factor to influence wash-off rates. Many pines have both low wash-off rates and small contact

The entire efficiencies of PM reduction by rainfall could be affected greatly by water-soluble fractional composition of PM. Ratios of water-soluble ions to total water-insoluble PM ranged from 7% to 50% on leaf surfaces (Xu et al., 2018). It was also found that water-insoluble PM accounted for more proportions in smaller size airborne PM (Dao et al., 2014; Velali et al., 2015). These water soluble ultra-fine PM can be dissolved promptly and removed with runoff (Beckett et al., 2000b; Freer-Smith et al., 2005). Watersoluble PM should be investigated with PM size distribution to further reveal reduction efficiency during rainfall. However, it is still challenging to partition water-soluble components from each size of PM accumulation on leaf surfaces. There are also lingering differences between lab and field studies. Initial accumulations might never reach the maximum in some places and thus PM wash-off rates would change. Initial PM accumulations in real environments often experience prior rainfall until they get the final wash-off (Wang et al., 2015). Field studies suggested that particles could not be washed off as easily from leaf surfaces as lab experiments above (Freer-Smith et al., 2005; Liu et al., 2018). It is limited in real rainfall by its lower intensity and distribution within the tree canopy (Xu et al., 2017). Complex canopy structures in big trees reduce raindrop energy and distribute rainfall volume on leaf surfaces in lower canopy, making it difficult to link rainfall rate and wash-off (Levia et al., 2006; Nanko et al., 2011). And size distribution of water insoluble particles differed among plant organs may also influence wash-off from tree surfaces (Xu et al., 2019). In addition, lab experiments could control rainfall characteristics to a great extent before raindrops touched leaf surfaces. Considering rainfall intensity and duration as rainfall characteristics is one-sided. Future experiments would better disentangle mechanisms by additional measurements of raindrop diameter, velocity, and hydraulic characteristics on tree surfaces. These mechanisms could greatly contribute to understanding rainfall in real environments. 5. Conclusion PM wash-off from leaf surfaces showed quite different patterns in size fractions, as found for wash-off masses, rates and proportions. Both rainfall intensities and plant species are important to consider in PM washing off dynamics from leaf surfaces. Rainfall intensities had a greater influence on wash-off masses and rates of large particles in diverse plant species and also for small particles in evergreen species during rainfall. Other small particles remained

