A Mass Balance Approach for Evaluating Leachable Arsenic and Chromium from an In-Service CCA-Treated Wood Structure

Abstract

Many existing residential wood structures, such as playsets and decks, have been treated with chromated copper arsenate (CCA). This preservative chemical can be released from these structures incrementally over time through contact with rainfall. The objective of this study was to evaluate the levels of arsenic and chromium leached from an in-service CCA-treated deck exposed to rainfall, as well as their possible impacts on soils and shallow groundwater. Two monitoring stations, one containing a CCA-treated deck and the other containing an untreated deck as a control, were constructed outside for this study. Rainfall, runoff water from the decks, soils below the decks, and infiltrated water through 0.7 m depth of soil were monitored for arsenic and chromium over a period of 3 years. The concentration of the CCA-treated deck runoff for arsenic (0.114 – 4.66 mg/L) and chromium (0.008 – 0.470 mg/L) were significantly (p < 0.001) higher than the untreated deck runoff (≤ 0.002 mg/L for both). During the 3 year monitoring period, 13% of the arsenic and 1.4 % of the chromium were leached from the amount initially present in the CCA-treated wood. Arsenic levels (< 0.1 – 46 mg/kg) in soils under the CCA-treated deck were significantly (p < 0.001) higher than under the untreated deck (< 0.1 – 2.7 mg/kg), while chromium levels were statistically the same below the two decks (2.4 – 9.6 mg/kg). Approximately 94% of the arsenic from the runoff was absorbed in the soils below the CCA-treated deck; the upper 2.5 cm of the soils captured 42% of the total. The infiltrated water concentrations for arsenic (< 0.001 – 0.085 mg/L) and chromium (< 0.001 – 0.010 mg/L) below the CCA-treated deck were both significantly (p < 0.001) higher than below the untreated deck (≤ 0.006 mg/L). The amounts of arsenic found in the infiltrated water below the CCA-treated deck represented 6% of total arsenic leached and less than 0.7% of the initial mass in the wood. The study demonstrated that exposure of a CCA-treated deck to rainfall resulted in elevated arsenic concentrations in both runoff and soil. Although only a relatively small fraction of the initial arsenic from the wood was found to infiltrate through the soil, these impacts were significant and caused the infiltrated water to exceed drinking water standards. The study suggests that potential exposures to arsenic exist indirectly through an environment that is contaminated with arsenic leached from in-service CCA-treated wood.

1. INTRODUCTION

Since the 1970s, chromated copper arsenate (CCA) has been the most common preservative used in the U.S. to treat wood intended for outdoor structures, such as playsets, decks, fences, utility poles, and marine docks (Lebow 1996; Solo-Gabriele and Townsend 1999; AWPA 2005). Recently, CCA was the subject of risk assessments by the U.S. Environmental Protection Agency (EPA) and U.S. Consumer Product Safety Commission (CPSC) for potential exposures to children who contact CCA-treated playgrounds and home decks (U.S. EPA. 2001; U.S. CPSC 2003; Dang et al. 2003; Zartarian et al. 2003). In response to these risk assessments, manufacturers of CCA began a voluntary transition from CCA to alternative wood preservatives, and as of January 1, 2004, new CCA-treated wood is no longer manufactured for residential uses in the U.S. (U.S. EPA 2002). Although CCA-treated wood has been phased-out for residential applications, many in-service CCA-treated structures currently exist in the U.S. due to the long service life of this treated wood which varies from 10 to 40 years (Lebow 1996; McQueen and Stevens 1998; Hingston et al. 2001; Alderman et al. 2003). Therefore ongoing and future exposure to toxic metals from CCA treated wood remains a possibility.

The risk assessments conducted by U.S. governmental agencies focused on evaluating direct impacts to individuals (particularly children) through dislodgeable arsenic, which is the amount of arsenic that can be rubbed off through contact with the wood and subsequently ingested through hand to mouth activity (Kwon et al. 2004; Shalat et al. 2006). The risk assessments did not focus on possible indirect exposures through general environmental contamination from leachable arsenic which is arsenic lost from the wood in contact with water (Warner and Solomon 1990; Cooper 1991; Weis et al. 1991; Van Eetvelde et al. 1995; Lebow 1996). Arsenic lost by leaching may result from the depletion of the arsenic from the wood by rainfall, and this leachable arsenic can potentially impact runoff, soils, and potentially groundwater.

