Environ. Sci. Technol. 2008, 42, 5247–5253
Heavy Metal Capture and Accumulation in Bioretention Media HOUNG LI AND ALLEN P. DAVIS* Department of Civil and Environmental Engineering, University of Maryland, College Park, Maryland 20742
Received October 24, 2007. Revised manuscript received April 21, 2008. Accepted April 23, 2008.
Heavy metal capture and accumulation in bioretention media were investigated through the use of a one-dimensional filtration equation for particulate metals, advection/dispersion/ adsorption transport equations for dissolved metals, and sequential extractions. Predicted spatial profiles and partitioning patterns of captured metals were compared to data derived from a bioretention cell in the District of Columbia. Zinc, lead, and copper profiles showed a high surface accumulation, significantly decreasing with the media depth. Surface street particle-enriched areas had the highest heavy metal levels, demonstrating a close relationship between capture of metals and runoff particles. Sequential extractions suggested that most captured metals were of anthropogenic origin. Solubleexchangeable bound metals from the sequential extraction correlated well with predicted aqueous dissolved metals; the more strongly associated metal fractions correlated with modeled runoff and media particulate metals. A simple risk evaluation indicatedthatleadisthelimitingmetalinbioretentionaccumulation. On the basis of information collected in this study, a shallow bioretention cell design is suggested for systems with a focus on metal capture.
Introduction As urban stormwater runoff flows across impervious areas, it collects and accumulates pollutants that are detrimental to the quality of local receiving water bodies. In the past several years, many best management practices (BMPs), such as bioretention, vegetated filter strips, and green roofs (1), have been developed and deployed to curtail the nonpoint source water pollution that traditionally accompanies urban growth and to attempt to restore the predevelopment hydrology. Many of these BMPs employ filtration/infiltration through porous media, allowing collected stormwater to percolate through some type of filter media as a treatment and to slow flow. The treated water is allowed to continue for groundwater recharge, or it is collected in an underdrain for discharge to local receiving waters. Among these BMPs, bioretention, also known as a rain garden, is increasingly being adopted (1). Bioretention generally consists of a porous soil media layer covered with a thin layer of hardwood mulch. A vegetation mix (grasses, shrubs, and small trees) is planted to promote evapotranspiration, biological activity, and pollutant uptake, as well as to maintain soil porosity and hydraulic conductivity. In studies to date, bioretention has demonstrated a good-toexcellent performance in improving water quality (2, 3). * Corresponding author phone: (301)405-1958; fax: (301)405-2585; e-mail:
[email protected]. 10.1021/es702681j CCC: $40.75
Published on Web 06/11/2008
2008 American Chemical Society
Mechanisms of pollutant removal include sedimentation, filtration, adsorption, precipitation, and biological transformations. A major advantage inherent to bioretention is an ability to remove both particulate and dissolved pollutants. However, issues related to long-term accumulation of pollutants captured in bioretention media have been addressed only minimally, and pollutant capture mechanisms are not adequately identified and quantified. Heavy metals are found in urban runoff originating from a variety of sources (4–6), and laboratory studies have indicated the excellent removal of metals using bioretention media (3). Characterization and speciation of metals in runoff are not wellunderstood. Particulate metals may be in elemental, mineral, or sorbed form; dissolved speciation may include a divalent metal form and complexes with dissolved organic matter and carbonate (7). Nonetheless, the fate and spatial profiles of captured toxic and persistent pollutants within bioretention media and the possibilities of re-entrainment into infiltrating runoff are of great concern with respect to BMP performance, as well as to design and maintenance issues. Heavy metal pollutant profiles that decrease significantly with depth have been noted in the media of stormwater infiltration basins (8), and the profile characteristics can have major implications for pollutant buildup, design recommendations, and the degree (replacement depth) of BMP maintenance. However, this pollutant profile needs to be authenticated for bioretention, and mechanisms that are responsible for the profile have not been resolved. This paper presents a quantitative theory for bioretention pollutant capture and the results of media analyses for heavy metals for a bioretention cell (supported by a 1.5 year water quality monitoring program) in the District of Columbia.
