Environ. Sci. Technol. 2008, 42, 536–542
“Phytoscreening”: The Use of Trees for Discovering Subsurface Contamination by VOCs A. SOREK,† N. ATZMON,‡ O. DAHAN,§ Z. GERSTL,† L. KUSHISIN,† Y. LAOR,4 U. MINGELGRIN,† A. NASSER,† D . R O N E N , §,⊥ L . T S E C H A N S K Y , † N . W E I S B R O D , § A N D E . R . G R A B E R * ,† Institute of Soil, Water and Environmental Sciences, The Volcani Center, Agricultural Research Organization, POB 6, Bet Dagan 50250 Israel, Institute of Plant Sciences, The Volcani Center, Agricultural Research Organization, POB 6, Bet Dagan 50250 Israel, Blaustein Institutes for Desert Research, Zuckerberg Institute for Water Research, Dept. of Hydrology and Microbiology, Ben Gurion University of the Negev, Sde Boker Campus 84990 Israel, Institute of Soil, Water and Environmental Sciences, Newe-Ya’ar Research Center, Agricultural Research Organization, POB 1021, Ramat Yishay 30095 Israel, and Hydrological Service and Water Quality Division, Israel Water Authority, POB 20365, Tel Aviv 61203 Israel
Received August 13, 2007. Revised manuscript received October 17, 2007. Accepted October 31, 2007.
We tested the possibility of using tree cores to detect unknown subsurface contamination by chlorinated volatile organic compounds (Cl-VOCs) and petroleum hydrocarbons, a method we term “phytoscreening”. The scope and limitations of the method include the following: (i) a number of widespread Cl-VOC contaminants are readily found in tree cores, although those with very high vapor pressures or low boiling points may be absent; (ii) volatile petroleum hydrocarbons were not wellexpressed in tree cores; (iii) trees should be sampled during activeevapotranspirationandfromdirectionsthatarewellexposed to sunlight; (iv) there is not necessarily a direct correlation between concentrations measured in tree cores and those in the subsurface; (v) detection of a contaminant in a tree core indicates that the subsurface is contaminated with the pollutant; (vi) many possible causes of false negatives may be predicted and avoided. We sampled trees at 13 random locations in the Tel Aviv metropolitan area and identified ClVOCs in tree cores from three locations. Subsequently, subsurface contamination at all three sites was confirmed. Phytoscreening is a simple, fast, noninvasive, and inexpensive screening method for detecting subsurface contamination, and is particularly useful in urban settings where conventional methods are difficult and expensive to employ.
Introduction Widespread and persistent subsurface contamination by volatile organic compounds (VOCs) such as trichloroethene * Corresponding author e-mail:
[email protected]. † Institute of Soil, Water and Environmental Sciences, The Volcani Center. ‡ Institute of Plant Sciences, The Volcani Center. § Ben Gurion University of the Negev. 4 Newe-Ya’ar Research Center. ⊥ Israel Water Authority. 536
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
(TCE), tetrachloroethene (PCE), and benzene has been reported in most industrialized countries in the world. These compounds, particularly chlorinated VOCs (Cl-VOCs), can persist for decades in soils, sediments, and aquifers (1). Extensive and expensive site investigations are generally required to characterize subsurface contamination by VOCs. Studies have demonstrated that trees and other plants can absorb VOCs from the subsurface, a characteristic which is exploited for phytoremediation purposes (2, 3). Trees exposed to organic chemicals in soils, sediments, and water can take them up via the transpiration stream, sorption on roots, and gas diffusion into roots (4–8). Uptake varies with climate, soil organic matter and water content, plant characteristics (transpiration rate and root lipids content), and the contaminant’s physical-chemical properties (especially octanol–water distribution coefficient; Kow) (8), as well as the presence of multiple contaminants (9). A maximum in the transpiration stream concentration factor as a function of log Kow was observed at a log Kow of about 2.5 for a number of environmentally important VOCs (4). Efforts have been made to use tree cores or branches to delineate subsurface contamination plumes, thereby reducing costs and the time involved in site characterization (5, 7, 10–13). Subsurface concentration, depth to groundwater, tree species, and in-tree dechlorination were all found to have a major impact on the concentration of chlorinated ethenes in tree cores (7). To date, studies of VOCs in tree cores or branches have been conducted at sites of known subsurface contamination by Cl-VOCs. The current contribution was motivated by the possibility of using trees to detect unknown subsurface contamination sources of both Cl-VOCs and petroleum hydrocarbons, particularly in urban habitats. We term the proposed method “phytoscreening”. In many urban centers, historical records on the location and activities of former industries frequently are missing or incomplete. As a result, a “shotgun” approach to subsurface contaminant detection is required, the cost of which in large, populous areas can be prohibitive. Moreover, in highly urbanized settings, conventional screening methods involving the installation of monitoring wells or soil gas vapor probes are excessively complicated by the dense urban surface and subsurface infrastructure. To employ phytoscreening as an inexpensive, noninvasive tool for detecting unknown subsurface VOC contamination, the scope and limitations of the methodology need to be clearly defined. This was carried out by systematic tree core and subsurface sampling at two sites of known VOC contamination in an urban setting, one contaminated by Cl-VOCs and the other by petroleum hydrocarbons. Following this, the utility of phytoscreening for identifying sites of unknown subsurface VOC contamination in the Tel Aviv metropolis was tested.
