1.0 Introduction

In 2000, Viacom merged with CBS Corporation, which was fommerly known as Westinghouse Electric Corporation. Throughout this plan Westinghouse Electric Corporation and CBS Corporation will be referred to as Viacom.

Groundwater monitoring is to occur over a five-year period. After the five-year period, a review of all monitoring results will be conducted according to CERCLA Section 121(c) and Section 300.430(fl(4)(ii) of the National Contingency Plan. The review will occur five years after construction began at Neal's Landfill. The remedial action began in April 1999. Interim Groundwater monitoring which was initiated before the remedial action has continued to the present. The interim monitoring will continue until the Long-term Groundwater Monitoring Plan is implemented.

The end of the five-year monitoring period will be April 2004. At that time, the Consent Decree (CD) parties will review the monitoring data and make any agreed upon changes to the Groundwater monitoring plan. Also, during the monitoring period, changes to the plan may be proposed and the plan changed with the consent of all the parties.

1.1 Site Location

1.2 Site History

Between 1958 and 1965, expansion of the landfill occurred in the topographic low areas adjacent to the east-west oriented ridge. During 1966 and 1967, electrical capacitors from the Westinghouse that contained PCBs were deposited at Neal's Landfill. During this time period, both areas adjacent to the ridge were being utilized for disposal. Other material deposited at Neal's Landfill from Westinghouse included capacitor parts, filter aids, and sawdust.

Landfill operations were typical of the time period with the absence of controls. No liner was used. Dumping was not controlled or recorded. Daily burning of the waste was common and scavenging was prevalent.

Investigations at the site began as early as 1976. Extensive site sampling began in the early 1980s (Reference 3). The site was placed on the NPL in 1983. Initial site cleanup activities began in 1983 governed by a set of interim remedial measures detailed in a stipulation and order.

A consent decree was negotiated and signed in 1985. This decree required additional interim measures at the site and detailed final remedial and closure activities including long temm monitoring. The final remedy was later modified in 1999. The following sections provide additional detail on the remedial measures implemented at the site.

1~2.1 Interim Remedial Measures (IRMs)

The second phase of the interim remedial measures, done under the terms of the Consent Decree (CD) (Reference 5), included:

• A stream bank and stream sediment removal project completed in October 1988. Sediments were removed from the entire 4,500 linear feet length of Conard's Branch to Richland Creek, and also from 300 linear feet of Richland Creek.
• Construction and operation of a spring water collection and treatment system. This National Pollutant Discharge Elimination System (NPDES) permitted facility treats base Groundwater flow collected from two springs and a fill-soil seep (North Spring, South Spring, and Southwest Seep). The plant is designed to treat flows up to 1 cubic foot per second (450 gpm) to a discharge level of 1 ppb of PCBs. Figure 2 shows the location of the Spring Treatment Facility on the site. The facility began operation in February 1990.

1.2.2 1999 Remedial Action

The remedy was designed to accomplish the following:

• To remove materials with greater than 500 ppm PCBs to an off site TSCAlandfill.
• To relocate all remaining waste in low-lying areas of the site that could be susceptible to Groundwater backflooding to higher ground.
• To reduce the final footprint area to be capped.
• To cap all remaining waste with a RCRA Subtitle C cover.
• To manage surface runoff and potential run-on into lined stoma water ditches and route this clean water into surface streams. This will minimize Groundwater recharge near the site, reducing spring flows and potentially reducing PCB loading to the springs.

1.2.3 Interim Groundwater Monitoring Program

Mobilization on site to perform the remedial actions occurred during the week of April 12, 1999. Excavation began on April 29, 1999. Interim monitoring has continued approximately every 60 days to present.

The results of the pre- and post-excavation interim Groundwater monitoring program are discussed in Section 1.5. The monitoring reports submitted to the government parties contain more complete data, including field measurements, field log sheets and the laboratory and validation reports for the PCB analytical results.

Spring and residential sample locations for these events are shown on Figure 4. The locations of the monitoring wells that were sampled are shown on Figure 5.

The Branam residence became vacant after the June 22, 1999 sampling and the residential well at that location was not available for further sampling. Monitoring events were performed throughout the excavation and consolidation activities. The on-site monitoring wells were sampled until April 13, 2000. At that time the CD parties agreed to eliminate the monitoring wells from the interim monitoring program.

The SOW also called for the interim Groundwater program to continue on a bi-monthly basis until the long-term Groundwater monitoring program is approved and implemented. However, Viacom has chosen to monitor monthly to enable a better assessment of data trends in the short term.

1.3 Topography and Surface Wafer Features

The major surface water features near the landfill include Richland Creek and two tributaries of Richland Creek flowing from the landfill, Conard's Branch and Southwest Seep Branch. Conard's Branch flows to the north from the northwest corner of the site entering Richland Creek 0.75 miles downstream. Southwest Seep Branch flows southwest along the southern boundary of the site and joins with Richland Creek about 2 miles northwest of the landfill.

During the 1999 remedial action, a system of surface drainage ditches was installed around the landfill. These ditches take rain water from the capped area and from higher ground around the capped area and direct it to Conard's Branch to the north and the Southwest Seep branch to the south.

In addition to surface streams there are a number of springs and seeps. To the north of the site are the South Spring, North Spring, the overflows of South Spring, Frog Seep, Deer Lick Seep and Bedsprings Seep. Figure 2 shows the origins of Conard's Branch being at the Overflow Springs directly upstream of the South Spring. South Spring is the main source of water for Conard's Branch. North Spring feeds into Conard's Branch further downstream, as shown in Figure 2. To the southwest there was the Southwest Seep, and Taylor and Branam Springs.

Karst features, such as sinkholes and swallowholes were identified at the Neal's Landfill site (Reference 7). These karst features were located in and around the fill area by site surveys and historic areal photography review. Some sinkholes were located under areas of fill. These areas were cleaned of all fill in 1999 and the fill was relocated either offsite or to higher ground where no karst features were identified.

South Spring, North Spring, and the overflow springs of South Spring represent related discharge points to one regional Groundwater system. This system's approximate drainage area probably encompasses at least 450 acres at low flow, and may receive stomm water overflow from the adjacent watersheds. South and North Springs are the perennial discharges for this system which receives about 80% of its discharge from the South Spring regional system and the other 20% from infiltrating water in the nearby hillsides. The overflow spring's discharges are ephemeral, and discharge the surfeit that South and North springs cannot carry.

Southwest Seep and the other seeps are local seeps that were developed at the landfill material/natural soil interface or the natural soil/bedrock interface and discharge rain water percolating through the fill material and soil horizons. They have small recharge contribution areas of a few acres or less. Southwest Seep generally only flowed during the wet weather periods of early winter to late spring and was landfill leachateseeping through the fill material and emerging at the fill/natural soil contact. When the landfill material was removed in 1999 there ceased to be a seepage plane at this location.

During low flow conditions, discharges from North and South Springs and (if it were flowing) the Southwest Seep are collected in sumps and sent to the spring treatment facility for treatment. The treated effluent from the facility is discharged to Conard's Branch. During high flow conditions a maximum of about 450 gpm is directed to the STF. The remainder of the discharge from North and South Springs and the overflow springs would overflow the sumps directly to Conard's Branch. Previously the overflow from Southwest Seep went to Southwest Seep Branch. Figure 6 shows the routing of the water from the North Spring, South Spring and Southwest Seep to the spring treatment facility.

1.4 Geology and Hydrogeology Summary
1.4.1 Site Geology

A 0 to 11 foot layer of native clay was between the waste fill and the limestone bedrock surface. The native material consists of a stiff-to-hard, medium-to-low plasticity, fractured silty clay.

The residual silty clay soils are underlain by the Ste. Genevieve Limestone. The Ste. Genevieve Limestone Fommation in this region has features typical of karst terrain, namely numerous sinkholes, swallowholes, solution cavities, caves, sinking or disappearing streams, and springs. The characteristics of the Ste. Genevieve Limestone vary with depth. Three different members make up the Ste. Genevieve Limestone, including (in descending order): the Levias Member, the Spar Mountain Member, and the Fredonia Member. Shale interbeds and chert beds are also present within the Ste. Genevieve. A geologic map of the area was made by an Indiana University graduate student, and is included in Appendix B.

Both the EPA and Viacom installed monitoring wells at Neal's Landfill beginning in 1982. The purpose of the monitoring wells was to collect geologic and Groundwater quality data and measure Groundwater elevations from the limestone bedrock aquifer. In the summer of 1982, seven coreholes, 1A, 2A, 3A, 4A, 5A, 6A and 8A were drilled by the EPA (Reference 8) primarily to collect detailed information on the bedrock stratigraphy. Monitoring wells were subsequently installed in these coreholes. Figure 5 shows the location of all the monitoring wells. Table 1 lists the details of each monitoring well.

