Assessment of heavy metal contamination in soil and produce of Philadelphia community gardens

ABSTRACT

Urban and ex-urban residents have been increasingly utilizing community gardens to supplement their diets, foster relationships with neighbors and learn new skills. Soils in urban-region community gardens, however, can be detrimental to human health if contaminated with metals. In this study, the soils of 20 Philadelphia-region community gardens (and produce from 6 gardens) were analyzed for heavy metal content (As, Cd, Cr, Co, Cu, Ni, Pb) to assess bioavailability, determine relationships with environmental and demographic variables and compare with published safe limits. About 58% of soil samples and 86% of produce samples were above their respective safe lead level. Metal concentrations in garden produce differed between produce types, with the highest concentrations of As, Cr, Co and Cu found in root and leafy vegetables and the lowest concentrations found in fruiting vegetables. Philadelphia-region community gardeners are exposed to unsafe levels of metals both from the soil and from consumed produce.

KEYWORDS:

• Heavy metals
• soil contamination
• vegetable contamination
• urban gardens
• plant uptake
• health risk
• inductively coupled plasma mass spectrometry

1. Introduction

Community gardens in urban areas have become popular ways to provide local residents with fresh produce while also providing jobs, education and prompting community engagement [1]. Prior research has shown that community garden participants notice an increase in their physical and mental well-being, and report increased social connections in their communities [2,3]. In addition, gardens improve air quality, retain stormwater, increase biodiversity, lessen the need for long-distance food distribution, diminish food waste and address social issues like food insecurity and access to green space [4–6]. Urban food deserts, in which people live over 1 mile away from a grocery store, disproportionately affect census tracts with high poverty rates, high minority percentages and lower levels of education [7]. The presence of community gardens in these tracts can alleviate the stress of food insecurity, reduce money spent on groceries and improve the nutrition of these populations.

An obstacle to the optimal functionality of urban community gardens, and the health of those that depend on them, is the contamination of soil and produce. Common soil contaminants include asbestos, organic compounds (agrochemicals, hydrocarbons, PAHs, petroleum products, VOCs) and heavy metals [8–11]. Heavy metals are naturally occurring in soils but typically at levels far below concern for human health. Anthropogenic activities can lead to the accumulation of heavy metals in the soils of urban locations [1,12]. Anthropogenic sources of heavy metals include fossil fuel combustion, corrosion of metal structures, application of fertilizers and pesticides, industrial activity and the disposal of municipal and industrial waste [13]. Past land use can affect contemporary metal concentrations, as metals can remain in the soil long after they are introduced. In Philadelphia, the legacy of industrial activity, such as lead smelting and the use of leaded-gasoline and lead-based paints, has contributed to the elevated levels of lead (Pb) and various other heavy metals in soils throughout the region [14,15].

Heavy metals in soil can be assimilated into plant tissues. Intawongse and Dean [16] observed increasing metal concentration in plants with increasing corresponding soil metal concentration, though uptake rates varied with plant type and metal. Plants accumulate heavy metals preferentially in root, leaf and stem tissues, and therefore the edible parts of leafy and root vegetables often have higher metal concentrations than those of fruiting vegetables [17,18]. Consumption of produce with high concentrations of heavy metals can be detrimental to human health. For example, arsenic (As) is known to cause skin disorders, diabetes, hyperpigmentation, high blood pressure, keratosis, cancer and can lead to impaired nerve function [19,20]. Consumption of cadmium (Cd) can cause stomach irritation, vomiting, diarrhea and deterioration of bones [21]. Consumption of small amounts of chromium (Cr) can lead to irritation of the stomach and small intestine, respiratory issues, anemia, sperm damage and cancer [22]. Prolonged exposure to Pb, even at low levels, can cause depression, memory loss, anemia, kidney damage, brain damage, high blood pressure, heart disease, reduced fertility, cancer, neurotoxicity, impaired hemoglobin synthesis and impaired reproductive functionality [23,24].