Methodology General Study Area—Tel Aviv Metropolis. The study area, encompassing metropolitan Tel Aviv, is located over the central part of the Coastal Plain aquifer of Israel, a phreatic sandy aquifer stretching along 2000 km2 of the Mediterranean coast. The study area (168 km2) has around 1.19 million inhabitants and a population density of about 7065 inhabitants/km2. Industrial chemicals and wastes were not comprehensively regulated in Israel until the last decade of the 20th century, such that local factories released wastes directly to the land surface, local waterways, underground sewage pits, or the general sewage system. The main industries utilizing Cl-VOCs include textiles, metals, electronic equip10.1021/es072014b CCC: $40.75
2008 American Chemical Society
Published on Web 12/12/2007
ment manufacturing, and wood products. These industries also make widespread use of petroleum hydrocarbons, as do the many gas stations dotted throughout the urban center. For many decommissioned factories, there is little or no information on their location or history of solvent use and disposal. It has become known only recently that groundwater underlying the Tel Aviv metropolis is extensively contaminated by Cl-VOCs (14, 15). Cl-VOCs Polluted Site. The test site contaminated with Cl-VOCs was a former military industry facility (IMI Magen) chiefly involved in metal-working and plating. First commissioned in 1949, it reached a total area of about 44 000 m2 before decommissioning in 1995. Industrial processes included metal cleaning with chlorinated solvents [mainly, TCE and 1,1,1-trichloroethane (1,1,1-TCA)], electroplating, and metal painting. TCE at concentrations up to 300 000 µg/L were detected in groundwater at the water table (between 18 and 20 m below the land surface), and other Cl-VOCs were also detected at levels of several thousand micrograms per liter (14, 15). The same Cl-VOCs were found at high concentrations in the vadose zone (14, 15). Tree cores were taken from four specimens on the IMI Magen site: a Dalbergia sisso (rosewood, 32 cm in diameter and 13 m tall) and a Ficus elastica (rubber tree, 46 cm in diameter and 15 m tall), located 3 m from each other, and a Cupressus sempervirens L. (cypress, 46 cm in diameter and 17 m tall) and a Ficus microcarpa (laurel fig, 37 cm in diameter and 12 m tall), also located 3 m one from the other. Dalbergia sisso is a fast growing, deciduous tree species originally from southeast India which has shallow lateral roots, deeply penetrating tap roots, and high transpiration rates. Ficus microcarpa and Ficus elastica are evergreen tree species with aggressive but shallow root systems, and Cupressus sempervirens is a Mediterranean evergreen conifer species, with a well developed and relatively deep root system. A vadose zone monitoring well, M12, was established midway between the rosewood and rubber tree, and another, M14, midway between the cypress and laurel fig. Wells were sampled by a multilayer sampler (MLS) that consists of 150 mL stainless steel cells filled with distilled water and closed at both ends with a dialysis membrane (Versapor membrane, PALL Corporation, 0.2 µm) (14, 15). The cells are lowered into the well using a sampler that can accommodate any number of cells, which are isolated from each other by flexible Viton seals that fit the inner diameter of the well screen. For obtaining a groundwater profile or a profile of gases from the unsaturated zone, the sampler is lowered into an observation well and retrieved from it after a given time interval (usually several weeks), and the chemical composition of the water in each cell is determined. In the saturated zone, VOCs in groundwater equilibrate with water in the cell, while in the unsaturated zone, VOCs in the gas phase equilibrate with water inside the cell. Vadose zone gas phase concentrations (Cg) are calculated from measured cell water concentrations of VOCs (Cw) according to Cg ) KH* × Cw (KH* being a dimensionless form of Henry’s constant). Further details about the wells and MLS sampler are found in the Supporting Information. The four trees were