Article pubs.acs.org/est
Volatilization of Trichloroethylene from Trees and Soil: Measurement and Scaling Approaches William Doucette,†,* Heather Klein,† Julie Chard,† Ryan Dupont,† William Plaehn,‡ and Bruce Bugbee† †
Utah Water Research Laboratory, Utah State University, 8200 Old Main Hill, Logan, Utah 84322-8200, United States Parsons, 1700 Broadway, Suite 900, Denver, Colorado 80290, United States
‡
S Supporting Information *
ABSTRACT: Trichloroethylene (TCE) volatilization from leaves, trunk, and soil was measured to assess the significance of these pathways from phytoremediation sites at Travis and Fairchild Air Force Bases. Measurements were scaled temporally and spatially to estimate the annual volatilization of TCE at the Travis (0.82 ± 0.51 kg/yr) and Fairchild sites (0.014 ± 0.008 kg/yr). Volatilization was primarily through the leaf (0.34 ± 0.16 kg/yr at Travis and 0.01 ± 0.06 kg/yr at Fairchild) and soil (0.48 ± 0.36 kg/yr at Travis, 0.003 ± 0.002 kg/yr at Fairchild) pathways. The larger volatilization estimate at Travis was expected because of the site’s higher TCE groundwater concentrations. Using groundwater data collected in 2004 and 2009, calculations show that over the 5 year period, 1.7 and 0.015 kg of TCE were removed each year at the Travis and Fairchild sites, respectively. On the basis of the scaled field measurements, volatilization from the leaves and soil may play a significant role in TCE removal at both sites. Daily and seasonal variations were not addressed during the limited daytime sampling events, but the methods described here provide a novel and practical framework for evaluating the potential importance of volatilization of TCE and similar compounds at phytoremediation sites.
■
INTRODUCTION Phytoremediation has been promoted as a low cost sustainable remediation alternative for the cleanup of shallow groundwater contaminated with chlorinated volatile organic compounds (CVOCs) like trichloroethylene (TCE).1 Although there is evidence of an active uptake component for some compounds,2 plant uptake of xenobiotic organics is primarily a passive process3 driven by transpiration. Volatilization from leaves and stems, referred to as phytovolatilization, can be a significant loss mechanism for TCE and similar CVOCs.4,5 The estimated atmospheric half-life of TCE is 7 days due to its reaction with hydroxyl radicals6 Few direct field measurements of phytovolatilization of CVOCs from leaves7−9 and trunks4,8,10 have been reported, and the results vary greatly. Differences between the age and species of plants tested and the methods used to quantify and scale CVOC phytovolatilization complicate comparisons among studies. Phytovolatilization has typically been measured using static (no flow) or dynamic (flow through) enclosures.7,9−11 When using enclosures, care must be taken to minimize unnatural changes in environmental conditions (i.e., humidity, temperature, and pressure) so that the measured volatilization fluxes represent those occurring under ambient conditions and are appropriate for scaling. © XXXX American Chemical Society
The presence of trees over contaminated groundwater has also been reported to increase the volatilization of VOCs from the soil12 with the enhancement attributed to the decrease in soil water content associated with the transpiring trees12−14 and the root systems creating “preferential pathways” for VOC vapors.12 In addition to the influence of trees, the volatilization of VOCs from the soil surface depends on depth to groundwater, groundwater VOC concentrations, temperature, changes in atmospheric pressure, soil moisture, and porosity of the soil.12−17 Although diffusion is considered the dominant transport mechanism of VOCs through soil, advection, driven by atmospheric temperature and pressure changes,12,13,15 can also influence surface flux. Two general approaches have been used to measure soil VOC emission rates.18 Micrometeorological techniques are used to calculate emissions based on vertical air VOC concentration measurements above the soil surface and local meteorological data.19,20 Static or dynamic enclosures deployed Received: October 11, 2012 Revised: April 19, 2013 Accepted: May 3, 2013
A
dx.doi.org/10.1021/es304115c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
phytoremediation demonstration site was planted with 1130 poplar trees of three varieties to intersect a shallow TCE groundwater plume. In 2001, groundwater concentrations of TCE ranged from 400−17,000 μg/L.30 The three male hybrid poplar clone species, selected for their high water use, availability at local nurseries, and existing presence on Base were 184−411 (Populus trichocarpa x deltoides), OP-367 (Populus trichocarpa x nigra), and Eridano (Populus deltoides × maximowiczii). Natural thinning has reduced the number of trees to 273 as of September 2009 in the closed canopy. In 2009, the average tree height was 9 m (30 feet), and the average circumference was 34 cm. The range of tree height in 2009 was 3−15 m (10−50 feet). Precipitation at Fairchild AFB occurs primarily in the winter as snowfall. The average monthly rainfall for the winter (December through February), spring (March through May), summer (June through August), and fall (September through November) months are 49 mm (1.9 in.), 33 mm (1.3 in.), 20 mm (0.8 in.), and 34 mm (1.3 in.), respectively.31 The average annual maximum temperature at nearby Spokane Airport is 14 °C, and the minimum temperature is 3 °C. The growing season is 153 days,27 and the annual ETo in the area is 1146 mm (45.1 in.).31 At Fairchild AFB, limited sampling in 2009 (three monitoring points) showed that TCE groundwater concentrations vary from 1.4 μg/L in the west to 190 μg/L in the northeast corner of the planting area.32 Groundwater