Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jan 15.
Published in final edited form as: Environ Pollut. 2024 Nov 22;365:125374. doi: 10.1016/j.envpol.2024.125374

Comparing conventional and phytoscreening methods to detect subsurface chemical contaminants: a test case of volatile organic compounds in an urban setting

Brendan F O’Leary 1,2,3, Carol J Miller 1,3, Kelvin Selegean 1, Glen Ray Hood 2,3,*
PMCID: PMC11634637  NIHMSID: NIHMS2039544  PMID: 39581366

Abstract

The nationwide prevalence of brownfields, with often unknown types and quantities of subsurface chemical contaminants, highlights the need for rapid, cost-effective, and noninvasive methods to reduce routes of exposure. In post-industrial cities such as Detroit, Michigan, anthropogenic volatile organic compounds (VOCs), known to negatively impact human health, are typically detected at brownfields through conventional methods, e.g. screening soil, and groundwater. Recently, the method of phytoscreening–the chemical analysis of plant tissues to provide evidence for belowground contamination–has become a viable alternative to conventional methods. However, few studies have been designed to directly compare conventional and plant-based methods of detecting VOCs. To fill this knowledge gap, we sampled and compared the concentration of six VOCs including BTEX, PCE, and TCE detected in conventional media (soil, soil vapor, groundwater, sewer vapor) and different plant tissue (tree core, leaf, root, shoot) at two brownfields sites in Detroit: an abandoned gas station with a leaking underground storage tank, and a former dry cleaning facility. Our results suggest that the concentrations of VOCs detected in plants are similar to or in some cases greater than conventional methods and can differ across the growing season. For example, leaves and roots detected, on average, a higher concentration of VOCs compared to shoots and tree cores, however, TCE and PCE were generally in higher concentrations in soil and soil vapor. Moreover, the frequency at which conventional versus phytoscreening methods failed to detect VOCs was similar at one site and higher at another, suggesting that phytoscreening may yield fewer non-detects at known sites of contamination. While additional work is needed to understand the relationship between concentrations of VOCs detected in soil versus co-located plant samples, our results suggest that phytoscreening may be a viable and reliable method to detect belowground chemical contaminants while reducing screening times and cost, and increasing access to private property.

Keywords: anthropogenic VOCs, brownfield, Detroit, Michigan, phytochemical screening, plant tissue

Graphical Abstract

graphic file with name nihms-2039544-f0001.jpg

Introduction

In the United States, areas with a rich industrial legacy face many historic environmental concerns, including a large number of brownfield locations commonly polluted with subsurface contamination (Albanese & Cicchella, 2012; Barlow et al., 2012; Squillace et al., 2004). As a result, subsurface contamination tends to cluster in highly populated cities (Elliott & Frickel, 2015; Murray & Rogers, 1999). For example, according to Michigan’s Department of Environmental Great Lakes and Energy (EGLE), Wayne County in southeastern Michigan, home to Detroit, a post-industrial city with a population of >630,000, has over 9,000 individually regulated brownfields. In comparison, adjacent counties Oakland and Macomb have a combined 6,500 brownfields despite comprising more than twice the land area of Wayne County. In fact, the brownfield problem is so pervasive in Detroit that in 1996, the city formed the ‘Detroit Brownfield Redevelopment Authority’ whose goal was to provide tax incentives to “promote the revitalization of environmentally and distressed and blighted areas within the boundaries of the City of Detroit”.

One common class of subsurface chemical contamination often present at brownfields is anthropogenic volatile organic compounds (VOCs) (Ma et al., 2020; Miller et al., 2020). This class of VOCs, which are near ubiquitous in U.S. cities (Coggon et al., 2021), originate from a diversity of sources, including gasification plants, aging infrastructures such as oil storage and gas stations, and manufactured products such as building materials, paints, cleaning agents, furnishings, adhesives, combustion materials, and floor and wall coverings (Becher et al., 2022; Kuroda & Fukushi, 2008). Unfortunately, exposure to VOCs can cause both acute symptoms (e.g., headaches, nausea, throat irritation) and long term health issues (e.g., asthma, cancer, and adverse birth outcomes) (Chauhan et al., 2014; Forand et al., 2012; Hsu et al., 2018; Johnson et al., 2015; Ruckart et al., 2014). Moreover, in Detroit, preterm birth rates are the highest among major U.S. cities, ranging between 9.9–14.5% (March of Dimes, 2023), and are statistically linked to exposure to VOCs (Cassidy-Bushrow et al., 2020, 2021). Therefore, there is a need for rapid, cost-effective, and sensitive methods to locate VOC ‘hotspots’ that inform remediation efforts to reduce routes of exposure (O’Leary et al., 2023; Partha et al., 2022). This study aims to compare conventional and plant-based “phytoscreening” methods to screen for belowground VOC contaminants at brownfields in Detroit.

Not all brownfields are created equal, however, making VOC monitoring difficult. Each brownfield site has multiple location-specific variables, including subsurface heterogeneity, overlapping chemicals of concern, complexity due to the surrounding built environment, and absent responsible parties that create physical and social barriers to documenting VOCs at contaminated locations (Attoh-Okine & Gibbons, 2001; Gerstein, 2018, 2017; Lee & Mohai, 2011; Wu et al., 2017). In addition, the conventional screening methods of drilling wells to collect water and sampling soil cores are labor-intensive, invasive, time-consuming, and require specialized training (Lutes et al., 2022; Wilson et al., 2018). Moreover, groundwater monitoring wells are sparsely and unevenly distributed (Larsen et al. 2008), which is problematic in urban environments where hydrology is highly altered and property is often private and access-limited (Sorek et al., 2008; Wilcox & Johnson, 2016).

One alternative method that can reduce screening cost and time and potentially mitigate the site-specific barriers described above is phytoscreening–the chemical analysis of plant tissues to provide evidence for belowground contamination (Andraski et al., 2005; Burken et al., 2011; Holm & Rotard, 2011; Vroblesky et al., 1999). The idea of phytoscreening is simple: vascular plants grow a subsurface network of roots that uptake and transport water, nutrients, and potentially contaminants, including VOCs (partially) dissolved in water to aboveground plant tissues where they can be sampled and screened for chemicals of concern (COC) (Hood et al., 2021; Ma & Burken, 2004). Compared to conventional methods, phytoscreening aboveground tissues is less invasive, more cost-effective, and does not require specialized training (Duncan et al., 2017; Hood et al., 2021). As a result, phytoscreening, which began in the late 1990s (Gordon et al., 1997; Newman et al., 1997; Vroblesky et al., 1999) continues to be used to screen for contaminants in several environmental scenarios including optimizing groundwater well placement (Vroblesky et al., 1999), mapping the subsurface distribution and concentrations of pollutants (Andraski et al., 2005; Wilson et al., 2018), and assessing the risk of human exposure to VOCs via vapor intrusion (Wilson et al., 2017, 2018).

The application of phytoscreening is not without limitations, however. First, tree cores are the most widely used tissue for phytoscreening (e.g., see Burken et al., 2011; Duncan & Brusseau, 2018; Leoncini et al., 2022; Wilcox & Johnson, 2016; Wilson et al., 2018). When compared to the effectiveness of other aboveground plant tissues (e.g., shoots, leaves, fruit), results are mixed. For example, Doucette et al. (2007), Duncan et al. (2017), and Gopalakrishnan et al. (2007) collectively found that cores sampled from 10 different tree species detected higher concentrations of VOCs than shoots, leaves, and fruit. However, Hood et al. (2021) found that leaves and shoots detected the semi-VOC 1,4-dioxane in higher concentrations than oak tree cores, albeit with a limited sample size. Second, sampling tree cores can adversely impact plant health (Tsen et al., 2016), and mature trees may not be present at all brownfield locations (Johnson et al., 2018; Swan et al., 2017). Third, studies suggest phytoscreening may be VOC compound- and plant species-specific (Filippini et al., 2022). In a meta-analysis conducted by Leoncini et al. (2022), the authors found that cores of coniferous and diffuse-porous trees showed higher detection potential of chlorinated ethylenes when compared to ring-porous species, but trichloroethylene (TCE) and tetrachloroethylene (PCE) are detected twice as often as Dichloroethane (DCE) despite the three VOCs often co-occurring at polluted sites. Moreover, Yung et al. (2017) found that oaks, poplars, willows, and elms detected TCE and PCE at concentrations several orders of magnitude higher than birch, ash, cherry, locust, basswood, and beech trees at a single contaminated site. Fourth, temporal (seasonal) variation in environmental conditions may affect the frequency and concentration in which VOCs are detected (e.g., Burken et al., 2011). Lastly, studies designed to compare the concentration of VOCs detected using phytoscreening versus conventional methods are generally lacking. When they do exist, the results are often summarized qualitatively (Doucette et al., 2007; Duncan et al., 2017; Gopalakrishnan et al., 2007) and lack the quantitative, statistical inference needed to reject the hypothesis that phytoscreening is as good as, if not superior to conventional methods when screening for subsurface contaminants.