There are a variety of published safe levels for metals in soils that range considerably. Regulatory values of metal concentrations exist for residential soils and as standards following remediation projects, but standards for agricultural soils were difficult to find and ranged widely both domestically and internationally. Soil remediation threshold values ranged from more lenient, like those determined by the Pennsylvania Department of Environmental Protection [25], to conservative, like those dictated by Finland’s Ministry of the Environment [26] (Table 1). Safe levels specifically for gardening soils, like those set by the City of Toronto’s Public Health, were generally more conservative than post-remediation regulatory values. Toronto Public Health [27] has two published soil screening values (SSVs) for varying levels of contamination. If metal concentrations are above SSV1, the site is of medium concern and if concentrations are above SSV2, the site is of high concern. The U.S. EPA [28] has set a safe level of Pb (and no other metals) in gardening soils that is over 3 times the value set by Toronto Public Health (Table 1). The validity of applying soil remediation and residential standards to gardening soils is unknown but included in this study as a basis of evaluation. Safe metal concentrations in vegetables are also elusive. The China Food and Drug Administration (translation provided by [29]), as well as a joint report from the Food and Agriculture Organization and the World Health Organization [30], has published recommended safe levels in four metals that differ with vegetable type (leafy, root, fruiting, etc.) (Table 2). The lack of established safety standards for metal contamination in both soils and produce makes regulation and enforcement of safe contaminant levels impossible on a broad scale, posing risk to urban gardening communities.

Table 1. Safe levels (µg/g) of As, Cd, Cr, Co, Cu, Ni and Pb in soils, as published by various sources. x indicates that no levels were reported.

Heavy metal	Finland	US EPA	PADEP	TSSV1	TSSV2
As	5	x	12	11	110
Cd	1	x	110	1	10
Cr	100	x	x	390	630
Co	20	x	66	23	170
Cu	100	7200	x	180	660
Ni	50	x	4400	34	340
Pb	60	100	500	34	340
V	100	x	15	x	x

Finland: values from Finland Ministry of the Environment [26].

US EPA: values from U.S. EPA [23].

PADEP: values from Pennsylvania Department of Environmental Protection [25].

TSSV1: Soil screening value 1 from Toronto Public Health [27].

TSSV2: Soil screening value 2 from Toronto Public Health [27].

Table 2. Maximum safe levels of As, Cr, Cd and Pb in fresh vegetables.

Heavy metal	Regulatory standard (µg/g)	Vegetable type
As	0.5^a	Fresh Vegetables
Cr	0.5^a	Fresh Vegetables
Cd	0.05^ab	Fresh Vegetables
	0.1^ab	Stem and Root Vegetables
	0.2^ab	Leafy Vegetables
Pb	0.05^b	Fruiting
	0.1^b	Bulb, Root and Tuber Vegetables
	0.1^b, 0.2^a	Leguminous Vegetables
	0.1^b, 0.3^a	Brassica Vegetables
	0.3^ab	Leafy Vegetables

a: China’s Maximum Levels for Contaminants in Foods as translated by the USDA Foreign Agricultural Service [29].

b: FAO/WHO [30] General Standard for Contaminants and Toxins in Food and Feed.

ab: denotes the standards were the same for that respective vegetable type.

The benefits of urban gardening must be weighed against potential exposure to pollutants such as heavy metals. Best practices in garden construction and operation can alleviate pollutant exposure [28], though practical and financial obstacles may exist for populations that benefit most from the services provided by urban gardens. In this study, we assess trace metal contamination of produce and soil in urban and suburban community gardens in the Philadelphia region. Philadelphia has long been considered one of the poorest big cities in the U.S., and 2021 estimates from the U.S. Census Bureau report that 23.1% of people in the city of Philadelphia are living in poverty [31]. A report by the Philadelphia Department of Public Health revealed that access to healthy food decreases considerably in areas of high poverty as well as in areas with a greater black population [32]. Philadelphians of lower socioeconomic status are also more likely to face environmental burdens and suffer from the city’s industrial legacies [33]. Urban gardens in Philadelphia can provide fresh produce and green space in communities lacking access to markets, but community garden produce may also be a route of contaminant exposure to these same communities.

In this study, we investigate how metal contamination in Philadelphia urban gardens relates to geographic and demographic factors and examine how heavy metals in produce from urban gardens compare with commercially grown produce obtained from grocery stores. Our findings inform managers and stakeholders about factors to be considered when weighing the benefits and potential risks of urban gardening.

