A Practical Method to Evaluate Ground Water Contaminant Plume Stability

Evaluating plume stability is important for the evaluation of natural attenuation of dissolved chemicals in ground water. When characterizing ground water contaminant plumes, there are numerous methods for evaluating concentration data. Typically, the data are tabulated and ground water concentrations presented on a site figure. Contaminant concentration isopleth maps are typically developed to evaluate temporal changes in the plume boundaries, and plume stability is often assessed by conducting trend analyses for individual monitoring wells. However, it is becoming more important to understand and effectively communicate the nature of the entire plume in terms of its stability (i.e., is the plume growing, shrinking, or stable?). This article presents a method for evaluating plume stability using innovative techniques to calculate and assess historical trends in various plume characteristics, including area, average concentration, contaminant mass, and center of mass. Contaminant distribution isopleths are developed for several sampling events, and the characteristics mentioned previously are calculated for each event using numerical methods and engineering principles. A statistical trend analysis is then performed on the calculated values to assess the plume stability. The methodology presented here has been used at various contaminated sites to effectively evaluate the stability of contaminant plumes comprising tetrachloroethene, carbon tetrachloride, pentachlorophenol, creosote, naphthalene, benzene, and chlordane. Although other methods for assessing contaminant plume stability exist, this method has been shown to be efficient, reliable, and applicable to any site with an established monitoring well network and multiple years of analytical data. Introduction Evaluating plume stability is important for the evaluation of natural attenuation of dissolved chemicals in ground water. U.S. EPA (1998) states that the primary line of evidence in evaluating natural attenuation is historical ground water chemistry data that demonstrate a clear and meaningful trend of decreasing contaminant mass and/or concentration over time at appropriate monitoring or sampling points. When characterizing ground water contaminant plumes, there are numerous methods for evaluating concentration data. Wiedemeier et al. (2000) discussed common approaches for evaluating plume stability using both graphical and statistical techniques. Graphical methods include the following: (1) the preparation of contaminant concentration isopleth maps; (2) plotting concentration data vs. time for individual monitoring wells; and (3) plotting concentration data vs. distance downgradient for several monitoring wells. Common statistical methods for evaluation of temporal and spatial trends include regression analysis (U.S. EPA 2006), the Mann-Whitney U-test (Mann and Whitney 1947), and the Mann-Kendall test (U.S. EPA 2006; Gilbert 1987). Graphical plume stability analysis by comparing isopleth maps over time can provide compelling visual evidence for natural attenuation. However, a comparison of apparent plume size over time does not always provide a complete analysis. Consider, for example, the case of a plume that discharges to a surface water body, or a plume geometry that is persistent over time. In this case, the plume area would remain relatively unchanged, whereas the overall plume average concentration and mass may be decreasing. The change in plume mass would not be necessarily reflected in the visual analysis of isopleth maps. However, a quantitative analysis of changes in overall plume concentration and mass would provide a better understanding of the plume stability. A common approach for evaluating plume stability is the use of statistical analysis techniques for single-well data. However, chemical concentration trends at individual monitoring wells may show different trends. For example, at a given site, there may be wells exhibiting decreasing Copyright a 2008 The Author(s) Journal compilationa 2008National GroundWater Association. Ground Water Monitoring & Remediation 28, no. 4/ Fall 2008/pages 85–94 85 trends in some wells while exhibiting indeterminate or even increasing trends in other wells. Furthermore, single-well trend analyses are somewhat dependent on monitoring well location and the number of wells. For example, a site may only have one or two wells exhibiting decreasing trends, with many wells exhibiting indeterminate or increasing trends or vice versa. One could ostensibly increase the number of wells exhibiting decreasing trends simply by installing more wells in the area of the plume exhibiting decreasing trends. However, evaluating trends in the overall plume area, average concentration, and mass provides a more thorough understanding of the stability of the entire plume as opposed to discrete locations within the plume. This article