Modeling Metal Capture A descriptive model illustrating heavy metal capture within a bioretention media column is depicted in Figure 1. The incoming runoff carries both dissolved and particulate-bound pollutants. Sansalone and Buchberger (9) reported that copper and zinc are mainly in a dissolved form, while lead is primarily particulate-bound in urban roadway stormwater runoff. Dissolved pollutants can adsorb to the surfaces of deposited street particles and the bioretention soil media. This process can be interpreted with the one-dimensional advection/dispersion/adsorption transport equation within the control volume (10)
( )
( ) ( )(
∂C ∂2C F ∂Cs ∂C -u )D ∂t ∂Z ε ∂t ∂Z2
)
(1)
where C and Cs are the dissolved and media-sorbed pollutant concentrations, respectively (Figure 1), t is time, Z is the vertical media depth, D is the hydrodynamic dispersion coefficient, u is the average runoff velocity, and F and are the bulk mass density and porosity of the media, respectively. A simple linear isotherm can be proposed for the adsorption term for heavy metals to evaluate media metal accumulation (10, 11) Kds )
( ) Cs C
(2)
where Kds is the distribution coefficient for bioretention soil media (L/kg). Previous studies have indicated that metal adsorption on mineral surfaces is nonlinear; however, linear expressions are used because of their simplicity and ease in comparison for regulatory or other purposes (11, 12). VOL. 42, NO. 14, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5247
As runoff enters and percolates through the media column, particulate-bound metals are trapped through filtration processes. A first-order relation can be used to describe the depth profile of captured particles within bioretention media (13), which is correlated to particulate-bound metal concentrations Cp ) qRm + KdsFC ) qRm0 exp(-λZ) + KdsFC ) Kdpm0C exp(-λZ) + KdsFC (3) where Cp is the concentration of metal sorbed to particulate matter, qR is the metal mass per unit runoff suspended solids mass (Figure 1), m is the concentration of runoff particles collected within the media, m0 is the concentration of runoff particles accumulated at the media surface, and λ is the filter coefficient. The first term in eq 3 is the particulate metal concentration associated with the captured runoff particles, and the second term is the concentration associated with bioretention media. Similar to eq 2, the partitioning between dissolved metals and runoff particles can be expressed as
( )
(4)
where Kdp is the distribution coefficient for runoff particles. The sum of the dissolved and particulate metal concentrations CT at any point in the column can be described by CT(Z, t) ) C + Cp ) C[1 + Kdpm0 exp(-λZ) + KdsF]
CT0 ) C0[1 + Kdpm0 + KdsF]
(6)
and 1 + Kdpm0 exp(-λZ) + KdsF 1 + Kdpm0 + KdsF
)
( )
)
(7)
where C0 is the dissolved influent metal concentration at the media surface. Eq 1with eq 2can be solved to find the aqueous concentration employing an initial condition (C ) 0) and boundary conditions of a constant influent dissolved metal concentration C0 and a semi-infinite and uniform media column (C ) 0 at Z ) ∞) (10).