Viacom also drilled and installed five open-hole monitoring wells, MW1 to MW5, in the summer of 1982 (Reference 9). These wells are cased in the unconsolidated deposits (soil) and a minimum of five feet in the bedrock, but are uncased in the remaining rock- hole.

Well 10S was installed by the EPA in the fall of 1982 and screened from 749 to 744 feet (amsl) (Reference 10). In addition, Well 11 was installed by the EPA in the spring of 1983 and completed as an open-hole monitoring well (Reference 11). These wells were installed at higher elevations than the EPA NAN series wells, but were still completed in, and thus monitor, the limestone bedrock aquifer.

Well 9A was installed by the EPA in the fall of 1982 to have a monitoring point in a hydraulically downgradient position (as determined by Groundwater elevations obtained in the summer of 1982) at the Neal's Landfill Site (Reference 10). The EPA installed well 1M in January 1983 to replace Well 1A (Reference 12).

These 15 wells have been used to monitor Groundwater elevations and the levels of chemical constituents in the Groundwater in the limestone bedrock aquifer beneath the Neal's Landfill site.

The EPA installed six additional monitoring wells in zones of perched water observed above shale lenses and dolomitic limestone lenses discovered during the coring at Wells 1A through 8A. These are Wells 1 S. 5S, and 8S, installed in the fall of 1982, and Wells 5SS, 2SS, and TOSS, installed in the spring of 1983 (Reference 11).

In November 1983, the United States Geological Survey (USGS) and P. LaMoreaux and Associates (PELA) also installed three monitoring wells, B-8, B-10 and B-15, in the refuse and soil deposits in the areas depicted on Figure 5 (Reference 13).

All monitoring wells mentioned above are located on Figure 5, and all bedrock monitoring well construction details are included in Table 1.

The geologic investigation work conducted at the Neal's Landfill site determined the following:

• Portions of the waste disposed of at the landfill were placed directly on bedrock of the Ste. Genevieve Limestone Formation.
• The limestone bedrock at the site is fractured and has been weathered and dissolved, particularly in the uppermost 10 to 30 feet of bedrock. The portion of the bedrock immediately underlying the soil is significantly weathered and represents an epikarstic zone.
• The limestone bedrock units underlying Neal's Landfill dip overall to the southwest.
• A shallow northwest-southeast trending stratigraphictrough was identified within the Fredonia Member of the Ste. Genevieve Limestone. The stratigraphic trough is coincident with a potentiometric trough within the Fredonia Member aquifer.
• Portions of the Fredonia Member below the water table are weathered and show evidence of solutional enlargement of some fractures, joints, and bedding planes.

1.4.2 Site Hydrogeology

The maximum depth of flow in karst terrain has been shown to be related to the dip of the rock beds and catchment or basin length. The steeper the dip and bigger the basin, the deeper the flow. The dip at Neal's Landfill is less than 1% and the basin is on the order of 400-500 acres. This suggests a shallow development of flow, within the first 200 feet of the valleys.

1.4.2.1 Groundwater Tracer Tests

The results of these tests showed that tracer injected in site monitoring wells located around the perimeter of the landfill primarily resurges at South Spring and North Spring of the Conard Branch system during both high and low flow. Tracer dye injected in wells at the southwest comer of the landfill was detected at Taylor and Branam Springs, indicating a Groundwater drainage divide is present in the southwest portion of the site.

Viacom does not believe that tracer injected in Well EPA-TM was recovered for either the low or high flow tests. Flourescein was injected in well EPA-1AA as well as other wells during the tests. Due to the times that the wells were injected with fluorescein and when break through of the fluourescein was detected, Viacom concluded that the tracer injected in Well EPA-TM was not recovered. This could indicate a drainage divide may be present east of the site. However, it is not known if any landfill leachate crosses over this divide. Viacom personnel were unable to obtain access to the Conard property immediately north of Neal's Landfill during the tracer study. Therefore, any potential Groundwater resurgence present on the Conard property was not monitored during either the high flow or low flow test. This may be why dye at well EPA-TM was not detected.

Figure 7 shows the dye injection points and the results of the high flow tracer test. Figure 8 shows the results of the low flow tracer test.

Dye tracer tests confirmed that the vast majority of groundwater flow beneath the landfill is controlled by a conduit system that eventually discharges to the surface at springs located within valleys to the northwest of the site, at South Spring, North Spring and Conard's Branch. A second component of landfill groundwater flow discharges to the southw

1.4.2.2 Off Site Dye Tracer Tests

Three additional dye tracer tests were performed by an Indiana University graduate student to delineate the drainage basin of South Spring. In July 1993, dye was injected at the Cave Creek sink swallow hole under low flow conditions. Recovery of the dye was at Richland Springs only and not at South Spring. Figure 9 shows the location of the creeks, sinks and springs. In March 1994, dye was injected under moderate flow conditions at Harshman sinkhole and the Taylor-Hollingsworth sinkhole, as shown in Figure 9. The dye from Harshman sink was recovered at Rogers Spring, only. The dye - from the Taylor-Hollingsworth sinkhole was recovered at South and North Spring at Neal's Landfill. In April 1995, a third tracer test was performed under high flow conditions. Two dyes were injected at high flow swallow holes at Cave Creek sink. Recovery of the dye was at Richland Springs only and not at South Spring. However, the dye was introduced at the very end of the storm flow. It may be that earlier components of the storm flow did cross over to the South Spring.

Figure 9 shows that these three tracer tests indicate that the general groundwater flow direction in the vicinity of Neal's Landfill is from the southeast to the northwest.

1.4.3 Groundwater Elevation Data

Groundwater elevation data has been recorded from the monitoring wells since the initial wells were constructed in 1982. Table 2 lists individual monitoring well groundwater elevations measured by hand from 1982 to present. It includes monitoring well groundwater elevation readings taken during the semi-annual site monitoring from 1989 until December 1998. The monitoring well levels taken during the interim groundwater monitoring program are also included. Data was taken bimonthly beginning in April 1999 until April 2000.

Table 3 shows individual groundwater elevation measurements taken during April 1998 in piezometers installed by EPA and Viacom, before the remedial action. The piezometers were located to measure water elevations where saturated waste materials were encountered. Water elevations were measured on April 2 and again on April 16, 1998 after a two-inch rainfall. EPA concluded that the occurrence of perched saturated zones in the landfill was random and that no identifiable saturated zone existed in the landfill (Reference 14). The piezometers listed in Table 3 were removed during the 1999 remedial action.

1.4.3.1 Potentiometric Contours

1.4.3.2 Continuous Groundwater Monitoring

Continuous monitoring was performed in three of the EPA piezometers and in MOO-5A from May to October 1998, as listed in Table 4. Based on this data EPA concluded that groundwater backflooding was apparently not impacting piezometer levels in these areas of the site (Reference 14).

After completion of the remedial action and cap construction, five piezometers were installed through the consolidated waste, as shown in Figure 38, to monitor for the presence of groundwater. Significant water has only been detected in one of the five piezometers, PZ-1, which collected up to five feet of water. The four other piezometers have been equipped with crest gages, which indicate the presence of water between observations, and the highest level the water achieved. Table 4a shows the crest gauge- measurements observed.

Since the beginning of 2000, various monitoring wells, including PZ-1, have been instrumented for continuous water level monitoring. Quarterly reports are issued, listing charts of continuous water level and temperature data. Cumulative rainfall is also charted. Table 4 lists the minimum and maximum water levels recorded in monitoring wells 5A, 10S, 11 and PZ-1, as reported each quarter through 2001.

1.4.3.3 Bedrock and Groundwater Elevation

The results of the tracers tests, as shown in Figures 7 and 8, and the potentiometric surface generated from monitoring well elevations, as shown on Figure 5, indicate that groundwater flows from 10S northwest across the landfill. Therefore the groundwater elevation down gradient from 10S, under the consolidated waste on site, would be less than the elevation at 10S.

The relative elevation of the pre-remediation fill material to the maximum groundwater elevations is discussed in the Neal's Stability Evaluations (Reference 16). Some of the pre-remediation fill in the southeast area of Neal's Landfill was found to be below the 773.3 feet amsl maximum groundwater elevation. During the 1999 remedial action, this material was removed and disposed of offsite or consolidated at a higher elevation under the RCRA subtitle C cap. Figure 42 is a cross-section of the landfill through 10S, 5A and 9A. It shows the maximum groundwater gradient in relation to the fill material. The stability evaluation concludes that the groundwater table has not been observed to be higher than the base of the fill under the final remediated landfill footprint.

