April 22, 2002

Response to Dye Tracer Studies at the Lemon Lane Landfill

Dear Mr. Hailer:

The United States Environmental Protection Agency (EPA) is in receipt of your March 6, 2002 correspondence and your February 20, 2002 e-mail concerning the dye tracer studies completed by Westinghouse, now doing business as Viacom Inc. The EPA has tasked its consultants Earth Tech and Tetra Tech to review the dye tracer studies. The following will address several issues regarding the ground water tracer test investigations at Lemon Lane Landfill, and their interpretation, as raised by your reviews.

EPA will respond to your comments considering relevant site-specific data, general tracer test practice, previous reviews of the data conducted by independent parties, and Westinghouse's own review and summary of the data (Westinghouse, July 1994). Some of the issues you raise are directly addressed in the 1994 document. Other specific referenced documents are listed at the end of this letter. EPA states specific comments contained in correspondences, and provides comments concerning several of the issues raised.

1. Location of Multiple Injection Points - Injection points are neither in the waste, nor in locations likely to intersect the natural flow paths that PCBs might take.

No tracer test will identify a specific flow path; the best that can be hoped for is to identify specific points along the flow path, or paths (wells, caves, karst windows) and the terminal emergent points of the ground water flow system at springs. The issue raised here is that tracers may have followed different flow paths than the PCBs released from the landfill. This is possible, but not supported by the data as a whole.

PCBs at the site appear to be released from storage in the vadose epikarst zone, primarily during storm events. This zone may extend vertically several tens of feet above the potentiometric surface that is associated with the principal phreatic conduit zones, and the springs in the area that drain these conduit zones. Tracers injected in wells open to the phreatic zone allow the identification of spring discharge points for phreatic ground water beneath the site, and by inference the locations where PCBs may emerge. EPA recognizes that ground water direction in the vadose epikarst zone may be different than that in the phreatic zone. This condition is fairly well documented and has to do with the formation of conduits under free-surface gravity flow conditions in the vadose zone versus pipe flow conditions in the phreatic zone.

Palmer (1986, 1999) has documented that vadose gravity flow in karst terrains developed in gently dipping strata tends to be strongly accordant to the local dip. On the other hand, phreatic flow conduits tend to develop parallel to strike of the rock, or non-concordant to the structure along the most hydraulically efficient flow path. Ground water flow in the epikarst may occur laterally for a considerable distance before the descending ground water reaches the saturated zone. The question thus becomes whether the magnitude of this lateral migration is large enough that the ground water and contaminants in the ground water are transported into an adjacent ground water basin, as would be defined from tracer injections in phreatic wells. For a site such as Lemon Lane Landfill that is located near the middle of a ground water basin (see Figure 1; basin delineation by Fitch, 1994), lateral flow in the vadose zone is probably not a huge issue.

The injection points used in the 1989 testing consisted of three existing monitoring wells surrounding the landfill, MW-1S, MW-10 and MOO-7. Each of these is, or was, completed into or very nearly into the top of phreatic zone. The optical brightener tracer injected at MOO-7 was qualitatively recovered at Illinois Central Spring. Later quantitative tracer work in 1996 confirmed that the MOO-7 area drains to Illinois Central Spring. The Fluorescein at MW-10 was visually observed at Illinois Central Spring (Westinghouse, 1994). These results clearly demonstrate that the phreatic zone in the southern portion of the landfill drains to Illinois Central Spring. But what is known about the direction of vadose flow in this area?

On October 31, 2001 Fluorescein tracer dye was injected into the epikarst at the 840 foot elevation at piezometer LF-6-8 in the southeastern portion of the landfill (Viacom, 2002). The tracer was injected about 25 feet above the phreatic zone. The epikarst in this area is known to contain high concentrations of PCBs. Free product PCB oil was recovered from LF-6 when it was installed. US EPA believes that this area of the landfill may be a significant source of PCBs released at Illinois Central Spring.