(
)]
(8)
where R ) [1 + (F/)Kd] is the retardation factor. The captured metal mass per unit particulate matter mass (including both runoff and media particles) at depth Z is denoted as M (mg/kg; M at the media surface is denoted as M0). At depth Z, the mass summation of the media and captured runoff particles per unit media bulk volume is media particle concentration ) F + m0 exp(-λZ)
(9)
As such, an overall metal distribution coefficient at depth Z, Kd/ (L/kg), is defined as a linear combination of the metal distribution coefficients of the runoff and media particles Kd/ )
FKds + m0 exp(-λZ)Kdp M ) F + m0 exp(-λZ) C
(10)
At the media surface (11)
/ where Kd0 is the Kd/ value at Z ) 0. Combining eqs 10and 11, the metal spatial profile in the bioretention media column over time is
( )(
Kd/ M ) / M0 Kd0
(5)
Correspondingly, at the media surface, the sum of the dissolved and particulate metal concentration CT0 is
( )(
[ (
C0 RZ - ut uZ RZ + ut erfc + exp erfc 2 D √ 2 RDt 2√RDt
/ M0 ) Kd0 C0
qR Kdp ) C
CT C ) CT0 C0
C(Z, t) )
C C0
)
(12)
Methodology Site Description and Media Collection. The bioretention cell is located along the Anacostia River in the District of Columbia, built during the summer of 2001. The cell (design drainage area of 0.077 ha), located in an active parking lot, is trapezoidal-shaped (sides of 2.9, 5.4, and 6.3 m), with a media depth of ∼1.1 m. The original media consisted of 50% (by vol) sand, 30% top soil, and 20% mulch. Media core samples were taken from the surface to 85-90 cm deep. The corer was submerged in 15% Na2EDTA solution (Fisher Scientific) overnight before it was used, flushed with deionized water, and air-dried. Ten- or 20-cm segments were separated with a precleaned knife to examine the vertical profile of metals in the media. An additional sample was taken from an obvious accumulation of deposited runoff particles near one runoff entrance. Nitrile gloves were worn during all sampling, and the collected samples were double
FIGURE 1. Descriptive model for dissolved/particulate metal capture in a bioretention media column. Metal partitioning occurs between the particulate matter phase (both media particles and runoff particles) and the aqueous phase; see text for details. 5248
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 14, 2008
TABLE 1. Sequential Extraction Scheme for Fractionation of Copper, Lead, and Zinc within Bioretention Media (14) solid/ liquid ratio (g/mL)
step
fraction
extraction procedure
F1
solubleexchangeable
0.1 M Sr(NO3)2 (4 h)
1:16
F2
sorbedcarbonate
1 M Na-acetate (pH 5, 5 h)
1:16
F3
oxidizable
5% NaOCl (pH 8.5, 1.5 h at 90-95 °C)
1:8
F4
reducible
0.2 M oxalic acid + 0.2 M NH4 oxalate + 0.1 M ascorbic acid (pH 3, 1.5 h at 90-95 °C)
1:25
F5
residual
aqua regia (HCl + HNO3, 1.5 h at 70 °C)
1:16
bagged. Media samples were air-dried, sieved at 2 mm, and heated at 103-105 °C for 3 days before further analysis. Media samplings were carried out in December 2004 and December 2005 at the same general locations. Characterization of Media Samples. Approximately 400 g of media from the second bioretention sampling was sent to the University of Delaware Soil Testing Program for characterization. The media pH spatial profile was determined by adding 25 mL of deionized water to 5 g of samples from different depths, stirring intermittently for 1 h, and then left to stand for 0.5 h. Copper, lead, and zinc were analyzed employing a media sequential extraction technique (Table 1) modified from the work of Ahnstrom and Parker (14). Unless otherwise noted, all chemicals used in the sequential extraction were American Chemical Society reagent-grade or better. All labware was acid-washed and thoroughly rinsed with deionized water. For each sample, ∼0.5 g of dried media was used. At the end of each extraction, the sample was centrifuged at 3000 rpm for 10 min and filtered using a 0.2 µm membrane disk filter (Pall Corporation). The supernatant was diluted to 50 mL with deionized water for subsequent metal analyses. Metal analyses were carried out on a PerkinElmer 5100ZL atomic absorption spectrophotometer. Copper and lead concentrations were determined on a furnace module, according to Standard Method 3110 (15). Zinc concentrations were determined on a flame module, according to Standard Method 3111 (15). Analytical standards were prepared using 1000 mg/L stock solutions for each metal (Fisher Scientific or VWR Scientific Products). Duplicates were employed for each media sample extraction. Reagent blanks for all extractants were analyzed in parallel with samples and found to be below the media quantification limits (0.2 mg/kg for Cu and Pb and 5 mg/kg for Zn) in all cases. During analyses, a standards check was performed for every 10 samples, and only deviations Cu> Zn for soils (11, 28, 29) and sediments (30). In this study, the percentages of the sums of F1-F4 fractions to the total metal accumulation (F1-F5) for the surface layers (in which most metal accumulation occurred) over time were 78% for copper, ∼100% for lead, and 72% for zinc. These high percentages suggest that the origins of the captured metals were mostly anthropogenic and could be attributed to metal accumulation from incoming runoff. Furthermore, the captured metals exhibited a strong association with the media, suggesting that they are not washed
