SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
NAPL CLEANUP ALLIANCE
CASPER REFINERY WORKING GROUP
CONFERENCE CALL

April 3, 2001
1:30 p.m.–3:00 p.m.

On April 3, 2001, the following members of the Remediation Technologies Development Forum’s (RTDF’s) Refinery Alliance met in a conference call:

Eva Davis, U.S. Environmental Protection Agency (EPA)
Jeff Hostetler, TriHydro Corporation
Randy Jewett, Texaco Group
John Meyers, ThermoRetec Consulting Corporation
Ali Tavelli, Wyoming Department of Environmental Quality
Kent Udell, University of California at Berkeley
Also present was Christine Hartnett of Eastern Research Group, Inc. (ERG).

INTRODUCTION

The Casper Refinery Working Group has formed to address contamination at a refinery site in Casper, Wyoming. Ali Tavelli said the workgroup is in the process of defining goals and developing a strategy for site investigation and cleanup. She said that John Meyers used an interesting approach at a site in Colorado. She asked Meyers to describe his work so that call participants could determine whether similar techniques should be employed at the Casper site.

INVESTIGATIONS PERFORMED AT A SITE IN COLORADO

Meyers noted that a cleanup project is being performed for two active refinery sites near Denver, Colorado. (The refineries are next to each other, and are being treated as one site.) The refineries, which have been operational since the 1930, have a combined capacity of about 90,000 barrels per day. A slurry wall has been installed and containment systems have been constructed. Despite these remediation efforts, contaminant concentrations in downgradient ground water are above regulatory standards. Therefore, in 1998, the site owners entered into a joint consent order with the state and EPA Region 8. The owners agreed to implement interim measures that: (1) prevent contaminants from migrating off site, and (2) recover free product to the maximum extent practicable. Meyers said that his discussion would focus on efforts to accomplish the latter objective.

Before initiating any investigatory or cleanup activities, Meyers said, the site owners worked with regulatory agencies to identify clear remediation goals. Rather than focusing on non-aqueous phase liquid (NAPL) removal, Meyers said, the decision was made to focus on reducing NAPL mobility. Once a goal had been chosen, a targeted investigation was performed to glean more information about the site’s subsurface conditions. This investigation had four objectives: (1) estimate the nature and extent of free product, (2) identify product type and distribution properties, (3) estimate the volume and mobility of free-phase product, and (4) design a conceptual product recovery system. Meyers described the techniques used during the investigation and the conclusions that emerged. These are the highlights of his discussion:

• The Rapid Optical Screening Technology (ROSTÔ) was used to obtain qualitative information about hydrocarbon location. About 90 ROSTÔ borings were installed across the 60-acre plume, Meyers said. The borings were drilled to a depth of 30 to 40 feet and spaced about 200 feet apart. (Meyers acknowledged that better resolution could have been achieved if spacing had been more dense. At this site, however, the 200-foot spacing was justified because: [1] the site’s subsurface is fairly homogeneous, and [2] denser spacing would not be practical at an active refinery site.) Meyers said that the ROSTÔ tool sends out a 390-angstrom wavelength into subsurface formations and measures fluorescence. Lighter-end petroleum products fluoresce more strongly than heavier-end products. The response, which is measured and displayed, can be processed through a software program to yield a picture similar to a chromatogram.
• A multi-step process was used to correlate ROSTÔ fluorescence intensity with oil saturation. Meyers said that conventional borings were installed next to 18 of the ROSTÔ borings. Cores were collected from these 18 locations, and samples were collected continuously from the surface to the smear zone. Acetate liners were used in the sampling process. When drawn to the surface, the liners were quickly capped and the cores were placed in liquid nitrogen. The cores were frozen to prevent fluid from draining out of the soils. Nevertheless, some fluid loss occurred. The cores were sent to a laboratory that specializes in petroleum analysis. Laboratory staff cut 57 plugs out of the 18 cores. All of the samples were analyzed for oil and water saturation, but only three were subjected to a more rigorous analysis of soil-fluid interactions. These three samples, each of which represented a different lithology, were analyzed for intrinsic permeability, capillary pressure, and water/oil relative permeability. Then the following steps were taken:
  
  
  • Step #1: The measured values of oil and water saturation were recorded and graphed. Meyers said that the laboratory reported oil and water saturations as a proportion of pore volume. He expected oil saturation plus water saturation to equal 100% in samples collected below the saturated zone. This was not the case, however; Meyers suspected that fluid loss was responsible for the puzzling results. Kent Udell proposed an alternative: perhaps gas was present in significant quantities in the saturated zone. (Udell reported seeing such a phenomenon at Alameda Naval Air Station.) Meyers said that oil saturation measurements at the site were lower than he expected (10% to 40%), and connate water levels (reaching 60% in some samples) were higher than he anticipated. These results surprised him, and he questioned their validity, at least until he saw that they were: (1) consistent with the capillary pressure data, (2) validated using a hydrocarbon bail-down test, and (3) consistent with what had been found at a site in North Dakota.
    