Large amounts of the chemicals in CCA-treated wood have been documented to be lost (i.e. depleted) during the long service lives of the product according to stake tests (AWPA 2005). Studies have documented depletion rates ranging from an average 25% lost after 20 to 43 years of exposure in temperate Sweden (Evans and Edlund 1993) and 22% after 44 months in tropical Hawaii (Jin et al. 1992). The data from stake tests is most often used to estimate performance against biological deterioration, but it is not designed to evaluate leaching rates (Lebow 1996). A few studies, however, have evaluated the leachates from CCA-treated wood under field conditions during comparatively long-term field monitoring: 6 months in Japan (Yamamoto et al. 1999); 10 months in Australia (Kennedy and Collins 2001) and 1 year in Florida (Khan et al. 2006). Those studies documented that approximately 5 % of arsenic was leached from the wood in one year (Kennedy and Collins 2001; Khan et al. 2006) resulting in arsenic leachate concentrations between 0.1 and 8.4 mg/L (Khan et al. 2006)

Earlier field surveys found that arsenic levels in most soil samples collected near in-service CCA-treated wood decks were noticeably higher than background levels. For example, 11 mg/kg of arsenic on average (4 - 42 mg/kg) was reported in residential CCA-treated wood equipped playgrounds in Florida (Shalat et al. 2006). Comparatively higher levels were documented in public parks, e.g. 76 mg/kg (1 - 305 mg/kg) in Connecticut (Stilwell and Gorney 1997) and 29 mg/kg (1-217 mg/kg) in Florida (Townsend et al. 2003). Those studies suggest that elevated arsenic levels in the soil were from arsenic leached from the wood.

Only a few studies discuss groundwater contamination associated with in-service CCA-treated wood structures. Although the results of earlier soil surveys described above may suggest that leachable arsenic introduced into the soils could be captured, limited studies to date have measured infiltrated water from soils located below CCA-treated decks. Based on the measurements of arsenic in soil profiles and ratios of arsenic to chromium, Townsend et al. 2003 suggested that some of the arsenic leached from the corresponding structures may have migrated through the soil profile. In South Florida, Khan et al. 2006 focused on evaluating the species of arsenic released from an in-service CCA-treated deck and arsenic concentrations in the infiltrated water below 0.7 m of sandy soil for a period of one year. Even though observed arsenic concentrations in the deck runoff was 0.6 mg/L on average, the arsenic concentrations in the infiltrated water were low near background levels (0.003 mg/L) towards the beginning of the study but increased to 0.018 mg/L by the end of the study (Khan et al. 2006). Results from Khan et al. 2006 suggest that leachates from CCA-treated structures could potentially impact groundwater even though the loss of arsenic is slow.

A common thread among much of the prior work has been to evaluate impacts of releases to different environmental media separately. Either loss from the wood was evaluated, or the focus may have been on impacts to soil or surrounding water. Rarely is more than one environmental component evaluated and as such there is a need for an entire mass balance approach to quantitatively establish the fate of the arsenic from CCA-treated structures. The primary objective of this study was thus to evaluate the fate of arsenic leached from CCA-treated wood using a mass balance approach. Information obtained from this study would be useful for expanded risk assessments of CCA treated wood focused on indirect exposures through general environmental contamination.

2. MATERIALS AND METHODS

Specifically, a monitoring system was constructed which would allow for the measurements of arsenic in the runoff, arsenic in the soils below a structure, and arsenic levels in shallow infiltrated water. The study utilized the same decks as Khan et al. 2006 and both studies overlapped in time, as the first year of Khan et al. 2006 coincided with the first year of the 3-year study reported herein. In addition to the longer time scale, the current study differs in that chromium was also measured and that arsenic and chromium were measured in the soil. This study focused on establishing a mass balance for the arsenic and chromium leached from the decks into all possible environmental media including runoff, soil, and infiltrated water. Mass balances were evaluated for arsenic during the 3 year period and empirical relationships were developed to estimate impacts during the longer service life (20 years) of treated wood products. Concentrations of arsenic and chromium in runoff water, soil, and infiltrated water exposed to CCA-treated wood were compared with risk based environmental guidelines as established for the State of Florida, USA.

2.1. Monitoring decks

Each deck used in the current study, as in Khan et al. 2006, consisted of four wooden supports for a top surface area of 3.3 m² (2.8 m² for the board surfaces and the remainder corresponding to the space between each board). One deck was made entirely of untreated wood as a control; the other's top surface boards (2.8 m² × 2 cm) were CCA-treated wood at 4 kg/m³ rated (3.5 kg/m³ measured) retention. Each deck was placed on top of a sand layer which was contained within its own plywood box. Plywood used for the box consisted of untreated wood and the plywood was lined with plexiglass. The soil placed in each box originated from local rock mining activities in Miami-Dade County, Florida which was characterized by low organic content (0.6%). The total bulk volume of soil below each deck was 4.2 m³ (porosity 35%; soil particle density of 2,690 kg/m³), resulting in a depth of 0.7 m of soil over a 6.0 m² surface area.