TABLE 2. List of Selected Parameters for Proposed Model in This Studya parameters
values
filter coefficient av input TSS media porosity media bulk density av runoff velocity hydrodynamic dispersion coefficient time metal distribution coefficients (L/kg) Cu Pb Zn a
λ ) 0.02 cm-1 m0 ) 0.00016 kg/L ) 0.47 F ) 1.36 kg/L u ) 0.96 cm/h D ) 42 cm2/h t ) 4.5 year for runoff particles, Kdp 9200 22660 3432
ref 13 33 33 34 35 for soil media, Kds 4799 171214 11615
9 and 22 9 and 22 9 and 22
See text for details.
out by subsequent wet weather flow. Minimal uptake by bioretention vegetation also may be expected (31). Metal Accumulation Modeling Parameters. Water quality data were used to define the model parameters. Input pollutant concentrations to the cell were similar to those typical of urban stormwater runoff (ref 1 and Supporting Information Table S-2). The input pollutant concentrations demonstrated significant variability since they are subject to variable precipitation patterns, drainage area land use, habitant activities, and other local factors. After treatment, ranges of the output pollutant concentrations were much smaller. Overall, the bioretention cell demonstrated good removal efficiencies for particulate matter and heavy metals, in agreement with other studies (3, 32), with the exception of copper. Median removal efficiencies for lead and zinc were 77 and 83%, respectively, with median discharge concentrations of 15 and 18 µg/L, respectively. Increases in copper concentrations were found, suggesting some washout from background media. Input and output TSS values were used to calculate the filter coefficient λ using (13) ln λ)-
Coutput Cinput Z
tions since the model assumes no associated metal at t ) 0, which was particularly obvious in the copper data. The steadystate and linear isotherm assumptions also are expected to introduce error. Both measurement and model indicate that lead has the sharpest spatial profile as compared to copper
(13)
The calculated λ values ranged from 0.01 to 0.06 cm-1 (median of 0.02 cm-1). Similarly, the median input TSS value (m0 ) 160 mg/L) for this facility also was used in model evaluation (Table 2). The field bulk density and porosity were estimated (Table 2) from the bioretention media texture and organic content using the method of Tomasella and Hodnett (33). The average runoff velocity u was estimated from the local annual precipitation (102 cm (34), cell area of 17 m2), design drainage area (770 m2), and media porosity assuming a drainage area runoff coefficient of 0.9. Literature values of the metal distribution coefficients for runoff particles Kdp (9) and soil media Kds (22) were used (mean values), as well as the hydrodynamic dispersion coefficient for sandy loam sites (35). Model Application. Figure 3 presents a comparison between the measured media metal contents for F1-F5 in December 2005 (averaged values are in dimensionless forms, normalized with respect to the surface pollutant concentrations). Also shown is the TSS concentration profile predicted by eq 13, which decreases exponentially but is not as sharp as measured metal profiles, indicating that the metals were not only captured through filtration but also by media and runoff particle adsorption. Predicted total media metal concentrations also are shown in Figure 3. Results indicate a successful prediction, reflected by the match between predicted and measured metal profiles. The background media metals introduce error in the predic-
FIGURE 3. Comparison among the measured media metals for each fraction, media metal level predictions, dissolved metal concentrations, and TSS profiles in the bioretention facility, December 2005 (t ) 4.5 years) for Cu, Pb, and Zn. (Figure legend: black: F1; white: F2; striped: F3; gray: F4; and white dots: F5). VOL. 42, NO. 14, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5251
TABLE 3. Estimates of Accumulated Heavy Metal Levels in Surface Media Layer of Bioretention Cella Cu total metal concentration of surface media, Dec 2004 (mg/kg) total metal concentration of surface media, Dec 2005 (mg/kg) av regulatory guidance values for residential soils (mg/kg) (36) a
Pb
Zn
29
399
111
75
660
532
3700
350
28000
See text for details.