8

X. Xu et al. / Environmental Pollution 255 (2019) 113234

stuck to leaf surfaces in early rainfall, thereby increasing reduction efficiency with rainfall. To renew PM accumulation on leaf surfaces, cypresses and evergreen broadleaved shrubs had large wash-off masses, while deciduous broadleaved shrubs had high cumulative wash-off rates. These results provide guidance for selection of plant species in urban environments for maximizing PM removal from the atmosphere. Acknowledgements This study was supported by National Key R&D Program of China (2016YFC0500802), Beijing Laboratory Project (2015BLUREE07), Beijing Municipal Education Commission (CEFF- PXM2018_ 014207_ 000024) and China Scholarship Council (No.201806510013). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.envpol.2019.113234. References Ali, H., Khan, E., Sajad, M.A., 2013. Phytoremediation of heavy metalsdconcepts and applications. Chemosphere 91 (7), 869e881. Beckett, K.P., Freer Smith, P.H., Taylor, G., 2000a. Particulate pollution capture by urban trees: effect of and windspeed. Glob. Chang. Biol. 6, 995e1003. Beckett, K.P., Freersmith, P., Taylor, G., 2000b. Effective tree species for local airquality management. J. Arboric. 26 (1), 12e19. Baker, E.A., Hunt, G.M., 1986. Erosion of waxes from leaf surfaces by simulated rain. New Phytol. 102 (1), 161e173. Brewer, C.A., Smith, W.K., 1997. Patterns of leaf surface wetness for montane and subalpine plants. Plant Cell Environ. 20, 1e11. Cai, M., Xin, Z., Yu, X., 2017. Spatio-temporal variations in PM leaf deposition: a meta-analysis. Environ. Pollut. 231, 207e218. Chen, L., Liu, C., Zou, R., Yang, M., Zhang, Z., 2016. Experimental examination of effectiveness of vegetation as bio-filter of particulate matters in the urban environment. Environ. Pollut. 208, 198e208. Dao, X., Wang, Z., Lv, Y., et al., 2014. Chemical characteristics of water-soluble ions in particulate matter in three metropolitan areas in the North China plain. PLoS One 9, e113831. Dzierzanowski, K., Popek, R., Gawro, H., et al., 2011. Deposition of particulate matter of different size fractions on leaf surfaces and in waxes of urban forest species. Int. J. Phytoremediation 13, 1037e1046. Freer-Smith, P.H., Beckett, K.P., Taylor, G., 2005. Deposition velocities to Sorbus aria,
’, Pinus nigra and  Acer campestre, Populus deltoides  trichocarpa ‘Beaupre Cupressocyparis leylandii for coarse, fine and ultra-fine particles in the urban environment. Environ. Pollut. 133, 157e167. Germer, S., Neill, C., Krusche, A.V., et al., 2007. Seasonal and within-event dynamics ^nia, of rainfall and throughfall chemistry in an open tropical rainforest in Rondo Brazil. Biogeochemistry 86, 155e174. Kim, K.H., Kabir, E., Kabir, S., 2015. A review on the human health impact of airborne particulate matter. Environ. Int. 74, 136e143. Levia, D.F., Frost, E.E., et al., 2006. Variability of throughfall volume and solute inputs in wooded ecosystems. Prog. Phys. Geogr. 30 (5), 605e632. Levia, D.F., Van Stan, J.T., Siegert, C.M., et al., 2011. Atmospheric deposition and corresponding variability of stemflow chemistry across temporal scales in a mid-Atlantic broadleaved deciduous forest. Atmos. Environ. 45, 3046e3054. Liu, C., Zhang, Z., Chen, L., et al., 2018. Impacts of rainfall on leaf PM detention in five commonly used urban tree species in northern China. Acta Ecol. Sin. 38 (7), 2353e2361 (in Chinese). Liu, L., Guan, D., Peart, M.R., et al., 2013. The dust retention capacities of urban vegetation-a case study of Guangzhou, South China. Environ. Sci. Pollut. Res. 20, 6601e6610. Lovett, G.M., Lindberg, S.E., 1984. Dry deposition and canopy exchange in a mixed oak forest as determined by analysis of throughfall. J. Appl. Ecol. 21 (3), 1013e1027. Mo, L., Ma, Z., Xu, Y., et al., 2015. Assessing the capacity of plant species to accumulate particulate matter in Beijing, China. PLoS One 10, e0140664. Nanko, K., Onda, Y., Ito, A., et al., 2011. Spatial variability of throughfall under a single tree: experimental study of rainfall amount, raindrops, and kinetic energy. Agric. For. Meteorol. 151 (9), 0e1182. Neinhuis, C., Barthlott, W., 1998. Seasonal changes of leaf surface contamination in beech, oak, and ginkgo in relation to leaf micromorphology and wettability. New Phytol. 138, 91e98. Nowak, D.J., Hirabayashi, S., Bodine, A., et al., 2013. Modeled PM2.5 removal by trees in ten U.S. cities and associated health effects. Environ. Pollut. 178, 395e402. Ould-Dada, Z., Baghini, N.M., 2001. Resuspension of small particles from tree surfaces. Atmos. Environ 35, 3799e3809.

