Urban Allotment Gardens for the Biomonitoring of Atmospheric Trace Element Poi I ution

Miguel lzquierdo-Díaz,* Peter E. Holm, Fernando Barrio-Parra, Eduardo De Miguel, Jonas Duus Stevens Lekfeldt, and Jakob Magid

Abstract

This study evaluates the results of the characterization of air pollution in urban green areas using edible plants.Tothis purpose, we examined the effect of location (i.e., three different levels of pollution), substrate (peat moss and vermiculite), and plant species (oilseed rape [Brassica na pus L.] and kale [Brassica o/eracea L.]) on the accumulation oftrace elements on leaves. A total of 36

samples of unwashed leaves were digested with HN03-H

20

2 and

analyzed for 27 elements by inductively coupled plasma mass spectrometry. Considering the location, plants exposed next to the road showed higher contents of traffic-related elements, and additionally, outdoors samples were enriched in marine aerosol ions. Cadmium and Pb concentrations did not exceed the European legal maximum levels for vegetables, so their consumption would be safe for human health. Results support the hypothesis that edible plants such as kale and rapeseed could be used as bioindicators of atmospheric pollution.

GLOBAL EXP ANSION of anthropogenic activities along with rapid world population growth and urbanization has brought a huge influx of toxic pollutants to soil,

water, and air. Contaminants are emitted into the environment in different ways and from different sources such as mining and industrial processes, fossil fuel burning, transportation, waste disposal, or agricultura! activities (Alloway, 2004). Ambient out­door air pollution is the direct cause of at least 3 million deaths globally every year due to acute lower respiratory disease in chil­dren and cardiovascular or pulmonary chronic diseases in adults (WHO, 2016). Furthermore, 90% of the world's population live in cities and towns exceeding the particulate matter levels recommended by the World Health Organization Air ~ality Guidelines (WHO, 2016). Trace elements are usually enriched in the finest fractions of urban particulate materials (Madrid et al., 2008). These particles are easily introduced in the atmosphere and hence may be inhaled by humans or the trace elements in their composition can be incorporated into plants by absorption from the soil or surface deposition ( Charlesworth et al., 2011). Trace element toxicity is an increasing concern dueto ecological impacts and potential adverse effects on human health ( Karaaslan and Yaman, 2013; Daud et al., 2015). Given these facts, there is a strong incentive to develop simple, economical, and sufficiently precise monitoring systems of outdoor air quality.

Biomonitoring is the regular and systematic use ofliving organ­isms to measure the environmental exposure to pollutants and to identify potential health risks (Severoglu et al., 2015). It allows continuous observation, resulting in an averaged and integrated response with a high spatial and temporal resolution that combines the effect of all environmental factors. Therefore, application of vegetative life as a passive sampling system for urban air monitor­ing is a simple, cost-effective, and reliable method compared with conventional physicochemical techniques (Guzmán-Morales et al., 2011; Sawidis et al., 2011 ), such as cascade irnpactors, cyclones, or high- or low-volume air samplers.

Many studies have highlighted the effectiveness of lichens and mosses as biological indicators of air quality (Augusto et al., 2013; Bargagli, 2016; Urosevié et al., 2017). Their growth rates are acceptably rapid and, above all, they need to take all

M. lzquierdo-Díaz, F. Barrio-Parra, and E. De Miguel, Prospecting & Environment Laboratory, Universidad Politécnica de Madrid, Alenza 4, 28003 Madrid, Spain;

P.E. Holm, J.D.S. Lekfeldt, and J. Magid, Dep. of Plant and Environmental Sciences, Faculty of Science, Univ. of Copenhagen, Thorvaldsensvej 40, 1871 Frederiksberg

C, Denmark. Assigned to Associate Editor Karen Bradham.

Abbreviations: KMO, Kaiser-Meyer-Olkin; PAH, polycyclic aromatic hydrocarbon;

PC, principal component; PCA, principal component analysis.

the nutrients from the atmosphere and therefore, they are very sensitive to environmental changes. These organisms, however, are infrequently found in cities because of the high degree of pol­lution and the degradation of their natural habitats, so there is an increasing interest in using vascular plants as trace elements biomonitors (Bonanno, 2014).

O ver the past few years, several studies have shown the poten­tial of trees to monitor air pollution. Pirre needles collected from specimens in roadside, urban, and industrial areas showed higher concentrations of several heavy metals than those sampled in ref­erence sites (Kord et al., 201 O; Sawidis et al., 2011; Karaaslan and Yaman, 2013; Daud et al., 201S).1he leaves ofother species, such as Phoenix dactylijera L. (Al-Khashman et al., 2011), Morus nigra