sampled a number of times at different sampling heights and from different directions. Tree cores were also sampled from the same species of trees at an agricultural research area located outside the study region for use as controls. Cores 15-20-cm-long were obtained with an incremental hand borer (5 mm o.d.) and, after removal of the outer bark layer, were rapidly dissected on-site into 5-cm-long segments from the outside to the inside of the tree. The core segments were wrapped in aluminum foil and placed in 8 mL glass vials capped with PTFE-lined silicone septa and polypropylene hole caps. Samples were stored in an ice-containing cooler until their same-day arrival in the
laboratory, where they were stored at -80 °C until being processed for determination of VOC content. Processing involved lightly crushing the frozen sample immediately after removal from the freezer and transferring the crushed sample to a headspace vial equipped with Teflon-lined septa and hole-cap. Petroleum Hydrocarbon-Polluted Site. Tree sampling was carried out at a gas station in operation since 1956 (Delek Ramat Aviv), where a lens of mixed petroleum hydrocarbons (gasoline, kerosene, and diesel) approximately 1-m-thick was discovered floating on the groundwater table (8 m below the land surface) in 2000 (16). The lens and plume of dissolved benzene, toluene, ethyl benzene, and xylenes (BTEX) were previously delineated by 12 monitoring wells (16). In the current study, two rosewood specimens (36 and 31 cm diameter, and about 13 m high) located directly above the delineated floating lens and one eucalyptus tree (Eucalyptus camaldulensis, 50 cm diameter and about 25 m high) on the fringe of the contaminated groundwater plume were sampled. Groundwater was also sampled by a bailer from one of the wells located near a rosewood tree (rosewood 2). Random Tree Sampling in the Metropolitan Area and Verification of Subsurface Contamination. Tree cores from 35 mature trees at 13 randomly selected locations in the urban Tel Aviv area were sampled. Mature Eucalyptus spp., Ficus spp., and Dalbergia sisso (rosewood) trees were sampled, as these are common species in the area. At places where VOC contamination was detected in the tree cores, monitoring wells were established (17). The wells were finished at 5 m below the water table, and groundwater was sampled at 4 m below the water table by a submerged pump, following the evacuation of three well volumes (17). VOC Extraction and Analysis. Tree Cores. Sample vials containing the crushed cores were incubated at 95 °C for 20 min in a CombiPal autosampler (CTC Analytics AG, Switzerland). An aliquot of the heated headspace was withdrawn and injected into a gas chromatograph/mass spectrometer (GC/MS; Agilent). Extraction efficiency was tested for Eucalyptus spp., Ficus spp., and Dalbergia sisso wood by spiking frozen, crushed wood; details and results are reported in the Supporting Information, Figure S1. In the following, results (corrected for extraction efficiency) are presented for the outermost 0–5 cm core segment. Error bars in the figures for tree cores represent the coefficient of variation of replicate spiked crushed wood (see Supporting Information). Data and discussion on the distribution of VOCs inside the core to a 20 cm depth are presented in the Supporting Information, Figures S2 and S3. Vadose Zone Samples. Water samples retrieved from the MLS were collected in headspace-free 40 mL vials preloaded with 0.1 mL of 6 N HCl, preserved in an ice-containing cooler, and refrigerated at 4 °C for no more than 14 days before analysis. Aliquots from the vials were placed in 10 mL headspace vials, which were incubated at 90 °C for 20 min and then analyzed by GC/MS following the same method as for the tree cores. Error bars in the figures for vadose zone samples represent the coefficient of variation of 10 replicate spiked water samples. Analysis. GC separation was performed on a J&W-VRX column. Analytical details are given in the Supporting Information. Quantification was performed using a specific ion monitoring method based on EPA SW-846 method 8260B.
