Optimization of Groundwater Pump and Treat Systems  Using Numerical Modeling and the Monte Carlo Approach
Manu Sharma, Gradient Corporation, Cambridge, MA
Andrew B. Bittner, Gradient Corporation, Cambridge, MA
Tarek Saba, Gradient Corporation, Cambridge, MA

Environmental Data Quality and the Search for Representativeness
Deana M. Crumbling, U.S. Environmental Protection Agency, Washington DC

A Simple Phase Equilibrium Model for Predicting the Historical Presence of NAPL: a Case Study  
John D. Moss and Daniel D. Titus, HRP Associates, Inc., Plainville, CT

Groundwater Effects from Highway Tire Shred Use
Mary O. Brophy, NYSDOT/SUNY School of Public Health, Binghamton, NY
Joseph Graney, Binghamton University, Dept of Geological Sciences and Environmental Studies, Binghamton, NY

Passive Diffusion Bag Sampler Results From Multiple DoD Installations
John Tunks, Parsons, Denver, CO
John Hicks, Parsons, Denver, CO
Javier Santillan, AFCEE/ERB, Brooks City-Base, TX  
Raphael Vazquez, AFCEE/ERT, Brooks City-Base, TX 

Chlorinated Solvent DNAPL Extent Characterization at the East Gate Disposal Yard (Egdy), Fort Lewis, Washington
Kira P. Lynch, Environmental Scientist, U.S. Army Corps of Engineers, Seattle, WA

Optimization of Groundwater Pump and Treat Systems  Using Numerical Modeling and the Monte Carlo Approach

Manu Sharma, Andrew B. Bittner and Tarek Saba, Gradient Corporation, 238 Main Street, Cambridge, MA  02142, Tel: 617-395-5000, Fax: 617-395-5001

Groundwater has been remediated at a number of sites across the country using the pump and treat approach.  Although these systems have proven to be ineffective in aquifers with lenses of low conductivity materials and when non-aqueous phase liquids are present, they can be an effective remedial technique in relatively permeable aquifers, if properly designed.  The use of numerical models and optimization techniques can facilitate the design of pump and treat remedial systems and can significantly increase contaminant removal efficiency and reduce costs and operation time.

Optimization techniques were developed and employed to design a groundwater pump and treat remedial system at a Superfund Site underlain by a highly productive and permeable aquifer.  A 3-Dimensional groundwater flow and solute transport model for tetrachlorethene was developed and calibrated to simulate Site conditions.  Once calibrated, optimization techniques were used to design the pump and treat system.  First, a preliminary system design was developed  consisting of 3 extraction wells and 3 injection wells operating at a total combined extraction rate of 450 gpm.  Second, best-guess initial extraction and injection well locations were defined based on engineering judgment.  The numerical groundwater and solute transport model was iteratively executed using the Monte Carlo approach with commercially available software to evaluate numerous combinations of injection and extraction well locations and to develop an extraction and injection rate schedule over time.  In all, over 1,500 simulations that required intense computational effort, but minimal modeling oversight time, were undertaken to arrive at the final design.  The optimal solution resulted in a remedial system that is projected to achieve the regulatory clean-up standards approximately 6 years faster (approximately 20% of total predicted remediation time) than the remedial system based on engineering judgment.

Environmental Data Quality and the Search for Representativeness

Deana M. Crumbling, U.S. Environmental Protection Agency, Technology Innovation Office, 1200 Pennsylvania Ave., NW, Washington, DC 20460  USA, Tel: 703-603-0643