L. (Daud et al., 2011 ), Ficus benjamina L. (Guzmán-Morales et al., 2011), andPlatanus sp. (Ili et al., 2016), have also been used. Along with leaves, sorne researchers have also analyzed bark or twigs from trees to find out that they had bioaccumulated trace elements ( Sawidis et al., 2011; Guéguen et al., 2012; Severoglu et al., 201 S; Esen et al., 2016). (Hemi)epiphytic plants have been used as well because they absorb toxic substances only from the air (Rodriguez et al., 2011; de Paula et al., 20 l S; Giampaoli et al., 2016) and grow faster than trees ( e.g., Lolium multijiorum Lam. [Klumpp et al., 2009] or Taraxacum ojficinale Weber [Kleckerová and Docekalová, 2014]). In addition to trace elements, the effi­cacy of trees far biomonitoring cesium radionuclides ( Cosma et al., 2016), platinum-group metals (Bonanno and Pavone, 201 S), and polycyclic aromatic hydrocarbons (PAHs) (Rinaldi et al., 2012) has been proven. EuroBionet, a European networkfor the assessment of urban air quality, has developed standard meth­ods to monitor trace elements and PAHs using ryegrass (Lolium

multijiorum) and curly kale (Brassica oleracea L. var. acephala),

respectively (Klumpp et al., 2002, 2009). The capacity to capture airborne particles from ambient air

varies considerably among plant species due to their different morphological foliar characteristics. Interception efficiency is mainly affected by cuticle and epidermal layer features (stomata, trichomes, and wax) and leaf surface area, geometry, and rough­ness (Kulshreshtha et al., 2009; Ram et al., 2014). Consequently, variability in the anatomy and structure ofleaves may lead to dif­ferences in the capacity to trap airborne particles.

With the expansion and consolidation of the urban green spaces' framework, a regional biomonitoring network could be devised that uses edible plants grown by urban gardeners as bio­monitors. Following this idea, the main aim of this study was to evaluare the potential of edible plant species in urban gardens far biomonitoring atmospheric particular matter pollution. Trace ele­ments concentrations in biomass may be influenced by (i) loca­tion, primarily dueto traffic density in the surroundings, (ii) type of substrate, and (iii) plant species. The best combination of those is assessed far future air quality monitoring of trace elements. A last objective of this study was to evaluare whether the trace ele­ment load on vegetables would be within safe limits far their con­sumption if they were cultivated in the sampling sites.

Materials and Methods Plant Cultivation

Seeds of kale (Brassica oleracea 'Nagoya Red') and oilseed rape (Brassica napus 'Silvershadow') were pregerminated far

1 O d in trays with two types of substrate: blande sphagnum peat moss (Pindstrup Substrate no. 2, Pindstrup Mosebrug A/S; composition shown in Supplemental Table S 1) and vermiculite (heat-expanded phyllosilicate mineral) in a growth chamber. Pregermination conditions were 16-h days and 8-h nights at 20 and l SºC, respectively. After pregermination, three seedlings were transplanted in 2-L plastic pots. A pseudo-hydroponic watering system was developed to avoid manual irrigation, con­sisting of a plastic box with six hales in the lid so that the pots remained in suspension. The basin was filled with distilled water when peat was used as substrate and with a micronutrient solu­tion (composition shown in Supplemental Table S2) when the substrate was vermiculite. Absorbent fiber strips were placed through the hales in the bottom of the pots and immersed in the reservo ir, allowing plants to suck water by capillarity. In this way, contaminants will come exclusively from the atmosphere (i.e., trace elements absorbed in the inner layer of the leave or airborne particles deposited on its surface).

After 32 d of cultivation with a last week of acclimatization ( changing conditions to 14-h days and 10-h nights at tempera­tures of 10 and SºC, respectively), two boxes ( one of each kind of substrate) each containing two pots ofkale and four of rapeseed plants (Fig. 1), were placed in the corresponding sites of study in Copenhagen. The difference in the number of pots contain­ing each species was due to the lower germination success rate of kale. T wo additional boxes were kept in the clima te chamber as control blank samples.

Study Sites Copenhagen, capital city of Denmark, has a population of

rv 1.3 million inhabitants. The city faces the 0resund Strait, which connects the N orth Sea with the Baltic Sea. The experimental loca­tions were selected according to their different expected levels of atmospheric pollution: Landboh0jskolens Have, the horticultura! garden in University of Copenhagen Frederiksberg campus park surrounded by a street with a low traffic density (SSº 40'S4.611 N, 12°32126.811 E), and the university students' urban garden, which isnext to anÁgaderoad with ahigh traffic density (SSº 41 110.911 N, 12°32139.211 E). The plants were exposed far a period of 2S d (7Nov.-2 Dec. 2016), duringwhich the maximum andminimum temperatures recorded were 10 and -4ºC, whereas the accumu­lated precipitation was 14 mm. Control samples were kept in an air-filtered climate chamber with similar outdoor weather condi­tions (14 h of daylight at lOºC and 10 h of darkness at SºC). Most previous studies reported longer exposure periods, ranging from 2 to 6 mo (Rodriguez et al., 2011; Rinaldi et al., 2012; Hassan and Basahi, 2013; Giampaoli et al., 2016) to years in the case of trees specimens (Kord et al., 2010; Sawidis et al., 2011; Karaaslan and Yaman, 2013; Daud et al., 201S). Although Klumpp et al. (2009) reported an exposure period far ryegrass cultures of28 d, curly kale was left far 8 wk (Klumpp et al., 2002). In this study, a short period was chosen because samples had already grown to the size that they are usually consumed.