Percy, K.E., Baker, E.A., 2010. Effects of simulated rain on epicuticular wax production, morphology, chemical composition and on cuticular membrane thickness in two clones of sitka spruce (picea sitchensis (bong.) carr.). New Phytol. 116 (1), 79e87. Petroff, A., Mailliat, A., Amielh, M., et al., 2008. Aerosol dry deposition on vegetative canopies. Part I: review of present knowledge. Atmos. Environ. 42 (16), 3625e3653.
ska, H., Wrochna, M., et al., 2013. Particulate matter on foliage of Popek, R., Gawron 13 woody species: deposition on surfaces and phytostabilisation in waxes e a 3- year study. Int. J. Phytoremediation 15, 245e256. Potter, C.S., Ragsdale, H.L., 1991. Dry deposition washoff from forest tree leaves by experimental acid rainfall. Atmos. Environ. part A. general Topics 25 (2), 341e349. Przybysz, A., Sæbø, A., Hanslin, H.M., et al., 2014. Accumulation of particulate matter and trace elements on vegetation as affected by pollution level, rainfall and the passage of time. Sci. Total Environ. 481, 360e369. Pullman, M.R., 2009. Conifer PM2.5 Deposition and Resuspension in Wind and Rain Events. Cornell University. Sæbø, A., Popek, R., Nawrot, B., et al., 2012. Plant species differences in particulate matter accumulation on leaf surfaces. Sci. Total Environ. 427e428, 347e354. Schaubroeck, T., Deckmyn, G., Neirynck, J., et al., 2014. Multilayered modeling of particulate matter removal by a growing forest over time, from plant surface deposition to washoff via rainfall. Environ. Sci. Technol. 48, 10785e10794. Song, Y., Maher, B.A., Li, F., et al., 2015. Particulate matter deposited on leaf of five evergreen species in Beijing, China: source identification and size distribution. Atmos. Environ. 105, 53e60. Sun, J., Yu, X., Li, H., et al., 2016. Simulated erosion using soils from vegetated slopes in the Jiufeng Mountains, China. Catena 136, 128e134. Tallis, M., Taylor, G., Sinnett, D., et al., 2011. Estimating the removal of atmospheric particulate pollution by the urban tree canopy of London, under current and future environments. Landsc. Urban Plan. 103, 129e138. Tomasevi
c, M., Vukmirovi
c, Z., Rajsi
c, S., et al., 2005. Characterization of trace metal particles deposited on some deciduous tree leaves in an urban area. Chemosphere 61, 753e760. US Environmental Protection Agency (US EPA), 2013. National Ambient Air Quality Standards (NAAQS). https://www.epa.gov/criteria-air-pollutants. Van Stan II., J.T., Siegert, C.M., Levia, D.F., et al., 2011. Effects of wind-driven rainfall on stemflow generation between codominant tree species with differing crown characteristics. Agric. For. Meteorol. 151 (9), 0e1286. Velali, E., Papachristou, E., Pantazaki, A., et al., 2015. Redox activity and in vitro bioactivity of the water-soluble fraction of urban particulate matter in relation to particle size and chemical composition. Environ. Pollut. 208, 774e786. Wang, H., 2012. Mechanisms of Plant Leaves Based on Leaf Surface Wettability on Rainfall Interception and Dust-Capturing. Xi’an University of Architecture & Technology in Chinese. Wang, L., Gong, H., Liao, W., Wang, Z., 2015. Accumulation of particles on the surface of leaves during leaf expansion. Sci. Total Environ. 532, 420e434. Weerakkody, U., Dover, W.J., Mitchell, P., et al., 2018a. The impact of rainfall in remobilising particulate matter accumulated on leaves of four evergreen species grown on a green screen and a living wall. Urban For. Urban Gree. 35, 21e31. Weerakkody, U., Dover, J.W., Mitchell, et al., 2018b. Quantification of the traffic generated particulate matter capture by plant species in a living wall and evaluation of the important leaf characteristics. Sci. Total Environ. 635, 1012e1024. World Health Organization (WHO), 2003. Health Aspects of Air Pollution with Particulate Matter, Ozone and Nitrogen Dioxide. Report 2003 on a WHO Working Group. http://www.euro.who.int/data/assets/pdf_file/0005/112199/ E79097.pdf. World Health Organization (WHO), 2013. Health Effects of Particulate Matter Final. Report 2013 on a WHO Working Group. http://www.euro.who.int/__data/ assets/pdf_file/0006/189051/Health-effects-of-particulate-matter-final-Eng. pdf?ua¼1. Xu, X., Zhang, Z., Bao, L., et al., 2017. Influence of rainfall duration and intensity on particulate matter removal from plant leaves. Sci. Total Environ. 609, 11e16. Xu, Y., Xu, W., Mo, L., et al., 2018. Quantifying particulate matter accumulated on leaves by 17 species of urban trees in Beijing, China. Environ. Sci. Pollut. Res. 25 (13), 12545. Xu, X., Yu, X., Mo, L., Xu, Y., Bao, L., Lun, X., 2019. Atmospheric particulate matter accumulation on trees: A comparison of boles, branches and leaves,. J. Clean. Prod 226, 349e356. Yang, J., McBride, J., Zhou, J., et al., 2005. The urban forest in Beijing and its role in air pollution reduction. Urban. For. Urban. Green. 3, 65e78. Yang, J., Wang, H., Xie, B., et al., 2015. Accumulation of particulate matter on leaves of nine urban greening plant species with different micromorphological structures in Beijing. Res. Environ. Sci. 28 (3), 384e392 (in Chinese). Yin, S., Shen, Z., Zhou, P., et al., 2011. Quantifying air pollution attenuation within urban parks: an experimental approach in Shanghai, China. Environ. Pollut. 159, 2155e2163. Zhao, X., 2018. Study on PM2.5 Adsorption Function of Urban Forest Vegetation: A Case in Shenyang. Shenyang Agricultural University in Chinese. Zhang, W.K., Wang, B., Niu, X., 2016. Dynamic research on particulates-absorbing capacities of common tree species leaves in Beijing over time. Acta Sci. Circumstantiae 36 (10), 3840e3847 (in Chinese).