and zinc, which is attributed to its higher Kd values associated with both media and runoff particles (Table 2). The dissolved metal component was calculated separately (Figure 3). This value is very small as compared to the particulate fraction. The extraction results indicate that soluble-exchangeable bound metals (F1) correlate well with the predicted aqueous dissolved metal fractions. Since F1 has the weakest particle association, a strong correlation may be expected (18). The low F1 values and predicted dissolved metal concentrations indicate the stability of the captured metals within the bioretention cell. As compared to lead and zinc, the median influent copper concentration was low (91 and 145 µg/L, respectively, visa`-vis 45 µg/L for Cu; see Supporting Information), indicating a potential for Cu desorption. Furthermore, on the basis of our field data and modeled findings, the majority of the metals was removed through filtration and adsorption in the top media layers. In the bottom layers, the runoff copper concentrations in the percolating runoff can be very low, prompting media copper desorption over time. Additionally, as previously mentioned, copper’s weaker association with soil media (as compared to lead and zinc) and its tendency to associate with dissolved organic matter also may promote washout. Environmental Significance and Outlook. Bioretention has demonstrated excellent performance in heavy metal removal, which can prevent downstream metal toxicity. However, the nonbiodegradable nature of heavy metals demands further examination for media metal accumulation. A simple risk evaluation for metal accumulation in bioretention is presented in Table 3. Regulatory guidance values for heavy metals in residential soils for 30 states in the U.S. were compiled, which are typically based on child soil exposure at home or other corresponding locations (36). The mean value from this compilation is used for each metal as a representative regulatory limit (Cu: 3700 mg/kg, range of 25-20 000 mg/kg; Pb: 350 mg/kg, 61-500 mg/kg; and Zn: 28 000 mg/kg, 20-170 000 mg/kg). Bioretention facilities are not necessarily located in residential areas but may be accessible to children. Measured total metal levels in the bioretention cell surface layer indicate that lead exceeded the mean limit, although most of the lead was tightly bound. While lead levels for the majority of the media in the cell may be below levels of concern, the uneven lead profile characteristics may dictate the need for surface media layer replacement to eliminate lead accumulation in the media. However, the degree (media replacement depth) and frequency of pollutant cleanup will depend on local runoff characteristics and bioretention cell conditions. Comprehensive runoff monitoring and media sample collection/ analyses for bioretention cells are costly and time-consuming. Employing the metal accumulation model with only a few simple inputs (such as inflow and outflow TSS concentrations, precipitation data, media soil texture and organic matter 5252
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 14, 2008
content, and cell size), the captured metal spatial profile over time can be estimated. The numerical model is based on steady-state and linear isotherm assumptions. If more comprehensive bioretention modeling is required, nonsteady-state and nonlinear isotherm numerical development based on this model structure may be employed to improve accuracy at the cost of greatly increasing model complexity and input data requirements. Results of this study also can be applied to bioretention design. Since most captured metals accumulated within 10-20 cm of the surface, a more shallow media design of 20-40 cm is proposed for a system with focus on metals capture. A reduction in the media depth will significantly reduce construction costs and increase the feasibility for BMP installation, such as in places with an elevated groundwater table and/or shallow stormwater infrastructure. The media selection should be pollutant dependent. A low copper content media is recommended to increase the copper treatment capacity.