Plant Harvest, Preparation, and Chemical Analysis After the exposure period, leaves were harvested, gathering the

three specimens of each pot to obtain a composite sample, which was stored in a paper bag. Samples were left unwashed, since it was intended to simulare the suitability far human consumption

••• •••

Fig. 1. The left panel is a schematic diagram ofthe experimental setup, each pot containing three seedlings. Sphagnum is shown as black filled circles, vermiculite as circles with gold stripes, K stands for kale, and R stands for rapeseed. The right panel shows one ofthe locations.

in the worst-case scenario, a procedure also applied in other bio­monitoring studies (Sawidis et al., 2011; Giampaoli et al., 2016; Ili et al., 2016). Samples were subsequently oven dried at 60°C during 2 d and milled using zirconia grinding balls. Of the result­ing powder 100 mg was mixed with 2.5 mL HN0

3 (70%) and

1 mL H20

2 (15%) and digested in a microwave pressure system

(UltraWAVE, Milestone Srl). Digested solutions were trans­ferred to volumetric flasks and diluted up to 50 mL with Milli-Q water (resistivity = 18.2 MO cm). A blank prepared with the same chemical reagents and a standard reference material ( apple [ Malus domestica Borkh J leaves NIST-SRM 1515) were included in each digestion batch far quality assurance and control pur­poses. Majar and trace analytical determinations were perfarmed by inductively coupled plasma mass spectrometry (7900 ICP­MS, Agilent), quantifying the concentration of 27 elements. Analytical precision, expressed as the relative SD from the SRM 1515 measurement replicares, was <20% far all elements, except Cr with a value of27%. Recovery rates ranged from 84 to 118% of certified values, except far Na and those elements with reponed values below O.OS mg kg-1 (As, Sb, Se, and Se), and dritt between samples was generally within ± 10%, except far Sb. The limit of detection was calculated as three times the SD of a minimum of eight replicare analyses of the calibration blank.

Together with macro- and micronutrients (i.e., Co, Cr, Cu, Mn, Ni, Se, V, and Zn), the suite of elements considered in this study included toxic elements like As, Cd, Pb, Sb, and Tl, whose potential noncarcinogenic and carcinogenic effects on human health include lung, tracheal, and bronchus cancer (inhalation) and skin lesions, proteinuria, and neurological and developmen­tal impairment, as well as liver, kidney, lung, and bladder cancer (oral exposure).

Statistical Analysis All statistical calculations were based on dry-weight con­

centrations and were perfarmed using R version 3.3.2 (R Core Team, 2017). T ukey's honest significant difference test was used far multiple mean comparisons. A principal component analy­sis (PCA) was undertaken to reveal possible groupings among trace elements. The component matrix was rotated using Kaiser's normal rotation method.

Results and Discussion Trace Element Concentrations

The average trace elements contents far the two species depending on location type and substrate are shown in Table l. Beryllium was under the limit of detection far almost all sam­ples and was therefare omitted from the quantitative analysis of results. Elements could be categorized into three groups accord­ing to the arder of magnitude of their determined concentra­tions. Essential plant macronutrient (Ca, K, Mg, and P) and Na values were > 103 mg kg-1

; micronutrients (B, Fe, Mn, and Zn), Al, and Sr ranged from 1 O and 103 mg kg-1

, whereas the concen­tration of other beneficia! ( Co, Cu, Ni, Se, and V) or toxic (As, Cd, Cr, Pb, Sb, Se, and Y) trace elements was < 1 O mg kg-1

•

Comparison of elemental mean concentrations revealed significant differences in the accumulation of trace elements depending on the exposure location (Table 2). Plants standing near the road showed higher concentrations of Al, As, Na, Pb, Sb, and V than those located in the park and in reference sites ( concentrations of Cr and Cu were also higher, but only relative to the reference site). All of them, with the exception ofNa, are traffic-related trace elements (Wei and Yang, 201 O; Carrero et al., 2013). We hypothesize that Na probably comes from NaCl salt scattered on sidewalks, which is common practice in Denmark to prevent ice farmation at this time of ye ar. In addition, growth chamber samples exhibited lower contents ofB, K, Mg, P, Se, Se, and Sr than outdoors sites. Most of these elements are recognized as majar seawater components, and it is therefare expected that marine aerosol, and consequently the particulate matter inter­cepted by outdoor vegetation, will be enriched in them (Dickson and Goyet, 1997; Zhao et al., 2015).