The goal of this study is to test the hypothesis that phytoscreening can detect six common VOCs at comparable or higher levels than conventional methods, including sampling soil, soil vapor, groundwater, and sewer vapor at two urban brownfield locations in Detroit, Michigan. Our study design allows us to fill the aforementioned knowledge gap by statistically comparing the concentrations of VOCs among different plant tissues, including leaves, roots, shoots, and tree cores, at various times during the growing season. Overall, with notable exceptions, our findings indicate that the levels of VOCs detected in plant tissues are comparable to those detected by conventional techniques and may vary across the growing season. As a result, phytoscreening could be an economical and convenient method for identifying subsurface chemical pollutants in situations where resources and site access are limited.

Methods

Description of brownfield locations

We sampled and measured the concentration of VOCs using conventional and phytoscreening methods in partnership with EGLE at two brownfields in neighborhoods located in Detroit, Michigan (Figure 1A). The first brownfield, located in east Detroit, an abandoned gas station, has a light nonaqueous phase liquid (LNAPL) VOC release from a leaking underground storage tank containing the following aromatic hydrocarbon COCs: benzene, toluene, ethylbenzene, and total xylene (the sum of m-, p-xylene and o-xylene), collectively referred to as BTEX. We refer to this location as the LNAPL site. Importantly, LNAPL compounds float on top of the water table, quickly biodegrade in aerobic conditions, and have short half-lives ranging from hours to days in typical environmental conditions (DeVaull, 2007). At the LNAPL site, the water table was observed at 0.6 meters below ground surface (bgs). The second brownfield, located in north Detroit, harbors an abandoned industrial dry-cleaning facility that has been releasing the following dense nonaqueous phase liquids (DNAPL) COC for several decades: trichloroethylene (PCE) and tetrachloroethylene (TCE) in addition to BTEX. We refer to this location as the DNAPL site. Unlike LNAPL compounds, DNAPL compounds sink below the water table to the bottom of the saturated zone and biodegrade slowly in anaerobic conditions (Yao & Chen, 2022). This difference in location and biodegradation of VOCs in the subsurface causes DNAPL plumes to typically have a larger area of groundwater impact when compared to LNAPL plumes (U.S. Environmental Protection Agency, 2015). Additionally, the water table was observed at 1.2 meters bgs. Given that the water table is shallow at both sites, the root zone of plant species samples in this study remain within the vadose (unsaturated) zone where they likely uptake VOCs in both groundwater and soil vapor.

Figure 1.

Figure 1.

Open in a new tab

(A) Two brownfield sites including a former gas station harboring a leaking underground storage tank (LNAPL) and a former dry cleaning facility (DNAPL), located in Detroit, in southeastern Michigan, USA. The dashed lines denote the city limits of Detroit. Note that the internal dashed lines denote Highland Park and Hamtramck, respectively, which are not part of the city of Detroit. (B) A conceptual model of media sampled for VOCs including conventional samples (soil, soil vapor, groundwater, sewer vapor) and phytoscreening (leaves, shoots, roots, and tree cores).

Sampling VOCs

To compare the concentration of VOCs using conventional and phytoscreening methods, we deployed two general sampling strategies that are summarized in Figure 1B. First, Michigan’s Department of Environment, Great Lakes and Energy (EGLE) performed a Phase II Environmental Site Assessment at the LNAPL site in September 2018 and at the DNAPL site in February 2019 using conventional sampling methods to measure VOC concentrations in soil. This assessment involved delineating the VOC plume by collecting hand augered and direct push soil samples following the US Environmental Protection Agency (EPA) field soil sampling method with methanol preservation (Lewis et al., 1994; Raza et al., 2018). To further delineate the VOC zone of impact, EGLE again sampled soil and groundwater, as well as sewer vapor and soil vapor at the LNAPL site and soil, groundwater, and soil vapor at the DNAPL site quarterly between 2021 and 2022 (Table 1). At both sites, the soil vapor sampling ports were installed by placing vapor pins in the sub-slab and 15 cm metal screen soil vapor monitoring points at 2.4 m below ground surface. The soil vapor and sewer vapor samples were then collected by directly filling an evacuated, 6-liter stainless steel canister (Pruitt, 2023). Similar methods were used to collect sewer vapor from manholes adjacent to the property. To collect additional groundwater samples, wells (152 cm screen length, 2.5 cm diameter, 0.025 cm screen size) were installed within the saturated zone previously identified by EGLE at both locations. The groundwater was sampled using low flow (minimal drawdown) methods and preserved in hydrochloric acid (Puls & Barcelona, 1996). Additional soil samples were taken from both sites using split spoon direct push cores and preserved with methanol (Lewis et al., 1994; Raza et al., 2018). All soil and groundwater samples were placed on ice in the field and cold stored in the laboratory prior to chemical analysis. Conventional sample locations at each site are mapped on Figure 2A, C. The data from these sampling events is publicly available through environmental reports via EGLE’s Remediation Information Data Exchange database at the following link: https://www.michigan.gov/egle/maps-data/ride.

Table 1.

The common and scientific names of plant species and the number of samples (n) collected from different plant tissues and conventional media to screen for VOCs in the fall and the following summer at two brownfield sites in Detroit, Michigan.

Site Common plant name (scientific name) Media n (fall) n (summer)
LNAPL Mulberry (Morus alba) Shoot 2 2
Grapevine (Vitis riparia) Leaf 2 2
Catalpa (Catalpa speciose) Core 1 1
Mulberry (Morus alba) Core 1 1
Green Ash (Fraxinus americanus) Shoot 2 2
Birch (Betula pendula) Shoot 2 2
Soil 24 6
Soil vapor 3 3
Sewer vapor 8 18
Groundwater 2 2
DNAPL American elm (Ulmus americana) Shoot 2 2
Root 1 1
Black cherry (Prunus serotina) Shoot 1 1
Leaf 2 2
Core 1 1
American elm (Ulmus americana) Shoot 1 1
Leaf 2 2
Soil 88 106
Soil vapor 3 3
Groundwater 4 2
Open in a new tab

Figure 2.

Figure 2.

Open in a new tab

Maps for the two brownfield sites detailing sampling locations and concentration ranges of total VOCs (combined BTEX, PCE, and TCE) detected using (A) conventional methods (soil, groundwater, soil vapor, and sewer vapor) and (B) co-located plant and soil samples at the LNAPL site harboring a leaking underground storage tank, and (C) conventional methods and (D) co-located plant and soil samples at the DNAPL site, home to a former industrial dry cleaning facility. Different media and different concentrations of total VOCs are designated by different shapes and colors, respectively. The blue outlines in each panel denote the property boundaries. Ùnits for the concentrations of total VOCs include μg/kg (plant and soil), μg/m3 (soil vapor, sewer vapor), and μg/L (groundwater). See Figures 3 and 4 for concentrations of the six individual VOCs.

Second, we collected co-located plant and soil samples from each of the two brownfields, once at the outset of the growing season before winter in November 2020 and once the following growing season in June 2021. The co-located sampling design enabled us to assess whether VOC concentrations in soil and plant tissues are correlated. Importantly, these soil samples were collected simultaneously with the plant tissue samples and are independent of those sampled by EGLE described above. The plants screened for VOCs consisted of seven mature (adult) species across the two sites and were chosen for several reasons: (1) they are common at vacant lots and brownfields in Detroit, (2) their proximity to the building at each site, and (3) their aboveground biomass was sufficiently large to collect multiple tissues types including shoot, tree core, leaf, root at each of the two sampling periods (Table 1). To compare the concentration of VOCs between plant and soil, co-located soil samples were collected directly underneath or immediately adjacent to the base of the plant at approximately 32 cm below the ground surface, within or adjacent to the plants root zone, using a hand auger. For both plant and soil samples, a minimum of ~10 grams of material was collected and for plant tissue cut into smaller pieces, immediately placed in 40 mL glass vials filled with 20 mL of methanol, and capped with a PTFE/silicone septa, following the methods of Hood et al. (2021). All samples were then placed on ice, transported to the laboratory, and held in a refrigerator at 4°C prior to processing. Due to logistical reasons, we only collected root tissue from a single plant (American elm) at the DNAPL site (Table 1). For information about the specific location at each brownfield property from which co-located soil and plant samples were collected as well as the plant species and sample sizes, see Figure 2B, D.