Figures 43 shows the location of three additional cross sections of the pre-remediation landfill that were generated by Tetra Tech from data obtained from the March 1998 boring program. The boring locations are shown on Figure 43. Figures 44, 45 and 46 are the three cross sections.. The locations and depths of the borings are shown on the cross sections. These cross sections also show the depth of the fill material before the remediation.

Only the eastern portion of Section A-A' on Figure 44 shows fill material below the 773.3 feet maximum groundwater elevation. Section B-B' does not show any fill material below 784 feet amsl. Section D-D' on Figure 46 shows fill materials in the southeast and northwest areas that were below 773 feet. As can be seen in Figure 10, the material in these areas was removed during the remediation.

1.4.4 Relationship of Well Level 5a and 10s and Conard's Branch Flow

1.5 Historical PCB Data
1.5.1 Groundwater Monitoring Wells

The EPA has recognized this situation. EPA guidance for groundwater monitoring in karst is contained in Groundwater Monitoring in Karst Terrain: Recommended Protocols and Implicit Assumptions" (Reference 17), and "RCRA Ground-Water Monitoring - - Technical Enforcement Guidance Documents (Reference 18). These documents state that the proper locations for monitoring the quality of groundwater in karst areas are springs, cave streams and monitoring wells that have been proven by dye tracing to intercept conduits carrying water from the monitored location.

Groundwater samples from the site monitoring wells have been analyzed for PCBs since the first wells were installed in 1982. Table 5 shows the monitoring well sample PCB analysis data for the site monitoring wells.

Groundwater data for Neal's Landfill was collected from seven monitoring wells biannually from 1989 through 1998. Sample results from monitoring wells EPA-6A, 8A and 9A are typically below detectable levels (bdl). The maximum detected total PCB concentration was 38 ppb from MW4 in November 1993. Monitoring well PCB concentrations detected during the biannual sampling program are presented in Table 5.

As discussed in Section 1.2.4, four monitoring wells were sampled as part of the interim groundwater monitoring program from April 22,1999 until April 13,2000. These monitoring wells are EPA-3A, EPA-5A, MW4 and MOO-5. EPA-3A had PCB levels as high as 36.0 ug/L on October 29,1999. MW4 had its highest value during the interim period on that date at 2.3 ug/L. MOO-5 was measured at up to 0.57 ug/L on April 22, 1999. EPA-5A was measured as high as 2.4 ug/L on December 17,1999. The monitoring well PCB results from the interim monitoring program are also included in Table 5.

1.5.2 Surface Water

As discussed in Section 1.3, there are two spring systems that receive waters potentially impacted by the site. One system is to the northwest of the site and includes South Spring, North Spring and a number of overRows and smaller springs, which feed the headwaters of Conard's Branch upstream of South Spring. All these waters flow into the Conard's Branch stream. As noted in Section 1.4.2.1, dye tracer testing showed this system to be the major exit for site groundwater.

To the southwest of the site there were three springs/seeps, Southwest Seep, Taylor Spring and Branam Spring. Dye tracer testing showed that some site waters from the southwest comer of the site flowed to Taylor and Branam Springs. Southwest Seep received some infiltration through the fill and flowed only after rain events. The drainage basin for this seep was removed as a part of the 1999 remedial action and Groundwater no longer emerges at this location since the remedy was completed.

After the removal of the Southwest Seep, the southwest spring system has been minimally impacted by PCBs from the site, as demonstrated by the interim monitoring data for Branam and Taylor springs, shown on Table 6.

The northwest spring system has been sampled for PCBs extensively both during base flow (or non-storm) and storm conditions. This system is consistently impacted by PCBs. The data for this system will be discussed separately for non-storm and storm conditions.

1.5.2.1 Non-stop Conditions

1.5.2.1.1 Stream Sampling

Surface water samples were obtained by the USEPA and analyzed for total PCBs (Reference 19) in October 1992 and July 1993, after the Interim Remedial Measures were completed. The results listed in Table 7 for 15 stream samples show a large improvement. PCBs were not detected in any of the 15 samples above the detection limit of 1 ug/l.

In May 1998, Viacom collected surface water samples from two locations along Richland Creek and one location along Conard's Branch in conjunction with the sediment sampling event as described in Section 1.5.3. The sampling locations are shown on Figure 15 and the analytical results listed in Table 7. The analytical results for the two samples from Richland Creek indicate that PCB concentrations were below the detection limit of 0.1 ppb. The analytical results for the sample from Conard's Branch indicated a PCB concentration of 0.46 ppb.

In August 2001, the USEPA collected surface water samples in one location in Conard's Branch half a mile from the site, and four locations in Richland Creek from 3 to 34.5 miles downstream of the site, as part of the fish sampling event discussed in Section 1.5.4. The sampling locations are shown on Figure 34aand 34b. The results are listed in Table 7. All four RichlandCreeksamples were non-detect at a detection Nlimit of 0.1 ppb. The Conard's Branch water samples at one half mile from site were less than 60% of the May 1998 sample result which was taken at approximately the same location.

1.5.2.1.2 Spring Emergence Sampling

Five spring emergence water samples were obtained by the USEPA and analyzed for total PCBs (Reference 19) in October 1992 and July 1993, after the IRMs were completed. Again, a major improvement in PCB levels was found. Results are listed in Table 6 for the 5 spring samples. PCBs were reported above the detection limit of 1 ug/l in two of the five samples, (the New Seep and South Spring) where expected, with total PCB concentrations for each sample of 1.3 ug/l.

Table 8 lists all the non-storm data, PCBs and conductivity, for the individual springs of the Northwest Spring system. Figure 16 shows all South and North Spring PCB data plotted vs. time. It appears that both the IRMs and the final remedial action at the site have had the effect of lowering PCBs. It also shows that the North Spring PCBs are usually a fraction of the South Spring PCBs and both follow the same trends.

Surface water data for Neal's Landfill has been collected in association with monitoring requirements for the spring treatment facility. An influent and effluent sample is collected twice each month. The influent consists of a combination of water from the Southwest Seep, South Spring, and North Spring. The sample of the influent to the treatment facility is collected from a pipe within the plant. The treatment facility effluent is sampled upstream of its discharge to Conard's Branch.

Table 9 lists the flow and the PCB concentration measured twice monthly in the STF influent and effluent flows from February 1990 to present. Figure 17 shows a chart of the STF influent flow on the bimonthly sample dates and the PCB concentration in the influent samples from the startup of the plant in 1990 until the present. This data, like the individual spring data, shows that the PCB content in the influent water has trended downward since the remedial action in 1999.

According to the Consent Decree, (Reference 5), the effluent from the treatment facility is to be less than 1 ug/l. The maximum effluent PCB concentration ever measured was 0.85 ppb on October 19, 1994. PCBs have only been detected above 0.2 ug/l on ten occasions, the last on June 6, 2001 when the result was 0.21 ppb. This is also the highest effluent sample result since the remediation in 1999.

1.5.2.1.4 PCB Correlation with Conductivity and Flow Rate

When evaluating contaminant loading at natural karst springs, Viacom has noted that PCB concentrations are normally a function of other spring parameters such as flow and conductivity. For example, at Lemon Lane Landfill the Illinois Central Spring PCB concentrations were found to be higher at very low base flows and lower during wetter periods with higher base flows. Therefore, to properly interpret spring PCB concentrations and detemmine if there are true trends over time, a simple plot of PCBs versus time at a sample location is not sufficient.

Figure 18 shows the relationship between the spring treatment facility influent flow rate and PCB content for the Northwest Spring System. Although the correlation is not strong, it appears that PCBs, under the low flow conditions are a function of flow. The regression curve for this data shows that as the influent flow rate increases, the PCB content tends to decrease. At the lowest end of low flows the PCBs tend to be higher. This is consistent with what Viacom has found at Illinois Central Spring at Lemon Lane.

Figure 19 shows PCBs vs. conductivity at South Spring. PCBs appear to be a function of conductivity with higher PCBs at higher conductivity. This is consistent with the PCB- flow correlation in that conductivity is inversely related to flow, with conductivity typically higher at lower flows and lower at higher flows.

The implications of the noted correlations are that to properly trend PCBs in this spring system, you must also monitor either flow or conductivity or both. The PCB relationships with flow and conductivity cause an apparent seasonal trend. In the dryer seasons of late summer and fall when flows are generally lower, PCBs and conductivities are relatively higher. Typically, in the wetter late winter and spring seasons, when flows are generally higher, PCBs and conductivities are relatively lower.