    
  • Step #2: An American Petroleum Institute (API) model was used to determine what the oil and water saturation should have been. Meyers said that 10 of the ROSTÔ borings were located right next to monitoring wells. Samples were collected to measure the thickness of free product within the monitoring wells. The API model was used with these data, along with a variety of other input parameters, to determine the thickness of product in the surrounding aquifer.
    
    
  • Step #3: An equation was used to correlate ROSTÔ intensities to oil saturations. Meyers said that oil saturation was calculated by multiplying slope by ROSTÔ intensity. The values generated were plotted on a graph.
    
    
  • Step #4: The oil saturation curve generated in Step #3 was laid over the curve generated in Step #2. Meyers noted that 10 of the data points on the ROSTÔ-generated oil saturation curve were collected from the same locations as the 10 monitoring wells that were used to generate the API-model–generated curve. The 10 points on the ROSTÔ-generated curve were adjusted to fit the API-model–generated curve, and correction factors were calculated for each of the 10 data points. The average correction factor was about 2.4.
    
    
  • Step #5: The final oil saturation curve was calculated. Meyers said that the averaged correction factor was applied to the other 80 ROSTÔ points to generate the final oil saturation profile.

• New maps were generated. Prior to the site investigation, Meyers said, site owners only had a vague understanding of subsurface conditions. This was because their maps simply showed where free product had been found; the maps gave the impression that a large lake of oil was present under the site. After the investigation, Meyers said, maps were created to depict the thickness of product in different parts of the aquifer. In addition, product mobility curves were generated to differentiate between areas of high and low NAPL mobility. Data for these mobility maps were generated using the following equation:

Mobility = (Oil conductivity) ´ (True thickness of product in the soil)
Viscosity 

Meyers said that the new maps allow investigators to identify areas with high saturation levels and decent mobility. This gives investigators a better chance of targeting important areas for remediation. The investigation has helped site owners, regulators, and remediation designers understand the magnitude of the problem at the site. Now that the problem is better quantified, it is easier to determine what can reasonably be accomplished during cleanup. Meyers said that the product mobility curve will be used to select a cleanup endpoint. It is crucial to pick an endpoint before remediation efforts are initiated, he said. Otherwise, how will investigators know when to shut off a remediation system? The NAPL evaluation also allowed investigators to make product recovery performance predictions that will provide a benchmark for comparing to actual performance.

Meyers said that efforts have been initiated to move forth with remediation activities at the Colorado site. After several different remediation schemes were evaluated, water flooding was selected as the preferred choice. The system’s design is being finalized and a pilot study will be initiated soon. Udell said that he is eager to see what results are obtained using an ambient water flood. It may be interesting, he suggested, to follow up with a hot water flood.

PATH FORWARD FOR THE CASPER REFINERY SITE

Tavelli asked whether the site investigation methodologies that Meyers discussed would be useful at the Casper site. Call participants reached a consensus that using the ROSTÔ tool was a good idea: it is relatively cost effective, relatively simple to implement, and it provides good qualitative data. It would also be useful, they said, to collect cores in the areas that are shown to have intense fluorescence. Obtaining ROSTÔ data and correlating it with oil saturations will allow investigators at the Casper site to refine their understanding of the subsurface.

Call participants also listed some other analytical techniques that might have merit:

• Carbon dioxide (CO₂) probes. Udell said that CO₂ probes were used to identify the location of free product at the Guadalupe Oil Field site. Collecting information about CO₂ concentrations also provides useful information about the level of biodegradation in the subsurface.
  
  
• Gore-sorbers. Jeff Hostetler said that Gore-sorbers can provide qualitative information about the subsurface.

Call participants agreed to think more about the analytical techniques that would be best to use at the Casper site. In the meantime, Tavelli and Hostetler will put together a draft statement summarizing the workgroup’s goals. This will be circulated to workgroup members for approval before being sent to the NAPL Cleanup Alliance members. Call participants talked briefly about the goals, and agreed that remediation will need to be accomplished in multiple stages at the Casper site. First, an active remediation technology will be needed to remove large quantities of NAPL to “cut the head off the dragon.” Then, once a certain endpoint is reached, the remediation system will be shut off and natural attenuation will be expected to reduce contaminants to concentrations below maximum contaminant levels. Two important questions will need to be addressed: (1) What endpoint should be used? and (2) How long a timeframe is reasonable to accomplish the entire cleanup?

Hostetler said that the workgroup needs to produce a work plan for the Casper site within the next couple of months. By the end of the year, he hoped, site investigation would be completed, results would be compiled, and potential remediation technologies would be identified.