Email: crumbling.deana@epa.gov

The first-generation data quality model that equated environmental data quality with analytical quality was a useful starting point for the site restoration community. However, in practice this model fails because it is blind to the complexities that are collectively termed “data representativeness.” To achieve policy goals of “sound science” in environmental projects, the environmental data quality model must be updated to explicitly consider those variables that impact our ability to generate data that are representative from both analytical and sampling standpoints. Because environmental matrices tend to be highly heterogeneous on a variety of spatial, temporal, and chemical scales, measured values may span orders of magnitude within a single site. The idea that “representativeness” reflects some “average property” is thus rendered meaningless, unless “representativeness” is grounded in the project decision-making process. A variety of decisions are typically made over the course of site investigation and cleanup, and each may require data sets with different representativeness. For example, a data set representative of risk assessment decisions (a statistically random data set representative of an average contaminant concentration over some specified exposure unit) will not be representative of cost-effective remedial design (requiring non-random data representative of contaminant locations, mass, and concentration extremes). Data representativeness is therefore meaningful only if sufficient up-front planning has defined the scale of intended decision-making, which is then used to guide data collection activities. Using non-representative data to make decisions leads to poor remedial designs and erroneous conclusions about exposure. Both waste vital public and private resources. Historically, cost considerations made it extremely difficult to manage sampling representativeness in routine projects. But this situation is rapidly changing with the development of new technologies and strategies for managing site cleanup. The U.S. EPA articulated the Triad approach as a practical framework that synthesizes progress in technology and science with the goal of evolving site cleanup practices into second-generation strategies. The Triad approach stresses the importance of systematically identifying and managing project decision uncertainties, including sampling representativeness for data sets. It highlights the contributions of emerging technologies (such as field analysis and decision-support software) and multidisciplinary expertise to the production of accurate conceptual site models that evaluate heterogeneities and other variables critical to successful site restoration. With the Triad as a performance-based technical foundation, project cost savings have been observed to range up to 50% as compared to traditional process-driven strategies achieving the same decision confidence.

KEYWORDS: representativeness, heterogeneity, uncertainty, sampling, data quality, sound science, site cleanup, Triad approach, field analysis

A Simple Phase Equilibrium Model for Predicting the Historical Presence of NAPL: a Case Study 

John D. Moss and Daniel D. Titus, HRP Associates, Inc., 167 New Britain Avenue, Plainville, CT 06062 Tel: 860-793-6899, Fax: 860-793-6871

Using a phase equilibrium partitioning equation, as presented in the CT DEP Remediation Standard Regulations, a theoretical window of potential contaminant concentration ranges in soil (i.e. Cnap), as a historical indication of free phase product (i.e. NAPL) was developed.  Variability in soils and commercial chemical products (e.g. different brands) allow for broad ranges in the relative contaminant adsorption capacity of soils.  As such, a sensitivity analysis was performed on the equation’s variables.  The analysis indicated that Cnap is most sensitive to the soil organic carbon-water partition coefficient (Koc), which is a contaminant variable, and the organic carbon fraction (foc), which is a soil variable.  Therefore, by incorporating possible ranges of Koc and foc, a series of curves for predicting the presence or absence of NAPL were developed.   The calculated theoretical boundary curves of the model, based on maximum and minimum Koc values, indicate a Koc curve above which NAPL was present, an area between the maximum and minimum Koc curves where NAPL may have been present, and a minimum Koc curve below which NAPL was absent.

Using the model presented above, site specific data have been evaluated relative to a 1981 tanker truck release of a 40/60% tetrachloroethylene and high molecular weight cutting oil solution to an area above an industrial septic field, which had been historically impacted by dissolved phase tetrachloroethylene.   Remedial response actions within 10 hours of the spill included excavation of visibly impacted soils.  Following the remedial response, multiple investigations were conducted to determine if residual tetrachloroethylene contamination in excess of 10 mg/kg in the leach field/spill area is due to episodic dissolved phase releases from the septic system or residual NAPL associated with the 1981 release.  The model presented was used to apportion NAPL-related versus dissolved phase contamination.

Groundwater Effects from Highway Tire Shred Use  

Mary O. Brophy, NYSDOT/SUNY School of Public Health, 44 Hawley Street, Binghamton, NY 13901 Tel: 607-721-1838, Email: mbrophy@dot.state.ny.us
Joseph Graney, Binghamton University, Department of Geological Sciences and Environmental Studies, Binghamton, NY 13901, Tel: 607-777-6347, Email: jgraney@binghamton.edu

Approximately 250,000 shredded tires have been used to construct a highway exit ramp as part of a demonstration project.  One up gradient well was installed before the tire shreds were put in place.  Two down gradient wells, and two tire fill sampling wells (that collect water infiltrating through the tire shreds) were installed during the project.  Levels of organic compounds and metals have been monitored since the project was completed in May of 2001.  Organic compounds were not detected in the leachate or down gradient ground water.  Arsenic, barium (Ba), cadmium, chromium, copper, iron (Fe), lead, manganese (Mn), mercury, selenium, silver and zinc (Zn)were quantified in both filtered and unfiltered samples.  Of most concern are the elevated levels of Ba, Fe, Mn and Zn in filtered samples when compared with water quality standards.  Zn and Ba are elevated in one of the tire fill wells whereas Fe and Mn have been consistently elevated in the down gradient wells.  The elevated Zn and Ba may be related to the use of shredded tires.  The elevated Fe and Mn may be associated with traffic on the adjacent interstate and ramp.  Other factors that may contribute to our understanding of these results include the use of salt as a deicer on the adjacent highway and ramp and the intrinsic hydrogeology of the site. Because of the association between increased manganese and some neurodegenerative diseases, such as Parkinson’s Disease and attention deficit disorder, it is important to evaluate the combined long term effects of tire shreds and runoff from roadways on groundwater quality before tire shreds are used more widely in highway construction.