presents practical techniques for calculating and evaluating trends in overall plume area, average concentration, mass, and center of mass. Rather than relying on complex computer models based on numerous assumptions, the techniques presented are relatively simple and based on the use of grid files generated from commonly used contouring and spreadsheet software. The examples presented in this article were generated using Surfer (Golden, Colorado) by Golden Software and Microsoft Excel (Redmond, Washington). The major advantage of using computer software to perform all calculations is that trends in the individual metrics (e.g., plume mass) are not subject to variability resulting from human bias. For example, it is common practice to generate hand-drawn isopleth maps. In this case, there is bias in the interpretation of individual isopleths as well as bias between monitoring events. This would occur when one operator interprets isopleth frequency and location differently than another operator who may come on to a project several years later. However, when isopleths are generated based on a grid file generated from a computer-based gridding algorithm, data are interpreted consistently from event to event, thereby eliminating both spatial and temporal bias. The plume stability analysis described in this article is an effective tool to clearly demonstrate the occurrence (or nonoccurrence) of natural attenuation of a contaminant plume. This method is efficient because it requires only the analytical data supplied by a laboratory to perform the assessment, and it is reliable because all calculations and interpretations are conducted by the software, thereby eliminating any operator bias and subjectivity. This method is relatively easy to implement and would be applicable to most contaminated sites with an established monitoring well network and multiple years of analytical data. The method presented in this article is a combination of graphical and statistical analysis techniques. Contaminant concentration isopleth maps are prepared for various sampling events. The overall plume area, average concentration, and mass are then calculated for each sampling event using innovative techniques. In order to demonstrate that a plume is shrinking, temporal trends in these calculated values should result in statistically significant decreasing trends. An increasing trend in any of these values would indicate that the plume is expanding. In addition to temporal trend analysis of plume characteristics, the center of plume mass is calculated. To further demonstrate plume stability, the location of center of plume mass should not show continual downgradient migration over time. As contaminant mass is removed via pumping or natural processes, the location of the center of plume mass should be relatively stable or exhibit a general receding trend. In order to demonstrate the use of these plume-wide natural attenuation parameters, this article describes a plume stability analysis that was conducted for a former wood-treating site in central Louisiana. The plume stability analysis was conducted for a naphthalene plume over the period 1987 through 2005. Description of Site The site is located in central Louisiana and is the former location of a wood-treating facility that treated utility poles from 1937 until it was closed in September 1997. Dimensional lumber products such as cross-ties and guard posts were also treated at the site. Creosote and pentachlorophenol were used for treating. Although the site is impacted by many polynuclear aromatic hydrocarbons (PAHs), the plume stability analysis was focused on naphthalene concentrations in ground water. Naphthalene was selected as the indicator of creosote in ground water for the site because it is the PAH constituent found in the highest concentration in creosote, is the most mobile PAH, and constitutes the majority of the contaminant mass present in the ground water plume at the site. The uppermost water-bearing unit at the site is the Bentley sand. Monitoring wells at the site are installed in both the upper part (Upper Zone) and the lower part (Lower Zone) of the Bentley sand. The naphthalene plume occurring in both aquifer zones was assessed as part of the plume stability analysis. There are 57 monitoring wells at the site, of which 53 are screened in the Bentley sand. Twenty-eight of these wells have been monitored on a quarterly basis as part of the ongoing monitoring and corrective action programs at the site, which began in 1987. A site diagram is shown in Figure 1. Methodology In order to conduct the plume stability analysis, a primary contaminant of interest must first be selected. Once an indicator contaminant is selected, then concentration isopleth maps are developed for each year of the monitoring record. If many years of data exist, then the analyst may use every other year in order to reduce the overall effort. The underlying grid files used to develop the