Results and Discussion Scope and Limitations of Phytoscreening Methodology. Tree core specimens collected from the control agricultural site yielded negative results for all VOC analytes at each sampling. VOCs detected in the unsaturated zone at the IMI Magen site, contaminated by Cl-VOCs, include TCE, 1,1,1TCA, 1,1-dichloroethene (1,1-DCE), 1,1-dichloroethane (1,1VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
537
FIGURE 1. Typical subsurface (lower pane) and tree core (upper pane) Cl-VOC results at the IMI Magen site sampled May 2005. Tree cores were sampled at a 90 cm height. Groundwater table is denoted by the line labeled wt. (A) Rosewood and rubber tree specimens were located by monitoring well M12, and (B) laurel and cypress tree specimens were located by monitoring well M14. Contaminants in the subsurface are found also in the tree stems, but cores from adjacent trees do not necessarily have similar contaminant concentrations. DCA), CHCl3, CCl4, and PCE. Representative profiles from the unsaturated zone gas phase of monitoring well M12 are shown in Figure 1A (bottom pane), displaying the general trend of increased concentration with depth in the subsurface. In the accompanying bar graph (Figure 1A, upper pane), contaminants identified in the nearby rosewood and rubber trees sampled on the same day as the unsaturated zone are shown. All the VOCs detected in the monitoring well are present in the tree cores, with the exception of 1,1-DCE, one of the dominant subsurface contaminants. In the M14 monitoring well, the major contaminants in the subsurface are 1,1,1-TCA, 1,1-DCE, and TCE (Figure 1B). While all of these components are observed in the cypress and laurel fig trees adjacent to monitoring well M14, it can be seen that 1,1-DCE is poorly expressed in the tree cores as compared with its concentration in the subsurface. PCE and CCl4 were identified at low levels in some of the trees but not in the subsurface at the May 2005 sampling event (Figure 1). However, both compounds have been identified in the subsurface at other times. The log Kow value of 2.13 for 1,1-DCE suggests that it is taken up and translocated by trees (4). Yet, 1,1-DCE was rarely detected in the sampled tree cores, despite its relatively high concentrations in the subsurface (Figure 1). The vapor pressure (Vp) of 1,1-DCE is substantially higher than that of the other contaminants, and its boiling point (Bp) is much lower (Vp ) 600 mmHg at 25 °C, Bp ) 31.7 °C; Table S1, Supporting Information). In the Tel Aviv region, air temperatures routinely reach values not far from the boiling point of 1,1-DCE, even in the winter, when the average daily maximum temperature is about 18 °C. Thus, we conjecture that, while 1,1-DCE is taken up into the transpiration stream, it is readily lost via volatilization both from the tree trunk 538
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
and the sampled tree cores. This result is also suggestive that other Cl-VOCs with particularly high vapor pressures and low boiling points, for example, vinyl chloride, may be difficult to detect by phytoscreening. While a number of the common Cl-VOCs, including 1,1-DCE, have been observed to undergo degradation under certain aerobic conditions (18–20), for the most part, aerobic degradation rates of 1,1-DCE do not exceed those of the other Cl-VOC solvents. As such, rhizosphere or in-tree degradation are considered unlikely to account for the conspicuous absence of 1,1-DCE from the tree cores. Large differences in VOC concentration between cores of trees located within 3 m of each other frequently are observed (Figure 1A,B). Such large differences in concentration between closely situated trees of different species confirm earlier reports (7) and demonstrate that no straightforward relationship between VOC concentrations in tree cores and VOC concentrations in the subsurface can be established. Furthermore, although VOC concentrations in the subsurface increase with depth (Figure 1), a relationship between concentrations in tree cores and the depth of the tree root system is not observed. For example, the cypress tree, a more deeply rooted species than the adjacent laurel fig, is expected to be exposed to higher subsurface concentrations (Figure 1B), yet the cypress exhibits much lower Cl-VOC concentrations than does the laurel fig. Variability in VOC concentration among trees is due in large part to inherent differences in transpiration rates between species and among specimens of the same species (21). Variability in VOC concentration within an individual tree is also observed along both vertical (Figure 2) and horizontal (Table 1) transects. Vertical VOC concentration profiles in the trees exhibited inconsistent concentration trends for a
FIGURE 2. Variations in Cl-VOC concentration as a function of sampling height in the (A) laurel and (B) rubber trees. Vertical VOC concentration profiles exhibit inconsistent concentration trends for a given compound, irregular extents of variability in concentration, and varying patterns for the different contaminants.
TABLE 1. Azimuthal Variations in Cl-VOC Concentrations in Tree Cores Obtained from a Single Height (90 cm) in Two Trees (Laurel Fig and Rubber Tree) at the IMI Magen Sitea laurel fig compound
S
N
W
1,1-DCA CHCl3 TCE
2.4 ( 0.4 1.8 ( 0.2 1.3 ( 0.1
2.9 ( 0.5 2.7 ( 0.5 3.3 ( 0.3 2.6 ( 0.3 4.7 ( 0.4 2.8 ( 0.2 rubber tree
1.0 ( 0.17 1.0 ( 0.1 1.0 ( 0.08
CHCl3 TCE
W 2.2 ( 0.2 1.3 ( 0.1
E 4.5 ( 0.4 4.8 ( 0.4
S 1.0 ( 0.1 1.0 ( 0.08
N 2.1 ( 0.2 1.5 ( 0.1
E
a The variations are given as the ratio of concentrations relative to the direction of the lowest concentration. The maximum observed variation in concentration as a function of azimuth is a factor of about 5. Error is indicated by (.