2. Materials and methods

2.1 Sample collection

Twenty community gardens (referred to as Gardens A – T) within five Southeastern Pennsylvania counties were selected for the study. Fifteen of the gardens were in the city of Philadelphia, and another five gardens were in the Western suburbs. The exact geographic location of sampling sites is not disclosed to protect the autonomy of gardens and maintain privacy. Soil samples were collected from all 20 community gardens. At six of the gardens (five urban and one suburban) produce samples were also collected. Produce from grocery store chains in the Philadelphia region was obtained. Garden and grocery samples were all collected between September and December 2021.

A total of 36 produce and 103 soil samples were collected. In the six gardens where both soil and produce were obtained, samples were collected and mailed by gardeners at each of the sites. Gardeners were provided with instructions on sample collection techniques to preserve the integrity of each sample and minimize heavy metal contamination. All samples were collected with plastic utensils and stored in polythene bags prior to analysis. Multiple soil samples were taken at each garden (n = 2–12 from individual garden). For each produce sample taken, a soil sample was taken from the first 10 cm of topsoil directly adjacent to that plant. All soils analyzed in this study are from actively used garden areas. Twenty different types of produce were obtained from six gardens. The seven most common produce types, which were collected from at least two gardens(basil, carrots, cherry tomatoes, collard greens, kale, romaine lettuce and tomatoes), were also purchased from grocery stores. Six commercially available produce samples (three different conventional brands and three organic brands) of each of the seven most common produce types were purchased from several different grocery store chains in the region. All vegetables sampled are considered either leafy (kale, collard greens, romaine lettuce, basil), root (carrot, potato, radish) or fruiting (tomato, pepper, squash).

2.2 Heavy metal analysis

Soil samples and the edible portion of each produce sample collected from gardens and grocery stores were dried (60°C for a minimum of 24 hours), ground and homogenized. Large organic matter and other debris were removed from the soil prior to analysis. Produce and soil samples were microwave digested in Teflon tubes with SCP Science 67–70% PlasmaPure nitric acid ($\mathrm{H}\mathrm{N}{\mathrm{O}}_3$) in a CEM $\mathrm{M}\mathrm{a}\mathrm{r}{\mathrm{s}}^{\mathrm{T}\mathrm{M}}$ 6 One Touch microwave digestion system. The digest method ramped the temperature to 175°C over a 20-to-25-minute period and held the temperature at 175°C for 15 minutes. A 0.5 g subsample of each soil and produce sample was digested in 10 ml (for soil) or 5 mL (for produce) $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$. The amount of $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$ used was adjusted for produce samples with less than 0.5 g of material to maintain an appropriate sample to acid ratio based on the equation:

$\mathrm{\%}\mathrm{H}\mathrm{N}{\mathrm{O}}_3=\frac{\mathrm{m}\mathrm{L}\mathrm{H}\mathrm{N}{\mathrm{O}}_3}{\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\left(\mathrm{m}\mathrm{L}\right)}$

A blank (with $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$ but no sample) was run during each digestion. Blank equivalents were calculated by multiplying average dilution-corrected metal concentrations for blank digests by the average dilution factor. These values ranged from −0.1365 µg/g to 6.8443 µg/g and were far below measurements made for both produce and soil samples, ensuring that no contamination was introduced during the digestion process (Table S1).

Following digestion, samples were diluted from 40% $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$ to 4% $\mathrm{H}\mathrm{N}{\mathrm{O}}_3$ concentration with 18.2 mΩ deionized water based on the equation:

$\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}{\mathrm{n}}_1\ast \mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}{\mathrm{e}}_1=\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}{\mathrm{n}}_2\ast \mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}{\mathrm{e}}_2$

The dilution factor (DF) was calculated based on the equation:

$\mathrm{D}\mathrm{F}=\left(\frac{\left(\mathrm{T}\mathrm{u}\mathrm{b}\mathrm{e}+\mathrm{S}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}+\mathrm{W}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\left(\mathrm{g}\right)\right)-\mathrm{T}\mathrm{u}\mathrm{b}\mathrm{e}\left(\mathrm{g}\right)}{1000}\hspace{0.17em}\right)/\mathrm{S}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\left(\mathrm{g}\right)$

In preparation for heavy metal analysis, samples were filtered at 0.45 µm. Arsenic (As), cadmium (Cd), chromium (Cr), cobalt (Co), copper (Cu), lead (Pb) and nickel (Ni) in produce and soil digests were analyzed on an Agilent 7900 Inductively Coupled Plasma Mass Spectrometer (ICP-MS). Samples were compared to standards of known concentration and adjusted for matrix effects and instrument drift using an internal standard. The ICP-MS reported limits of detection for each metal (Table S1). Values that fell below reported detection limits were considered to be the limit value itself.

