Sampling Unsaturated-Zone Water for Trichloroethene at Picatinny Arsenal, New Jersey zyxw James A. Smith,* H. Jean Cho, Peter R. Jaff6, Cecilia L. MacLeod, and Susan A. Koehnlein ABSTRACT A new method of collecting samples of unsaturated-zone water for quantitative analysis for a volatile organic compound, trichloroethene (TCE), was compared to three other, previously described sampling methodologies in the laboratory and in the field. In the laboratory, prepared water samples containing TCE in a known concentration (20 pg/L) were sampled repeatedly by using each of the four methods to quantify method precision and accuracy. To compare the four methods in the field, unsaturated-zone water above a TCE-contami- nated water-table aquifer was transferred from a depth of 2 zyxwvuts m to land surface with 0.1s-m-long suction lysimeters attached to 1.8s-m lengths of stainless-steel tubing. Statistical analyses of the laboratory and field data indicate that the new method, which involves collecting the water samples in gas-tight glass syringes, is superior to the other three meth- ods for the quantitative sampling and analysis of TCE on the basis of its high precision and accuracy and ease of use. This method was used to collect additional samples from the field site to quantify the spatial variability of TCE concentrations in the unsaturated-zone water. Re- sults of analysis of variance of the data indicate that the spatial con- centration variability is important, and that differences in TCE concentration are statistically significant for horizontal distances less than 3.6 m. NOWLEDGE of unsaturated-zone water chemistry K is required to evaluate how groundwater is af- fected by activities such as the irrigation treatment of wastewater (Stevenson, 1978), the application of pes- ticides to farmland (Smith and Carsel, 1986), and the migration of hazardous waste beneath land-disposal facilities (Morrison and Szecsody, 1987). Similarly, quantifying the concentrations of volatile organic compounds in unsaturated-zone water above a con- taminated water-table aquifer can be requisite for es- timating the net flux of these compounds from the ground-water plume (Cho and Jaffk, 1990). The most common method of bringing unsaturated- zone water to land surface for subsequent chemical analysis is through the use of ceramic porous-cup suc- tion lysimeters. The original design of the lysimeter was proposed formally in 1904 (Briggs and McCall, 1904); since that time, many researchers have modi- fied the original design and compared different types of lysimeters and sampling procedures for specific ap- plications (Wagner, 1962; Parizek and Lane, 1970; Wood, 1973; Harris and Hansen, 1975; Levin and Jackson, 1977; Chow, 1977; Long, 1978; Silkworth and Grigal, 1981; Morrison, 1982; Morrison and Szecsody, 1985; Suarez, 1986; Peters and Healy, 1988). All lysimeters require the application of a negative pressure (suction) more negative than the pressure of the unsaturated-zone pore water to induce water flow from the soil into the lysimeter. The suction required J.A. Smith, C.L. MacLeod, and S.A. Koehnlein, zyxwvuts US. Geol. Surv., 810 Bear Tavern Rd., Suite 206, W. Trenton, NJ 08628; H.J. Cho and P.R. Jaffi, Dep. of Civil Eng. and Operations Res., Princeton Univ., Princeton, NJ 08544. Received 22 July 1991. *Corresponding author. Published in J. Environ. Qual. 21:264-271 (1992). to collect a sample can prevent the accurate quantifi- cation of the concentrations of volatile organic con- taminants and pesticides in the sample because of degassing of the solute (Wood et al., 1981; Pettyjohn et al., 1981; Everett and McMillion, 1985; Suarez, 1986). This adverse effect has been observed in the sampling of saturated-zone water with peristaltic pumps in laboratory (Ho, 1983; Barcelona et al., 1984) and field experiments (Imbrigiotta et al., 1988a). To overcome this problem, Wood et al. (1981) and Pettyjohn et al. (1981) proposed a method of collect- ing water samples for volatile organic solute analysis from suction lysimeters that included an adsorbent trap to capture any outgassed organic vapor. Replicate sampling and analysis of water samples containing TCE collected in the laboratory with this method in- dicated that essentially 100% of the TCE in the stan- dard solution was recovered (Wood et al., 1981). This sampling methodology is cumbersome to implement in the field, however, and is not well suited for routine automated laboratory analysis. Other researchers have approached the problem of sampling unsaturated-zone water for volatile constit- uents indirectly. Suarez (1986) has described a mod- ification of the lysimeter and a modification of the sampling procedure to minimize CO, degassing and subsequent changes in the sample pH; Torstensson (1984) has described a methodology for sampling both saturated- and unsaturated-zone water with porous-cup lysimeters and sealed, vacuum-evacuated glass sam- pling vials. Neither of these methods has been eval- uated for its effectiveness in quantifying the concentrations of volatile organic solutes in unsatu- rated-zone water, although the system proposed by Torstensson (1984) has performed well relative to conventional bladder pumps and bailers used to sam- ple saturated-zone water for volatile organic com- pounds (Blegen et al., 1988). The objective of this work is to identify an accurate and precise method for sampling and analyzing un- saturated-zone water for volatile organic compounds. To accomplish this objective, a new sampling method (based on a procedure used to sample unsaturated- zone soil gas) was compared in the laboratory and the field to modified versions of the three methods de- scribed previously by Wood et al. (1981), Pettyjohn et al. (1981), Torstensson (1984), and Suarez (1986). This paper presents the results of these comparisons along with results of an analysis of the spatial con- centration variability of TCE in the unsaturated-zone water at the field site. Description of Field Site Picatinny Arsenal, located in north central New Jer- sey (Fig. l), was selected as the field site for this study. The area shown in Fig. 1 is located in a gla- 'The use of trade, product, brand, or company names in this report is for identification pur oses only and does not constitute endorsement by the US. GeoLgical Survey. 264 Published April, 1992