given compound, irregular extents of variability in concentration, and varying patterns for the different contaminants (Figure 2). A number of other comparisons of concentrations in cores collected at 30 and 90 cm heights also failed to reveal any consistent trends in VOC concentration as a function of height, with differences reaching as much as a factor of 5 (Supporting Information, Figure S3). Variations in VOC concentration up to a factor of about 5 were also found in cores taken at a single height (90 cm) from the four major compass directions (Table 1). Similarly large azimuthal variations in VOC concentrations around tree stems were reported earlier (5, 7, 11) and may be in part attributed to azimuthal variations in transpiration rates (22) that result from differences in the direction of sun exposure or heterogeneity in soil structure. Another important factor which may contribute to the large variations in VOC concentration around the trunk in a single tree is that root systems are spread out in the subsurface, where VOC concentrations are known to be very variable in space and time (15). VOC concentrations in tree cores also display strong seasonal variations due to seasonal variations in transpiration rates (23), which fall off significantly in the winter. In semiarid
climates such as Israel’s, in the winter wet season, lateral roots of trees acquire water by the uptake of recent precipitation contained in the upper soil layers, while tap roots derive water from the underlying water table. The stems thus obtain a mixture of the two water sources. As the dry season approaches, dependence on recent rainwater in the shallow subsurface decreases while that on groundwater from the deeper subsurface increases (24). Consequently, winter and summer samplings may represent both different rates of tree metabolism and water uptake from different depths in the subsurface. Seasonal changes in VOC concentrations in tree cores were examined by comparing samples taken from the rosewood and laurel fig from the same direction (S) and at the same height (90 cm) over five consecutive seasons. A clear seasonal trend in TCE concentration can be seen for both trees in Figure 3, with much higher concentrations recorded in the dry hot season (Oct. ’03, Sept. ’04, and May ’05) than in the wet cold season (Jan. ’04 and Feb. ’05). Subsurface TCE concentrations in the nearby monitoring wells for February ’05 and May ’05 are shown in the inset graphs. Concentration differences in the subsurface between these sampling dates are very small compared to the order of magnitude differences in concentrations observed in the tree cores, and therefore, the differences in subsurface concentrations cannot account for the seasonal differences observed in the tree cores. The shape of the TCE concentration curve along each of the two subsurface profiles (inset graphs, Figure 3) indicates that differences in the depth at which the roots take up water in the winter and the summer also cannot account for the order of magnitude reduction in the concentration of TCE in the winter. The fact that the seasonal variations repeat themselves regularly in two different trees over the course of five seasons suggests that this temporal variability is not random, in contrast to the apparently random spatial variability observed in the vertical and horizontal sampling (Figure 2, Table 1; Figure S3, Supporting Information). Altogether, it is assumed that the lower wintertime concentrations in the tree cores reflect lower rates of transpiration (23), accompanied by outward diffusion of the contaminants from the tree stem (5, 25, 26). The strong seasonal variations in VOC concentrations in tree cores have great significance for the phytoscreening method, as they VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
539
FIGURE 3. Seasonal variations in TCE concentration in the (A) rosewood and (B) laurel fig TCE concentrations in tree cores. These concentrations are high in the summertime and low in the wintertime. Concentrations of TCE in the subsurface for two of the samplings (Feb 2005, wintertime, and May 2005, summer time) are shown in the inset graphs. Groundwater table is denoted by the line labeled wt. TCE concentrations in the subsurface do not differ substantially between the wintertime and summertime samplings. demonstrate conclusively that Cl-VOCs do not accumulate inside trees over extended periods of time. A positive result in a tree core for Cl-VOCs indicates that the contamination is currently located in the subsurface and is not the remainder of some past contamination event which has already dissipated from the subsurface. Results of tree sampling for BTEX compounds at the gasoline station underlain by a lens of petroleum hydrocarbons at 8 m below the land surface are presented in Figure 4. Relatively low concentrations of BTEX (generally below 100 ng/g) were identified in the tree cores, although the two rosewood specimens were located directly