Sample processing

To determine the concentration of BTEX, PCE, and TCE in conventional samples at both the LNAPL and DNAPL sites, EGLE used the following EPA methods: (1) 5035A which utilizes a purge-and-trap gas chromatography-mass spectrometer (GC/MS) for soil samples, (2) 8260C which measures purgeable organic compounds in water by capillary column GC/MS and (3) TO-15 which uses GC/MS to analyze soil vapor and sewer vapor samples. Similarly, to determine the concentration of VOCs in plant and soil samples, chemical analysis was performed by Ann Arbor Technical Services (Ann Arbor, MI) following modified EPA method 5035A described in Duncan et al. (2017) and Hood et al. (2021). In brief, the methanol extracts were first purged (EPA method 5030B) and then analyzed via purge-and-trap GC/MS following EPA method 8260C for VOCs. Following the analysis, plant tissue was removed from each vial, placed in a ceramic vessel, and then oven dried at 105°C for 24 hours following modified EPA method 2540B to determine total solids and sample dry weight. The concentrations of BTEX, PCE, and TCE detected using both conventional and phytoscreening methods are reported in parts per billion (ppb) for soil (μg/kg), plant tissue (μg/kg), and groundwater (μg/L). For soil gas and sewer gas, concentrations are given in parts per billion by volume (ppbv) followed by micrograms per cubic meter in brackets in the text [μg/m3] to align with reporting methods of previous studies (Guo et al., 2020; Holton et al., 2015; Lin et al., 2022) and EGLE’s risk-based screening levels for these compounds (Tables 2 and 3). In the analyses described below, total VOCs (TVOC) represent the sum of all six COC.

Table 2:

The mean concentration of six VOCs and total VOCs (± S.E.) detected by conventional methods of screening soil, soil vapor, and sewer vapor, and phytoscreening tree cores, leaves, and shoots at the LNAPL site harboring a leaking underground storage tank. The sample size for each type of media (n) is given in parentheses in the second column. For each factor (media and season), values within columns not sharing a etter (a, b) differ significantly (P < 0.05), as determined by multivariate analysis of variance (MANOVA) followed by a Tukey’s post-hoc analysis. Bolded values are concentrations that are above the Environmental Protection Agency’s (EPA) risk-based screening level for each COC. Units for the concentrations of total VOCs are part per billion (ppb) or an equivalent (ppbv).

LNAPL Media (n) Benzene Toluene Ethylbenzene Total Xylene Tetrachloroethylene Trichloroethylene Total VOC
Media Core (4) 35 ± 20ab 34 ± 15ab 4.1 ±1.1ab 8.2 ± 2.2ab 4.1 ± 1.1ab 5.6 ± 1.3ab 91 ± 35ab
Leaf (4) 15 ± 8.3ab 39 ± 21ab 49 ± 28ab 100 ± 56ab 22 ± 8.6a 24 ± 12a 250 ± 130ab
Shoot (12) 7.7 ± 2.1ab 8.0 ± 1.7ab 8.1 ± 2.4ab 17 ± 4.4ab 6.1 ± 1.8ab 4.1 ± 0.71ab 51 ± 9.5ab
Soil (30) 450 ± 290 a 200 ± 150a 5,100 ± 5,000 a 120,000 ± 80,000 a 44 ± 6.1a 43 ± 6.2a 120,000 ± 81,000a
Soil vapor (6) 3.6 ± 2.1b 29 ± 18ab 2.6 ± 1.4b 10 ± 5.6ab 1.2 ± 0.28b 0.70 ± 0.25bc 47 ± 28ab
Sewer vapor (26) 5.1 ± 0.43 b 6.19 ± 1.4b 9.5 ± 0.93 ab 9.2 ± 0.95b 13 ± 3.5 a 0.85 ± 0.29 c 44 ± 5.5b
Season Fall (44) 310 ± 200a 140 ± 100a 3,400 ± 3,300a 77,000 ± 53,000a 33 ± 4.7a 32 ± 4.9a 81,000 ± 55,000a
Summer (37) 3.4 ± 0.49b 4.9 ± 1.2b 5.1 ± 0.84b 8.8 ± 1.8b 9.6 ± 2.6b 1.5 ± 0.33b 33 ± 5.2b
Open in a new tab

EGLE’s residential volatilization to Indoor Air Pathway Screening Levels: All units are expressed in units of ppb for soil (μg/kg) or ppbv and μg/m3 (given in brackets) for soil vapor and sewer vapor: Benzene (soil: 1.7 μg/kg, soil vapor: 34 ppbv [110 μg/m3], sewer vapor: 1 ppbv [3.3 μg/m3]); Toluene (soil: 3,700 μg/kg, soil vapor: 45,000 ppbv [170,000 μg/m3], sewer vapor: 1,400 ppbv [5,200 μg/m3]), Ethylbenzene (soil: 12 μg/kg, soil vapor: 78 ppbv [340 μg/m3], sewer vapor: 2.3 ppbv[10 μg/m3]), Xylenes (soil: 280 μg/kg, soil vapor: 1,800 ppbv [7,600 μg/m3], sewer vapor: 53 ppbv [230 μg/m3]), Tetrachloroethylene (soil: 62 μg/kg, soil vapor: 200 ppbv [1,400 μg/m3], sewer vapor: 6.0 ppbv[41 μg/m3]), and Trichloroethylene (soil: 0.33 μg/kg, soil vapor: 13 ppbv [67 μg/m3], sewer vapor: 0.37 ppbv [2 μg/m3]).

Table 3.

The mean concentration of six VOCs and total VOCs (± S.E.) detected by conventional methods of screening soil, soil vapor, and groundwater, as well as phytoscreening tree cores, leaves, shoots and roots at the DNAPL site, a former industrial dry-cleaning facility. The sample size for each type of media (n) is given in parentheses in the second column. For each factor (media and season), values within columns not sharing a letter (a, b, c, d) differ significantly (P < 0.05), as determined by multivariate analysis of variance (MANOVA) followed by a Tukey’s post-hoc analysis for BTEX compounds and PCE/TCE separately. Note that we did not sample soil vapor or groundwater for BTEX. Bolded values are concentrations that are above the Environmental Protection Agency’s (EPA) risk-based screening level for each COC. Units for the concentrations of total VOCs are part per billion (ppb) or an equivalent (ppbv).

DNAPL Media (n) Benzene Toluene Ethylbenzene Total Xylene Total BTEX Tetrachloroethylene Trichloroethylene Total PCE/TCE
Media Core (2) 33 ± 30ab 33 ± 10ab 4.2 ± 0.79ab 6.0 ± 0.86abc 77 ± 40ab 110 ± 31ab 4.3 ± 0.86b 110 ±31bc
Leaf (8) 13 ± 4.4b 20 ± 3.3ab 8.5 ± 5.7b 20 ± 11bc 61 ± 22b 18 ± 9.6b 5.2 ± 2.8b 23 ±12c
Shoot (8) 9.0 ± 2.8b 11 ± 1.4b 2.3 ± 0.22b 3.4 ± 0.00c 26 ± 4.1b 13 ± 6.1b 2.6 ± 0.32b 16 ±6.2c
Root (2) 28 ± 27ab 36 ± 13ab 11 ± 0.05ab 53 ± 16ab 130 ± 24ab 3,500 ± 3,400ab 57 ± 33a 3,500 ± 3400ab
Soil (194) 48 ± 0.77 a 48 ± 0.76a 47 ± 0.91 a 47 ± 1.1a 190 ± 3.1a 36,000 ± 25,000 b 64 ± 7.5 a 36,000 ± 25,000b
Soil vapor (6) Not sampled 57,000 ± 21,000 a 100 ± 55 a 58,000 ± 21,000a
Groundwater (6) 150 ±150 b 19 ± 18 b 170 ± 170c
Season Fall (104) 46 ± 1.3a 46 ± 1.3a 44 ± 1.7a 45 ± 1.7a 150 ± 7.6 a 67,000 ± 46,000a 48 ± 3.6a 68,000 ± 46,000a
Summer (121) 43 ± 1.6b 44 ± 1.3a 42 ± 1.7b 43 ± 1.9a 161 ± 6.6b 1,900 ± 920a 69 ± 12a 2,100 ± 920a
Open in a new tab

EGLE’s residential volatilization to Indoor Air Pathway Screening Levels: All units are expressed in units of ppb for soil (μg/kg) or ppbv and μg/m3 (given in brackets) for soil vapor and sewer vapor: Benzene (soil: 1.7 μg/kg, soil vapor: 34 ppbv [110 μg/m3], sewer vapor: 1 ppbv [3.3 μg/m3]); Toluene (soil: 3,700 μg/kg, soil vapor: 45,000 ppbv 170,000 μg/m3], sewer vapor: 1,400 ppbv [5,200 μg/m3]), Ethylbenzene (soil: 12 μg/kg, soil vapor: 78 ppbv [340 μg/m3], sewer vapor: 2.3 ppbv[10 μg/m3]), Xylenes (soil: 280 μg/kg, soil vapor: 1,800 ppbv [7,600 μg/m3], sewer vapor: 53 ppbv [230 μg/m3]), Tetrachloroethylene (soil: 62 μg/kg, soil vapor: 200 ppbv [1,400 μg/m3], sewer vapor: 6.0 ppbv[41 μg/m3]), and Trichloroethylene (soil: 0.33 μg/kg, soil vapor: 13 ppbv [67 μg/m3], sewer vapor: 0.37 ppbv [2 μg/m3]).