Beginning as early as 6.7 hours after the injection, the tracer was instrumentally and visually observed in five monitoring wells (00370, 00300, 00300A, NN300 and NN300A) southwest of LF-6. The dye was visually and instrumentally detected at Illinois Central Spring beginning 21 hours after injection.

The observed dye concentrations in the five wells (up 9,500 ug/1), and the fact that samples for dye analysis were collected without purging suggests that these wells are located very close to a primary conduit draining the southeast portion of the landfill. Potentiometric data and recent pump test results indicate that all of these wells, with the exception of 00300A, are highly interconnected in the phreatic zone at the 795 to 800 elevation. Moreover, ground water pumping from wells open at this level is known to instantaneously impact Illinois Central Spring flow. Well 00300A is perhaps best interpreted as a deep vadose well because the water level in this well is significantly higher than the other slightly deeper wells in the area that are open to the 795-800 foot phreatic zone.

EPA's interpretation of the October 31 tracer test data is that vadose flow in the epikarst from the PCB-contaminated area at LF-6 is southwest (down dip as suggested by Palmer's model) to somewhere close to 00300A, and at this point reaches the 795 to 800 foot phreatic zone. And, we know from this trace, and the aformentioned traces from MOO-7 and MW-10 (which bracket this area) that the phreatic zone Groundwater flow is emergent at Illinois Central Spring. We are thus confident that the epikarst drainage in the southern part of the landfill at LF-6 ultimately emerges at Illinois Central Spring, based on the most recent tracing data, and that the earlier traces employing wells peripheral to the landfill correctly evaluated the potential PCB discharge points from Lemon Lane Landfill.

EPA also points out that the use of wells, or infiltration galleries, adjacent to landfills is common practice among the best ground water tracing practitioners in the world. See Rabold and Ewers (1995) and Ewers and DeFosset (2000) for recent examples of such traces in Indiana. The approach is inherently different than tracing a point source discharge. See Earth Tech (1998, 1999) for recent examples from Indiana of point source traces. Margin injections are useful when contamination may be associated with a large site area, and are often more successful because the conditions of the injection are better controlled. Marginal injections also avoid ancillary problems such as disposition of excavated waste material and potential flushing of contaminants into the subsurface.

2. The tracer flyer can be adsorbed, delayed, or degraded in the subsurface, particularly in landfill waste debris. Absence of detections may be due to these interferences.

Dyes are adsorbed, delayed and degraded in the subsurface, and these are the primary reasons why dye recovery is never 100 percent. Some dyes are more susceptible to certain types of losses than others. In EPA's view the quantities of tracer injected at Lemon Lane Landfill have been adequate, and in some cases excessive, based on current practice.

3. The timing and frequency of sampling, and tracer collection and measurement were inadequate.

It should be remembered that the 1989 tracer test was designed as a qualitative trace, employing passive detectors for dyes. Bromide was co-injected as a conservative tracing agent. The instrumental work on dyes that was done by James F. Quinlan, and later by Indiana University (Westinghouse, 1994) lacked a rigorous analytical protocol, and was largely qualitative in nature.

EPA interpreted these results conservatively in the design of the 1990 trace, wherein all locations where dye was qualitatively observed in 1989 were again monitored in 1990.

4. No justification was given for the 3X criteria used to establish a bromide detection. The criteria for bromide significance was such that virtually all the sampling points had bromide events.

One of the shortfalls with the 1989 tracer test was the lack of adequate background determination for bromide. EPA believes that the 3X background criteria for interpreting a positive result provided a very conservative approach to interpreting the data. In 1989, there were no established criteria, and researches even today use various criteria for establishing a positive detection. EPA notes that Crawford and Associates (1997) utilize a 10X above the highest background criteria for interpreting a positive spectrofluorometric result. Ewers and DeFosset (2000) suggest a value of 4X to 10X.