above the hydrocarbon lens, with benzene, toluene, and xylene concentrations in the groundwater from a nearby monitoring well being 1100, 2400, and 860 µg/L, respectively (sampling date December ’03). In comparison, TCE concentrations in cores from the rosewood tree on the IMI Magen site were more than an order of magnitude higher (700-1800 ng/g; Figures 1 and 3), while exposed to 50–300 µg TCE/L in the gas phase of the unsaturated zone (equivalent to 125–750 µg TCE/L water). Vadose zone sediments below the sampled trees consist of sand and gravel (16), such that the low VOC concentrations in the tree cores cannot be attributed to a hydraulic barrier between the contamination and the tree roots. The log Kow, solubility, boiling and melting points, and vapor pressures of the four BTEX compounds are in the same range as those of the Cl-VOC compounds that are readily observed in tree cores (Table S1, Supporting Information), and the transpiration stream concentration factor for benzene in poplar cuttings was reported to be the same as for TCE (4). We speculate that the low content of BTEX in the tree cores may be due to BTEX degradation in the rhizosphere (27) or in the tree itself (28). Our data suggests that phytoscreening may be poorly suited for petroleum hydrocarbons, but this finding should be confirmed by additional study, for instance, with other tree species and at other sites. The results of sampling trees and the subsurface for VOCs at the two test sites, as well as results from previous studies (5, 7, 13), help to establish the scope and limitations of phytoscreening as a preliminary screening tool for VOCs in 540
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
FIGURE 4. Concentrations of BTEX compounds in tree cores from three trees at the gas station underlain by a lens of petroleum hydrocarbons floating on the groundwater: (A) October 2003, (B) October 2004, and (C) June 2005. Compounds are only present at low concentrations or not at all in the tree cores, despite the presence of an underlying lens of petroleum hydrocarbons. the subsurface: (i) A number of widespread Cl-VOC contaminants (PCE, TCE, cis- and trans-DCE, 1,1-DCA, 1,1,1-
TCA, CHCl3, and CCl4) are readily found in tree cores from a wide variety of tree species that are exposed to the contaminants in the subsurface. (ii) Contaminants with particularly high vapor pressures or low boiling points (such as 1,1-DCE) may be absent from tree cores. (iii) BTEX compounds were poorly expressed in tree cores in this study. (iv) Trees should be sampled during periods of active evapotranspiration and from directions that are well exposed to sunlight. (v) There is not necessarily a direct correlation between concentrations measured in tree cores and those measured in the subsurface. (vi) the detection of a contaminant in a tree core indicates with a very high degree of certainty that the subsurface is contaminated with the detected pollutant; that is, there are virtually no false positives. (vii) Some causes of false negatives (i.e., cases where a pollutant is not detected in the tree core although it is present in the root zone) such as low transpiration rates due to season, lack of sun exposure, or poor health of the tree, are predictable and avoidable. Other possible causes of false negatives, for instance, contaminant breakdown in the tree or rhizosphere (e.g., BTEX), or rapid outward diffusion from the tree stem (e.g., 1,1-DCE), may be anticipated on the basis of contaminant physical-chemical properties. (viii) Trees have a relatively large sampling volume (i.e., the root zone) compared with conventional subsurface sampling devices, but it is difficult to know the size or precise location of that volume, and it is expected to vary from tree to tree and from environment to environment as a function of climate, soil type, tree type, and tree age. Results of Phytoscreening in the Metropolitan Tel Aviv Area. Of 13 randomly selected locations for tree sampling, trees with measurable VOC contents were found at three sites, none of which were previously known to be contaminated. At the first location, PCE concentrations of up to 10 000 ng/g were measured in tree cores from several Eucalyptus spp. trees (average 53 cm in diameter, 25 m high) along a sidewalk in an urban residential part of the metropolis. There are no contemporary industries or factories in the area, although three water supply wells within 0.5 to 1 km of the site had been previously closed due to elevated PCE concentrations. There were no extant municipal records of industrial activity near the site of the trees. However, a survey of the long-term residents of the neighborhood revealed that a large dry-cleaning facility had stood at the site adjacent to the eucalyptus trees from the mid-1930s to the mid-1980s. The factory used PCE intensively in its cleaning operations. Newspaper archives revealed numerous complaints about air pollution originating from the factory throughout the 1970s