2.3 Demographic and environmental factors

Demographic and environmental data were collected for the area surrounding each of the garden sampling sites. Data were derived at census tract level. Data on traffic proximity, lead paint and percent of people of color were taken from the U.S. EPA’s EJ Screen tool [34]. The lead paint indicator is the percentage of houses built prior to 1960, when lead paint was still used. Proximity to traffic was measured by the total length of nearby roads and density of traffic. Datasets for Pennsylvania local and state roads were taken from Pennsylvania Spatial Data Access [35,36]. Data on population density was taken from the U.S. Census’s 2020 Census Demographic Data Map Viewer [37]. Analysis was performed in ArcMap 10.8.2 to calculate total lengths of roads in census tracts.

2.4 Data analysis

Statistical analysis was performed using IBM SPSS Statistics software, Version 27 (α = 0.05). Data were uploaded to IBM SPSS Statistics software, and various tests were run using sample metal concentrations in conjunction with corresponding environmental and social data, which were assigned to each sample based on garden location. A Pearson’s correlation was performed to determine the relationship between metals found in the produce and the corresponding soil in which that plant was growing. Pearson’s correlations were also used to distinguish relationships among metals found in produce samples and then separately for metals in the soils. A multiple linear regression was performed to determine how four factors (proximity to traffic, percentage of housing units built prior to 1960, percent population of color within census tract, and whether soils came from a raised bed or from the ground) influence each metal concentration in the soil. A t-test was used to determine differences in urban and suburban soil metals. A t-test was used to determine a difference between metal concentrations in garden produce and grocery produce. A Kruskal–Wallis test was used to assess the difference in metal concentrations in leafy, root and fruiting vegetables.

A transfer factor (TF) denotes the amount of metal in soil that is assimilated by a plant growing in that soil [38]. The equation of Cui et al. [39] was used to determine the TF in which the dry weight of the produce was related to the dry weight of the metal in the soil:

$\mathrm{T}\mathrm{F}=\frac{\mathrm{M}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{b}\mathrm{l}\mathrm{e}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t}(\mathrm{d}\mathrm{r}\mathrm{y}\mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t})}{\mathrm{M}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}\mathrm{f}\mathrm{r}\mathrm{o}\mathrm{m}\mathrm{r}\mathrm{o}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{o}\mathrm{f}\mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t}(\mathrm{d}\mathrm{r}\mathrm{y}\mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t})}$

3. Results

3.1 Heavy metals in soils

Heavy metal concentrations in garden soils tended to be positively correlated across the different metals (Table 3). For instance, concentrations of Pb were positively correlated with both As and Cd (p < 0.05; Table 3). Chromium, Cu and Ni in the soil appeared to form a group of heavy metals that tended to be positively correlated across soil samples (p < 0.05; Table 3). Cadmium had the lowest mean soil concentration among the metals we analyzed, and Pb demonstrated the highest average concentration (Table S2). Thirteen soil samples (12.5%) from urban and suburban gardens had Cd concentrations that exceeded the most conservative safe limit. In contrast, 56 samples (71.1%) exceeded the lowest recommended safe limit of Pb, while 9 samples (8.7%) exceeded the least strict safe limit (Table S2). The mean Pb concentration in Garden C exceeded the EPA standard by more than 3.5 times and the Finnish standard by an order of magnitude (Table S2). All but two urban community gardens had unsafe soil Pb concentrations. A few gardens had average soil Cd concentrations above the City of Toronto’s SSV1, while many gardens had average Pb concentrations above SSV1 and some that exceeded SSV2 (Figure 1).