Statistical analysis

To compare the concentration of each of the six VOCs and total VOCs among conventional (soil, soil vapor, groundwater, and sewer vapor) and phytoscreening methods, we first performed a series of multivariate analyses of variance (MANOVA). A single MANOVA was performed for the LNAPL site. However, the analysis at the DNAPL site required two MANOVAs because not all COC (specifically BTEX) were screened by EGLE at the DNAPL site: one for the BTEX compounds, including the total concentration of BTEX, and another for PCE/TCE, including total concentration of chlorinated ethylenes. Typically, reporting COC below the detection limit are treated as left-censored data during statistical analysis (Helsel, 2011), and therefore do not meet the assumptions of normality of parametric statistical tests. We, therefore, performed each MANOVA on log (ln) transformed data considering non-detect samples both at the method detection limit (MDL) and one-half the MDL (Jia et al., 2008; Limmer et al., 2011, 2015). There was no qualitative difference in the statistical results in these datasets, thus we used the MDL, which represents a more conservative approach and entails a higher level of environmental exposure risk. In each MANOVA, we treated the factors ‘media’ (shoot, tree core, leaf, root, soil, soil vapor, groundwater, sewer vapor) and ‘season’ (spring and fall) as fixed factors. Importantly, while conventional samples by EGLE were sampled quarterly, we pooled the spring/summer and fall/winter quarters into two seasons for analysis. Following MANOVA, Tukey’s HSD post-hoc analyses were performed to determine significant pairwise differences in VOC concentrations among media. All statistical analysis was performed in R ver. 4.3.2 (R Core Team, 2013) and the results are presented throughout as back-transformed means ± standard errors (SE).

Lastly, we performed correlation analyses to determine if and to what extent VOC concentrations were correlated between co-located plant and soil samples. At each of the two sites, and for both sites pooled, we calculated Spearman’s rank correlation coefficient between TVOCs in plants and soil. Due to limited sample size, in each analysis we pooled data across tissue type and season. Similar to the MANOVA, MDL was used for samples below detection limit.

Results

Overall, our results suggest that in most cases, the concentrations of VOCs detected in plants are statistically similar in magnitude to conventional methods and can differ across the growing season at both brownfield locations in Detroit. The concentration of TVOCs detected from conventional media, including soil, soil vapor, and sewer vapor sampled by EGLE, are mapped in Figure 2A (LNAPL site) and Figure 2C (DNAPL site), while Figures 2B and 3D map TVOC concentrations of plant and co-located soil samples for the LNAPL and DNAPL sites, respectively. While we present the statistical analyses of the data below, we first summarize several qualitative patterns here.

Figure 3.

Figure 3.

Open in a new tab

Scattered boxplots of compound-specific concentration for plant tissue type (leaf, core, shoot, and root) at the LNAPL site (A, B, C) and at the DNAPL site (D, E, F, G). The mean is represented by the black ‘×’ in each panel. Note that root samples were not taken at the LNAPL site.

At least one VOC was detected and measured in every plant sample at both sites with two exceptions: a Green ash shoot and a core of a mulberry tree at the LNAPL site collected in the summer. Interestingly, the frequency at which conventional versus phytoscreening methods failed to detect any VOC was marginally lower at the LNAPL site (8% versus 10%), but higher at the DNAPL sites (37% versus 0%). At the LNAPL site, soil samples had the highest TVOC concentration, which is most likely the result of sampling at the location of the leaking underground storage tank in the northwest corner of the property (Figure 2A), followed by plant tissue, soil vapor, and sewer vapor. In fact, the average soil concentration of benzene, ethylenzene and xylene at the LNAPL site always exceeded EGLE’s risk based screening levels (RBSL), while the average concentration of benzene, ethylbenzene, TCE detected in sewer vapor also exceeded RBSL (Table 2).

Similarly, at the DNAPL site, soil also had the highest concentration of TVOCs followed by soil vapor, plant tissue, and groundwater. While the point source of contamination is unknown at the DNAPL site, the higher concentrations of VOCs in soil compared to other media likely represent sampling efforts by EGLE concentrated at or near the zone of impact inside of the building versus our own sampling efforts along the outside edge of the building foundation where plants were present (Figure 2C, D). The average TCE and PCE concentrations at the DNAPL site also exceeded RBSL for soil, soil vapor, and groundwater (Table 3).

LNAPL site (abandoned gas station)

At the LNAPL site, MANOVA revealed that the concentration of VOCs differed significantly among media (F42, 303.64 = 18.24; P < 0.00001) (Figures 3A–C, 4A) (Table S1). However, post-hoc analysis showed that 35 of the 105 pairwise comparisons (33%) differed across media (Table 2). For example, the concentration of PCE was highest in sewer vapor (mean = 13 ± 3.5 [88 ± 27], range = 1.8–94 ppbv [12–640 μg/m3]) and soil (mean = 44 ± 6.1, range = 0.28–74 μg/kg), which differed from the conventional method of screening soil vapor (mean = 1.2 ± 0.28 [8.0 ± 1.9], range = 0.30–2.1 ppbv [10–660 μg/m3]), as well as phytoscreening shoots (mean = 6.1 ± 1.8, range = 1.7–20 μg/kg) and tree cores (mean = 4.1 ± 1.1, range = 1.7–6.9 μg/kg). Interestingly, leaves detected PCE (mean = 22 ± 8.6, range = 6.9–45 μg/kg) and TCE (mean = 24 ± 12, range = 3.4–52 μg/kg) at intermediate levels, but not BTEX (Figures 3A–C, 4; Table 2). Despite the differences in concentration of PCE and TCE among media, TVOC concentrations, which include BTEX, PCE, and TCE, did not differ among media (Figures 3A, 4A, Table 2).

Figure 4.

Figure 4.

Open in a new tab

Scattered boxplots of total VOC concentration detected in groundwater, sewer vapor, soil, soil vapor, and plants at two sampling points in the fall and summer at (A) the LNAPL site and (B) the DNAPL site. The mean is represented by the red ‘×’ in each panel. Units for the concentrations of total VOCs are include ppb (plant, soil, groundwater), ppbv (soil vapor, sewer vapor).

Additionally, at the LNAPL site, MANOVA detected a difference in the concentration of each of the six VOCs and total VOCs between seasons (F7, 64 = 7.14; P < 0.0003) (Figure 4A) (Table S1). In all but one case (PCE), the average COC were detected in concentrations at least an order of magnitude higher in the fall versus the summer (Table 2). As a result, MANOVA revealed a significant media × season interaction (F35, 271.65 = 3.54; P < 0.001) (Table S1), suggesting that only certain media detect VOCs in statistically higher concentrations during the fall, while other media detect higher concentrations during the summer. For example, conventional methods sampling soil vapor in the fall detected higher concentrations of total xylenes (mean = 20 ± 8.1 [85 ± 35]; range = 8.1–35 ppbv [35–150 μg/m3]) than in summer months (mean = 0.70 ± 0.33 [3.0 ± 1.5]; range = 0.28–1.4 ppbv [1.2–5.9 μg/m3]) and phytoscreening leaves in the fall detected total xylene (mean = 200 ± 4.5, range = 200–210 μg/kg) in higher concentrations than in the summer (mean = 6.2 ± 3.6, range = 2.6–9.7 μg/kg) (Figure 4A, Table 2). Taken collectively, these results suggest that while sampling in the fall generally yields higher concentrations, when screening is required in the summer months, careful consideration should be given to the methods and media used to detect COC.

DNAPL site (former dry cleaning facility)

At the DNAPL site, MANOVA revealed that the concentration of VOCs differed significantly among media for BTEX (F25, 707.32 = 91.07; P < 0.00001) (Table S2) as well as TCE and PCE (F 21, 600.69 = 94.94; P < 0.00001) (Figures 3B–D, 4B) (Table S3). Unlike the LNAPL site, post hoc analysis indicated that there was at least one significant pairwise difference in mean concentration for each of the six COC, including total BTEX and total PCE/TCE (Table 3). As a result, 50 of the 121 pairwise comparisons (41%) were significantly different, 12 of which were observed for BTEX and 38 for PCE/TCE. For example, the concentration of both PCE and TCE was highest in soil vapor and soil, followed by roots and then groundwater. While BTEX compounds were not screened for soil vapor and groundwater at the DNAPL site, similar to PCE and TCE, soil sampling detected these compounds in higher concentrations than plants, albeit at a statistically non-significant level in all but one case: total xylene (Table 3).