Each deck was equipped with a rainfall gauge, a deck runoff collection unit, and an infiltrated water collection system. The deck runoff unit consisted of a gutter that drained the lowermost board from each deck. The gutter was covered with a polyethylene liner to prevent direct rainfall from entering the gutter. Water from the gutter was collected in a plastic reservoir (5-L) located below each deck. The infiltrated water collection system, located below the 0.7 m sand layer, consisted of geotextile underlaid by gravel, an impervious liner which facilitated the drainage of the water towards a perforated pipe, and ultimately to a graduated reservoir (130 to 260-L).

2.2. Sample collection and analyses

The CCA-treated and untreated decks were monitored over a period of three years from September 14, 2002 through September 23, 2005. Water samples (including rainfall, deck runoff, and infiltrated water) from each deck were collected in pre-acid washed high density polyethylene (HDPE) bottles (120-mL) everyday for the first month, and twice a week for an additional 13 months. For months 14 through 24, samples were collected once a week, and then collected once a month during the third year of monitoring.

pH was measured (Model 525A, Orion Research Inc., Beverly, MA, USA) immediately after sample collection and then acidified by adding 1 ml of 1:1 nitric acid (HNO₃) and stored in a refrigerator until sample analysis. One soil core sample was collected every six months below the center of each deck. At the conclusion of the project (at month 36), 10 cores were collected below the CCA-treated deck in order to survey the horizontal and vertical distributions of arsenic and chromium concentrations. Surface samples (upper 2.5 cm) were evaluated for all 10 cores. These 10 surface samples were collected in areas not covered by the deck (3 samples), below the edges of the deck (3 samples), and under the deck (4 samples). The entire depth profile (2.5 cm increments) was analyzed for 2 of the 10 cores, both of which were collected under the deck. The soil core sampling apparatus included a 2.9 cm diameter unslotted stainless probe fitted with a pre-acid washed plastic liner that was either 30 or 60 cm long (Forestry Suppliers, Inc., Jackson, MS, USA).

Solid samples (including sawdust and soil) were processed by digesting 1 g in nitric acid and hydrogen peroxide according to a modified version of US EPA Method 3050B (US EPA 1996). The modification of the digestion procedure was the omission of the hydrochloric acid (HCl) addition as it would have impacted the subsequent arsenic analysis by atomic absorption (AA); Shibata et al. 2004 compared the digestates with and without HCl addition, and did not find a significant difference in results for arsenic analysis. Digested sawdust was analyzed using AA (Perkin Elmer Model AA800, Wellesley, MA) with flame atomization (FLAA). Digested soil samples and water samples were analyzed with graphite furnace atomization (GFAA). Samples were analyzed for total arsenic and chromium in triplicate for FLAA and duplicate for GFAA. The relative standard deviations of the replicates were less than 10%. The analytical detection limits for the FLAA and GFAA were 1 mg/L and 0.001 mg/L, respectively, for both arsenic and chromium. The methodological detection limits for solid samples were 100 mg/kg with FLAA and 0.1 mg/kg with GFAA.

2.3. Mass balance calculations

Mass balances of arsenic and chromium were calculated as follows. The initial mass of arsenic and chromium (mg) in the CCA-treated wood were calculated from the product of the concentration in the sawdust (mg/kg), the volume of the wood (0.056 m³ = 2.8 m² × 0.02 m), and the density of the wood (513 kg/m³). Amounts of arsenic and chromium (mg) leached by runoff were calculated as the product of the deck runoff concentration (mg/L), the surface deck area (2.8 m²), and the rainfall (mm) minus an initial abstraction of rainfall (mm) by the wood. For this study, a 2-mm initial abstraction was used for computation purposes. This value was based on a laboratory test, although the initial rainfall abstraction by wood can be variable under the field conditions.

Amounts of arsenic in the soil below the CCA-treated wood deck were calculated from the product of the soil arsenic concentration (mg/kg), the volume of soil (m³), and the bulk density of the soil (1,750 kg/m³). Amounts of arsenic and chromium recovered from the infiltrated water were calculated from the product of the infiltrated water concentrations (mg/L) and the volume of infiltrated water collected in the reservoir under the deck. During times when the infiltrated water overflowed the collection reservoir or during missed sampling dates due to maintenance of the reservoir, the infiltrated volume was estimated based on the product of the rainfall depth and the catchment area (6.0 m²). This value was then adjusted for moisture evaporation from soil (34%). This evaporation percentage was calculated from the observed rainfall amounts and collected infiltrated water volumes excluding overflows for the monitoring period.