For the other elements, differences due to the location were not significant, probably because of the larger influence of the type of substrate and/ or the species. Plants tissues grown in ver­miculite had higher contents ofMg, Mo, and Ni, which could be attributed to a higher bioavailability in that substrate compared with sphagnum peat, since they are components of the nutrient solution and are therefare present in a free soluble farm. Barium, Co, Cr, Li, V, and Y concentrations were also higher, but their source must be the original vermiculite substrate, since theywere

VI e o

'.;::;

~ e ~ e o u ~ e Q)

E Q)

Qj Q) u

~ e "' Q)

E u

'.;::; Q)

E .s ~ .....:

e N

>-

>

Q) Vl

u Vl

.!J c..

c..

z

"' z

l..Ji ~ M 0 ~ ~ ~ M l..Ji ~ ~ ~ \O \O M q M 00 l..Ji ~ \O l..Ji ~ o o o o o ~ ~ o M

0 ~ ~ N M ~ ~ 00 0 0 00 0 g 0 0 0 0 0 0 N ~ N O V O O O O O O O O O O

8 8 8 8 g ~ ~ ~ ~ ~ ~ ~ o o o o o o o o o o o o 00 M \O M \O 0 N N 0 l..Ji M o o o o o o o o o o o o o o o o o (j o g g V V

0 0 0 0 0 0 N M N g g g g g g o o o o V V V V V V O O O O

g g 8 8 ~ o ~ g 00 ~ ~ ~ o o o o o o o o o o o o

~ '° ~ g ~ ~ ~ ~ ~ ~ ~ ~ O O N N O O 00 ~ o o 00 M

~ ~ 00 ~ 00 ~ ~ N ~ ~ ~ ~ ~

~ :E c~~~~~~~~~M~gs ""O

1 \O N M N 00 N l..Ji M O l..Ji

~ ~ ~ ~ O ~ ~ ~ ~ ~ ~ M ~ ~ ::2: O'l M.._ l..Ji.._ \O.._ 'tj".._ M.._ M l..Ji.._ 'tj".._ N.._ N.._

Q) LL

:::¡ u

"' u

VI

<C

+­Q)

o.. E "' Vl

e o

'.;::;

"' u o --'

E

N

00 N N

00 o N M

l..Ji '° 00 o '° ~ o o l..Ji ~ '° o ~ l..Ji ~ ('(") '° ~ q o o o o o o o o o N l..Ji 0 \O N 00 ~ \O l..Ji ~ ~ l..Ji 0 l..Ji N N 0 ~ l..Ji ~ 00 ~ o o o o o o o o o o o o

0 'tj" M M \O 0 l..Ji \O ~ N \O ~ l..Ji 0 0 \O N 0 0 00 M 0 0 o o o o o o o o o o o o

N M M 0 'tj" \O M \O 'tj" 00 N N N M N M M M N M M

8 o o o 8 8 o o 8 g ~ 8 o o o o o o o o o o o o

N N l..Ji 0 0 l..Ji \O 00 ~ M \O NN\.0000

N

Q) ""O ·¡;; ""O ro o ce

ce'

not added to the irrigation mixture (Papadopoulos et al., 2014). In contrast, several elements were enriched in the leaves of speci­mens grown in peat, among them trace elements that could have been interpreted as typically representative of urban environ­ments, such as Cd and Zn (Wei and Yang, 2010; Charlesworth et al., 2011). Metal ions are associated with different constituents in growth media, and thus the quantity available far plant uptake in peat (free elements in soil solution or in exchangeable frac­tions) may exceed the amount dissolved in the nutrient solution used far vermiculite watering.

Regarding plant species, rapeseed presented significantly higher levels of Cd, Co, Cu, Fe, Mn, and Zn, while only the concentrations of Ba and Sr were higher in samples of kale leaf. Even though kale has curly leaves and a wider surface area, which implies a higher capacity far trapping airborne particles, the con­centration relative to the total biomass might be lower. Thus, it seems that rapeseed plants would be more sensitive as bioindica­tors of trace elements atmospheric pollution.

Principal Component Analysis To assess the sampling adequacy of the factorial model,

Kaiser-Meyer-Olkin (KMO) and Barlett's sphericity tests were conducted. The farmer compares the observed correlation coef­ficients between variables and their partial correlations, whereas the latter is used to verify that the matrix of correlation coeffi­cients is different from the identity matrix. Acceptability criteria are set at KMO index > 0.5 far the KMO test and at p value < O.OS far the Barlett's sphericitytest. In this study, the KMO index was 0.672 and the p value < 0.001, respectively. Communalities related to the elements were all >0.8, except far Se (0.68) and Y (0.74), which are also acceptable values. The first faur principal components (PCl to PC4, witheigenvalues oflO.l, 8.0, 2.8, and 1.9, respectively) explained 88% of the total data variability. For a better understanding and to maximize the variance, a varimax rotation was perfarmed that resulted in the faur components and correlation coefficients shown in Table 3.

Samples could be separated rather well depending on the loca­tion and similarly on the type of substrate (Fig. 2 and 3). Principal Component 1 seems to group samples grown in the vermiculite and watered with the nutrient solution and those elements (Ba, Co, Li, Mg, Mo, Ni, and Y) associated with them. It also possible to subdivide the plants that have grown in vermiculite outdoors from those of the climatic chamber, which could be associated with the traffic-derived elements Ba, Co, and Ni. Principal Component 2 differentiates vegetables that have grown in sphagnum media and elements (Ca, Cd, Cu, Fe, Mn, P, and Zn) with a higher concentra­tion in this substrate. Principal Component 3 is essentially made up of traffic-related metal( oid)s (Al, As, Cr, Pb, Sb, and V, and to a lesser extent, Na, Se, and Se), showing positive coefficients far plants exposed near the road and negative far samples grown in a clean environment. As noted befare, Na is likely to come from salt spread on adjacent roads and sidewalks to prevent ice farmation, rather than traffic pollution. Principal Component 4 clustered sea­water ions (B, K, Se, and Sr), which could also include Mg and Na, since they show loadings >0.4, as well as samples that stayed in the growth chamber.