Figure 1. Average (± SD) concentrations of arsenic (As, panel a), cadmium (Cd; panel b), chromium (Cr, panel c) and lead (Pb, panel d) in soils collected from twenty urban and suburban community gardens in the Philadelphia region. The average for urban (n = 15) and suburban (n = 5) gardens is shown. The City of Toronto Public Health soil screening value (TSSV) level 1 (for As, Cd and Pb) and level 2 (for Pb) are indicated for reference (City of Toronto Public Health data).

Table 3. Pearson correlation between metals in soil.

Metal	As	Cd	Cr	Co	Cu	Ni	Pb
As	1	0.253	−0.298	−0.183	−.500**	−.492**	.337*
Cd		1	0.007	−0.021	0.192	.360*	.879**
Cr			1	−0.034	.652**	.426**	−0.071
Co				1	−0.083	.561**	0.112
Cu					1	.509**	0.177
Ni						1	0.294
Pb							1

*Indicates significance of <0.05.

**Indicates significance of <0.001.

Concentrations of Cd, Cr, Cu and Pb were significantly greater in the soils of urban community gardens than those of suburban community gardens (p < 0.05). Concentrations of As and Pb were significantly greater in soil from in-ground gardens, while concentrations of Cr and Co were significantly greater in raised beds (p < 0.05).

A combination of four social and environmental variables (proximity to traffic, percentage of housing units built prior to 1960, percent population of color within census tract, and whether soils came from a raised bed or from the ground) were able to predict between 5.7% and 62.8% of the variation in concentrations of As, Cd, Co, Cu, Ni and Pb in urban community garden soils (Table 4). These variables did not significantly predict soil Cr concentrations. The relationship was particularly strong for soil Pb concentrations, where these four variables were each significant and accounted for 62.8% of the variability in soil Pb concentrations. Increased concentrations of Pb were found in locations with higher proximity to traffic, higher percent people of color, a larger fraction of houses built prior to 1960 (and thus subject to lead paint) and from sites with in-ground gardens (as opposed to gardens with raised beds) (Figure 2).

Figure 2. The relationship between community garden soil lead concentrations and traffic proximity [panel a [34]], housing age as indicated by the EPA’s lead paint index [panel B [34]], the percent people of color [panel c [34]], and the average (± SD) concentration of soil lead in raised beds versus in-ground gardens (panel d).

Table 4. Significant coefficients from multiple linear regression of five variables and metal content in soil of urban community gardens.

Metal	% People of color	Traffic proximity	Raised or in-ground	Lead paint indicator	Constant	R2	p
As	ns	ns	0.908	4.337	0.945	0.165	0.001
Cd	ns	ns	ns	1.067	0.044	0.148	0.003
Cr	ns	ns	ns	ns	ns	ns	ns
Co	0.048	ns	−0.782	−6.251	11.309	0.228	<0.001
Cu	−1.245	−0.005	25.874	ns	141.327	0.162	0.005
Ni	ns	0.001	ns	ns	20.748	0.057	0.038
Pb	3.834	0.006	114.096	413.297	−507.501	0.628	<0.001

ns denotes no significance.

3.2 Heavy metals in produce

Cobalt in urban garden produce samples had the lowest mean concentration among all the measured trace metals, while Cu had the highest (Table S3). Out of the 36 total garden produce samples, 31 exceeded recommended safe levels for Pb, and 20 exceeded safe levels for Cd. The percentage of samples that exceeded maximum safe levels was 8.3%, 55.6%, 19.4% and 86.1% for As, Cd, Cr and Pb, respectively. All analyzed metals that had published safe limits in vegetables (As, Cd, Cr, Pb) were found to be exceeded. The highest Pb concentration of 36.3 µg/g was found in radish from Garden C. Garden C had the highest average Cd, Cr and Pb concentrations out of all the gardens (Figure 3).

Figure 3. Average (± SD) concentrations of arsenic (As; panel a), cadmium (Cd; panel a), chromium (Cr; panel a) and lead (Pb; panel b) in produce collected from six urban (n = 5) and suburban (n = 1) community gardens in the Philadelphia region. The average for all garden samples and the average for all produce samples collected are shown.