3. RESULTS AND DISCUSSION

3.1. Rainfall and deck runoff concentrations

Rainfall concentrations were consistently below the detection limit of 0.001 mg/L for arsenic and chromium, except for two samples that detected 0.002 and 0.003 mg/L of chromium. Likewise for the untreated deck, arsenic and chromium concentrations in the runoff were less 0.001 mg/L, except for three samples that measured at 0.002 mg/L for chromium (Table 1).

Table 1.

Summary of average pH and concentrations of arsenic and chromium in water samples from untreated and CCA-treated (CCA) decks

		Rain	Runoff leachate		Infiltrated water
Parameter			Untreated	CCA	Untreated	CCA
pH	Minimum	3.55	4.60	3.42	4.42	6.69
	Maximum	7.47	7.71	8.28	8.68	9.27
	Average	4.98	6.31	6.33	7.91	8.11
	Stdev.a	0.90	0.65	0.66	0.58	0.35
As (mg/L)	Minimum	< 0.001	< 0.001	0.114	< 0.001	< 0.001
	Maximum	< 0.001	< 0.001	4.660	0.004	0.085
	Average	< 0.001	< 0.001	0.807	< 0.001	0.017
	Stdev.	< 0.001	< 0.001	0.668	< 0.001	0.019
Cr (mg/L)	Minimum	< 0.001	< 0.001	0.008	< 0.001	< 0.001
	Maximum	0.003	0.002	0.470	0.006	0.010
	Average	< 0.001	< 0.001	0.088	< 0.001	0.003
	Stdev.	< 0.001	< 0.001	0.070	< 0.001	0.002

^aStdev = standard deviation.

From the CCA-treated deck, arsenic concentrations in the runoff ranged from 0.144 to 4.66 mg/L with an average of 0.807 mg/L and chromium ranged from 0.008 to 0.470 mg/L with an average of 0.088 mg/L. The arsenic and chromium levels in the CCA-treated deck runoff were significantly (p < 0.001) higher than that measured in rainfall and the runoff from the untreated deck. The concentrations of arsenic and chromium were significantly correlated (r = 0.867; p = 0.05). However, the levels of arsenic and chromium in the CCA-treated deck runoff were significantly (p < 0.001) different, by roughly factor of 10, although the initial concentrations in the CCA-treated wood were similar (1840 mg/kg of arsenic and 1670 mg/kg of chromium). The difference between arsenic and chromium concentrations in the leachates could be due to the more effective fixation of chromium in the treated wood (Cooper et al. 1995).

The concentrations of arsenic and chromium in the CCA-treated wood deck runoff varied (Figure 1), but generally decreased over time. Based on average concentrations over half year intervals, the runoff levels with time were expressed (r = 0.994 for arsenic and 0.864 for chromium; p = 0.05) as:

$$C_{\mathit{runoff},t}=a\cdot t^b$$ (1)

Where C_runoff,t was the average concentration of CCA-treated deck runoff (mg/L) at year t, and t was the time of weathering (year), a and b were empirical coefficients. The coefficients a for arsenic and chromium were 0.84 and 0.097, respectively, and the b for arsenic and chromium were −0.73 and −0.56, respectively.

Figure 1.

Arsenic and chromium in the runoff leachate from the CCA-treated wood deck versus time

The average pH of rainfall was 4.98. The average pH of runoff water was 6.31 and 6.33 from the untreated and CCA-treated decks respectively. These pH values were statistically different from rainfall pH (p < 0.001), but not statistically different from each other.

This increase in pH from rainfall to the deck runoff is likely due to the buffering capacity of wood. Such results are consistent with observations from other field-based studies (Shibata et al. 2006; Khan et al. 2006). Overall correlations between pH and the arsenic and chromium concentrations in the deck runoff were not statistically significant. Although Shibata et al. 2006 showed that rainfall pH has an impact on arsenic leaching during the first three months of weathering, the wood evaluated previously by Shibata et al. was in the form of mulch (shredded chips). The process of shredding the wood in the earlier work by Shibata et al. could have resulted in the exposure of chemicals contained within the interior portions of the wood. This exposure of the chemical could have resulted in the observed correlations with rainfall pH during the previous mulch study

3.2. Soil concentrations

For soil samples collected under the untreated deck, the arsenic concentrations ranged from below the limit of 0.1 mg/kg to 2.7 mg/kg with 0.5 ± 0.8 mg/kg on a depth average (or a core sample average), which was consistent with reported background levels of 1.3 ± 3.8 mg/kg in Florida (Chen et al. 1999). For the CCA-treated deck, the arsenic concentrations ranged from less than 0.1 to 46 mg/kg with a depth average of 3.6 ± 6.3 mg/kg. The arsenic levels below the CCA-treated deck were significantly (p < 0.001) higher than from below the untreated deck (Table 2). The observed arsenic concentrations were within the reported values in soil near in-service CCA-treated wood structures (Stilwell and Gorney 1997; Townsend et al. 2003; Shalat et al. 2006).