5. Ancillary Data Collection - Two locations, ICG-2 and ICG-3 were not monitored for bromide.

Both of these stations had very high background levels of bromide, apparently due to leaching from a road salt pile in the area. It is doubtful that meaningful data could have been collected at either of these locations. Both of these locations were retained for the subsequent 1990 tracer test that utilized a dye with low background rather than bromide as the tracing agent. The dye was not observed above background levels at either location during the 1990 tracer investigation.

Further, no PCBs were detected in sediment at either location in 1991, nor were PCBs detected in water during multiple sampling events in 1992 (Westinghouse, 1995). EPA considers these data to be definitive evidence of a lack of connection to Lemon Lane Landfill.

6. Estimates of Tracer Recover - There is no indication of how much bromide was injected. Without a calculation of recovery it is not possible to estimate whether the study has monitored all the exit points, whether the bromide is transported in deeper subsurface flow or has accumulated in sequestered zones.

Refer to Westinghouse (1994). The total amount of bromide injected during the 1989 test was 196,128 grams. The amount reported to be recovered at the Quarry monitoring location was reported to be 193,786 grams, or 98.8% of the total injected. EPA believes this value to be bias high because the bromide peak at Quarry was not sampled and a background value of zero was utilized in the recovery calculations. The 98.8% calculated value considers all bromide recovered during a 57 day sampling period. But, only about 17 percent of the bromide was recovered during the first 70 hours of the test that was sampled at 2 to 14 hour interval. Subsequent storms continued to flush high bromide values through the ICS system, and a large part of the injection was slowly released from the well bore injection sites during the 57 day period.

Clear, unambiguous breakthrough curves for bromide occurred only at Illinois Central Spring and the Quarry monitoring location. Several ambiguous detections of bromide occurred during this test. In general, these detections were single sample events, elevated only slightly above background, occurred a long time after the injection, or were a combination of the above. As such, they were regarded as highly questionable evidence of hydraulic connection to Lemon Lane Landfill.

7. Assessment of Connections - The WEGS report (Table 2 - Matrix or Results) indicates that only 4 of 21 possible exit points had strong connections. The very few exit points identified as likely and the four others as possible suggest that only a few exit points may carry contamination from the site. Review of the bromide data and the limited qualitative dye detections that were reported, indicates a different conclusion.

EPA recognizes several limitations to the 1989 investigation:

• The 1989 test was qualitative for dyes using charcoal
• Colored elutriates were detected that were difficult to interpret
• No background dye measurements were obtained
• Only a single background measurement for bromide was made
• Detection of Direct Yellow was made by spectrofluorophotometer, but not reported on cotton li> Single sample dye detections were made at several locations, with a lack of breakthrough, or at a very long time after injection
• SSFP analyses were only made several months after the fact, with no background determination
• Bromide sample collection began after the peak at the principal tracer recovery location.

Despite these limitations the bromide data indicated an overwhelming majority of the tracer appeared at the Illinois Central Spring and the Quarry monitoring location. Sporadic detections of dye and bromide at other locations are questionable in demonstrating hydraulic connection to the site, and suggest, conservatively, minor and ephemeral connection to the site. But, because of the several ambiguities in the data, the governmental parties requested that a storm event tracer test be rerun, this time attempting to apportion the distribution of tracer among springs.

The 1990 fluorescent tracer apportioning showed that the Quarry monitoring location (monitored instead of Illinois Central Spring because it was located further downstream) received the overwhelming majority of the tracer recovered. ICG-1 spring was found to receive tracer at the time, which upon subsequent investigation turned out to be due to surface stream backflooding from the Quarry Spring Branch. Three springs to the north of site, Slaughterhouse, Packing House Road (PH Road), and Packing House Culvert(PH Culvert) had similar tracer detections as a breakthrough curve, indicating they were three distributary resurgences of the same conduit. Tracer concentrations at Quarry were three orders of magnitude above those at Slaughterhouse and the associated springs. Sporadic and near-background level detections of tracer at 11 other springs indicated possible minor and ephemeral connection with ground water in the vicinity ofthe landfill.