Figure 20 shows the PCBs at South Spring for the periods just prior to final remedy construction, during construction, and since construction, plotted against conductivity. The PCBs, if in a real descending trend, should show a family of curves that descend with respect to the same conductivity levels when plotted for successive time periods. Inspection of Figure 20 shows that PCBs did appear to elevate during the remedy construction period, and that there does appear to be a descending trend since remedy completion. Additional data will be collected to confirm and monitor the trend.

1.5.2.2 Storm Conditions

Since the inception of the project, Viacom has monitored a number of storms at the Northwest Spring system. Table 10 is a summary showing the number and characteristics of the storms monitored. Appendix A shows summary data for these events.

Viacom has monitored up to three stations during storms. This includes South Spring, North Spring and Conard's Branch. In events where PCBs have been monitored at all three locations, Viacom has observed that North Spring typically has less PCBs than either South Spring or Conard's Branch, and that the PCBs are typically similar between South Spring and Conard's Branch. The Conard's Branch station monitors the combined flow of all the overflow springs plus the storm overflow from South Spring.

During storm events of significant magnitude, the overflows upstream of South Spring will all flow and these overflows will move by far the most water during major events. The combination of high flows and significant PCB levels makes the Conard's Branch location the most logical place to monitor the bulk of the PCB mass moved during rain events. Sampling done by the USGS and EPA shows that all the overflow springs have a similar PCB profile during storms and that this profile matches quite closely the Viacom data collected at the Conard's Branch station by Viacom.

The best sampling coverage for a pre-remedy storm was the April 1998 event. This event represents a typical moderate sized spring stomm event for the Bloomington area. Figure 21 presents a summary of the relevant data taken at Conard's Branch for that event. Because PCB concentrations can be impacted by storm water dilution it is somewhat easier to view the PCB discharge history during an event by examining the cumulative PCB mass discharge history for the event, such as depicted in Figure 22.

By inspection of both figures 21 and 22, it appears that PCBs begin a rapid rise as the flow rises and conductivity drops. Then PCB mass discharge rate stabilizes for much of the flow peak period. There is also a second brief period of high PCB mass discharge on the receding limb. Then the PCB mass discharge rate tails off late on the receding limb. The total PCB mass discharged at the Conard's Branch station during this event is approximately 76 grams.

Since completion of the remedy in 1999, Viacom has monitored seven additional storm events. To make comparisons and draw conclusions with storm data from one storm to another is a difficult task. Viacom's experience has shown that stomms of varying magnitude and intensity will have different PCB discharge levels and that although the PCB response can be generalized, making specific conclusive comparisons is typically complicated by a lack of data from precisely comparable events.

In reviewing the list of storms monitored (Table 10), Viacom believes that the most comparable events to the April 1998 storm are the April 2000, and the two February 2001 storms. This is because the May and June 2000 storm flow responses are too small, the October 2000 event did not have complete sampling coverage, and the November 2000 event was a double peak storm with data for only the second peak.

The April 2000 event was slightly larger than the 1998 event and the two February 2001 events were both smaller. However, these storms had fairly complete PCB sampling coverage and were single peak events where all the overflows contributed significant flow. Based on the data available, these are the most comparable to the April 1998 event.

Figures 23, 24, and 25 show the available relevant data for the three comparable events. Figures 26, 27, and 28 show the PCB cumulative mass for these events. Figure 29 shows PCB cumulative mass for all the comparable storms on a time scale that begins the first flow response during the storms.

In reviewing these figures there are several obvious conclusions. First, the April 2000 event put out a much larger mass of PCBs than the other events. This was caused by a very large PCB loading during the early rising limb period of the storm. Second, the February 2001 storms put out much less PCBs than the April 1998 event or the April 2000 event.

In further examining this data, there is another consistent finding. The receding limb PCBs when viewed on a flow basis appear to have declined. This is best viewed by looking at a plot of PCB vs. flow for all receding limb data for all the larger events (everything but the May and June 2000 events). This is shown in Figure 30. Note that the most recent data for the February 2001 storms appears to be the lowest receding limb PCB data on a flow basis. The impact of the lower PCBs on the receding limb of all post remedy storms is also evident on Figure 29 where the slopes of the PCB mass curves are much lower than the slope of the April 1998 event after 10 to 20 hours.

A possible explanation for the April 2000 event and its relatively high PCB mass discharge is that the remedy at the site had been completed the preceding fall. The low flow PCB data at the Northwest Spring system appeared to experience a high during and immediately after construction as discussed in section 1.5.2.1. It may be that because the interim cap was removed and the system was disturbed during the remedy that the conduit system was loaded up with PCBs and it took a few large storm events to flush out the PCBs. The lower PCBs seen in the February storms may be a more reliable indication of improvement (a lowering of PCBs discharged during storms) resulting from the remedy. This would be in line with the apparent improvement in low flow PCBs being observed in the recent data from the Northwest Springs.

There is a fair amount of uncertainty in comparing these storm events. First, there is uncertainty in the measurements made during the storm. For example, storm flows are monitored at the Conard's Branch culvert by using a water level recorder. A rating curve was made several years ago to convert water depth to flow. This data may be inaccurate at times due to variations in up culvert flow conditions as well as depth sensor calibration. The PCB measurements are also a source of uncertainty. All the storm data was collected at the same locations, by the same methods and analyzed by the same lab. However, Viacom has observed up to a 100% difference in duplicate samples taken during storm events over the years. These differences are attributable to both normal lab variations and actual heterogeneity in conditions over very short periods of time in the stream. There is also uncertainty in the comparability of the storms. None of the storms are exactly the same and even two storms that are nearly the same can have somewhat different responses. There is also some uncertainty over the best way to analyze contaminant loading and storm data from karst systems. There is little literature available on the subject.

The uncertainties can best be addressed by sampling multiple storms over an extended period of time and continuing to improve the quality of the measurements made. By this method enough of a database will be developed to lend more credibility to the conclusions.

1.5.2.2.1 Comparison of Historical Peak Flows Recorded at Neal's Landfill

Storm flow records for 1983-1934 as recorded by USGS, 1993-1994 as recorded by Viacom, and year 2000 as recorded by Viacom are shown in the attached Tables 11, 12, and 13, respectively.

The USGS records were made at a flume device that no longer exists but was not more than 20 feet downstream from where the Viacom flow measurements were taken. The Viacom flow measurements were taken at the Canard road crossing and are based on a rating curve developed for the 3.5 feet CMP oval culvert with concrete headwalls and wingwalls.

The tables show the measured peak flow for a storm event beginning on the date shown. The flows were recorded hourly for the Viacom data and every 15 minutes for the USGS data. The tables also record the total rain that fell before the peak flow was recorded. In some instances, where site hourly rain data was not available, total daily rainfall at Monroe County airport as reported to the NWS was substituted. Pre-flow refers to total South Spring and combined overflow springs flow prior to the beginning of the rain that produced the hydrograph peak for the USGS and 1993-94 Viacom data. The 2000 Viacom data does not include the amount of South Spring that was flowing into the treatment facility. The data was entered into a multiple linear regression computer program (STATISTICA) with measured peak flow as the dependent variable and total rain and pre-storm flow as the independent variables. At the bottom of the tables the regression equation is listed and the peak flows calculated for that equation is listed in the predicted peak flow column. Plots of the measured vs. predicted peak flows are also attached showing the R2 values for each set of data.

Figure 31 shows a plot of the peak flow calculated for each data set based on their respective regression equations and an assumed pre-storm flow of 200 gpm for each '/: inch increment of total rainfall up to 5 inches. The 1983-84 and 1993-94 data plots virtually on top of each other, while the 2000 data shows more than a 50% reduction in peak flows compared to those earlier years.

The reasons for the peak flow reductions are not clear. A preliminary assessment of major changes in the basin over the last 5 years show the following as potential causes:

• A large flood control basin was constructed in the Cave Creek drainage area. While this system has been previously dye tested to detemmine if it was impacting site flows, and the dye test showed it was not, the testing was not all inclusive of the highest flow conditions.
• The changes in surface flow patterns around the landfill as a result of the source control remedy. This included re-routing of about 100 acres of drainage to the surface that could have contributed to spring flow. The surface water controls implemented as part of the remedy were designed to reduce spring flows, but the magnitude of the reduction is larger than expected.
• A significant expansion of the quarrying activity at the nearby Roger's Quarry.

How and if the above changes may have impacted both peak and total flows at the Northwest Spring system will be separately evaluated as part of a Groundwater investigation plan to be conducted by Viacom in 2002/2003.