Concentrations of Cd and Pb followed the expected pattern of concentration across type of produce (leafy > root > fruiting). Root vegetables had the highest average As, Cr and Co concentrations. Kruskal–Wallis tests found strong evidence (p < 0.001) of a difference between the mean ranks of metal concentration across vegetable type (leafy, root, fruiting) for As and Cr and moderate evidence (p < 0.05) for that of Co and Cu concentrations. The tests for As and Cr indicated a significant difference between leafy and fruiting vegetables and between root and fruiting vegetables (but not between leafy and root vegetables). Dunn’s pairwise tests revealed that As and Cr concentrations in both leafy vegetables and root vegetables were greater than those in fruiting vegetables (p < 0.001 with Bonferroni correction). A Dunn’s test for Co indicated a significant difference between leafy and fruiting vegetables (p < 0.05), with leafy vegetables having greater concentrations. A Dunn’s test for Cu indicated a significant difference between root and leafy vegetables and between root and fruiting vegetables (p < 0.05). Dunn’s pairwise test revealed that both leafy and fruiting vegetables had significantly greater Cu concentrations than root vegetables.

Linear regressions suggested that several of the same variables that were significant predictors of trace metal concentration in soils also were related to metal concentrations in community garden produce. We found that the total length of local roads within 0.25 square miles of each garden explained 18.9%, 13.5% and 18.5% of Cd, Cr and Pb concentration variability in garden produce (p < 0.05). The prevalence of housing units built prior to the 1960s, which are more likely to contain traces of lead paint, explained 17.8% of variability in Cd concentrations and 18.5% of Pb variability in garden produce concentrations (p < 0.05). The percentage of people of color in a census tract explained 28.2%, 20.6%, 14.3% and 24.7% of variability in Cd, Cr, Co and Pb concentrations in garden produce (p < 0.05).

We found that Cu had the highest mean concentration among the metals measured in produce procured from grocery stores, while Pb had the lowest. Multiple commercially available produce samples exceeded vegetable maximum safe levels for Cd (Table S3). No grocery produce exceeded safe limits for Pb, and only one sample exceeded the safe level of As. To assess differences in metal content in garden produce (n = 36) and grocery produce (n = 43), a t-test was performed. It revealed that concentrations of Cr, Pb and Ni were significantly greater in garden produce than grocery produce (p < 0.05). Average concentrations of Cr, Co, Cu, Ni and Pb were all greatest in garden produce. A greater number of produce samples from gardens were above maximum safe levels than samples from grocery stores.

3.3 Metal transfer

A Pearson’s correlation indicated that concentrations of Cd and Pb in produce and corresponding concentrations in the soil were strongly positively correlated (p < 0.001; Table 5). The relationship between Pb concentrations in the soil and Pb in the produce of gardens differed by vegetable type, with root vegetables having higher concentrations of Pb at a given soil Pb concentration, followed by leafy, and then fruiting vegetables (Figure 4).

Figure 4. The relationship between community garden soil lead concentrations and corresponding produce lead concentrations with distinctions for leafy (n = 13), root (n = 5) and fruiting (n = 18) vegetables.

Table 5. Pearson correlation coefficients between corresponding soil and produce samples.

Metal	Pearson coefficient
As	-
Cd	0.585**
Cr	-
Co	0.412*
Cu
Ni	-
Pb	0.454**

*Indicates significance of <0.05.

**Indicates significance of <0.001.

Transfer factors (TF) measure the relationship between metal concentrations in soil and concentrations in produce grown in that soil as a measure of the transfer of metal from soil to produce. TF values range from zero, indicating little metal transfer from the soil to the plant, to one (or above), indicating either significant transfer from the soil, bioaccumulation within the plant, and/or alternate sources of metal to the plant. In our study of Philadelphia-region gardens, TFs ranged from 0.0003 to 1.089 across measured trace metals for the six most commonly sampled produce types (Table 6). The average pattern for TFs was Cd > Cu > Ni > As > Co > Cr > Pb. Basil had the greatest uptake of Cr and Co from the soil, while carrots had the greatest uptake of As, Cd and Pb from the soil. Cherry tomatoes had the greatest uptake of Cu and the lowest uptake of As and Pb, while tomatoes had the lowest uptake of Cr, Co and Ni (Table 6).