It should be kept firmly in mind that tracer detection is not PCB detection. Tracers are highly soluble in water and tracer injection on the periphery may be more widely distributed than contaminants from the site itself. Both EPA and Westinghouse followed a very conservative approach in the interpretation of the tracer test data. Following the 1990 tracer test, Westinghouse initiated sampling at 33 locations potentially connected to the landfill based on a conservative interpretation of the tracer test data. This sampling was conducted in February, June, September and November 1992. PCBs were detected at Illinois Central Spring, the Quarry monitoring location, ICG-1, Robertson Spring and Detmer A Spring (Westinghouse, 1995). EPA believes that the later two locations are located in a separate groundwater basin and that the PCBs are related to drainage from the former ABB plant site. PCBs were not detected at the large majority of springs that showed possible, ephemeral connection to the site.

Illinois Central Spring, Quarry, ICG-1, Slaughterhouse, PH Road, and PH Culvert springs were sampled for 8 consecutive weeks beginning in August 1995, and then monthly until July 1996 during low flow. Only Illinois Central and Quarry had PCBs detected in these low flow water samples. Nine of the thirteen peripheral springs (Stoney West, Stoney East, WN-1, WS-2, ICG-6, Snoddy, Bypass 37, Urban, and Crestmont) that showed potential connection to the landfill based on sporadic tracer detections in at least one water sample or activated carbon detector in 1987 or 1990 were included in the 1995 sampling program. The Jack's Defeat East, Kirby Road, Clear Creek Upstream and Detmer A locations were excluded because they were associated with other karst drainage basins.

Three locations,WN-1, WS-2, ICG-6, were dry at the time of sampling. One spring, labeled Snoddy, which had a PCB detection of 0.15 ppb, was actually the surface water flowing where commercial construction had obliterated the previous spring location. The rest of the samples from the peripheral springs were all non-detect at 0.1 ug/1.

Sediment sampling was also conducted at the nine aforementioned peripheral springs in 1995. Samples were collected at the spring orifice, and four downstream locations where sediment was observed to collect. Except for a 0.67 ppm detection on a split sample at Stoney West, and a similar low-level detection on an Earth Tech split sample at Urban Spring, no PCBs were detected at the peripheral springs. These included the WN-1, WS-2 and ICG-6 locations that were dry. PCBs were detected at Slaughterhouse, PH Road, and PH Culvert Springs as well as Illinois Central and Quarry Springs, the directly connected springs where unambiguous tracer breakthrough has repeatedly occurred.

Based on the data as a whole, EPA does not dispute the fact that minor and ephemeral connections may exist between ground water in the vicinity of Lemon Lane Landfill and a few small springs in the area. However, EPA is convinced that substantially all of the PCBs releases derived from Lemon Lane Landfill and emergent in karst spring flow occur via the Illinois Central Spring flow system and it's discharge point at Illinois Central Spring. The three springs of the Slaugherhouse Spring Complex appear to have a temporary high-flow connection to Lemon Lane Landfill and transmit a minor PCB discharge. See the following response.

8. Frequently cited as the basis for the separation of contamination issues into source control and water treatment is the "fact" that 98.9 % of the water moving off site is thorough six exit points. This 1989 WEGS report is cited as the basis for this quantitative determination of site water movement. A review of this report does not show any conclusion containing that number, nor is there sufficient data in the investigation to come to that conclusion.

Refer to Westinghouse (July 1994), Table 6, and the discussion related to ICG-1 Spring contained herein (Item 12, below). Westinghouse estimated that 4,898.753 grams of Rhodamine WT tracer (about 18 percent of the injected total) was recovered during the 1990 high-flow tracer test. A total of 64.135 grams was estimated to have been recovered at ICG-1. Since the dye detected at this location is believed to be related to backflooding of water from the Illinois Central Spring branch, counting this as recovered dye is effectively counting it twice. So, subtracting the 64.135 grams from the 4,898.753 total leaves 4,834.618 grams as recovered. Of this quantity, 4,744.234 grams (98.13%) was recovered at the Quarry monitoring location, and 30.183 grams (0.62%) was recovered at Slaughterhouse Spring complex (Slaughterhouse, Packinghouse Culvert, and Packinghouse Road). The total recovered at Quarry and Slaughterhouse is thus 98.13% plus 0.62%, or 98.8%.