1.5.3 Sediment Sampling

U.S. Fish and Wildlife Services sampled sediment in Conard's Branch and Richland Creek during a fish sampling event in August of 1982. Sediments were sampled from the Neal's Landfill site to 34 miles downstream of the site in Richland Creek. PCB analytical results of the samples are listed in Table 13a. Samples from Conard's Branch ranged from 37.6 ppm at the site to 2.69 ppm before its confluence with Richland Creek. Sediments in Richland Creek varied from 2.21 ppm at its confluence with Conard's Branch down to 0.31 ppm at 2 miles downstream of the site. At 11 miles and 34 miles from the site in Richland Creek sediments were non-detect for PCBs.

O'Brien and Gere sampled sediments in Conard's Branch, Southwest Seep Branch and Richland Springs Branch per the October 28, 1983 Exhibit A to the Stipulation and Order (Reference 4). Tables 14 and 15 list the sample results. Figure 33 shows the sample locations. Conard's Branch had detectable levels of PCB throughout, with a maximum level of 2535 ppm at sample location 140, at the EPA flume directly downstream of South Seep. The highest levels in the Southwest Seep branch were 20.6 ppm at the emergence of Southwest Seep. At about the point where Taylor Spring enters the Southwest Seep branch, sample location 13, sediment results became non-detect for the remainder of the branch. Branam Spring enters Southwest Seep branch downstream where the sediment was ND.

Based on the 1 983 sample results, Viacom removed the sediments from Conard's Branch and Richland Creek in 1987 and 1988 as part of the IRMs. Removal of sediments in the Southwest Seep branch was not required.

Sediment samples were collected in Richland Creek by Viacom in November of 1988 and analyzed for PCBs. This sampling was performed after the IRM sediment removal was completed. Ten samples were collected at the confluence of Conard's Branch with Richland Creek and ten samples were collected in the vicinity of the SR43 bridge, which spans Richland Creek. PCBs were not detected (~1.0 PPM) in any of the 20 sediment samples.

Sediment samples were collected in April 1991 by the U.S. Fish & Wildlife Service and analyzed for PCBs (Reference 20). Two samples were collected from the upper end of Conard's Branch, one from the overflow basin of the spring treatment facility and one from the midpoint of Conard's Branch. Reported total wet weight PCB concentrations were 13.99 mg/kg and 2.85 mg/kg, respectively. Results of the 1988 and 1991 Neai's Landfill sediment sampling are presented in Table 16.

In October 1992, 25 sediment samples and two duplicate samples were collected from Richland Creek, Conard's Branch, and the springs and seeps adjacent to Neal's Landfill. Split samples were analyzed independently by both the Indiana Department of Environmental Management (IDEM) and Viacom. PCBs were detected in 13 out of 25 samples. The maximum detected concentration reported by IDEM was 12.0 mg/kg for a sample collected adjacent to EPA-37 (Viacom's split sample, analyzed by the same contract laboratory, reported a total PCB concentration of 6.6 mg/kg for the same sample). The maximum detected concentration reported by Viacom was 21 mg/kg for a sample collected from Conards Branch upstream of the spring treatment facility (IDEM reported total PCB concentrations of 1.2 mg/kg and 0.82 mg/kg, respectively for the same sample and a duplicate). PCBs were identified as predominately Aroclor 1248, with minimal amounts of Aroclor 1242 and Aroclor 1260 identified in some samples. Results of the 1992 Neal's Landfill sediment sampling are also presented in Table 16. Locations of the 1988,1991,1992 and 1993 sediment samples are shown in Figure 34.

In May 1998 Viacom collected sediment samples from 26 locations along Conardrs ~ Branch, three locations along Richland Creek, and one location at the Southwest Seep. The sediment sampling locations are shown in Figure 15. PCB content of the samples ranged from <0.1 for two sediment samples to 48 ppm at the eastern most overflow spring, as shown in Table 17.

In August 2001, the USEPA collected sediment samples in one location in Conard's Branch, half a mile from the site, and four locations in Richland Creek, from 3 to 34.5 miles downstream of the site, as part of the fish sampling event discussed in Section 1.5.4. The sampling locations are shown on Figure 34a and 34b. Sample results are shown in Table 13a. Sediments in Conard's Branch at one half mile from the site ranged from 0.10 to 0.26 ppm. Sediments in Richland Creek at 3 miles from the site ranged from <0.05 to 0.11 ppm. All sediment samples further downstream in Richland Creek were non-detect at a detection limit of 0.05 ppm.

1.5.3.1 Year to Year Comparisons in Sediment PCB Levels

Comparing sediment results from 1998 to 2001 shows an additional significant decline in PCB levels. Sediment PCB levels in 1998 in Conard's Branch from 0.5 miles from site to the confluence with Richland Creek ranged from 1.3 to 4.4 ppm. The sediment PCB level in 2001 at 0.5 miles from site ranged from 0.1 to 0.23 ppm for an order of magnitude reduction. The highest PCB level of 0.11 ppm measured in Richland Creek sediments in 2001 was at 3 miles from site. This is much less than PCB levels measured in 1998 at 0.26 to 0.49 ppm in Richland Creek from just downstream of the confluence with Conard's Branch to 5.5 miles from site.

1.5.4 Fish Sampling

1.5.4.1 Factors Affecting Fish Sample Results

a. The distance of the sample location from the source. Fish collected closer to the source should have a higher burden of PCBs.

b. The type of fish. Inter-species differences in food and habitat preferences effect the accumulation of PCBs.

c. The lipid content of the fish sample. Generally, PCBs accumulate in the body fat of fish and the more fat in the fish the higher the equilibrium PCB content to be expected. Usually a specific determination should be made concerning lipid normalization of data prior to making comparisons or evaluating trends.

d. The size (as measured by length or weight) and age of the fish. Generally, larger fish accumulate more PCBs.

e. The type of sample analyzed (i.e., filet or whole body). Generally, there is less PCBs in a filet sample than a whole body sample. Also of concern would be the post collection handling of the fish. For example, was the fish deputized? How were filets prepared?

f. The time of year the fish is collected. Fish annual feeding and breeding cycles can impact contaminant levels.

g. The type of analytical procedures used. For example, packed column Aroclor analysis versus capillary column Aroclor analysis versus capillary column sum of congener analysis.

h. The differences between one lab's techniques and another even when both labs are using the same general procedure. This can be examined by perfomming split sampling/analysis of field dupes. However, in the absence of concurrent field dupes, this can only be estimated.

i. The time difference between sampling events. Fish can retain PCBs in their bodies for extended periods of time even if PCB levels in the stream have dropped. This is because PCBs can have a long half-life in fish. The half-life has been estimated to be up to 6 years for larger fish. The half-life for any particular fish will depend on the species, the size, lipid content, age and other site-specific factors. Therefore, even if PCB levels in the stream have dropped, if the fish samples are collected shortly thereafter (after only a few years), the PCB levels in the fish may not have appreciably changed.

j. Differences in the physical habitat at the sampling location over time. If there have been any major changes in the local habitat at the sampling location, this could impact the fish behavior and contaminant loading.

k.The PCB source strength. Fish receive their loading of PCBs from several sources. These include: the water home dissolved PCBs, PCBs in their food, PCBs in suspended and settled solids (sediments) incidentally ingested.

l. The number of fish samples analyzed. Obviously, to account for normal data scatter and the high number of confounding factors, more data will provide greater confidence.

It is obvious from the above list that to correctly depict trends in the PCB content of fish requires a well designed sampling plan that takes into account all the possible confounding factors. The sampling plan could accommodate these factors by specifying the following:

• Sample consistent fish types and consistent locations
• Collect multiple samples per target species per location
• Consistently specify sample preparation
• Sample the same time of year
• Target fish of the same age/size each time
• Utilize the same lab and same analytical protocols each time
• Allow sufficient time between events to allow fish to equilibrate with new source strengths

1.5.4.2 Fish Sampling Events

Table 18 presents summary data for all known fish sampling events related to the Neal's Landfill site. Figures 34a and 34b present the sample locations.