depths at the phytostabilization study area range from 3.6 to 6 m (12 to 20 feet) bgs. Groundwater flows to the northeast throughout the site with base-wide groundwater velocities ranging from 12 to 22 m (40 to 72 feet) per year.32 Sample Collection. Field sampling was limited to three daytime events between June and October 2009 at each site preventing an adequate assessment of seasonal and diurnal impacts. During the first sampling event, tree core samples were collected from approximately 20 trees and analyzed for TCE. In addition, nine soil surface volatilization samples were collected from within (three) and just outside (six) the planted area and analyzed for TCE. This initial sampling was performed to verify that the trees contained TCE, to determine the spatial distribution of TCE within the trees at the site, and to evaluate the significance of TCE volatilization directly from the soil surface. The focus of the second and third sampling events was to collect leaf and trunk volatilization samples from a small subset trees found to contain TCE during the initial sampling event. The accessibility of the trees for volatilization sampling and range of measured tree core concentrations were the main selection criteria used. Groundwater data was not specifically collected for the volatilization assessment. Data collected to fulfill base regulatory requirements was used under the assumption that the samples were collected using the required standard sampling and QA/QC protocols and that they are representative of the groundwater systems below these sites. Tree Cores. Tree cores were collected using a hand driven 5.15 mm (0.2 in.) increment borer and were placed directly in preweighed 20 mL headspace vials containing 10 mL of an acidified saturated sodium chloride solution. Each vial was then capped using a Teflon-coated butyl rubber septa and aluminum screw top lid. The vials were reweighed prior to analysis to determine the mass of fresh plant tissue collected. A HewlettPackard 7890A gas chromatograph (GC)/5973C mass spectrometer (MS) operated in selected ion monitoring
on the soil surface use the product of concentration and airflow through the device to calculate the emission rate.21,22 Scaling leaf volatilization measurements from small enclosures to entire trees or tree stands has been performed using leaf area7 or from estimates of the volume of transpired water.9,11 Scaling based on transpiration is preferred because it minimizes the problem of dissimilar volatilization rates between shaded and sunlit leaves. For scaling volatilization from trunks or soil, the individual measured area fluxes are typically extrapolated to the entire site. In most phytoremediation applications, contaminant concentrations vary widely within the site and cost considerations limit the number of samples that can be collected. This spatial variability complicates scaling to the entire site because concentration impacts volatilization from trees and soil. The Thiessen approach,23,24 used in hydrology and meteorology applications, may provide a practical approach for extrapolating small numbers of discrete and highly divergent volatilization samples to an entire phytoremediation site. The goal of this study was to develop a simple approach for cost-effectively estimating the total annual mass of TCE volatilized from phytoremediation sites at Travis and Fairchild Air Force Bases using refined measurement and scaling methods.
■
MATERIALS AND METHODS Site Descriptions. Travis AFB is located approximately 5 km (3 miles) east of Fairfield, California, midway between the cities of San Francisco and Sacramento. The 0.9 ha (2.24 acre) phytoremediation demonstration site (SD036) was planted with 480 red ironbark eucalyptus (Eucalyptus sideroxylon “Rosea”) between 1998 and 2000 to intersect a shallow TCE groundwater plume. In 1998, groundwater concentrations in the planted area ranged from 260 to 17,000 μg/L TCE.25 Red ironbark eucalyptus, a broadleaf evergreen, was selected because of its high water use, availability at local nurseries, and established presence on Base. As of October 2009, the tree stand had closed and naturally thinned to 388 living trees. In 2009, the average tree height was 10 m (33 feet), and the average circumference was 38 cm. The range of tree heights in 2009 was 2−18 m (6−59 feet). Average monthly rainfall depths for the winter (December through February), spring (March through May), summer (June through August), and fall (September through November) months are 124 mm (4.9 in.), 46 mm (1.8 in.), 2 mm (0.08 in.), and 36 mm, (1.4 in.) respectively.26 Average maximum/minimum temperatures at nearby Vacaville in January and July are 13/3 °C and 35/13 °C, respectively.26 The growing season in nearby Sacramento is 289 days,27 and annual reference evapotranspiration (ETo) in the area is 1255 mm (49.4 in.).28 Sampling conducted at Travis AFB in 2009 (13 monitoring points) indicated that TCE groundwater concentrations varied across the site with the lowest levels (9000 μg/L) found at the north end.29 Within the phytoremediation site, groundwater is located at approximately 6−12 m (20−40 feet) below ground surface (bgs) during the summer, but levels can fluctuate seasonally by 0.6− 1.2 m (2−4 feet). Groundwater flow at the site is to the southeast at a velocity of 21 m (69 feet) per year.29 Fairchild AFB is located approximately 15 km (9 miles) west of Spokane, Washington. In 2001, the 0.4 ha (1 acre) B
dx.doi.org/10.1021/es304115c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 1. Schematic of the leaf volatilization sampling system.
and collecting samples using the same times and flow rates used in the field gave an average recovery of 80 ± 19% (n = 4). Postspike blanks showed minimal carryover (