Table 2.

Arsenic and chromium concentrations in time series in soil cores collected below the untreated (UN) and CCA-treated (CCA) decks

	Arsenic												Chromium
	0.5 year		1 year		1.5 years		2 years		2.5 years		3 years		3 years
Depth (cm)	UN	CCA	UN	CCA	UN	CCA	UN	CCA	UN	CCA	UN	CCAb	UN	CCA
0.0 - 2.5	<0.1	4.5	<0.1	11.5	<0.1	12.2	<0.1	13.4	<0.1	16.2	1.9	25.4	3.7	6.1
2.5 - 5.0	<0.1	1.3	0.1	5.7	<0.1	7.0	<0.1	4.4	<0.1	9.8	1.7	12.6	4.7	5.4
5.0 - 7.5	<0.1	0.6	0.2	3.3	<0.1	3.5	<0.1	1.0	<0.1	6.4	1.6	8.7	3.5	4.9
7.5 - 10.0	0.9	0.5	<0.1	0.2	<0.1	2.5	0.1	1.7	<0.1	4.4	1.7	6.0	4.5	4.4
10.0 - 12.5	0.4	0.4	0.9	<0.1	<0.1	0.7	<0.1	1.0	<0.1	2.1	1.8	4.9	6.2	4.2
12.5 - 15.0	0.4	0.3	0.4	0.1	<0.1	0.2	<0.1	0.7	<0.1	1.4	1.3	5.1	4.4	4.8
15.0 - 17.5	0.6	0.4	0.2	<0.1	0.2	<0.1	<0.1	0.8	0.1	0.9	1.5	4.9	4.1	4.7
17.5 - 20.0		0.2	0.1	<0.1	<0.1	0.4	<0.1	1.0	<0.1	0.8	1.8	3.7	3.3	4.1
20.0 - 22.5				<0.1	<0.1	<0.1	<0.1	0.8		0.7	2.0	2.8	3.9	4.1
22.5 - 25.0				<0.1		0.6				0.6	1.9	1.5	3.0	3.9
25.0 - 27.5											2.7	1.0	3.7	3.6
27.5 - 30.0											1.7	0.8	3.5	3.6
30.0 - 32.5											2.1	0.7	2.4	4.8
32.5 - 35.0												0.8		3.2
35.0 - 37.5												1.0		4.1

^bCCA, the arsenic concentrations for this column was the average from core samples A and B (See Figure 2)

The spatial distributions of arsenic in surface samples (0 – 2.5 cm) within the 6 m² area of soil foundation of the CCA-treated wood deck ranged widely from 0.5 to 46 mg/kg at the end of three years (Figure 2). Directly under the CCA-treated deck the average surface soil concentration was 17 ± 19 mg/kg (samples A to D). The arsenic concentration near the edge of the deck was 3.2 ± 4.0 mg/kg on average, which was comparatively lower than concentrations in soils collected directly under the deck. At locations 30 to 45 cm away from the CCA-treated deck, the arsenic concentration was 0.9 ± 0.5 mg/kg on average

Figure 2.

Horizontal distribution of arsenic and chromium (mg/kg) in the 6 m² area of surface soils below the CCA-treated deck after 3 years of CCA-treated deck installation. The shaded part was directly under the deck (3.3 m²) and dripping lines are shown by the dotted lines. Soil locations indicated by *A and *B correspond to locations where metals were analyzed throughout the soil profile.

Over time, the arsenic concentrations in surface soils increased from 0.3 mg/kg (3 year average of surface soil under the untreated deck) to 4.5 mg/kg after 6 months of CCA-treated deck installation, and continuously increased to 25 ± 29 mg/kg at the end of the 3-year monitoring period (Table 2). Utilizing these data, surface soil concentrations with respect to time under the CCA-treated deck was expressed as (r = 0.953; p = 0.05):

$$C_{\mathit{soil},t}=a\cdot t^b$$ (2)

Where C_soil,t was the approximate arsenic surface soil concentration under the CCA-treated deck (mg/kg) at year t, and t was the time of weathering (year), a and b were empirical coefficients. The coefficients a and b for arsenic were 8.9 and 0.8 respectively as determined from the data provided in Table 2. This equation corresponds to the net concentration in the surface soils and thus represents the combined effects of arsenic input from the deck and arsenic loss due to leaching from the soil. Based on Equation 2, the arsenic concentration in the sandy soil under the deck would be estimated to be at 98 mg/kg 17 years after the end of monitoring period (or after 20 years of the CCA-treated deck installation).