A chronological list of fish sampling events follows:

• On September 30, 1981, Indiana State Board of Health (ISBH) collected fish samples at 2 locations in Richland Creek, and 30 fish comprising 4 composite samples of different species were collected at 3 miles from the site. PCB levels ranged from 2.3 ppm for 12 shiners to 7.4 ppm for 4 suckers. At 34.5 miles from the site, 6 northern suckers were collected which had a combined PCB content of 0.14 ppm.
• In August of 1982 the US Fish and Wildlife Services (FWS) performed an extensive fish sampling event. Fish were sampled in 10 locations ranging from the site boundary in Conard's Branch to 34 miles from the site in Richland Creek and 89 mostly composite fish samples were analyzed. PCB levels ranged from 279 ppm for 3 darters collected 0.4 miles from the site to <0.01 ppm for four silver redhorse fish collected at 34 miles from the site.
• In May 1987 ISBH collected fish at 2 locations in Richland Creek and 6 creek chubs collected 0.25 miles upstream of the confluence with Conard's Branch contained 0.37 ppm PCBs. Composite samples of 3 different fish species were collected at 3 miles downstream of the site and analyzed at 0.29 to 0.36 ppm PCBs.
• In April 1991 FWS collected 2 composite samples in Richland Creek at 2 miles from the site (Reference 20). PCB levels were 1.25 ppm for 3 creek chubs and 1.95 ppm for 3 white suckers.
• In 1992 US FWS collected 3 fish samples from unknown locations along Richland Creek.
• In September 1993 ISBH collected composite samples of 4 different fish species in Richland Creek at 1 mile from the site. PCB levels ranged from 0.076 ppm for 21 rock bass to 1.9 ppm for 7 white suckers.
• In May 1998 Viacom performed a major fish sampling event. Fish were collected in 3 locations and 15 fish samples were analyzed from 1 location in Conard's Branch 0.5 miles from the site. PCB levels ranged from 1.4 ppm for 1 white sucker filet to 25 ppm for 2 whole creek chubs. The other 2 sample locations were in Richland Creek at 1 mile and 5.5 miles from the site. At the 1 mile location the highest PCB level analyzed from 23 fish samples was 1.5 ppm for 1 whole stoneroller. The highest PCB level for the 20 fish samples from 5.5 miles was 1.9 ppm for 1 large mouth bass filet. Individual samples were as low as 0.03 ppm at 5.5 miles from site.
• In August of 2001, the EPA performed another major fish sampling event (Reference 25). Fish were collected from 6 locations ranging from 0.5 miles from the site in Conard's Branch to 34.5 miles from the site in Richland Creek, and 67 fish samples were analyzed. The majority of the samples were single fish samples. PCB content ranged from 5.9 to 9 ppm at 0.5 miles from the site. All fish sampled at 3 miles downstream and further were < 1 ppm. Al112 samples from 34.5 miles from site were non-detect.

1.5.4.3 Analysis of Fish PCB Data

Second, the data has been analyzed for temporal trends by comparing specific sampling locations and species over time.

1.5.4.3.1 Lipid Content Conrelation

1.5.4.3.2 Comparison of Fish PCBs-Overall Mean of All Fish at Each Location

In reviewing all the past data collected on fish there are three periods where a large number of fish were taken from the drainage system. This is the 1982 data collected by the USFWS, the 1998 data collected by Viacom and the 2001 data collected by the USEPA.

In-between the 1982 and 1998 collections, a number of interim remedial measures were taken, including the installation of the spring treatment facility for treatment of the Northwest spring water and sediment removal from Conard's Branch and Richland Creek. These measures significantly reduced the PCB loading to the stream and the fish.

Between the 1998 and 2001 fish sampling events the remedial action removal and capping was performed at the site. As discussed in Section 1.5.2, water quality became worse during and directly after the remedial actions possibly because the interim cap was removed and the system was disturbed during the remedy so that the conduit system temporarily received more PCBs. It may have taken a few large storm events to flush out the PCBs. Later data now seems to show an improvement in water quality.

Fish PCB levels were compared from location to location as well as from year to year over these remedial actions. Table 19a shows the average PCB content in all fish samples collected, regardless of the species, at each sample location, during each of the three major sampling events. Different fish species as well as both whole fish and filet samples are averaged together at each location.

Table 19a compares fish data at varying distances from 0 to 34.5 miles downstream from the source (Neal's Landfill), as well as comparing data at specific locations from 1982, 1998 and 2001.

The highest average PCB levels were in fish in Conard's Branch, closest to the site, for all three sampling events. As the distance along Richland Creek increases, the average PCB level in fish decreases, somewhat sporadically in 1982 but progressively in 1998 and 2001.

Figure 36b shows a plot of average PCB content in all fish samples collected at each location for the 2001 sampling event. It shows a reduction in overall mean PCB content with distance from the landfill. At sample location 6, furthest downstream, all the individual fish sample results were non-detect.

Table 19a also shows that significant reductions in PCB levels occurred from 1982 to 2001. PCB levels in Conard's Branch in 1982 averaged 133 ppm at 0.4 miles from the site. By 1998, the overall average at 0.5 miles from the site dropped to less than one tenth of that value at 11.85 ppm. By 2001 the overall average further dropped by almost an additional 50% to 6.93 ppm. In Richland Creek, at one mile from the site, directly downstream of the confluence with Conard's Branch, the overall average PCB in fish dropped from 6.73 ppm in 1982 to 1.35 ppm by 2001. In 1982, the overall average PCB was 1.56 ppm at 6 miles from the site. By1998 the PCB level at 5.5 miles dropped to less than a third of the 1982 level to 0.45 ppm.

The only year to year anomaly in this table is the apparent increase in PCB levels at 1 mile from the site from 1998 to 2001. The increase is due to the levels of PCBs obtained in White Suckers during 2001 at this sample location. The 5 whole white sucker samples taken at this location in 2001 ranged from 1.2 to 3.6 ppm and averaged 2.1 ppm PCBs compared to 2 fillets and 1 whole white sucker taken here in 1998, which averaged 0.71 ppm. If the white suckers are deleted from the comparison at this location, all other fish for 1998 average 0.55 ppm compared to 0.6 ppm in 2001. The 0.6 value for 2001 is composed of the average of 8 whole longear sunfish which were the only other fish besides white suckers collected at this location in 2001. The 2 longear sunfish samples from 1998, 1 fillet and 1 whole, averaged 0.66 ppm. The 1998 average is made up of 23 individual samples, including 4 crayfish samples which average only 0.2 ppm.

Even while averaging together different fish species and different types of samples, Table 19a does show that PCB levels in fish do decrease as distance from the site increases. It also shows that PCB levels in fish have decreased over the years since both the interim remedial and source control measures were performed.

1.5.4.3.3 Fish Data Comparison - Specific Species

The above analysis shows substantial declines in-average PCB levels with distance from the site and with time when averaging together all species and sample types collected at each sampling location. While this is a valid method of comparison, averaging together PCB results from different species may add greater variance and uncertainty to the comparisons. If possible, a species specific data comparison should be made. In this section, comparisons are made with both distance and time using PCB results only for a given species of fish at a given location (i.e., creek chubs at 0.5 miles from the site). To further reduce the variance and uncertainty, comparisons should be made only of samples prepared in the same way, (i.e., whole body.) However, because of the limited sample groups available, especially with the 1982 data, samples prepared in different methods, whole body and fillets (edible portions) are averaged together.

The best data sets for specific species comparison analyses are:

• A group of 27 Whole body" samples of longear sunfish, which were obtained in 2001 from sample locations 2, 4, 5, and 6.
• A group of 14 creek chubs, which were collected in 1998 over 3 sample locations.
• Rock bass, creek chub and longear sunfish samples taken at 1 mile from the site in 1982.

The PCB content for the 2001 group of longear sunfish was averaged at each sample location and plotted. The data and plots for longear sunfish are shown in Figure 36a and are typical for the other specific species studied. All these species showed a definite decrease in PCB content with increasing distance from the site.

Comparisons of specific species of fish were also made from the 1982,1998 and 2001 data to determine if specific, comparable data supports a significant decrease in PCB loading in fish over time. The comparison was made for creek chubs, longear sunfish and rock bass between 2001,1998 and 1982. Table 19 presents a comparison of the data in summary tabular form.

Creek chub sample results from the closest sampling location (0.4 to 0.5 miles from the site) in Conard's Branch were compared from 1982 to 1998 to 2001, as shown in Table 19. An 87% reduction occurred from 1982 to 1998 and an additional 56% reduction occurred from 1998 to 2001 at this location. This is similar to the overall average PCB reductions shown in Table 19a. Creek chubs were also compared at sample locations up to 5.5 miles from the site. As shown in Table 19, PCB levels in creek chubs at 1 mile from the site shows a 93% reduction from 1982 to 1998.

Longear sunfish show a 90% reduction in the level of PCBs from 1982 to 2001 at 1 mile from the site. Rock bass also show a 90% reduction in PCBs from 1982 to 1998 at 1 mile from the site.

The conclusion from this analysis is that both the interim and final remedial measures have resulted in significant declines in PCB levels in fish. Additional data should be collected in a couple of years to determine if this downward trend continues. There is also an ongoing risk assessment for this site being done by EPA with input from the other parties to determine risk to human health and ecological receptors.