Similar to the LNAPL site, at the DNAPL site, MANOVA detected a seasonal difference in the concentration of BTEX (F5, 190 = 5.39; P <0.00001) (Table S2) but not PCE and TCE (F3, 209 = 0.50; P = 0.96) (Table S3). However, a post-hoc analysis revealed that the differences were nuanced: both benzene (mean = 46 ± 1.3; range = 0.50–63 ppb) and ethylbenzene (mean = 44 ± 1.7; range = 0.50–50 ppb) were detected in higher concentrations in the fall, while total BTEX (mean = 170 ± 6.6; range = 2.0 – 300 ppb) was detected in higher concentrations in the summer (Figure 4B, Table 3). Interestingly, we did not find a significant difference among seasons for the concentration of toluene, total xylene, PCE, TCE, or combined PCE/TCE. As a result, the BTEX MANOVA showed a significant media × season interaction (F20, 631.11 = 6.33; P < 0.00001) (Table S2), while the interaction term in the PCE/TCE MANOVA was non-significant (F18, 591.63 = 1.14; P = 0.15) (Table S3), suggesting that the COC potentially present at a site along with time of year that media is sampled should be used when determining screening methods.

Discussion

In their study that coined the term phytoscreening, Sorek et al. (2008) declared that the method is “…simple, fast, noninvasive, and inexpensive … and is particularly useful in urban settings where conventional methods are difficult and expensive to employ”. Despite this assertion, in the subsequent 16 years, few studies have directly statistically compared conventional versus phytoscreening methods to directly justify this claim. We stress that our results are not merely an incremental contribution to the phytoscreening literature, rather our study helps fill this important knowledge gap. We found statistical supports that highlights the following: (1) in most cases phytoscreening can detect each of six common anthropogenic VOCs at concentrations comparable to or in some cases greater than conventional methods, (2) the frequency of non-detects using phytoscreening is similar to or less than conventional methods, (3) among plant tissues, leaves and roots detect VOCs in the highest concentrations, and (4) most VOCs are detected in higher concentrations in the fall compared to the summer. Overall, our study generally supports the hypothesis that phytoscreening detects BTEX, PCE and TCE in similar or higher concentrations that conventional methods.

Our study did reveal important differences in the concentration of VOCs detected among conventional and phytoscreening methods between sites perhaps related to the history of and difference between the chemicals released at each brownfield. During their Phase II site assessment, EGLE anticipated detecting BTEX leaking from an underground storage tank at the LNAPL site, an abandoned gas station, and PCE and TCE at the DNAPL site, a former dry cleaning facility. The concentration of BTEX at the LNAPL site did not statistically differ among plant tissue types, although several conventional samples were detected well above RBSL, highlighting the general human health concerns at this location. Regardless, we attribute this pattern to all samples being exposed to the same chemical release of BTEX on-site (Table 2). However, PCE and TCE, which were not initial COC at the LNAPL site, were generally detected in lower concentrations across media well below the RBSL, with one exception: sewer vapor. We postulate that the lower concentrations of PCE and TCE in all media (excluding sewer vapor) at the abandoned gas station come from a source traveling through the sewer pipes from an off-site location. Indeed, EGLE’s sewer samples taken ~150 meters southwest of the study area in Figure 2A measured at concentrations above RBSL for each compound (Table 2). Conversely, at the DNAPL site, the concentration of all VOCs, including BTEX, which were not original COC at this location, differed among media. However, only PCE and TCE were detected in concentrations above RBSL in soil, soil vapor, and groundwater samples. According to Limmer et al. (2015), VOC concentrations increase in plants with decreasing volatility. Perhaps the chemical properties of PCE and TCE, both compounds that biodegrade slowly and have lower volatility compared to BTEX compounds, account for their higher concentration at the DNAPL site more generally, and in plants specifically.

Phytoscreening is variable among plant tissues and between seasons

Despite our results, plants are not a “be-all-end-all” method for screening for belowground chemical contaminants. First, not all plant tissues detected VOCs in comparable concentrations. Our study found that the concentration of each VOC was lowest in shoots at both sites, while leaves had the highest concentration at the LNAPL site. In contrast, the roots detected the highest concentration at the DNAPL site, followed by tree cores (Figure 3, Table 2, 3). This result is perhaps not unexpected given that roots are in direct contact with polluted soil in the unsaturated zone at the site of plant uptake. Second, we found that the concentration of VOCs observed in media varies temporally. Surprisingly, most media collected in the fall detected higher concentrations of COC than in the summer (Figure 4, Tables 2, 3). Importantly, this pattern holds when comparing EGLE’s conventional samples collected across seasons in fall 2018 and spring 2019 to our own plant and soil samples collected across seasons in fall 2020 and spring 2021 and our own samples collected in fall 2020 versus spring 2021. During the growing season in the spring and summer, the concentration of contaminants in plant tissue typically increases compared to the fall and winter months due to two reasons: environmental conditions and transpiration rates (Limmer et al., 2014). For example, Holm and Rotard (2011) found that tree cores detected chlorinated organic compounds in concentrations two orders of magnitude greater during warm, dry periods compared to colder, wet conditions. Similarly, at contaminated sites in Tel Aviv, Israel, and Rolla, Missouri, USA, Sorek et al. (2008) and Limmer et al. (2014) found that the concentration of TCE and PCE in tree cores were approximately 10 to 100 times lower in the winter than in other seasons. Both Tel Aviv and Rolla experience mild winters with low to modest amounts of rain. In more northern climates that experience cold, wet winters where most deciduous plants lose their leaves prior to the onset of winter, this effect may be more extreme. In these cases, it has been suggested that plants should be sampled during active transpiration (Sorek et al., 2008).

Our study did not consider the effect of transpiration, which is known to increase the concentration of VOCs (Sorek et al., 2008; Limmer et al., 2014). For example, Limmer et al. (2014) detected PCE and TCE in concentrations an order of magnitude greater in the summer compared to the winter months in each of four consecutive years in red oak and bald cypress trees. The authors attributed this difference to a rapid increase in transpiration rates in the spring and decreased rates in the fall. However, the movement of VOCs between the plant and the atmosphere can also be bi-directional; plants release VOCs through their stomata during transpiration, and VOCs in the air surrounding a plant may enter the leaves via the same pathway (Dela Cruz et al., 2014). In Detroit and other large industrial cities, highly volatile BTEX compounds are commonly present in ambient air from several sources, including combustion emissions (O’Leary & Lemke, 2014). In fact, in a previous study, leaves were used to sample ambient air for BTEX compounds (Wetzel & Doucette, 2015), suggesting the presence of ambient VOCs influences the VOC concentrations in plants.

VOC concentrations may be less variable in plants than conventional screening methods

While we know that not all plants are equal in their phytoscreening capabilities (e.g., (Duncan et al., 2017; Yung et al., 2017), an interesting outcome of our study is that the variation in concentrations of VOCs observed in plant samples pooled across tissue types, seasons and sites (variance = 2.0 × 105 ppb) was approximately 100,000 times smaller than other media pooled in the same manner (variance = 2.0 × 1010 ppb). While we acknowledge that because of their physiochemical properties the VOCs will likely have different distributions among media (Rivett et al., 2011), the observation of reduced variance in plants is important because due to the high screening costs, decisions about allocating funds for clean-up are often made using a small number of samples and outliers are often common when screening for belowground chemical contaminants (Simmons, 2023). Our results suggest phytoscreening may reduce the frequency of outliers, including both non-detects and samples of extraordinarily high concentrations, potentially due to the large root zone when compared to an isolated conventional sample. Ultimately, decreased variance of phytoscreening versus conventional methods may produce more consistent results and better inform management decisoions about the allocation of limited funds to cleanup efforts.

Can plants predict the concentration of VOCs in the subsurface?

A major open question in the phytoscreening literature is understanding if and in what scenario(s) the concentration of VOCs in plants predicts the concentration of the same chemicals observed in conventional media sampled in the subsurface. While we admit that our sample size was limited, correlation analyses failed to detect a significant relationship between TVOCs in co-located soil and plant samples at both sites combined (rs = −0.2508, P = 0.17, n = 31) (Figure S1A), LNAPL site (rs = −0.4040, P = 0.08, n = 20) (Figure S1B), and at the DNAPL site (rs = −0.2069, P = 0.54, n = 11) (Figure S1C). At the LNAPL site, this pattern is driven, in part, by five soil samples being below the detection limit for many VOCs. Regardless, several studies have found similar results including Wilcox and Johnson (2016) that used tree cores to screen for TCE at the site of a former electroplating facility. The authors failed to find a correlation between TCE concentrations and nearby soil samples, which they attributed to variation in three key variables: geology of their field site, structure of tree roots, and transpiration rates across the growing season.