This study clearly showed the impact of time although earlier studies which collected soil samples below different structures showed no correlation between the arsenic levels in the surface soils and age of the CCA-treated wood structure (Stilwell and Gorney 1997; Townsend et al. 2003). The reason for this apparent difference was because the earlier studies evaluated different decks at a single point in time. In addition, each of these decks could have been subject to different soil (Bodek 1988; Schnoor 1996) and wood characteristics (Shibata et al. 2006) which would mask the effects of time on soil impacts. In this study, since the soil below one structure was consistently sampled, the differences in soil, hydro-climatologic, and wood characteristics were better controlled; thus the impact of time on soil concentrations could be readily observed.

The vertical distribution of arsenic concentrations under the CCA-treated deck ranged widely (Table 2). The arsenic levels in soils from all core samples were the highest at the surface and decreased exponentially with depth of soil (r = 0.788 to 0.974; p = 0.05). Although the impacts to surface soils were the most apparent, gradual accumulations of arsenic levels in deeper soils were also observed over time (Figure 3).

Figure 3.

Vertical distribution of arsenic in the soil below the untreated and CCA-treated (CCA) decks. For the soils under the untreated wood deck (a.k.a. untreated average), the concentrations of samples for 0.5 to 3 years were used. For soils collected below the CCA-treated deck at year 3, the average concentration of the surface soil (0 – 2.5 cm) samples and the average concentration soils (> 2.5 cm) from core samples collected under the CCA-treated deck were used.

For chromium, the depth averaged concentrations below the untreated deck (3.9 ± 0.9 mg/kg on average) and CCA-treated deck (4.6 ± 1.4 mg/kg on average) were not statistically different. The surface soil chromium levels with an average of 5.6 ± 1.6 mg/kg at each sampling point were strongly correlated with arsenic levels (r = 0.883; p = 0.05). However, those levels did not range widely unlike the arsenic distribution because the higher chromium background concentrations and the amounts of chromium introduced into the soils were smaller relative to that for arsenic.

3.3. Infiltrated water concentrations

The average pH values of the infiltrated water below the untreated and CCA-treated decks were 7.91 and 8.11, respectively, and were significantly different (p = 0.003). The pH of the infiltrated water was also significantly (p < 0.001) different from the values observed from the deck runoff.

Below the untreated deck, the arsenic concentrations of the infiltrated water ranged from the detection limit of 0.001 mg/L to 0.004 mg/L, and the chromium ranged from less than 0.001 to 0.006 mg/L (although most of samples were less than 0.001 mg/L). These levels were consistent with reported background levels in groundwater in areas not impacted by natural sources of arsenic and chromium (WHO 1996). The arsenic and chromium levels were not statistically different.

For the CCA-treated deck, the arsenic concentrations ranged from less than 0.001 to 0.085 mg/L with an average of 0.017 mg/L, and the chromium ranged up to 0.010 mg/L with an average of 0.003 mg/L. The arsenic and chromium levels below the CCA-treated wood deck were significantly (p < 0.001) higher than below the untreated deck. Below the CCA-treated wood deck, the arsenic and chromium levels were almost the same during the first nine months of monitoring; however, the arsenic concentrations increased with time after this period, while the chromium levels were consistently low (Figure 4). Overall levels of arsenic were significantly (p < 0.001) higher than chromium. The large differences between the arsenic and chromium levels occurred partly because larger amounts of arsenic were introduced to the soils from the CCA-treated wood relative to chromium.

Figure 4.

Arsenic and chromium in the infiltrated water below CCA-treated wood deck versus time

Based on average concentrations over half year intervals, the arsenic level in the infiltrated water through 0.7 m of sandy soil below the CCA-treated deck was expressed with respect to time (r = 0.961; p = 0.05) as:

$$C_{\mathit{inifiltrated},t}=a\cdot t^b$$ (3)

Where C_{infiltrated,t} was the infiltrated water concentration for arsenic (mg/L) at year t, and t was the time of weathering (year), a and b were empirical coefficients. The coefficients a and b were 0.011 and 1.7, respectively.