isopleth maps are then used to calculate plume area, average concentration, mass, and the geographic location of the center of mass. Trends in each of these metrics may then be evaluated using standard statistical analysis techniques. Detailed descriptions of each step are discussed subsequently. Generation of Contaminant Concentration Isopleth Maps Naphthalene concentration isopleth maps representing the naphthalene plumes in both the Upper and the Lower aquifer zones were developed for selected sampling events. J.A. Ricker/ Ground Water Monitoring & Remediation 28, no. 4: 85–94 86 Analytical data reported for each monitoring well varied seasonally, with some wells reporting quarterly results and others reporting less consistently. To remove seasonal variations, an annual average concentration of each year was calculated starting with 1987 through 2005. Due to the volume of data, only the odd years between 1987 and 2003 were used in the analysis. Naphthalene concentration isopleth maps were developed for each of the Upper and Lower aquifer zones for each of the years presented previously. The isopleth maps were each delineated to a naphthalene concentration of 10 lg/L. The raw monitoring well data were log transformed prior to generating the grid file in Surfer. When using contouring software such as Surfer, a grid file is generated simply importing raw data from a spreadsheet in X, Y, Z format, where X and Y refer to the well coordinates and Z refers to the log-transformed concentration. The grid file may be created using any number of gridding methods available within the software; however, kriging was used to generate the grid files presented in this article. Although it is beyond the scope of this article, care should be taken in the selection of grid mesh size. In general, the finer the grid mesh, the more accurate the gridbased calculations will be. Once the grid file is created, the grid is then inverse log transformed back to ‘‘real world concentrations’’ using a grid operation utility within Surfer. The isopleth maps are then generated by Surfer by direct interpolation of the grid file. The analytical data used to generate the Lower Zone naphthalene isopleth map for 2005 are summarized in Table 1. Figure 2 shows an example of analytical data represented by a Surfer grid file, and Figure 3 shows an example of the data represented by an isopleth map based on the underlying grid file. Calculation of Plume Area, Average Concentration, and Mass The plume area, average concentration, and mass are calculated based on the grid file generated by Surfer. In order to calculate these parameters, the plume boundary concentration must first be defined. In many cases, the individual contaminant cleanup level is used, and in the current example, the naphthalene plume boundary is defined by its respective site-specific cleanup level of 10 lg/L. Once the plume boundary is defined, the plume area is determined using the grid volume calculation utility within Surfer. This utility generates a grid volume report that includes, among other parameters, the positive volume and positive planar area of the grid file within the specified boundary concentration. The planar area is based on units used to generate the grid file. In the current example, the area of the plume, which is determined directly from the grid volume report, is 117,915 m (11.8 ha). Figure 1. Site diagram. J.A. Ricker/ Ground Water Monitoring & Remediation 28, no. 4: 85–94 87 The average plume concentration is calculated using grid operations based on principles used in cut-and-fill computations commonly applied in civil engineering applications. The positive volume of the grid file (within the specified boundary concentration) is included in the grid volume report. The units of the grid volume in the current example are lg/L m. The units of volume of the grid in this case are meaningless by themselves; however, when the grid Figure 2. Grid representation of 2005 Lower Zone naphthalene data. Table 1 2005 Lower Zone Naphthalene Data Well X Y Concentration (lg/L) Log Concentration DB-15 499,256 99,655 NS — DB 18 499,191 99,566 9400 3.97 DB-21 498,826 99,626 560 2.75 DB-22 498,610 99,350 7600 3.88 DB-23 499,213 99,031 NS — DB-24 498,686 98,830 NS — DB-25 498,127 98,508 190 2.28 DB-26 497,934 98,867 ND 2.00 DB-27 498,152 99,275 ND 2.00 DB-28 498,472 99,668 ND 2.00 DB-29 498,803 100,290 30 1.48 DB-33 500,120 98,384 NS 2.00 DB-34 497,931 99,369 NS 2.00 DB-35 497,720 98,054 ND 2.00 DB-36 497,540 97,495 ND 2.00 DB-37 497,586 98,191 ND 2.00 DB-38 497,815 98,017 ND 2.00 DB-39 498,279 98,192 ND 2.00 DB-40 498,540 98,304 ND 2.00 DB-41 497,382 97,282 ND 2.00 DB-42 496,942 98,579 ND 2.00 DB-43 496,756 97,206 ND 2.00 DB-44 498,022 100,539 ND 2.00 Notes: ND 1⁄4 naphthalene not detected; NS 1⁄4 well not sampled. 497800 498000 498200 498400 498600 498800 499000 499200 499400 499600 X Coordinate 97800 98000 98200 98400 98600 98800 99000 99200 99400 99600 9980