1.5.5 Residential Wells

In November 1985, the Indiana University School of Public and Environmental Affairs completed a private well user survey around four Bloomington sites for Viacom (Reference 21). Of 134 properties within a 5,000 feet radius of the Neal's Landfill site, at least 47 residential wells were in use. These well locations are depicted on Figure 37.

Sampling of 35 selected private wells was conducted in 1986 and 30 of the wells were non-detect for PCB. The other 5 wells had PCBs at concentrations less than 0.01 ug/l, which is well below the drinking water standard of 0.5 ug/l, as shown in Table 20.

Viacom obtained and reviewed logs of wells within 1,000 feet of the site from the Indiana Department of Water Resources. In general, the well logs indicate that the Ste. Genevieve and the St. Louis Limestone formations are the bedrock units utilized for water supplies.

The Branam and Conard's residential wells were monitored for the interim Groundwater monitoring program during and after the 1999 remedial action. From February 24,1999 to present the Conard's well samples were analyzed at BDL 9 times. The Branam well was measured at BDL twice, until June 22,1999, after which the residence was unoccupied. Table 21 lists data from these sampling events.

In July 1999 the USEPA sponsored the sampling of residential wells surrounding Neal's Landfill for PCBs and VOCs. IDEM and Tetra Tech completed the sampling. Fourteen residential wells and the Anderson spring, located at the Vaughn Martin household, were sampled. All the residential wells and the spring were non-detect for PCBs and VOCs. Table 22 fists the locations that were sampled.

Comparisons of specific species of fish were also made from the 1982,1998 and 2001 data to determine if specific, comparable data supports a significant decrease in PCB loading in fish over time. The comparison was made for creek chubs, longear sunfish and rock bass between 2001, 1998 and 1982. Table 19 presents a comparison of the data in summary tabular form.

Creek chub sample results from the closest sampling location (0.4 to 0.5 miles from the site) in Conard's Branch were compared from 1982 to 1998 to 2001, as shown in Table 19. An 87% reduction occurred from 1982 to 1998 and an additional 56% reduction occurred from 1998 to 2001 at this location. This is similar to the overall average PCB reductions shown in Table 19a. Creek chubs were also compared at sample locations up to 5.5 miles from the site. As shown in Table 19, PCB levels in creek chubs at 1 mile from the site shows a 93% reduction from 1982 to 1998.

Longear sunfish show a 90% reduction in the level of PCBs from 1982 to 2001 at 1 mile from the site. Rock bass also show a 90% reduction in PCBs from 1982 to 1998 at 1 mile from the site.

The conclusion from this analysis is that both the interim and final remedial measures have resulted in significant declines in PCB levels in fish. Additional data should be collected in a couple of years to determine if this downward trend continues.

1.5.5 Residential Wells

In November 1985, the Indiana University School of Public and Environmental Affairs completed a private well user survey around four Bloomington sites for Viacom (Reference 21). Of 134 properties within a 5,000 feet radius of the Neal's Landfill site, at least 47 residential wells were in use. These well locations are depicted on Figure 37.

Sampling of 35 selected private wells was conducted in 1986 and 30 of the wells were non-detect for PCB. The other 5 wells had PCBs at concentrations less than 0.01 ug/l, which is well below the drinking water standard of 0.5 ug/l, as shown in Table 20.

Viacom obtained and reviewed logs of wells within 1,000 feet of the site from the Indiana Department of Water Resources. In general, the well logs indicate that the Ste. Genevieve and the St. Louis Limestone formations are the bedrock units utilized for water supplies.

The Branam and Conard's residential wells were monitored for the interim Groundwater monitoring program during and after the 1999 remedial action. From February 24, 1999 to present the Conard's well samples were analyzed at BDL 9 times. The Branam well was measured at BDL twice, until June 22,1999, after which the residence was unoccupied. Table 21 lists data from these sampling events.

In July 1999 the USEPA sponsored the sampling of residential wells surrounding Neal's Landfill for PCBs and VOCs. IDEM and Tetra Tech completed the sampling. Fourteen residential wells and the Anderson spring, located at the Vaughn Martin household, were sampled. All the residential wells and the spring were non-detect for PCBs and VOCs. Table 22 lists the locations that were sampled.

As detailed above, contaminated fill and soil with PCB content above 500 ppm were removed from the site. The remaining waste has been consolidated at an elevation that should not be wetted by groundwater and covered with a RCRA subtitle C cap.

Five new piezometers (PZ1 -PZ5) were installed through the clay barrier layer of the landfill cap at the locations shown on Figure 38. All new piezometers were drilled to bedrock.

As indicated above, during both normal flow conditions and rainfall events, discharges from North and South Springs are treated at the Viacom Neal's Landfill Spring Treatment Facility up to a rate of 450 gpm. The treated effluent from the facility is discharged to Conard's Branch.

PCB concentrations in the spring flows and the fish in the associated streams have decreased over time. Flows in the Northwest Spring system are also reduced, apparently due to changes in the basin.

Viacom has plans to establish the appropriate deed restriction for the site. Inspection, maintenance and repairs of the cap are performed as required based on a quarterly long-temm inspection program (Reference 24).

2.0 Long-term Monitoring Approach

The consent decree contained provisions and requirements for the long-term groundwater monitoring at the Neal's Landfill site. The provisions were based on a model of porous media flow and required a suite of onsite and offsite wells. This model is not appropriate for the karst geology at this site. Therefore, the original groundwater- monitoring provisions were deleted from the SOW requirements agreed to by the parties for the removal action.

The SOW for the site required the implementation of an interim groundwater monitoring program as discussed in section 1.2.3. The SOW also required the submittal of a long- temm groundwater monitoring plan. Viacom has the responsibility of preparing the long- term monitoring plan.

A review of the site history, hydrogeology and historical PCB data shows that

• The remediation of the site in 1999 was substantial. Materials greater than 500 ppm PCBs were removed. The vast majority of the remaining materials with PCB content <500 ppm were consolidated under a RCRA subtitle C cap at an elevation above the highest groundwater level measured on site.
• All waste materials that were potentially subjected to backflooding were also removed and placed on higher ground under the final cap (even those materials not containing PCBs).
• High and low flow tracer tests show that the direction of groundwater flow under the site is from southeast to northwest. Tracer tests also show some of the groundwater from the southwest portion of the site goes to springs to the southwest.
• The site is located in karst.As such, USEPA guidance suggests the most accurate indicator of Groundwater conditions will be springs or stream water samples. The existing wells are of highly questionable utility in determining Groundwater quality. Figure 39 and 40 compare PCB sample results from the South Spring to the PCB results from the monitoring wells during the Interim Monitoring Program. PCB levels in the wells do not follow the trend of PCB levels in the South Spring.
• The southwest spring system consisted of the Branam and Taylor Springs and the Southwest Seep. The Southwest Seep has dried up after remedy construction because its drainage area was excavated and removed. As shown on Table 6, PCB data from Branam and Taylor springs shows lime to no PCB impact. Only six samples out of 79 are above detection, with the highest results at 0.2 ug/L for Taylor Spring in July and September of 2001. All 14 sample results since then have been BDL, except one for Taylor Spring, which was at the detection limit of 0.1 ug/L.
• The most recent PCB data for the Northwest Spring System shows that the PCBs are a function of flow and conductivity during low flow and a function of time with respect to storm progress during storm events.
• Based on the low flow data for the Northwest Spring System, the PCB levels increased during the remedy construction and for a few months after construction. Based on the most recent data at low flows, the PCBs in the Northwest Spring System have since trended down and are now at or below pre-remedy construction levels.
• Levels of PCBs discharged via the Northwest Spring System during storms appear to have increased immediately after remedy construction but now also appear to be declining.
• PCB levels in fish in the streams fed by the Northwest Spring System have been measured over the years and as recently as 2001. Based on a comparison with fish samples taken prior to any remedial action at the site, the PCB levels have dropped substantially, and continue to trend down. This is to be expected based on the removal of sediments, water treatment and source control actions taken.
• Stream sediments were removed in 1987 and remain substantially below the levels found pre-removal.
• Peak storm flows at the Northwest Spring System appear to be much lower relative to those measured prior to 1998. This is probably due to the surface water control portions of the remedy and other basin wide changes. Based on the above, the goals of long term monitoring should be:
• To continue to evaluate trends of PCB levels in the waters of the Northwest and the Southwest Spring Systems.
• To evaluate the continued effectiveness of the cap system to prevent wetting of fill material.
• To continue to monitor fish in Conard's Branch and Richland Creek.