Table 6. Average transfer factors for specific vegetables calculated from samples across all gardens (µg/g).

Produce	As	Cd	Cr	Co	Cu	Ni	Pb
Basil	0.090	0.255	0.015	0.019	0.231	0.023	0.010
Collard Greens	0.007	0.236	0.006	0.009	0.043	0.030	0.004
Kale	0.011	0.193	0.011	0.007	0.061	1.089	0.006
Carrot	0.018	0.490	0.012	0.010	0.164	0.025	0.015
Cherry Tomato	0.002	0.252	0.005	0.006	0.241	0.016	0.003
Tomato	0.003	0.478	0.004	0.006	0.257	0.013	0.003

4. Discussion

Community gardens in the Philadelphia region were found to have elevated soil concentrations of As, Cd, Cr, Cu and Pb. Metal concentrations found in this study were within the range of concentrations found by other studies in the Philadelphia region [14,40]. Urban gardens generally had greater metal concentrations in their soils than suburban gardens (Figure 1). Sources of metals in urban soils and street dust include industrial activity, traffic emissions, weathering of buildings and atmospheric deposition [13]. We found that As and Cd were both positively correlated with Pb in garden soils, Cu and Cr were positively correlated, and Ni was positively correlated with all metals except for As and Pb (Table 3). Silva et al. [41] similarly found correlations between Cd and Pb and between Cr and Ni in urban soils. The correlation between concentrations of metals in garden soils is likely indicative of overlapping sources of entry into the garden environment, as suggested in Manta et al. [42]. Traffic proximity, which was measured by a combination of total road length and traffic density in the area surrounding each garden, was significant in predicting increased Pb and Ni soil concentrations within the multiple regression analysis (Table 4; Figure 2A). Traffic density has previously been found to influence soil concentrations of Cr, Cu and Zn, with larger concentrations found near high density traffic [43]. Concentrations of As, Cd and Pb in community garden soils have also been found to decrease with increased distance from a road edge [44]. While our ability to predict garden soil metal content was somewhat limited for most metals (R²<0.25; Table 4), soil Pb levels were strongly predicted (R² = 0.63) by the three variables we investigated in the neighborhood surrounding each garden (percent people of color, traffic proximity and the age of the housing) in addition to whether the garden was a raised bed or in-ground garden (Table 4).

Fourteen of the gardens in this study had soil Pb levels in excess of the City of Toronto’s Public Health soil screening level 1 (SSV1) indicating medium concern, while two gardens exceeded the SSV2 value and are of high concern. One of the gardens (garden C) and several samples from other gardens had soil levels above 400 µg/g, which the EPA considers unsafe for human contact [45]. Lead naturally occurs in soils at concentrations less than 50 µg/g. We found concentrations of Pb exceeding 10 µg/g in produce from two of the gardens. Gardeners and residents, including children, in Philadelphia-region community gardens are exposed to unsafe levels of metals both from the soil and from the produce they consume. Produce grown in these gardens take up metals from the soil, which accumulate in their tissues at sometimes unsafe concentrations for human health (Figure 3). Concentrations of Cd and Pb are the most concerning. All gardens where produce was obtained had multiple samples with unsafe Pb concentrations (Figure 3; Table S3), even though two of these same gardens had soil Pb concentrations within safe levels (Figure 1D; Table S2). All gardens had at least one produce item with concerning Cd levels, even when the majority of gardens had soil Cd concentrations within safe levels. In contrast, while several soil As concentrations were concerning according to recommended safe levels, As concentrations in vegetables were mostly within the recommended range. These discrepancies raise questions of validity of published recommendations and regulations for soils, especially for those soils utilized for gardening. The bioavailability of metals in soil could also be a reason for these differences. Cd was found to have the highest transfer factor, suggesting that even small concentrations of Cd in the soil will be taken up by plants. Clarke et al. [44] found similar bioavailability patterns for Cd, As and Pb in community gardens as a result of the varying geochemical properties of these metals.