Passive Diffusion Bag Sampler Results From Multiple DoD Installations

John P. Tunks, Parsons, 1700 Broadway, Ste. 900, Denver, CO. 80290, Tel: 303-764-8740, Fax: 303-831-8208  
John Hicks, Parsons, 1700 Broadway, Ste. 900, Denver, CO. 80290, Tel: 303-764-1941, Fax: 303-831-8208 
Javier Santillan, AFCEE/ERB, 3207 Sidney Brooks, Brooks City-Base, TX  78235-5344, Tel: 210-536-5207, Fax: 210-536-5989
Raphael Vazquez, AFCEE/ERT, 3207 Sidney Brooks, Brooks City-Base, TX  78235-5344, Tel: 210-536-1431, Fax: 210-536-4330

Groundwater sample collection using passive diffusion bag samplers (PDBSs) represents a relatively new technology that employs passive sampling methods for monitoring volatile organic compounds (VOCs) in groundwater.  The potential benefits and cost savings associated with using PDBS for long-term monitoring are significant, as no purge waters are generated, and labor requirements for sampler installation and retrieval are minimal.  Results of a field-scale PDBS demonstration performed at 14 Department of Defense installations between May 2001 and May 2002 will be presented.  The primary objective of the PDBS demonstration is to assess the effectiveness of the PDBS method by comparing groundwater analytical results for VOCs obtained using the current (conventional) sampling method with results obtained using the PDBS method.  The comparison of the conventional and diffusion sampling results will allow assessment of the appropriateness of implementing diffusion sampling for VOCs at each sampled well.  Details will include a general description of the work performed, the common findings for all installations sampled, and an analysis of the effectiveness of the technology.  A list of operational parameters that promote the usability of PDBS, and a list of operational parameters that indicate when poor performance is likely to occur will be presented.  A cost and performance analysis also will be presented that includes implementation costs, cost comparison to conventional sampling, sampling cost

Chlorinated Solvent DNAPL Extent Characterization at the East Gate Disposal Yard (Egdy), Fort Lewis, Washington

Kira P. Lynch, Environmental Scientist, U.S. Army Corps of Engineers, Seattle District, 4735 East Marginal Way South, Seattle, WA, 98134, USA, Tel: 206-440-3209, Email: kira.p.lynch@usace.army.mil

The EGDY contains mixed solvent and petroleum hydrocarbon dense non-aqueous phase liquids (DNAPLs) disposed in trenches as drummed waste or directly as liquid waste.  The DNAPLs are the source of a 13,000 feet long TCE plume that contaminates an upper, unconfined aquifer and a lower, confined aquifer to depths of 220 feet below ground surface (bgs).  From 1997 to 2002 the EGDY DNAPLs were characterized in two phases using a variety of tools (e.g., historical aerial photographs, EM-61 geophysical survey, soil gas sampling, exploration trenching, direct-push multi-level groundwater sampling, a membrane interface probe, and analysis of continuous rotosonic cores).  During the EGDY characterization, samples were analyzed in field laboratories or fixed labs with rapid turnaround times to expedite the on site decision making process.  The data collection tools used and the number and location of data collection points for each tool were decided by a team of scientists and engineers as the investigation progressed based upon a continuously updated conceptual site model.  The first phase of the characterization resulted in the location of past disposal areas and several hot spots where DNAPL may have been present in groundwater.  During the second phase of the EGDY characterization, DNAPL data were collected to support the design of thermal treatment of the suspected DNAPL areas identified during the first phase.  The final result of the EGDY characterization effort was the location of three large and several small volume solvent DNAPL sources within the EGDY.  DNAPL was encountered at all three large volume DNAPL areas to a maximum depth of forty-six feet bgs.  The areal extent of the three large DNAPL areas are 0.6, 1.2 and 0.4 acres with an estimated volume of 26,000, 52,000 and 13,000 gallons of mixed solvent/hydrocarbon DNAPL, respectively.  Approximately 100,000 yd³ of NAPL contaminated soils will be thermally remediated by soil heating beginning in 2003.  This presentation will describe how the Triad approach was used to develop a conceptual site model for this DNAPL site, and reduce overall uncertainty in the characterization.

Top