The largest observed arsenic level in the infiltrated water was very near the smallest observed concentration in the deck runoff during the monitoring period. Based on the Equation 1 and 3, the arsenic concentrations in the deck runoff and infiltrated water would be the same levels after 6 years of the CCA-treated deck installation. These results indicated that accumulated arsenic in the soil under the CCA-treated wood deck could migrate down to 0.7 m below the ground surface in a relatively short period (several months). This would be significant as groundwater supplies, in particular supplies within Florida are reported to be as low as 0.6 m (Merritt 1996).

3.4. Mass balance

A total 4,040 mm of rainfall was recorded during the three year monitoring period. Average annual rainfall was 1,350 mm, which was slightly smaller than the historical average annual rainfall of 1,490 mm in Miami Florida (NOAA 2004). The total volume of the deck runoff from the 2.8 m² top surface deck was estimated at 10,700 L, with an assumption that the amount of deck runoff was equivalent to the total rainfall minus 2-mm of initial abstraction by the wood. Considering the entire surface area of the sandbox, the total volume of water estimated to impact the soil was 23,400 L. The cumulative volume of infiltrated water recovered was approximately 15,100 L. It should be noted that the infiltrated water was diluted since the surface area of sand box was almost twice the size of the top surface area of the deck. Approximately 6% of the rainfall was estimated to have been abstracted by the wood, and 32% evaporated from soil.

The initial amounts of arsenic and chromium in the CCA-treated wood deck were 53,200 and 48,300 mg respectively based on the sawdust analysis. A total 6,450 mg of arsenic and 668 mg of chromium were leached out from the CCA-treated wood deck during the three years. The cumulative masses of arsenic leached were 7% after the first year, with an increase to 10% after two years and 13% after three years. The fraction of the chromium leached was much smaller (1.4% after 3 years). The fractions leached during the first year as observed in the current study were consistent with earlier studies that reported only 5% lost for arsenic (Kennedy and Collins 2001; Khan et al. 2006) and 1% lost for chromium (Kennedy and Collins 2001) for one year. The long term cumulative % arsenic leached from the CCA-treated wood deck could be expressed as:

$$\frac{M_n}{M_0}(\%)=\frac{R\cdot {\displaystyle \sum _{i=1}^n}\left(0.84\cdot {t_n}^{-0.7}\right)}{M_0}\ast 100\%$$ (4)

Where M_n was the mass of arsenic (mg) leached from CCA-treated wood runoff, M₀ was the initial amount of arsenic (mg) in the CCA-treated wood, R was runoff water volume (L), n was the number of years after the outdoor installation, 0.84·t_n^−0.7 was the average arsenic runoff concentration (mg/L) at time t_n from the Equation 1, and 100 the conversion factor for percentage (%). Based on the Equation 4 with M₀ of 53,200 mg and R of 3,890 L (South Florida's average annual rainfall of 1487 mm (NOAA 2004), 6% rainfall abstraction, and 2.8 m² of the top deck surface area), 32% of the initial mass of arsenic could be leached from the CCA-treated wood deck and 68% would still remain in the wood after 20 years of installation.

The mass of arsenic recovered in the infiltrated water was estimated at 392 mg after three years. Arsenic recovered in the infiltrated water represented 0.7% of the initial arsenic in the CCA-treated wood deck, and 6% of total leached in the runoff. Accordingly, 94% of the arsenic leached from the CCA-treated deck was captured by soils after three years. Based on the average arsenic concentrations; 17 mg/kg in 3.3 m² directly under the deck (sample A to D) and 3.2 mg/kg in 2.7 m² outside the deck (sample E to J), approximately 86% of total arsenic introduced into the soil within the monitoring box was captured in the soils directly under the deck and 14% by soils not under the deck.

With regards to soil depth, approximately 42% of the total amount of arsenic leached by the deck runoff was captured in the surface soils (0 – 2.5 cm), 52% was captured in the deeper soils, and 6% was recovered at depth from the infiltrated water. It should be noted that the soils received only 13% of the initial mass of arsenic from the CCA-treated deck during the three year study period. Accordingly, another 20% of the initial mass of arsenic was estimated to be released in to the environment and thus larger amounts of arsenic may travel downward in sandy soil during a subsequent 17 years of service life.

3.5. Comparison to guideline levels

For the evaluation of runoff water from the decks, Florida Surface Water Quality Classification (SWQC) of 0.05 mg/L for arsenic and 0.1 mg/L for chromium (FDEP 1996) were used for the comparison. The untreated deck runoff water never exceeded the FL SWQC for arsenic and chromium.