and early 1980s, until the facility was closed and relocated to another area. The site was then developed for residential high-rise buildings. After the tree sampling results were reported to the appropriate authorities, a shallow groundwater monitoring well (water table about 18 m below the land surface) was established approximately 50 m upgradient of the sampled trees (17). The well revealed PCE in the groundwater at a concentration of 117 µg/L and TCE at 5 µg/L (17), thus confirming the subsurface contamination of the site, as predicted by phytoscreening. A second contaminated site was also identified on the basis of high concentrations of PCE (maximum of 4000 ng/ g) and lower concentrations of other VOC contaminants (cis1,2-DCE, 1,1,1-TCA, and TCE) in Eucalyptus spp. trees (average 52 cm in diameter, 25 m high) in a city park. A survey of the area revealed the site of a former IMI facility within 50 m of the trees, and an adjacent large industrial complex in existence since the 1950s, currently home to dozens of small industries, many of which use solvents on a regular basis for metal treatment and painting. Half of this industrial complex was replaced about 15 years ago by highrise residential buildings. A shallow groundwater monitoring
well (water table about 12.5 m below the land surface) installed at the site of the former IMI factory subsequent to the discovery of contaminants in the nearby trees revealed concentrations of about 4700 µg/L PCE; 410 µg/L TCE; 731 µg/L cis-1,2-DCE; and tens of micrograms per liter of 1,1DCE, 1,1-DCA, and trans-1,2-DCE in the groundwater (17). The third site at which Cl-VOCs were found in tree cores was a tree-lined boulevard located within a few hundred meters of the IMI Magen site, but in the upgradient groundwater flow direction. Dalbergia sisso (rosewood) trees (average 36 cm diameter, 11 m high) along the boulevard had concentrations of TCE ranging between 56 and 150 ng/ g. A subsequent investigation revealed that an electric appliances manufacturing factory had been located adjacent to the position of several of the sampled trees from the 1950s to 2001 (17). In a shallow groundwater monitoring well (water table about 19 m below the land surface) installed at the northern end of the tree-lined boulevard (downgradient of the aforementioned factory), about 150 m north of the nearest sampled tree, 2620 µg/L TCE, 71 µg/L PCE, and 21 µg/L of 1,1-DCE were detected in the groundwater (17). The successful application of phytoscreening, a simple, fast, noninvasive, and inexpensive screening method for detecting previously unknown subsurface Cl-VOC contamination, is thus demonstrated. Taking into consideration the scope and limitations of the methodology, we find that phytoscreening is a valuable preliminary subsurface screening tool, particularly in urban settings where it is difficult and expensive to install monitoring wells, and where passive gas monitoring devices may be subject to intentional or unintentional interference. Although very useful in screening large urbanized areas to pinpoint previously unknown contamination, the lack of straightforward correlation between concentrations of pollutants in tree cores and the adjacent subsurface limits the utility of tree core analysis for plume delineation or for monitoring temporal changes in pollutant concentrations in the subsurface.
Acknowledgments This study was supported by The Water Authority of the Israel Ministry of National Infrastructures, The Agricultural Research Organization of the Israel Ministry of Agriculture and Rural Development, and The Office of the Chief Scientist of the Israel Ministry of Environmental Quality. Contribution No. 610/07 from the ARO, The Volcanic Center, Bet Dagan, Israel.
Supporting Information Available A description of the multilayer sampling (MLS) methodology; a description of the method of extraction efficiency determination; a figure depicting the extraction efficiency results (Figure S1); details of the analytical procedure; a description of contaminant distribution inside the tree cores; a figure showing the distribution of contaminants inside the tree core (Figure S2); a table detailing selected physical-chemical data for the main contaminants of interest (Table S1); a figure showing VOC concentrations in tree cores taken at 30 and 90 cm height, along the entire 20 cm long core (Figure S3); and a list of references. This information is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Pankow, J. F.; Cherry, J. A. Dense Chlorinated Solvents and Other DNAPLs in Groundwater; Waterloo Educational Services Inc.: Ontario, Canada, 1996. (2) Newman, L. A.; Doty, S. L.; Gery, K. L.; Heilman, P. E.; Muiznieks, I.; Shang, T. Q.; Siemieniec, S. T.; Strand, S. E.; Wang, X. P.; Wilson, A. M.; Gordon, M. P. Phytoremediation of organic contaminants: A review of phytoremediation research at the University of Washington. J. Soil Contam. 1998, 7, 531–542. VOL. 42, NO. 2, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