EPA disagrees. The 1989 qualitative high flow tracer test was repeated in 1990 at the request of the Governmental Parties to the Consent Decree to determine the proportional distribution of flow to the tracer monitoring locations utilized in 1989. Because of the many ambiguous and low level tracer results obtained in 1989, a tracer dye with generally low background interference was utilized. An expanded background data set consisting of four samples from each monitoring station was utilized in 1990. Dye recovery was based on spectrofluorometric analysis of a single injected dye, which prevented analytical interference between co-injected tracer dyes, and facilitated dye quantitation. Sampling frequency was such that a complete dye concentration curve could be generated at each monitoring station where breakthrough occurred.

10. Illinois Central Spring was not monitored

During the 1990 tracer test, Westinghouse elected to perform tracer monitoring at a culvert downstream from both Illinois Central Spring, and the various Quarry springs. There were two reasons for this. First, it was known at that time that tracer dyes appeared at both Illinois Central Spring, and Quarry A and B Springs. But, the hydrologic relationship between Illinois Central Spring and the Quarry Springs was not completely understood. It was known that water from Illinois Central Spring entered the Illinois Central swallowhole area downstream of the emergence, and that the flow reappeared at Quarry B or Quarry A and B depending on flow conditions, a short distance downstream. But, it was not known if Quarry A and B were also supplied by a direct underflow from the Illinois Central Spring conduit system whereby tracer would be missed if only Illinois Central Spring was monitored. Hence, the monitoring point was moved downstream below Illinois Central and the Quarry springs to be sure all of the flow from the system was monitored. A culvert at the monitoring location, referred to as Quarry, provided an excellent location for determination of flow.

11. Dye recovery was only 18%

EPA does not regard this as an exceptionally low dye recovery given the tracer dye used and the type of dye injection. The tracer dye used in the 1990 investigation was Rhodamine WT. This dye was utilized because it generally has low background levels and had not been previously utilized in tracing investigations at the site. However, the dye is also subject to higher losses due to adsorption than other tracer dyes (Aley, 1999). Worthington (1991) cites Rhodamine WT tracer recoveries from several quantitative tracer tests in the Canadian Rocky Mountains. Recoveries ranged from 15 to 47 percent for five tests wherein samples were collected for a sufficient period of time to define a dye recovery curve. The 18 percent recovery is toward the low end of the values reported by Worthington.

Low dye recovery may also be associated with tracer injected in wells. Tracer injected into wells may be recovered at a low rate due to incomplete flushing of the dye in retention in the well bore and adjacent subsidiary conduit zones. EPA notes that during pump and tracer tests conducted in October 2001, visually "pink" colored water was recovered from monitoring wells in the vicinity of MW-10, indicating residual dye remaining 11 years after the original dye injections. Viacom also suggests that weir overtopping at the Quarry monitoring location, and "quenching" effect due to high dye concentrations during instrumental analysis of the Quarry samples also contributed to the low recovery.