Several additional explanations may contribute to the non-significant correlation observed in our study. First, soil measurement reflects a concentration of VOCs at a single point in time, whereas the plant samples could represent continuous uptake and accumulation of VOCs (Collins & Finnegan, 2010; Weyens et al., 2010). Second, based on observations of the plant’s root system and age, most plants sampled in our study likely cover a much larger and deeper area than a single soil sample within the saturated zone. Third, BTEX compounds are more volatile than chlorinated compounds such as PCE and TCE, which tend to sink below the water table. As a result, sampling soil cores at 32 cm may be too shallow for accurate VOC measurements, especially at the DNAPL site. For example, Schumacher et al. (2004) noted significant correlations between the concentration of PCE in tree cores and soil, but the soil was sampled at depths between 1.2 and 5 meters. Similarly, in their recent meta-analytical analysis, Leoncini et al. (2022) found that the concentration of three chlorinated ethylenes, including PCE, TCE, and dichloroethylene (DCE) as well as total chlorinated ethylene in tree cores was significantly correlated with the concentration of these VOCs in groundwater. However, the correlation is not a 1:1 ratio: factors such as increased volatility and proximity to shallow, permeable aquifers increase the detection probability of VOCs. Moreover, coniferous trees that transpire throughout the year and thus continuously uptake groundwater (Trapp, 2007), and diffuse-porous trees, many of which have large vessels that increase conductivity (flow and transport) of water through xylem tissue (Westhoff et al. 2008) increase the ability to detect VOCs via phytoscreening. Our study did not include coniferous or diffuse-porous plant species, but we hope to incorporate these species in future analyses.

Plants other than trees show promise in phytoscreening application

We purposefully ignored species identity in our study and chose plants common in vacant brownfields near each property’s building structures. While tree species (e.g., poplar and willows) are the most common plants used for phytoscreening and phytoremediation (Leoncini et al., 2022), not all brownfields harbor trees. Moreover, sampling cores can negatively affect tree health, and arborists recommend not coring trees more than once (Tsen et al. 2015). Thus, there is a need to develop plants other than trees, including vines, shrubs, and grasses, as phytoscreening tools. In this regard, we found that Riverbank grapevine (Vitis riparia), sampled at the LNAPL site, detected toluene, ethylbenzene, total xylene, PCE, and TCE in the highest concentration compared to the other plant species. Given that grapevine (1) is one of the most common species in southeastern Michigan (Moore & Wen, 2016), (2) can grow roots up to six meters deep (Richards, 2011) reaching the aquifer below the city in many locations and (3) shoots and leaves have been used to detect the semi-VOC 1,4-dioxane (Hood et al. 2021), this species may be an ideal candidate for phytoscreening to locate VOC hotspots throughout Detroit. Nevertheless, our results suggest that regardless of the species screened, plants including non-tree species may be a general, low-cost tool to detect VOCs, particularly at polluted urban sites where plant diversity may be limited.

Conclusions

Ultimately, our study fills a knowledge gap by directly comparing conventional and plant-based screening methods. Our results demonstrate that phytoscreening can detect six common VOCs at a greater frequency and at comparable (or higher) concentrations than conventional methods in urban brownfields, leaves and roots detected VOCs in the highest concentrations that shoots and tree cores, and VOC concentrations are in higher concentrations observed in the fall compared to the summer. Our research also highlights the utility of collaborative efforts between academic institutions and government agencies. Given the utility of phytoscreening, at present, we are working with EGLE at several brownfields across Detroit to develop further this method as a noninvasive, low-cost alternative to detect the presence or infer the absence of VOCs. To accomplish this goal, we are creating a geospatial database of VOC concentrations in soil and plants across Detroit to determine what plant species and plant tissues are ideal for detecting different COC and predicting the location of previously unknown hotspots of VOC contamination in the city. Our ultimate goal is to use phytoscreening in combination with stable isotope techniques (e.g., Nisi et al., 2016) to identify belowground source waters used by plants to identify sources of contamination at polluted sites.

Supplementary Material

1

Highlights.

  • The frequency of plant samples below detection limits ≤ conventional methods.

  • Plants generally detect VOCs at similar concentrations than conventional methods.

  • The concentrations of VOCs in leaves and roots ≥ shoots and tree cores.

  • Most VOCs were detected in higher concentrations in the fall.

  • The concentration of VOCs in co-located plant and soil samples was not correlated.