In summary, comparison between mean concentrations showed that sorne of the elements had more than one source, but PCA data allowed us to identify which factors were more

Table 2. Significantly different average trace element concentration between factors (e).

Factor Groupt Al As B Ba Ca Cd Co Cr Cu Fe K Li Mg Mn Mo Na Ni P Pb Sb Se Se Sr V Y Zn

Location R > P ** ** * ** ** ** R>G ** ** ** P>G

Substrate S >V

S<V

Species R > K

R< K

**

**

*

* *

** ** ** ** **

* ** ** *

*, ** Significant at the O.OS and 0.01 probability levels, respectively.

** **

** ** ** ** ** ** ** ** ** * ** ** ** ** **

** ** ** ** ** ** ** ** ** * **

** * *

t Location: G, growth chamber; P, park; R, roadside. Substrate: S, sphagnum; V, vermiculite. Species: R, rapeseed; K, kale.

relevant in certain cases. As a result, concentrations of Cr, Na, Sr, and V are influenced more by the location than by the type of substrate, as discussed above. Similarly, Cu is more strongly related to peat than to the other factors analyzed. On the other hand, differences between species seem to have a smaller influ­ence, so both could be used far biomonitoring.

Legal Values in Foodstuff and Animal Feed Cabbage is used as an ingredient in traditional Danish dishes

and is also commonly used in the gastronomy of many other countries. The European Commission has set maximum levels far certain contaminants in faodstuffs (European Commission, 2006). Of all the elements analyzed in the present study, only Cd and Pb have established values, which far leaf vegetables on a wet

Table 3. Rotated principal component (PC) loadings.

Component PC1 PC2

Eigenvalue 6.5 6.5

Variance (%) 2S.l 2S.l

Accumulated (%) 2S.l S0.2

Al 0.31 -0.01

As -0.07 0.14

B 0.31 0.29

Ba 0.62 -0.49

Ca -0.6S 0.67

Cd -0.38 0.88

Co 0.89 0.23

Cr 0.44 -0.24

Cu 0.38 0.79

Fe -0.lS 0.91

K 0.24 -0.02

Li 0.78 -0.47

Mg 0.82 -0.19

Mn -0.07 0.98

Mo 0.76 -0.46

Na -0.57 0.32

Ni 0.87 -0.19

p -0.14 0.82

Pb O.OS 0.03

Sb 0.07 0.03

0.31

0.28

0.37

0.11

PC3 PC4 Communality

S.9 3.9

22.7 1 S.O

72.9 87.9

0.86t 0.01 0.83

0.86 0.3S 0.88

0.31 0.7S 0.83

0.24 0.38 0.82

0.04 0.31 0.97

-0.08 0.08 0.92

0.28 -0.13 0.93

0.68 0.37 0.84

0.28 0.04 0.86

0.17 -0.04 0.89

0.32 0.84 0.86

0.26 0.18 0.93

0.32 0.42 0.98

-0.07 0.01 0.97

0.37 0.1 S 0.94

0.48 0.46 0.87

0.27 0.08 0.88

0.02 0.49 0.94

0.89 0.27 0.87

0.90 0.21 0.86

0.59

0.4S

Se

Se

Sr

V y

-0.53 0.32 0.23

0.32

0.74

0.68

o.os 0.41

0.16

0.68

0.8S

0.90

0.88

0.74

0.96

0.38 -0.06 0.8S

0.76 0.04 -0.01

Zn -0.21 0.94 -0.01

t Underlined values are elements with loadings ;o,0.6 (absolute value).

basis are 0.20 and 0.30 mg kg-1, respectively. A moisture content

of 84% (USEPA, 2011) and the worst case, which corresponds to plants growing near the road, were assumed to compare with legal values. Since average kale Cd and Pb concentrations were O.OS and 0.06 mg kg-1

, respectively, vegetables cultivated using an unpolluted substrate should be within safe levels far human consumption, although plants were only exposed to the outside air far a limited time of their total growing peri o d.

Contrary to cabbage, in the case of rapeseed, it is the oil extracted from the seeds that is used far human consumption. Leaves are considered a byproduct that is often transfarmed into fadder far livestock, and thus trace element concentrations in rapeseed leaves should be compared with regulatory levels of undesirable substances in animal feed. These regulatory levels, which include As together with Pb and Cd, have been estab­lished far a feed material with a default moisture content of 12% (European Parliament and the Council of the EU, 2002). The maximum permissible limits far As, Pb (green fadder), and Cd (feed materials of vegetable origin) are 2, 40, and 1 mg kg-1

,

respectively. Average As, Pb, and Cd concentrations in rapeseed in this study were 0.02, 0.12, and 0.32 mg kg-1

, respectively, two orders of magnitude below their corresponding threshold values far As and Pb. In the case of Cd, however, one sample slightly exceeded the permissible limit with a concentration of 1.04 mg kg-1

, a fact that warrants a regular control of rapeseed grown in urban gardens near roads if the produce is used far animal feed.