EPA believes that a meaningful comparison of dye recovery may be made by comparing the Rhodamine WT recovery in 1990 with Fluorescein recovery calculated from data in Fitch (1994). In May 1992, Fitch injected 341 grams of Fluorescein at Martin Sink, 400 feet southwest of Lemon Lane Landfill. The injection was made during a small storm event by flushing the dye into an open cave using 1,500 gallons of potable water. Breakthrough of the dye plume began at Illinois Central Spring about 19 hours after injection. Continuous flow records were maintained at Illinois Central Spring, and a downstream gage at Quarry culvert. Samples for dye analysis were collected at one-hour intervals until the dye peak, at three to four hour intervals during the receding portion of the dye breakthrough curve, and then daily for 19 days. Calculated dye recovery for the initial 330 hours of the test, and until the dye concentration fell below the 1 ug/1 background range, was 43.6 percent at Illinois Central Spring. For a comparable time frame, the dye recovery at Quarry culvert was 37.3 percent. The principal differences between this test and the 1990 RhodamineWT test are that a relatively non-adsorbing dye was utilized in 1992, the dye injection was made into an open cavern, and large amounts of water were utilized to flush the dye into the conduit system. EPA believes that all of these factors attributed to the better dye recovery during the 1992 test.

12. The highest dye recoveries occurred at two springs hydrologically downgradient from Illinois Central Spring, but the water treatment plant is located at another spring.

Refer to Figure 1. The fluorescent tracer apportioning from the 1990 high flow trace showed Quarry Springs (monitored instead of Illinois Central because it was located further downstream) receiving the overwhelming majority of the tracer recovered. ICG-1 spring was found to receive tracer at the time, which upon subsequent investigation turned out to be due to surface stream backflooding from the Quarry Spring Branch (see below). Three springs to the north of site (Slaughterhouse, Packing House Road, and Packing House Culvert) had similar tracer detections as a breakthrough curve, indicating they were three distributary resurgences of the same conduit. Tracer concentrations at Quarry were 3 orders of magnitude above that of Slaughterhouse and associated springs. From the 1990 test, six springs were regarded to be directly connected to the site. These sites were Slaugherhouse, Packinghouse Road, Packinghouse Culvert, ICG-1, Quarry and Illinois Central.

The highest dye recoveries during the 1990 test were at the Quarry monitoring location (4,788 grams), and a downstream spring referred to as ICG-1 (64.1 grams). EPA assumes that the issue raised here is the detection of tracer at these two locations downstream from Illinois Central Spring. Samples were not collected from Illinois Central Spring for the reasons mentioned above. All previous tracings that had detected tracer at the Quarry location, had also detected the tracer at Illinois Central Spring.

Viacom believes that the tracer detections reported at the ICG-1 location during the 1990 high-flow trace are related to back flooding of the sampling location by the Illinois Central Spring branch, and do not as reported prior to 1996, indicate a hydrologic connection to the Illinois Central Spring flow system. EPA also notes that the lack of a connection to the Illinois Central conduit system is suggested by the lack of PCBs in ICG-1 discharge.

On April 23, 1996 Westinghouse conducted a high-flow time-of-travel dye injection of Illinois Central Spring Branch and West Fork Clear Creek. Tracer was injected at Third and Adams Street and visually followed downstream. The tracer was observed by Westinghouse to backflow into the ICG-1 culvert discharge point at the location where previous dye monitors and water samples for dye analysis had been collected.

The Quarry monitoring location is known to be directly connected to Illinois Central Spring via the Illinois Central Spring Swallowholes (see response above). The tracer test conducted in October 2001 provided an opportunity to evaluate whether or not there is a common underflow connection from the Lemon Lane Landfill conduit system to both Illinois Central Spring and Quarry B Spring. Both Illinois Central Spring and Quarry A and B springs contain PCBs. Under low flow conditions, water discharging at Illinois Central Spring flows south a short distance and sinks at the Illinois Central Swallowholes. This flow is known to reappear a short distance downstream at Quarry B Spring. Quarry A Spring is dry under very low flow conditions. Tracer dyes and PCBs emergent at Illinois Central Spring follow this pathway and emerge also at Quarry B Spring. The issue you raise here is whether or not a direct underflow conduit exists that connects the Illinois Central Spring conduit system, and Quarry B Spring. Such a conduit would allow PCBs to bypass the Illinois Central Spring Treatment Plant and emerge downstream at Quarry B Spring. EPA has two lines of evidence that suggest that such a conduit does not exist.