Acknowledgments The authors thank Navid Ariaban, Walter Caldwell, Chen Chiu, Ameya Pradhan, and Jim Stagge for help with sample collection. This research was supported by the District of Columbia Department of Health, Environmental Health Administration.
Supporting Information Available Details regarding monitored bioretention cell, including a map and a photo (Figures S1 and S2), media characteristics (Table S1), and water quality monitoring results (Table S2). This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Davis, A. P. Green engineering principles promote low impact development. Environ. Sci. Technol. 2005, 39 (16), 338–344. (2) Davis, A. P.; Shokouhian, M.; Sharma, H.; Minami, C. Laboratory study of biological retention for urban storm water management. Water Environ. Res. 2001, 73 (1), 5–14. (3) Davis, A. P.; Shokouhian, M.; Sharma, H.; Minami, C.; Winogradoff, D. Water quality improvement through bioretention: Lead, copper, and zinc removal. Water Environ. Res. 2003, 75 (1), 73–75. (4) Davis, A. P.; Shokouhian, M.; Ni, S. Loading estimates of lead, copper, cadmium, and zinc in urban runoff from specific sources. Chemosphere 2001, 44, 997–1009. (5) Hur, J.; Yim, S.; Schlautman, M. A. Copper leaching from brake wear debris in standard extraction solutions. J. Environ. Monit. 2003, 5, 837–843. (6) Councell, T.; Duckenfield, K. U.; Landa, E. R.; Callender, E. Tirewear particles as a source of zinc to the environment. Environ. Sci. Technol. 2004, 38 (15), 4206–4214. (7) Dean, C. M.; Sansalone, J. J.; Cartledge, F. K.; Pardue, J. H. Influence of hydrology on rainfall-runoff metal element speciation. J. Environ. Eng. 2005, 131 (4), 632–642. (8) Dechesne, M.; Barraud, S.; Bardin, J. Experimental assessment of storm water infiltration basin evolution. J. Environ. Eng. 2005, 131 (7), 1090–1098. (9) Sansalone, J. J.; Buchberger, S. G. Partitioning and first flush of metals in urban roadway storm water. J. Environ. Eng. 1997, 123 (2), 134–143. (10) Bedient, P. B.; Hanadi, S. R.; Newell, C. J. Ground Water ContaminationsTransport and Remediation, 2nd ed.; PrenticeHall: New York, 1999; pp 172-173. (11) Sauve´, S. Chemical speciation of metals in soils. In Bioavailability of Metals in Terrestrial Ecosystems: Importance of Partitioning for Bioavailability to Invertebrates, Microbes, and Plants; Allen, H. E., Ed.; SETAC: Pensacola, FL, 2002; pp 7-38. (12) Sparks, D. L. Environmental Soil Chemistry; Academic Press: San Diego, 1995. (13) Iwasaki, T. Some notes on sand filtration. J. Am. Water Works Assoc. 1937, 29, 1591–1597. (14) Ahnstrom, Z. S.; Parker, D. R. Development and assessment of a sequential extraction procedure for the fractionation of soil cadmium. Soil Sci. Soc. Am. J. 1999, 63, 1650–1658.