Acknowledgments

We thank Sarah Black for help with fieldwork and Beth Vens from EGLE for help accessing field sites. This work was supported by the National Science Foundation (Grant Number 1735038), National Institute of Environmental Health Sciences (Grant Number P30 ES020957), National Institute of Environmental Health Sciences (Grant Number P42 ES030991), and Wayne State University’s Office of the Vice President for Research.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Albanese S, & Cicchella D (2012). Legacy problems in urban geochemistry. Elements, 8(6), 423–428. [Google Scholar]
  2. Andraski BJ, Stonestrom DA, Michel RL, Halford KJ, & Radyk JC (2005). Plant-Based Plume-Scale Mapping of Tritium Contamination in Desert Soils. Vadose Zone Journal, 4(3), 819–827. https://doi.org/ 10.2136/vzj2005.0052 [DOI] [Google Scholar]
  3. Attoh-Okine NO, & Gibbons J (2001). Use of belief function in brownfield infrastructure redevelopment decision making. Journal of Urban Planning and Development, 127(3), 126–143. [Google Scholar]
  4. Barlow JRB, Kingsbury JA, & Coupe RH (2012). Changes in Shallow Groundwater Quality Beneath Recently Urbanized Areas in the Memphis, Tennessee Area1. JAWRA Journal of the American Water Resources Association, 48(2), 336–354. 10.1111/j.1752-1688.2011.00616.x [DOI] [Google Scholar]
  5. Becher J, Englisch C, Griebler C, & Bayer P (2022). Groundwater fauna downtown – Drivers, impacts and implications for subsurface ecosystems in urban areas. Journal of Contaminant Hydrology, 248, 104021. https://doi.org/ 10.1016/j.jconhyd.2022.104021 [DOI] [PubMed] [Google Scholar]
  6. Burken JG, Vroblesky DA, & Balouet JC (2011). Phytoforensics, Dendrochemistry, and Phytoscreening: New Green Tools for Delineating Contaminants from Past and Present. Environmental Science & Technology, 45(15), 6218–6226. 10.1021/es2005286 [DOI] [PubMed] [Google Scholar]
  7. Cassidy-Bushrow AE, Burmeister C, Lamerato L, Lemke LD, Mathieu M, O’Leary BF, et al. (2020). Prenatal airshed pollutants and preterm birth in an observational birth cohort study in Detroit, Michigan, USA. Environmental Research, 189, 109845. https://doi.org/ 10.1016/j.envres.2020.109845 [DOI] [PubMed] [Google Scholar]
  8. Cassidy-Bushrow AE, Burmeister C, Birbeck J, Chen Y, Lamerato L, Lemke LD, et al. (2021). Ambient BTEX exposure and mid-pregnancy inflammatory biomarkers in pregnant African American women. Journal of Reproductive Immunology, 145, 103305. https://doi.org/ 10.1016/j.jri.2021.103305 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chauhan SK, Saini N, & Yadav VB (2014). Recent trends of volatile organic compounds in ambient air and its health impacts: A review. Int. J. Technol. Res. Eng, 1(8), 667. [Google Scholar]
  10. Coggon MM, Gkatzelis GI, McDonald BC, Gilman JB, Schwantes RH, Abuhassan N, et al. (2021). Volatile chemical product emissions enhance ozone and modulate urban chemistry. Proceedings of the National Academy of Sciences, 118(32), e2026653118. 10.1073/pnas.2026653118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Collins CD, & Finnegan E (2010). Modeling the Plant Uptake of Organic Chemicals, Including the Soil−Air−Plant Pathway. Environmental Science & Technology, 44(3), 998–1003. 10.1021/es901941z [DOI] [PubMed] [Google Scholar]
  12. Department of Environment Great Lakes and Energy. (2024). Remediation Information Data Exchange (RIDE). Retrieved April 11, 2024, from https://www.michigan.gov/egle/maps-data/ride
  13. DeVaull GE (2007). Indoor vapor intrusion with oxygen-limited biodegradation for a subsurface gasoline source. Environmental Science & Technology, 41(9), 3241–3248. [DOI] [PubMed] [Google Scholar]
  14. Doucette WJ, Chard JK, Fabrizius H, Crouch C, Petersen MR, Carlsen TE, et al. (2007). Trichloroethylene Uptake into Fruits and Vegetables: Three-Year Field Monitoring Study. Environmental Science & Technology, 41(7), 2505–2509. 10.1021/es0621804 [DOI] [PubMed] [Google Scholar]
  15. Duncan CM, & Brusseau ML (2018). An assessment of correlations between chlorinated VOC concentrations in tree tissue and groundwater for phytoscreening applications. Science of The Total Environment, 616–617, 875–880. https://doi.org/ 10.1016/j.scitotenv.2017.10.235 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Duncan CM, Mainhagu J, Virgone K, Ramírez DM, & Brusseau ML (2017). Application of phytoscreening to three hazardous waste sites in Arizona. Science of The Total Environment, 609, 951–955. https://doi.org/ 10.1016/j.scitotenv.2017.07.236 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Elliott JR, & Frickel S (2015). Urbanization as Socioenvironmental Succession: The Case of Hazardous Industrial Site Accumulation. American Journal of Sociology, 120(6), 1736–1777. 10.1086/681715 [DOI] [PubMed] [Google Scholar]
  18. Filippini M, Leoncini C, Luchetti L, Emiliani R, Fabbrizi E, & Gargini A (2022). Detecting vinyl chloride by phytoscreening in the shallow critical zone at sites with potential human exposure. Journal of Environmental Management, 319, 115776. 10.1016/J.JENVMAN.2022.115776 [DOI] [PubMed] [Google Scholar]
  19. Forand SP, Lewis-Michl EL, & Gomez MI (2012). Adverse birth outcomes and maternal exposure to trichloroethylene and tetrachloroethylene through soil vapor intrusion in New York State. Environ Health Perspect, 120(4), 616–621. 10.1289/ehp.1103884 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gerstein M (2018). Toxic vapor tests lag for thousands of Michigan sites. The Detroit News Feb 7, 2018. Retrieved from: https://www.detroitnews.com/story/news/local/michigan/2018/02/07/toxic-vapors-michigan-hundreds-evacuation/110212116/
  21. Gerstein M Mich. DEQ chief: ‘Vapor intrusion’ poses health threat. The Detroit News Feb 16, 2017. Retrieved from: https://www.detroitnews.com/story/news/local/michigan/2017/02/16/mich-deq-chief-vapor-intrusion-poses-health-threat/98020512/
  22. Gopalakrishnan G, Negri MC, Minsker BS, & Werth CJ (2007). Monitoring Subsurface Contamination Using Tree Branches. Groundwater Monitoring & Remediation, 27(1), 65–74. https://doi.org/ 10.1111/j.1745-6592.2006.00124.x [DOI] [Google Scholar]
  23. Gordon M, Choe N, Duffy J, Ekuan G, Heilman P, Muiznieks I, et al. (1997). Phytoremediation of Trichloroethylene with Hybrid Poplars. In Phytoremediation of Soil and Water Contaminants (Vol. 664, pp. 13–177). American Chemical Society. https://doi.org/ doi: 10.1021/bk-1997-0664.ch013 [DOI] [Google Scholar]
  24. Guo Y, Dahlen P, & Johnson PC (2020). Development and Validation of a Controlled Pressure Method Test Protocol for Vapor Intrusion Pathway Assessment. Environmental Science & Technology, 54(12), 7117–7125. 10.1021/acs.est.0c00811 [DOI] [PubMed] [Google Scholar]
  25. Helsel DR (2011). Statistics for censored environmental data using Minitab and R (Vol. 77). John Wiley & Sons. [Google Scholar]
  26. Holm O, & Rotard W (2011). Effect of Radial Directional Dependences and Rainwater Influence on CVOC Concentrations in Tree Core and Birch Sap Samples Taken for Phytoscreening Using HS-SPME-GC/MS. Environmental Science & Technology, 45(22), 9604–9610. 10.1021/es202014h [DOI] [PubMed] [Google Scholar]
  27. Holton C, Guo Y, Luo H, Dahlen P, Gorder K, Dettenmaier E, & Johnson PC (2015). Long-Term Evaluation of the Controlled Pressure Method for Assessment of the Vapor Intrusion Pathway. Environmental Science & Technology, 49(4), 2091–2098. 10.1021/es5052342 [DOI] [PubMed] [Google Scholar]
  28. Hood GR, Papuga SA, Socrates C, Rankin K, & Hwang K (2021). On the phytoscreening potential of insect-induced plant galls. Plant and Soil, 467(1), 569–578. 10.1007/s11104-021-05068-8 [DOI] [Google Scholar]
  29. Hsu C-Y, Chiang H-C, Shie R-H, Ku C-H, Lin T-Y, Chen M-J, et al. (2018). Ambient VOCs in residential areas near a large-scale petrochemical complex: spatiotemporal variation, source apportionment and health risk. Environmental Pollution, 240, 95–104. [DOI] [PubMed] [Google Scholar]
  30. Jia C, Batterman S, & Godwin C (2008). VOCs in industrial, urban and suburban neighborhoods, Part 1: Indoor and outdoor concentrations, variation, and risk drivers. Atmospheric Environment, 42(9), 2083–2100. 10.1016/j.atmosenv.2007.11.055 [DOI] [Google Scholar]
  31. Jiao Y, Kramshøj M, Davie-Martin CL, Albers CN, & Rinnan R (2023). Soil uptake of VOCs exceeds production when VOCs are readily available. Soil Biology and Biochemistry, 185, 109153. https://doi.org/ 10.1016/j.soilbio.2023.109153 [DOI] [Google Scholar]
  32. Johnson AL, Borowy D, & Swan CM (2018). Land use history and seed dispersal drive divergent plant community assembly patterns in urban vacant lots. Journal of Applied Ecology, 55(1), 451–460. https://doi.org/ 10.1111/1365-2664.12958 [DOI] [Google Scholar]
  33. Johnson DA, Drake C, Joseph CL, Krajenta R, Hudgel DW, & Cassidy-Bushrow AE (2015). Influence of neighbourhood-level crowding on sleep-disordered breathing severity: mediation by body size. J Sleep Res. 10.1111/jsr.12305 [DOI] [PubMed] [Google Scholar]
  34. Kuroda K, & Fukushi T (2008). Groundwater Contamination in Urban Areas. In Takizawa S (Ed.), Groundwater Management in Asian Cities: Technology and Policy for Sustainability (pp. 125–149). Tokyo: Springer Japan. 10.1007/978-4-431-78399-2_7 [DOI] [Google Scholar]
  35. Lee S, & Mohai P (2011). Racial and socioeconomic assessments of neighborhoods adjacent to small-scale brownfield sites in the Detroit region. Environmental Practice, 13(4), 340–353. [Google Scholar]
  36. Leoncini C, Filippini M, Nascimbene J, & Gargini A (2022). A quantitative review and meta-analysis on phytoscreening applied to aquifers contaminated by chlorinated ethenes. Science of The Total Environment, 817, 153005. 10.1016/J.SCITOTENV.2022.153005 [DOI] [PubMed] [Google Scholar]