First, we note that the calculated mass of Fluorescein dye recovered at Illinois Central Spring versus the Quarry monitoring location during the 1992 tracer investigations discussed above (data in Fitch, 1994). Calculated dye recovery for the initial 107 hours of the test, and until the dye concentration fell into the 1 to 2 ug/l background range, was 84 percent at Illinois Central Spring. For a comparable time frame, the dye recovery at Quarry culvert was 71 percent. The calculations indicate that less dye was recovered at Quarry. Perhaps some dye was lost due to adsorption and photodecomposition along the short surface and subsurface flow route from Illinois Central to Quarry, but as a practical matter the amount of dye recovered was not greater at Quarry. There is clearly no evidence that additional dye appeared at Quarry via some direct underflow route.

Secondly, a tracer test conducted on October 31, 2001 demonstrates that no direct underflow route to Quarry B exists. During the 2001 tracer test the Illinois Central Spring Treatment Plant was shut down and all water from Illinois Central Spring was pumped to storage. The purpose of this was to stop both the flow of water (and tracer dye) into the Illinois Central Swallowhole. Flow also enters the Swallowhole from several seeps in the Illinois Central Spring branch upstream from the Illinois Central Swallowhole, and downstream from the Illinois Central Spring Treatment Plant intake. To stop this flow as well, US EPA installed pumps upstream of the Swallowhole and pumped all water entering the Swallowhole to plant storage. Thus, all discharge into the Illinois Central Swallowhole was stopped for the initial portion of the tracer test. Any dye emerging at Quarry B Spring would indicate that Quarry B Spring was supplied, at least in part, by an underflow conduit system directly from Lemon Lane Landfill. Conversely, the absence of dye would conclusively demonstrate that no underflow connection exists, and that the PCBs detected at Quarry B Spring could be attributed solely to the discharge of PCB-contaminated water into the Illinois Central Swallowhole, or residual contamination remaining in the section of karst conduit between the Swallowhole and the spring. Samples for dye analysis were collected at the pump location upstream of the Illinois Central Swallowhole (referred to herein as Swallowhole Seep) and at Quarry B spring for a period of about 36 hour after dye injection. Fluorescein was detected in only one of 42 samples collected at Quarry B Spring.

This detection (0.26 ug/l) occurred well before the arrival of the dye at Illinois Central Spring, and was below background levels obtained prior to the dye injection. Fluorescein was not detected in any of 13 samples collected at the Swallowhole Seep location.

EPA thus believes that no underflow condition exists at Quarry B or Swallowhole Seep. Quarry B is supplied primarily by water from Illinois Central Spring water sinking at Illinois Central Swallowholes and some additional drainage basin, perhaps several acres in size, that supports base flow to the spring when there is no flow to the swallowhole. Treating the ground water emergent at Illinois Central Spring minimizes the amount of water to be treated because it is the farthest known upstream location where the PCBs within this flow system may be captured and treated.

Finally, with respect to Illinois Central Spring and the Slaughterhouse Spring system being the principal PCB discharge points, EPA cites review comments by two karst experts. Both the 1989 and 1990 high flow tracer test data were reviewed by Dr. Richard L. Powell, WW Engineering and Science. In a letter dated April 17, 1991, Dr. Powell concluded:

"In summary, I doubt the validity of the identifcation of low levels of fluorescence at some of the spring locations that is claimed to indicate a widespread dispersal of Groundwater from Lemon Lane as the reports indicate. I have no doubt that all or nearly all of the cavernous flow from Lemon lane flows to Illinois Central and Quarry Spring, and that some of the associated shallower karst Groundwater flows several directions to a few other springs."

The tracer test results from Lemon Lane Landfill were also reviewed for the Citizens Opposed to PCB Ash (COPA) by Gareth Davies, an independent karst consultant and ground water tracing expert. Mr. Davies concluded that:

"From a combination of geomorphological, contaminant, PCB, hydraulic gradient, and discharge data, it can therefore be hypothesized that the most likely discharge locations for ground water from beneath the landfill are the Illinois Central and Quarry Springs."