In terms of human health, products grown in an urban area using an uncontaminated substrate would comply with faod safety legislation far direct human consumption, although it would be desirable that health protection agencies establish regulatory values far the rest of elements included in the study. Other aspects should also be considered, such as longer exposure periods, possible seasonal variations in air pollution ( e.g., winter temperature inversions, which may lead to an accumulation of pollutants in the lower layers of the atmosphere, or drought periods that prevent rain from removing airborne pollutants by coagulation), and the fact that products may be washed befare consumption, resulting in the removal of part of the particulate matter deposited on the leaf surface, although the latter action would be advisable, as it would lead to safer levels.

Conclusions Analyses of 27 elements in unwashed leaves of oilseed rape

and kale grown in two substrates (vermiculite and peat) and placed in urban locations with three different levels of atmo­spheric pollution have revealed that, as expected, plants growing near congested roads accumulated larger quantities of elements

• • • .. 10 ·

o •• • "-ll'

• • ... ... .. ... .M.. ... ... ...

-10 · -5 10 15

RPC1

... • ...... ... 5- ... •• • • ...

"' .4 .... ~ o ll' •

• -5 - .. • ·;

10 20 RPC3

Fig. 2. Rotated principal component (RPC) sample seores plotted against Principal Component 1 (PC1) vs. PC2 and PC3 vs. PC4 grouped by sub­strate (vermiculite: triangles; sphagnum: circles) and by location (growth chamber: blue; park: green; roadside: red).

1.0 -

~

'd 0.5 -

Se

~ ii As ~ RISO i

N *~ o "-ll'

o.o

Gr llii ill

--0 .5 - lli • -0-5 0 0 05

RPC1

~ 0.75 -

i ' ~ ••

0.50- ~ ('j "-ll'

lliª llM9 * ª

Se ii.' As

l?b 0.25- .¡

llMº ¿S,b

"11

0.00 ,Mn tu

llͺ

º-ºº 0 25 0 50 0.75

RPC3

Fig. 3. Rotated principal component (RPC) trace elements loadings for Principal Component 1 (PC1) vs. PC2 and PC3 vs. PC4 (vermiculite: red squares; sphagnum: green circles; traffic-related: orange triangles; seawater: blue diamonds).

associated with traffic sources. Compared with the reference site, both park and roadside samples exhibited higher concentrations of sea salt ions. Although sorne specific elements showed pref­erential enrichment depending on the substrate on which the plant had grown, increased concentrations of most pollutants in areas with high traffic could be detected with both the substrates tested in this study. Lastly, since plant species seems to have a lower impact on results, kale could be used far biomonitoring purposes, given that its cultivation is more common

These findings lead to the conclusion that the analysis of atmospheric particles absorbed and deposited on edible plants could be an easy, economic, and reliable technique to biomoni­tor atmospheric particulate pollution in urban environments. The results of this study also indicare that although the potential use of plants grown in urban gardens as biomonitors of atmo­spheric pollution is based on their capacity to collect a sufficient amount of contaminant on their leaves, the concentration of trace elements in those leaves fall below the existing European regulatory limits far faodstuff and animal fadder (with the pos­sible exception in the latter case of Cd in urban gardens near roads with elevated traffic levels). In fact, the routine analyses of trace element concentrations as part of the biomonitoring effort would also allow the detection of gardens that should implement protective measures.

To validare the method far global application, future research should address sorne uncertainties arising from the limited data­base available in this study, which did not incorporare inland locations or more levels of traffic density. The possible influence of different weather conditions, such as other seasons and other regional climates, should also be explored.

Supplemental Material Sphagnum peat moss substrate and nutrient solutions chemi­

cal compositions used in the present study can be faund in the supplemental material.

Acknowledgments M. Izquierdo is grateful to the Consejo Social de la Universidad Politécnica de Madrid for the Ph.D. research stay fellowship ( l 4th Call) and to the Caracterización, Remediación, Modelización y Evaluación del Riesgo de Suelos Contaminados (CARESOIL-CM, S2013/ MAE-2739) research grant of the Regional Government of Madrid (Comunidad de Madrid, Spain ).

References Al-Khashman, O.A., A.H. Al-Muhtaseb, and K.A. lbrahim. 2011. Date palm

(Phoenix dactylifera L.) leaves as biomonitors of atmospheric metal pollu­tion in arid and semi-arid environments. Enviran. Pollut. 159: 1635-1640. doi: 10.1O16/j.envpol.2011.02.045

Alloway, B.J. 2004. Contamination of soils in domestic gardens and al­lotments: A brief overview. Land Contam. Reclam. 12:179-187. doi: 10.2462/09670513.658

Augusto, S., C. Máguas, and C. Branquinho. 2013. Guidelines for biomonitoring persistent organic pollutants (POPs), using lichens and aquatic mosses A review. Enviran. Pollut. 180:330-338. doi:l0.1016/j.envpol.2013.05.019