  37. Lewis TE, Crockett AB, & Siegrist RL (1994). Soil sampling and analysis for volatile organic compounds. Environmental Monitoring and Assessment, 30, 213–246. [DOI] [PubMed] [Google Scholar]
  38. Limmer MA, Balouet J-C, Karg F, Vroblesky DA, & Burken JG (2011). Phytoscreening for Chlorinated Solvents Using Rapid in Vitro SPME Sampling: Application to Urban Plume in Verl, Germany. Environmental Science & Technology, 45(19), 8276–8282. 10.1021/es201704v [DOI] [PubMed] [Google Scholar]
  39. Limmer MA, Holmes AJ, & Burken JG (2014). Phytomonitoring of Chlorinated Ethenes in Trees: A Four-Year Study of Seasonal Chemodynamics in Planta. Environmental Science & Technology, 48(18), 10634–10640. 10.1021/es502680p [DOI] [PubMed] [Google Scholar]
  40. Limmer MA, West DM, Mu R, Shi H, Whitlock K, & Burken JG (2015). Phytoscreening for perchlorate: rapid analysis of tree sap. Environmental Science: Water Research & Technology, 1(2), 138–145. 10.1039/C4EW00103F [DOI] [Google Scholar]
  41. Lin N, Zhong L, Godwin C, & Batterman S (2022). Be alert for vapor intrusion of 1,4-dioxane from contaminated groundwater. Science of The Total Environment, 825, 153713. https://doi.org/ 10.1016/j.scitotenv.2022.153713 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Lutes C, Stewart L, Truesdale R, De Loera J, Zimmerman JH, & Schumacher B (2022). Cost Comparison of Soil Vapor Extraction and Subslab Depressurization for Vapor Intrusion Mitigation. Groundwater Monitoring & Remediation, 42(4), 43–53. https://doi.org/ 10.1111/gwmr.12510 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Ma J, McHugh T, Beckley L, Lahvis M, DeVaull G, & Jiang L (2020). Vapor Intrusion Investigations and Decision-Making: A Critical Review. Environmental Science & Technology, 54(12), 7050–7069. 10.1021/acs.est.0c00225 [DOI] [PubMed] [Google Scholar]
  44. Ma X, & Burken J (2004). Modeling of TCE Diffusion to the Atmosphere and Distribution in Plant Stems. Environmental Science & Technology, 38(17), 4580–4586. 10.1021/es035435b [DOI] [PubMed] [Google Scholar]
  45. March of Dimes. (2023). March of Dimes Premature Birth Report Card. Retrieved from https://www.marchofdimes.org/report-card
  46. Miller CJ, Runge-Morris M, Cassidy-Bushrow AE, Straughen JK, Dittrich TM, Baker TR, et al. (2020). A Review of Volatile Organic Compound Contamination in Post-Industrial Urban Centers: Reproductive Health Implications Using a Detroit Lens. International Journal of Environmental Research and Public Health, 17(23), 8755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Moore M, & Wen J (2016). Flora of North America - Vitis riparia (Vol. 12). OXFORD UNIVERSITY PRESS. [Google Scholar]
  48. Murray KS, & Rogers DT (1999). Groundwater Vulnerability, Brownfield Redevelopment and Land Use Planning. Journal of Environmental Planning and Management, 42(6), 801–810. 10.1080/09640569910830 [DOI] [Google Scholar]
  49. Newman LA, Strand SE, Choe N, Duffy J, Ekuan G, Ruszaj M, et al. (1997). Uptake and Biotransformation of Trichloroethylene by Hybrid Poplars. Environmental Science & Technology, 31(4), 1062–1067. 10.1021/es960564w [DOI] [Google Scholar]
  50. O’Leary BF, & Lemke LD (2014). Modeling spatiotemporal variability of intra-urban air pollutants in Detroit: A pragmatic approach. Atmospheric Environment, 94, 417–427. https://doi.org/ 10.1016/j.atmosenv.2014.05.010 [DOI] [Google Scholar]
  51. O’Leary BF, Hill AB, Linn C, Lu M, Miller CJ, Newman A, et al. (2023). Exploring the association of Brownfield remediation status with socioeconomic conditions in Wayne County, MI. Environmental Science and Pollution Research, 30(21), 60768–60776. 10.1007/s11356-023-26666-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Partha DB, Cassidy-Bushrow AE, & Huang Y (2022). Global preterm births attributable to BTEX (benzene, toluene, ethylbenzene, and xylene) exposure. Science of The Total Environment, 838, 156390. 10.1016/J.SCITOTENV.2022.156390 [DOI] [PubMed] [Google Scholar]
  53. Pruitt L (2023). Soil Gas Sampling. Laboratory Services & Applied Science Division, U.S. Environmental Protection Agency, Athens, Georgia. [Google Scholar]
  54. Puls RW, & Barcelona MJ (1996). Low-flow (minimal drawdown) ground-water sampling procedures. Office of Research and Development, U.S. Environmental Protection Agency. [Google Scholar]
  55. R Core Team. (2013). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013; http://www.R-project.org/.
  56. Raza N, Hashemi B, Kim K-H, Lee S-H, & Deep A (2018). Aromatic hydrocarbons in air, water, and soil: Sampling and pretreatment techniques. TrAC Trends in Analytical Chemistry, 103, 56–73. [Google Scholar]
  57. Richards D (2011). The Grape Root System. In Horticultural Reviews (Vol. 5, pp. 127–168). 10.1002/9781118060728.ch3 [DOI] [Google Scholar]
  58. Rivett MO, Wealthall GP, Dearden RA, & McAlary TA (2011). Review of unsaturated-zone transport and attenuation of volatile organic compound (VOC) plumes leached from shallow source zones. Journal of Contaminant Hydrology, 123(3–4), 130–156. [DOI] [PubMed] [Google Scholar]
  59. Ruckart PZ, Bove FJ, & Maslia M (2014). Evaluation of contaminated drinking water and preterm birth, small for gestational age, and birth weight at Marine Corps Base Camp Lejeune, North Carolina: a cross-sectional study. Environ Health, 13, 99. 10.1186/1476-069x-13-99 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Schumacher JG, Struckhoff GC, & Burken JG (2004). Assessment of Subsurface Chlorinated Solvent Contamination Using Tree Cores at the Front Street Site and a Former Dry Cleaning Facility at the River-front Superfund Site, New Haven, Missouri, 1999–2003. US Department of the Interior, US Geological Survey. [Google Scholar]
  61. Simmons K (2023). Soil Sampling. Laboratory Services & Applied Science Division, U.S. Environmental Protection Agency, Athens, Georgia. [Google Scholar]
  62. Sorek A, Atzmon N, Dahan O, Gerstl Z, Kushisin L, Laor Y, et al. (2008). “Phytoscreening”: The Use of Trees for Discovering Subsurface Contamination by VOCs. Environmental Science & Technology, 42(2), 536–542. 10.1021/es072014b [DOI] [PubMed] [Google Scholar]
  63. Squillace PJ, Moran MJ, & Price CV (2004). VOCs in Shallow Groundwater in New Residential/Commercial Areas of the United States. Environmental Science & Technology, 38(20), 5327–5338. 10.1021/es0349756 [DOI] [PubMed] [Google Scholar]
  64. Swan CM, Johnson A, & Nowak DJ (2017). Differential organization of taxonomic and functional diversity in an urban woody plant metacommunity. Applied Vegetation Science, 20(1), 7–17. https://doi.org/ 10.1111/avsc.12266 [DOI] [Google Scholar]
  65. Trapp S (2007). Fruit Tree model for uptake of organic compounds from soil and air. SAR and QSAR in Environmental Research, 18(3–4), 367–387. 10.1080/10629360701303693 [DOI] [PubMed] [Google Scholar]
  66. Tsen EWJ, Sitzia T, & Webber BL (2016). To core, or not to core: the impact of coring on tree health and a best-practice framework for collecting dendrochronological information from living trees. Biological Reviews, 91(4), 899–924. https://doi.org/ 10.1111/brv.12200 [DOI] [PubMed] [Google Scholar]
  67. U.S. Environmental Protection Agency. (2015). OSWER Technical Guide for Assessing and Mitigating the Vapor Intrusion Pathway from Subsurface Vapor Sources to Indoor Air, EPA’s 2015. Retrieved from https://www.epa.gov/vaporintrusion
  68. Vroblesky DA, Nietch CT, & Morris JT (1999). Chlorinated Ethenes from Groundwater in Tree Trunks. Environmental Science & Technology, 33(3), 510–515. 10.1021/es980848b [DOI] [Google Scholar]
  69. Wetzel TA, & Doucette WJ (2015). Plant leaves as indoor air passive samplers for volatile organic compounds (VOCs). Chemosphere, 122, 32–37. https://doi.org/ 10.1016/j.chemosphere.2014.10.065 [DOI] [PubMed] [Google Scholar]
  70. Weyens N, Truyens S, Dupae J, Newman L, Taghavi S, van der Lelie D, et al. (2010). Potential of the TCE-degrading endophyte Pseudomonas putida W619-TCE to improve plant growth and reduce TCE phytotoxicity and evapotranspiration in poplar cuttings. Environmental Pollution, 158(9), 2915–2919. https://doi.org/ 10.1016/j.envpol.2010.06.004 [DOI] [PubMed] [Google Scholar]
  71. Wilcox JD, & Johnson KM (2016). Trichloroethylene (TCE) in tree cores to complement a subsurface investigation on residential property near a former electroplating facility. Environmental Monitoring and Assessment, 188(10), 587. 10.1007/s10661-016-5603-x [DOI] [PubMed] [Google Scholar]
  72. Wilson JL, Limmer MA, Samaranayake VA, Schumacher JG, & Burken JG (2017). Tree Sampling as a Method to Assess Vapor Intrusion Potential at a Site Characterized by VOC-Contaminated Groundwater and Soil. Environmental Science & Technology, 51(18), 10369–10378. 10.1021/acs.est.7b02667 [DOI] [PubMed] [Google Scholar]
  73. Wilson JL, Samaranayake VA, Limmer MA, & Burken JG (2018). Phytoforensics: Trees as bioindicators of potential indoor exposure via vapor intrusion. PLOS ONE, 13(2), e0193247. 10.1371/journal.pone.0193247 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Wu H, Tiwari P, Han SS, & Chan T-K (2017). Brownfield risk communication and evaluation. Journal of Property Research, 34(3), 233–250. 10.1080/09599916.2017.1320684 [DOI] [Google Scholar]
  75. Yao Y, & Chen Q (2022). Numerical Models of Vapor Intrusion. In Vapor Intrusion Simulations and Risk Assessments (pp. 35–47). Springer. [Google Scholar]
  76. Yung L, Lagron J, Cazaux D, Limmer M, & Chalot M (2017). Phytoscreening as an efficient tool to delineate chlorinated solvent sources at a chlor-alkali facility. Chemosphere, 174, 82–89. 10.1016/J.CHEMOSPHERE.2017.01.112 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2039544-supplement-1.docx (112KB, docx)

RESOURCES