From the CCA-treated deck, the chromium levels in individual runoff samples often exceeded the FL SWQC (FDEP 1996) but the average concentrations after one year were below the guideline. The arsenic levels consistently exceeded the guideline during the three year monitoring period. Although the arsenic concentrations decreased with time, which was consistent with earlier laboratory tests (Weis et al. 1991; Merkle et al. 1993; Breslin and Adler-Ivanbrook 1998) and a field study (Kennedy and Collin 2001), the empirical relationship (equation 1) estimates that the average arsenic concentration in the runoff after 20 years of the new CCA-treated deck installation (rated at 4 kg/m³ and measured at 3.5 kg/m³) would be 0.094 mg/L, which is still higher than the guideline level.

In Florida, the Soil Cleanup Target Levels (SCTLs) are used as guidelines to determine if recycled materials can be land applied, and these SCTLs are frequently used to evaluate soil contamination. Two SCTL thresholds are provided for Florida: one for residential areas (2.1 mg/kg for arsenic and 210 mg/kg for chromium) and one for industrial areas (12 mg/kg for arsenic and 470 mg/kg for chromium) (FDEP 2005). The chromium levels were consistently below the guidelines for all samples collected. For the arsenic below the untreated deck, most soil samples were below the residential guideline except for two samples (2.1 and 2.7 mg/kg), although these levels were within reported background levels of arsenic in Florida (1.3 ± 3.8 mg/kg, Chen et al. 1999). For surface soils below the CCA-treated deck, arsenic levels exceeded the FL SCTLs after six months of installation, and exceeded the arsenic industrial guideline about after one year of installation. However, at locations 30 to 45 cm away from the deck, the arsenic levels were below the residential guideline. This result was consistent with earlier soil surveys that showed arsenic levels decreased with distance from CCA-treated wood structures (Lebow et al. 2000; Stilwell and Graetz 2001; Chirenje et al. 2003). In this study, although the arsenic levels decreased with soil depth, which was consistent with earlier studies (Stilwell and Gorney 1997; Lebow et al. 2000; Townsend et al. 2003; Chirenje et al. 2003), the soils collected at a 20 cm depth after 3 years of the deck installation exceeded the residential guideline due to the accumulation of arsenic.

In Florida, groundwater concentrations are typically compared to Groundwater Cleanup Target Levels (GWCTLs) which are set at 0.01 mg/L for arsenic and 0.1 mg/L for chromium (FDEP 2005). These concentrations are consistent with the U.S. EPA Drinking Water standards; updated for arsenic from 0.05 mg/L to 0.01 mg/L in 2006 (U.S. EPA 2006). In this study, the infiltrated water concentrations below the untreated deck were consistently lower than the FL GWCTLs for arsenic (0.01 mg/L) and chromium (0.1 mg/L) (FDEP 2005). Below the CCA-treated deck, chromium levels were consistently below the guideline while arsenic levels exceeded the FL GWCTL for arsenic after 10 months of the deck installation. Since it took less than one year to surpass the guideline level in a 0.7 m depth of soil, deeper groundwater could be impacted by arsenic leached from in-service CCA-treated wood during longer service lives. It should be noted that the sandy soil used for this study is believed to have a relatively low arsenic sorption capacity, thus potentially allowing arsenic to migrate deeper leading to groundwater contamination; however this soil is representative of the soil types found in South Florida, in an area characterized by relatively shallow groundwater. This study showed that the infiltrated water that passes through 0.7 m of sand would fail the guideline even if less than 1% of the initial amounts of arsenic in the deck reached the water table.

5. CONCLUSIONS

This study showed that in-service CCA-treated wood structures release leachable arsenic that possibly could contaminate surrounding environments at levels that exceed regulatory guidelines. Considering the increasing trend for soil concentrations under CCA-treated wood, arsenic exposure through contaminated soils to children may be more significant at older playgrounds. Studies have shown that dislodgeable arsenic levels decrease over time (Shibata 2006), and thus, as a playground ages, the primary exposure route to children could potentially shift from direct contact with wood to direct contact with soil. This shift towards increasing soil arsenic concentrations should be considered in the expanded risk assessments aimed at evaluating children's exposures at CCA-treated wood playgrounds. In addition to direct exposure, community exposures may potentially occur if runoff water and infiltrated water contaminate surface and groundwater drinking water supplies. In summary, this study suggests the necessity of updating the current risk assessments. The risk assessments should focus not only on children exposed to dislodgeable arsenic by touching in-service CCA-treated wood, but should also focus on all individuals who may be exposed to arsenic indirectly through an environment that is contaminated with leachable arsenic.