Bargagli, R. 2016. Moss and lichen biomonitoring of atmospheric mer­cury: A review. Sci. Total Enviran. 572:216-231. doi:l0.1016/j. scitotenv.2016.07.202

Bonanno, G. 2014. Ricinus communis as an element biomonitor of atmospheric pollution in urban areas. Water Air Soil Pollut. 225:1852. doi:l0.1007/ s 11270-013-18 52-2

Bonanno, G., and P. Pavone. 2015. Leaves of Phragmites australis as poten ti al at­mospheric biomonitors of platinum group elements. Ecotoxicol. Enviran. Saf. 114:31-37. doi:l0.1016/j.ecoenv.2015.01.005

Carrero, J.A., l. Arrizabalaga, J. Bustamante, N. Goienaga, G. Arana, and J.M. Madariaga. 2013. Diagnosing the traffic impact on roadside soils through a multianalytical data analysis of the concentration profiles of traffic­related elements. Sci. Total Enviran. 458-460:427-434. doi:l0.1016/j. scitotenv.2013.04.047

Charlesworth, S., E. De Miguel, andA. Ordóñez. 2011. A review of the distribu­tion of particulate trace elements in urban terrestrial environments and its application to considerations of risk. Enviran. Geochem. Health 33: 103-123. doi: 10.l007/sl0653-010-9325-7

Cosma, C., A.R. lurian, R. lncze, T. Kovacs, and Z.S. Zunié. 2016. The use of tree bark as longterm biomonitor of137Cs deposition. J. Enviran. Radioact. 153: 126-133. doi: 10.1O16/j.jenvrad.2015.12.019

Daud, M., N. Khalid, S. Waheed, M. Wasim, M. Arif, and J.H. Zaidi. 2011. Morus nigra plant leaves as biomonitor for elemental air pollution moni­toring. Radiochim. Acta 99:243-252. doi:l0.1524/ract.2011.1814

Daud, M., M. Wasim, N. Khalid, and S. Waheed 2015. Pinus roxburghii plant needles as a three-season biomonitor for elemental air pollution moni­toring along roadside. Radiochim. Acta 103:737-743. doi:l0.1515/ ract-2015-2421

de Paula, P.H.M., V.L. Mateus, D.R. Araripe, C.B. Duyck, T.D. Saint'Pierre, and A. Gioda. 2015. Biomonitoring of metals for air pollution assess­ment using a hemiepiphyte herb (Struthanthus flexicaulis). Chemosphere 138:429-43 7. doi: 10.1O16/j.chemosphere.2015.06.060

Dickson, A.G., and C. Goyet. 1997. Physical and thermodynamic data. In: Handbook of methods for the analysis of carbon dioxide parameters in sea water. 2.13. ORNL/CDIAC-74. USDOE, Washington, DC. p. 1-22.

Esen, A.N., M. Kubefová, S. Haciyakupoglu, and J. Kucera. 2016. Instrumen­tal neutron activation analysis of plant tissues and soils for biomonitor­ing in urban areas in lstanbul. J. Radioanal. Nucl. Chem. 309:373-382. doi: 10.1007Is10967-016-4750-4

European Commission. 2006. Commission Regulation (EC) no. 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs. Eur. Commission, Brussels.

European Parliament and the Council of the EU. 2002. Directive of the Euro­pean Parliament and the Council of7 May 2002 on undesirable substances in animal feed 2002/32/EC. Pub!. Office Eur. Union, Luxembourg.

Giampaoli, P., E.D. Wannaz, A.R. Tavares, and M. Domingos. 2016. Suit­ability of Tillandsia usneoides and Aechmea fasciata for biomonitoring toxic elements under tropical seasonal climate. Chemosphere 149:14-23. doi: 10.1O16/j.chemosphere.2016.01.080

Guéguen, F., P. Stille, M.L. Geagea, and R. Boutin. 2012. Atmospheric pol­lution in an urban environment by tree bark biomonitoring-Part 1: Trace element analysis. Chemosphere 86:1013-1019. doi:l0.1016/j. chemosphere.2011.11.040

Guzmán-Morales, J., O. Morton-Bermea, E. Hernández-Álvarez, M.T. Rodrí­guez-Salazar, M.E. García-Arreola, and V. Tapia-Cruz. 2011. Assessment of atmospheric metal pollution in the urban area of Mexico City, using Ficus benjamina as biomonitor. Bull. Enviran. Contam. Toxico!. 86:495-500. doi: 10.l007/s00128-0l l-0252-9

Hassan, LA., and J.M. Basahi. 2013. Assessing roadside conditions and vehicular emissions using roadside lettuce plants. Pol. J. Enviran. Stud 22:387-393.

Ili, P., N. Keskin, and F. Sari. 2016. lnvestigation of the particulate matters on the leaves of Platanus sp. in Denizli, Turkey using FESEM-EDS. Fresenius Enviran. Bull. 25:2393-2403.