It should be emphasized that the results of the initial tracing experiments performed at the site suggested that the discharge locations for the site were principally Illinois Central Spring and Quarry Springs (McCann and Krothe (1992). The high-flow results suggested dye had migrated to many surface and ground-water locations far beyond these springs.

There would be no logical reason why ground water would flow to these other distal locations given low gradient, long, and thus inefficient flow paths, rather than use the more efficient pathways to Illinois Central Spring and Quarry Spring, where most of the PCBs are being discharged. The third most efficient pathway based upon hydraulic gradient would be the Slaughterhouse Spring Complex, also a site where PCBs have been detected.

Based upon the tracing results it is logical to conclude that the springs where both dyes and PCBs were detected are the discharge locations for Lemon Lane Landfill. The detection of dyes elsewhere but without PCBs in most cases suggests that the source of those dyes must be different than the source of the PCBs

Finally, after review of data provided by you by our karst consultants, the EPA does not agree with your interpretation of the Groundwater flow at the Lemon Lane Landfill. If you would like to discuss dye tracer test issues with John Bassett of Earth Tech, the EPA will set up a meeting in Bloomington, Indiana.

Thomas Alcamo
Chemical Engineer

References

Aley, Thomas, 1999, Groundwater Tracing Handbook: Ozark Underground Laboratory, Protem, Mo, 35p.

Crawford and Associates, Inc., 1997, Karst Groundwater Investigation Research Procedures. 8 p., figures, tables, appendices, December 1.

Earth Tech, 1998, Fluorescent Dye Tracing Investigation, North Harrison Community School Corporation, Wastewater Treatment Plant, NPDES Permit IN0038890, 16p, figures, tables, attachments, August.

Earth Tech, 1999, Sinkhole Dye Tracing Investigation, South Harrison Community School Corporation, Junior / Senior High School Wastewater Treatment Plant, Elizabeth, Indiana NPDES Permit IN0056839, figures, table, attachments, September.

Ewers, Ralph O. and Kevin DeFosset, 2000, Dye Tracing Studies, Washington County Landfill, Salem, Indiana: report prepared for Indiana Department of Environmental Management, 28 p., figures, tables, July 7.

Fitch, James R., 1994, A Karst Groundwater Study to Delineate the Quarry Spring Basin Groundwaters Near the Lemon Land Landfill, West-Central Bloomington, Indiana: unpublished MA thesis, Indiana University Department of Geological Sciences, 140p.

Palmer, Arthur N., 1986, Prediction of Contaminant Paths in Karst Aquifers m Proc. Environmental Problems in Karst Terranes and Their Solutions, National Water Well Association Symposium, Bowling Green, Kentucky, p. 32-53.

Palmer, Arthur N., 1999, A Statistical Evaluation of Structural Influence on Solution-conduit Patterns m Karst Modeling: Karst Waters Institute Special Publication 5, p. 187-195.

Rabold, Richard and Dr. Ralph O. Ewers, 1995, Hydrogeologic Site Characterization Report forthe INAAP RCRA Landfill, Charlestown, Indiana: report prepared for ICI Americas, Inc., Indiana Army Ammunition Plant, 49 p., figures, tables, maps, October 27.

Viacom, 2002, Lemon Lane Landfill, Karst Aquifer Test Reports from October-November 2001, 21 p., figures, tables, March 13.

Westinghouse, 1994, Summary Report - Lemon Lane Landfill Dye Tracer Tests: unpublished report, 43 p., as, tables, July.

1995, Field Sampling Plan for the Lemon Lane Groundwater Monitoring Investigation, QAPjP,vol. II, 38p, figures, tables, March 15.

Worthington, Stephen, 1991, Karst Hydrogeology of the Canadian Rocky Mountains: unpublished PhD dissertation submitted to the Department of Geography; McMaster University, Hamilton, Ontario, Canada, 370p.