(15) APHA; AWWA; WPCF. Standard Methods for the Examination of Water and Wastewater, 17th ed.; APHA: Washington, DC, 1995. (16) Lau, S.; Stenstrom, M. K. Metals and PAHs adsorbed to street particles. Water Res. 2005, 39, 4083–4092. (17) Brown, J. N.; Peake, B. M. Sources of heavy metals and polycyclic aromatic hydrocarbons in urban stormwater runoff. Sci. Total Environ. 2006, 359, 145–155. (18) Turer, D.; Maynard, J. B.; Sansalone, J. J. Heavy metal contamination in soils of urban highways: Comparison between runoff and soil concentrations at Cincinnati, OH. Water, Air, Soil Pollut. 2001, 132, 293–314. (19) Logan, T.; Miller, R. Background Levels of Heavy Metals in Ohio Farm Soils; http://www.ag.ohio-state.edu/∼ohioline/rc275/ index.html (accessed May 2006). (20) Deng, X.; Jennings, A. A. Evaluating an electrokinetically driven extraction method for measuring heavy metal soil contamination. J. Environ. Eng. 2006, 132 (4), 527–537. (21) NRSC, Soil Geochemistry Spatial Database; http://soils.usda.gov/survey/geochemistry (accessed May 2005), sampling in 2003 at Baltimore County, MD. (22) Sauve´, S.; Hendershot, H.; Allen, H. E. Solid-solution partitioning of metals in contaminated soil: Dependence on pH, total metal burden, and organic matter. Environ. Sci. Technol. 2000, 34, 1125–1131. (23) Wen, B.; Hu, X.; Liu, Y.; Wang, W.; Feng, M.; Shan, X. The role of earthworms (Eisenia fetida) in influencing bioavailability of heavy metals in soils. Biol. Fertil. Soils 2004, 40, 181– 187. (24) Smolders, E.; Degryse, F. Fate and effect of zinc from tire debris in soil. Environ. Sci. Technol. 2002, 36 (17), 3706–3710. (25) Norrstro¨m, A. C.; Jacks, G. Concentration and fractionation of heavy metals in roadside soils receiving de-icing salts. Sci. Total Environ. 1998, 219, 161–174.
(26) Ba¨ckstro¨m, M.; Karlsson, S.; Allard, B. Metal leachability and anthropogenic signal in roadside soils estimated from sequential extraction and stable lead isotopes. Environ. Monit. Assess. 2004, 90, 135–160. (27) Preciado, H. F.; Li, L. Y. Evaluation of metal loadings and bioavailability in air, water, and soil along two highways of British Columbia, Canada. Water, Air, Soil Pollut. 2006, 172, 81–108. (28) Covelo, E. F.; Andrade, M. L.; Vega, F. A. Heavy metal adsorption by humic umbrisols: Selectivity sequences and competitive sorption kinetics. J. Colloid Interface Sci. 2004, 280 (1), 1–8. (29) Sauve´, S.; Manna, S.; Turmel, M. L.; Roy, A.; Courchesne, F. Solid-solution partitioning of Cd, Cu, Ni, Pb, and Zn in the organic horizons of a forest soil. Environ. Sci. Technol. 2003, 37, 5191–5196. (30) Howari, F. M.; Banat, K. M. Assessment of Fe, Zn, Cd, Hg, and Pb in the Jordan and Yarmouk River sediments in relation to their physicochemical properties and sequential extraction characterization. Water, Air, Soil Pollut. 2001, 132, 43–59. (31) Sun, X.; Davis, A. P. Heavy metal fates in laboratory bioretention systems. Chemosphere 2007, 66, 1601–1609. (32) Hsieh, C.; Davis, A. P. Evaluation and optimization of bioretention media for treatment of urban storm water runoff. J. Environ. Eng. 2005, 131 (11), 1521–1531. (33) Tomasella, J.; Hodnett, M. G. Estimating soil water retention characteristics from limited data in Brazilian Amazonia. Soil Sci. 1998, 63 (3), 190–202. (34) National Weather Service. http//:www.nws.noaa.gov (August 2006). (35) Batta, R. K.; Murty, V. V. In situ determination of the hydrodynamic dispersion coefficient and its correlation under laboratory and field conditions. J. Hydrol. 1982, 59, 139–147. (36) Petersen, E.; Jennings, A. A.; Ma, J. Screen level risk assessment of heavy metal contamination in Cleveland area commons. J. Environ. Eng. 2006, 132 (3), 392–404.
ES702681J
VOL. 42, NO. 14, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5253