Karaaslan, N.M., and M. Yaman. 2013. Determination of nickel and chromium in Pinus nigra l., Cedrus libani, and Cupressus arizonica leaves to monitor the effects of pollution in Elazig (Turkey). lnstrum. Sci. Technol. 41:335-348. doi: 10.1080/10739149.2012.726685

Kleckerová, A., and H. Docekalová. 2014. Dandelion plants as a biomonitor of urban area contamination by heavy metals. lnt. J. Enviran. Res. 8: 157-164. doi: 10.22059/IJER.2014.705

Klumpp, A., W. Ansel, G. Klumpp, N. Belluzzo, V. Calatayud, N. Chaplin, et al. 2002. EuroBionet: A pan-European biomonitoring network for urban air quality assessment. Enviran. Sci. Pollut. Res. lnt. 9: 199-203. doi: 10.1007 / BF.02987489

Klumpp, A., W. Ansel, G. Klumpp,J. Breuer, P. Vergne, M.J. Sanz, et al. 2009. Airborne trace element pollution in 11 European cities assessed by ex­posure of standardised ryegrass cultures. Atmos. Enviran. 43:329-339. doi: 10.1O16/j.atmosenv.2008.09.040

Kord, B., A. Mataji, and S. Babaie. 2010. Pine (Pinus eldarica Medw.) needles as indicator for heavy metals pollution. lnt. J. Enviran. Sci. Technol. 7:79-84. doi: 10.1007 /BF.03326119

Kulshreshtha, K., A. Rai, C.S. Mohanty, R.K. Roy, and S.C. Sharma. 2009. Par­ticulate pollution mitigating ability of sorne plant species. lnt. J. Enviran. Res. 3: 137-142. doi: 10.22059 /IJER.2009.42

Madrid, F., M. Biasioli, and F. Ajmone-Marsan. 2008. Availability and bioacces­sibility of metals in fine particles of sorne urban soils. Arch. Enviran. Con­tam. Toxico!. 55:21-32. doi:l0.l007/s00244-007-9086-1

Papadopoulos, A., K. Giouri, E. Tzamos, A. Filippidis, and S. Stoulos. 2014. Natural radioactivity and trace element composition of natural clays used as cosmetic products in the Greek market. Clay Miner. 49: 53-62. doi: 10.1180/claymin.2014.049.1.05

R Core Team. 2017. R: A language and environment for statistical computing. R Found Stat. Comput., Vienna.

Ram, S.S., S. Majumder, P. Chaudhuri, S. Chancla, S.C. Santra, P.K. Maiti, et al. 2014. Plant canopies: Bio-monitor and trap forre-suspended dust par­ticulares contaminated with heavy metals. Mitig. Adapt. Strategies Glob. Change 19:499-508. doi:l0.1007/sl 1027-012-9445-8

Rinaldi, M.C.S., M. Domingos, A.P.L. Dias, J.B.N. Esposito, and J.D. Pagliuso. 2012. Leaves of Lolium multiftorum "Lema" and tropical tree species as biomonitors of polycyclic aromatic hydrocarbons. Ecotoxicol. Enviran. Saf. 79:139-147. doi:l0.1016/j.ecoenv.2011.12.013

Rodriguez,J.H., S.B. Weller, E.D. Wannaz, A. Klumpp, and M.L. Pignata. 2011. Air quality biomonitoringin agricultura! areas nearby to urban and indus­trial emission sources in Córdoba Province, Argentina, employing the bio­indicator Tillandsia capillaris. Eco!. Indic. 11:1673-1680. doi:l0.1016/j. ecolind.2011.04.015

Sawidis, T.,J. Breuste, M. Mitrovic, P. Pavlovic, and K. Tsigaridas. 2011. Trees as bioindicator of heavy metal pollution in three European cities. Enviran. Pollut. 159:3560-3570. doi:l0.1016/j.envpol.2011.08.008

Severoglu, Z., 1.1. Ozyigit, l. Dogan, G. Kurmanbekova, G. Demir, LE. Yalcin, and G.K. Kari. 2015. The usability of Juniperus virginiana L. as a biomoni­tor ofheavy metal pollution in Bishkek City, Kyrgyzstan. Biotechnol. Bio­technol. Equip. 29:1104-1112. doi:l0.1080/13102818.2015.1072478

Urosevié, M.A., G. Vukovié, P. Jovanovié, M. Vujicié, A. Sabovljevié, M. Sabovljevié, and M. Tomasevié. 2017. Urban background of air pollu­tion: Evaluation through moss bag biomonitoring of trace elements in botanical garden. Urban For. Urban Green. 25:1-10. doi:l0.1016/j. ufug.2017.04.016

USEPA. 2011. Exposure factors handbook. 2011 ed. EPA/600/R-09/052F. USEPA, Washington, DC.

Wei, B., and L. Yang. 201 O. A review of heavy metal contaminations in urban soils, urban road dusts and agricultura! soils from China. Microchem. J. 94:99-107. doi: 10.1O16/j.microc.2009.09.014

WHO. 2016. Ambient air pollution: A global assessment of exposure and bur­den of disease. World Health Org., Geneva.

Zhao, R., B. Han, B. Lu, N. Zhang, L. Zhu, and Z. Bai. 2015. Element compo­sition and source apportionment of atmospheric aerosols over the China Sea. Atmos. Pollut. Res. 6: 191-20 l. doi:l 0.5094/ APR.2015.023