Environmental justice issues, such as particulate matter exposure, urban heat islands and reduced tree cover, disproportionately affect low income and minority populations across the United States [46]. We found that Philadelphia-region gardens located in tracts with a high percentage of people of color had the greatest soil concentrations of Co and Pb (Table 4; Figure 2C) and greater produce concentrations of Cd, Cr, Co and Pb. Metals, such as Cr and Pb, have been found in urban soils adjacent to painted structures, such as railings, bridges, paved roads and even playgrounds [47,48]. Caballero‐Gómez et al. [49] found that elevated blood lead levels in children were correlated with housing code violations and the age of the housing (indicative of the use of Pb-based paint), and higher blood Pb levels were disproportionately found among black communities in Northern Philadelphia. We likewise found that the percentage of housing units built prior to 1960 (a lead paint indicator) was significant in predicting soil As, Cd and Pb content (Table 4; Figure 2B). Produce grown in these soils contains levels of Cd and Pb that are unsafe for human health. We found that grocery produce has significantly lower concentrations of Cr, Pb and Ni than produce from community gardens (Figure 3). While consumption of grocery store produce might lower exposure to toxic heavy metals, community gardens provide multiple benefits that should not be dismissed. Gardens provide fresh produce, which can be especially important in communities where access to produce is limited by distance, cost and other socioeconomic factors. Community gardens also help create green space in neighborhoods that are often lacking in sufficient green space. Gardens can help reduce runoff [50] and provide cooling [54] in urban environments. Urban gardens are also associated with a number of social benefits, including community cohesion and mental well-being [54, 2,51].

Living in urban environments can expose residents to environmental pollution from multiple sources, of which urban gardening is one potential source. We found that urban garden soils contained greater concentrations of Cd, Cr, Cu and Pb than suburban garden soils, which translated into higher Pb, Cr and Ni concentrations in community garden produce than produce procured from grocery stores (Figure 4; Table S3). Best practices can help minimize heavy metal content in gardening soil and produce. For instance, gardening in raised beds is a proven technique to reduce potential metal exposure [52]. Our findings confirm that raised beds significantly reduced concentrations of As and Pb in garden soil (Figure 2; Table S2). When filling raised beds, soil should be sourced from a location low in contaminants. In the gardens sampled in this study, the source of the soil used in raised beds was unknown and variation in source soil likely accounts for some portion of the variability in measured metal content and the elevated levels of Cr and Co found in several of the raised beds. Additionally, recommendations for gardeners often include guidance around preferentially growing fruiting produce, rather than root and leafy vegetables, as metals are typically retained in the leaves and roots of plants [27,28]. In our study, fruiting vegetables were found to have significantly lower mean ranks of As and Cr concentrations in comparison to those of root and leafy vegetables, and the relationship between soil and produce Pb was considerably lower for fruiting produce (Figure 4). These findings are consistent with the existing literature on produce type and metal uptake [17,53]. We also found that the transfer factors for fruiting vegetables were generally lower than those of root and leafy vegetables for most metals (with the exception of Cu; Table 6).

4.1 Conclusion & recommendations

The ecosystem and social services provided by urban gardens are important to maintain, and the exposure of gardeners and residents to pollutants should be limited as much as possible. When growing in urban locations, gardeners must take care to protect themselves against exposure to and consumption of heavy metals. The use of raised beds with soil sourced from a trusted location, locating gardens at a distance from roads, preferentially planting fruiting vegetables, and replacing contaminated soils are methods of reducing metal exposure that are supported by our results and existing literature on good gardening practices [28].

Ultimately, there is a need for increased governmental and nongovernmental agency involvement to ensure gardeners have access to information about safe gardening practices and the resources to establish and maintain community gardens using best practices. This is especially important in socioeconomically disadvantaged neighborhoods, where the services provided by community gardens have the largest impact, while the resources for safe gardening are also the most limited. There is additionally a need for more stringent guidelines, like those of Finland and Canada, to protect urban gardeners from heavy metal exposure.

Acknowledgments

The authors thank the Philadelphia-region community gardens who allowed us to test their soil and produce.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. Raw data can be provided upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/26395940.2023.2209283

Additional information

Funding

This work was supported by Villanova University’s Falvey Memorial Library Scholarship Open Access Reserve (SOAR) Fund [N/A]. This work was also supported by the National Science Foundation Division of Biological Infrastructure Major Research Instrumentation Grant [1726705].