541
(3) Collins, C.; Laturnus, F.; Nepovim, A. Remediation of BTEX and trichloroethene - Current knowledge with special emphasis on phytoremediation. Environ. Sci. Pollut. Res. 2002, 9, 86–94. (4) Burken, J. G.; Schnoor, J. L. Predictive relationships for uptake of organic contaminants by hybrid poplar trees. Environ. Sci. Technol. 1998, 32, 3379–3385. (5) Vroblesky, D. A.; Nietch, C. T.; Morris, J. T. Chlorinated ethenes from groundwater in tree trunks. Environ. Sci. Technol. 1999, 33, 510–515. (6) Ma, X. M.; Burken, J. G. VOCs fate and partitioning in vegetation: Potential use of tree cores in groundwater analysis. Environ. Sci. Technol. 2003, 37, 2024–2024. (7) Vroblesky, D. A.; Clinton, B. D.; Vose, J. M.; Casey, C. C.; Harvey, G.; Bradley, P. M. Ground water chlorinated ethenes in tree trunks: Case studies, influence of recharge, and potential degradation mechanism. Ground Water Monit. Rem. 2004, 24, 124–138. (8) Collins, C.; Fryer, M.; Grosso, A. Plant Uptake of Non-Ionic Organic Chemicals. Environ. Sci. Technol. 2005. (9) Graber, E. R.; Sorek, A.; Tchehansky, M.; Atzmon, N. Competitive uptake and sorption of trichlorethene and 1,1,1-trichloroethane by seedlings and wood. Environ. Sci. Technol. 2007, in press. (10) Ma, X. M.; Burken, J. G. VOCs fate and partitioning in vegetation: Use of tree cores in groundwater analysis. Environ. Sci. Technol. 2002, 36, 4663–4668. (11) Schumacher, J. G.; Struckhoff, G. C.; Burken, J. Assessment of subsurface chlorinated solvent contamination using tree cores at the Front Street site and a former dry cleaning facility at the Riverfront Superfund site, New Haven, Missouri; United States Geological Survey: Reston, VA, 2004; p 35. (12) Struckhoff, G. C.; Burken, J. G.; Schumacher, J. G. Vapor-phase exchange of perchloroethene between soil and plants. Environ. Sci. Technol. 2005, 39, 1563–1568. (13) Gopalakrishnan, G.; Negri, M. C.; Minsker, B. S.; Werth, C. J. Monitoring subsurface contamination using tree branches. Ground Water Monit. Rem. 2007, 27, 65–74. (14) Laor, Y.; Ronen, D.; Graber, E. R. Using a passive multilayer sampler for measuring detailed profiles of gas-phase VOCs in the unsaturated zone. Environ. Sci. Technol. 2003, 37, 352–360. (15) Ronen, D.; Graber, E. R.; Laor, Y. Volatile organic compounds in the saturated-unsaturated interface region of a contaminated phreatic aquifer. Vadose Zone J. 2005, 4, 337–344. (16) Risk Assessment and Corrective Action Plan, Delek Gas Station, Ramat Aviv, Israel; Ministry of National Infrastructures: Jerusalem , 2005; p 66.
542
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 2, 2008
(17) Risk Assessment and Development of Remedial Plans for Soil and Groundwater Contamination in the Dan Regio Area and Provision of Other Consultant Services, Task 4 and 5 Summary Report; Ministry of National Infrastructures: Jerusalem , 2007. (18) Klier, N. J.; West, R. J.; Donberg, P. A. Aerobic biodegradation of dichloroethylenes in surface and subsurface soils. Chemosphere 1999, 38, 1175–1188. (19) Shim, H.; Ryoo, D.; Barbieri, P.; Wood, T. K. Aerobic degradation of mixtures of tetrachloroethylene, trichloroethylene, dichloroethylenes, and vinyl chloride by toluene-o-xylene monooxygenase of Pseudomonas stutzeri OX1. Appl. Microbiol. Biotechnol. 2001, 56, 265–269. (20) Verce, M. F.; Gunsch, C. K.; Danko, A. S.; Freedman, D. L. Cometabolism of cis-1,2-dichloroethene by aerobic cultures grown on vinyl chloride as the primary substrate. Environ. Sci. Technol. 2002, 36, 2171–2177. (21) Cohen, Y.; Cohen, S.; Aviles, T. C.; Schiller, G. Variations in the radial gradient of sap velocity in trunks of forest and fruit trees. Plant Soil 2007, in press. (22) Cermak, J.; Cieciala, E.; Kucera, J.; Hallgren, J. E. Radial velocity profiles of water flow in trunks of Norway spruce and oak and the response of spruce to severing. Tree Physiol. 1992, 10, 367– 380. (23) Nietch, C. T.; Morris, J. T.; Vroblesky, D. A. Biophysical mechanisms of trichloroethane uptake and loss in baldcypress growing in shallow contaminated groundwater. Environ. Sci. Technol. 1999, 33, 2899–2904. (24) Dawson, T. E.; Pate, J. S. Seasonal water uptake and movement in root systems of Australian phraeatophytic plants of dimorphic root morphology: a stable isotope investigation. Oecologia 1996, 107, 13–20. (25) Ma, X. M.; Burken, J. Modeling of RE diffusion to the atmosphere and distribution in plant stems. Environ. Sci. Technol. 2004, 38, 4580–4586. (26) Ma, X. M.; Burken, J. G. TCE diffusion to the atmosphere in phytoremediation applications. Environ. Sci. Technol. 2003, 37, 2534–2539. (27) Orwell, R. L.; Wood, R. L.; Tarran, J.; Torpy, F.; Burchett, M. D. Removal of benzene by the indoor plant/substrate microcosm and implications for air quality. Water Air Soil Pollut. 2004, 157, 193–207. (28) Topp, E.; Scheunert, I.; Korte, F. Kinetics Of The Uptake Of C-14-Labeled Chlorinated Benzenes From Soil By Plants. Ecotox. Environ. Safe. 1989, 17, 157–166.
ES072014B