2.1 Northwest Spring System Monitoring

In order to estimate flows, a weir has been installed in Conard's Branch just downstream of the culvert under the road near the spring treatment facility. During non-storm sampling events, data from this weir will be combined with spring treatment facility flows to provide a total estimate of Northwest Spring System flow.

The North Spring overflow, which only flows during higher storm conditions, will not be included in this flow measurement. The North Spring overflow may be instrumented, temporarily, to obtain flow data for some specific higher storm flow events.

The trend in PCB levels from these springs will be evaluated on a flow and conductivity basis during non-storm conditions.

A minimum of two large storm events per year will also be monitored at the Conard's Branch culvert during 2002 and 2003. At least one event will be monitored in the spring of 2004, near the end of the five year review period. Parameters measured will be conductivity, flow, TSS and PCBs at 30 minute intervals during the rising limb of the storm to the peak flow, if possible. The criteria for the size of the stomm are that all overflows are flowing and the peak flow measurement at the weir immediately downstream of the culvert is at least 4000 gpm.

2.2 Southwest Spring System Monitoring

Branam and Taylor springs of the Southwest Spring System will be sampled twice annually until the end of the five year review period. Efforts will be made to sample these intermittent springs during times when they would be flowing, such as after significant rainfalls or during wet periods. Samples will be analyzed for total PCBs, TSS and conductivity.

2.3 Piezometer Monitoring

Recent monitoring of the piezometers shows that only PZ-1 has shown the presence of significant water. As part of the long-term monitoring plans datalogging transducers will be maintained in PZ-1 to collect continuous data.

Piezometers that have been essentially dry have had crest gauges installed. These are checked every two weeks to determine if water has entered the wells. If significant water develops in any other piezometer, then a continuous water level recorder will be added to that piezometer.

2.4 Well Monitoring

During the pre-remedial semi-annual monitoring from 1989 to 1998, water levels were taken at 12 site wells, open to or screened in the Fredonia Member. Each set of water levels was used to construct a potentiometric map. Twenty maps were generated that define the site water levels through a variety of hydrologic conditions. They will serve as a historical benchmark against which any major impacts, such as the limestone quarrying in Rogers Quarry should be apparent. These wells are:

All these wells still exist, except EPA4A. Viacom plans on taking water levels of the eleven remaining wells, as well as the Conard residential well, on a quarterly schedule. Each set of data will be used to construct the potentiometric map of the site. These future maps will be most comparable to the historic maps.

2.5 Fish / Sediment Sampling

Sediment and fish will be sampled in Conard's Branch and Richland Creek near the end of the review period.

3.0 Dab Analysis and Reporting

Long-term monitoring at Neal's Landfill will be done until the end of the five year monitoring period. The sample points will be sampled until April of 2004. At the end or the five year period, the CD parties will evaluate all the monitoring data and decide if monitoring will cease, continue as is, or continue in a modified form. Also, during the base monitoring period, changes to the plan may be proposed and the plan changed with the consent of all the parties.

4.0 Sample Collection

For springs at non-storm conditions, an unfiltered grab sample will be taken by hand dipping a 1 liter bottle into the main exit point of the spring orifice or center of the stream as applicable. The samples will be sent for total PCB and TSS analysis. The PCB analysis will be to a detection level of 0.1 ppb to an approved lab for this project whose procedures will be in accordance with the requirements of Test Methods for Evaluation of Sold Waste: PhysicaUChemical Method" (EPA SW-846, latest edition) analytical method 8082. TSS samples will also be sent to an approved lab and analyzed in accordance with procedures that meet the requirements of EPA 160.1 from EPA 600/4-79-030 latest edition.

During each sample cycle, a duplicate sample and field blank will also be taken. The duplicate will be a second sample of water taken from one of the sample points. The field blank will be an identical sample bottle filled with Dl water at the sitewhile taking the samples. During non-storm conditions field parameters of conductivity and temperature will also be measured.

During storm conditions, the Conard's Branch samples will be taken with an ISCO auto sampler at the upstream side of the Conard's Branch culvert. Each sample will be approximately 500 ml and taken into a glass container in the base of the sampler on a half hour basis until after the peak flow and then the frequency will be increased as determined by the sampling crew. When auto samplers are used to collect storm samples, the sampler will be iced. The maximum amount of time that a sample can be left in the sampler before it is removed, sealed and refrigerated will be limited to 12 to 16 hours.

The storm sample PCB analysis will be to a detection level of 0.3 ppb at an approved lab for this project whose procedures will be in accordance with the requirements of Test Methods for Evaluation of Solid Waste: Physical/Chemical Methodn (EPA SV~846, latest edition) analytical method 8082. TSS samples will also be sent to an approved lab and analyzed in accordance with procedures that meet the requirements of EPA 160.1 from EPA 600/~7~030 latest edition. Each storm sample will also be measured for conductivity by hand in the field before sealing for shipment to the lab.

If sample results are obtained below the detection limit (BDL) at the proposed detection limit of 0.3 ppb for storm samples, two 500 ml sample bottles, instead of one, will be filled sequentially at each sample time. The larger sample quantity will allow lower detection limits to be reached.

5.0 Sample Custody Procedures
5.1 Sample Identification System

A sequential sample numbering system will be used to identify each sample, including duplicates and blanks. Each sample will be assigned a unique sample number. The field activity leader will maintain a listing of sample identification numbers in the field logbook.

Each sample number will consist of six digits as illustrated by the following example:
NL0001.

The AL" is the site code and refers to Neal's Landfill. The four digits are the sequential number. The sample number will be added to the respective field notebook, sample label, and chain-of-custody form.

5.2 Initiation of Field Custody Procedures

For all samples, Region V chain-of-custody protocols, as described in the National Enforcement Investioabons Center (NEIC) Policies and Procedures, EPA-330/9-DDI-R, Rev. June 1985, will be followed. Custody procedures are described in Section 5.0 of the QAPj P.

5.3 Field Activity Documentation and Logbook

A field logbook, as discussed in FP-1 of the QAPjP, will be initiated at the start of the Field Sampling Program and maintained to record on site activities. The field logbook is a controlled document that becomes part of the permanent site file. The field logbook will consist of a bound notebook with consecutively numbered pages that cannot be removed. The logbook cover will indicate the following:

• Project name
• Project Geologist's and Field Activities Leader's name Sequential book number
• Project start date
• Project end date

• Summaries of daily site activities
• Arrival and departure of site visitors
• Arrival and departure of equipment
• Start and completion of sampling activities
• Sample pickup including chain-of-custody fomm number, carrier, date, and time
• Equipment calibration and repair
• Decontamination procedures used
• Health and safety issues
• Levels of personal protection

Refer to Section 5.1.2 and FP-1 of the QAPjP (Reference 22) on Field Logbook Record Keeping.

5.4 Sample Shipment and Transfer of Custody

6.0 Sample Container Preparation, Sample Preservation, and Maximum Holding Time
6.1 Bottle Requirements

In addition, the data for field blanks, etc., will be monitored for contamination per Section 3.1 of the QAPjP. Corrective actions will be taken as soon as a problem is identified and include discontinuing the use of a specific bottle lot, contacting the botth supplier(s) for retesting the representative bottle from a suspect lot, resampling the suspected samples, and validating the data, taking into account that the contaminants could be introduced by the laboratory (i.e., common lab solvents, sample handling artifacts, etc.). If a bottle QC problem occurs, a determination of whether the bottles and data are still usable will be made.

Amber glass bottles with tenon liners will be used for PCB water samples.

6.2 Sample Preservation and Holding Time

6.3 Sample Handling, Packaging, and Shipment

7.0 Decontamination Procedures

This section provides the general guidelines for the decontamination of sampling and monitoring equipment and sample bottles. FP-2 of the QAPjP discusses decontamination procedures.

The following equipment will be on site:

• Distilled water
• 10 percent by volume isopropanol and water solution
• Non-phosphate detergent
• Scrub brushes; squirt bottles for alcohol and water; plastic bags and plastic sheets
• Drums or carboys for disposal of waste

7.1 Sampling Equipment Decontamination

All sampling equipment and monitoring equipment (e.g. temperature and conductance probes) will be decontaminated between sampling locations by the following procedures:

1. Wash contaminated equipment contact surfaces with nonphosphate detergent.
2. Rinse with tap water.
3. Spray rinse with 10 percent alcohol solution.
4. Rinse with distilled water and air dry.

7.2 Sample Bottle Decontamination

Sample bottles or containers filled in the field will be decontaminated before being packed for shipment or handled by personnel without derrnal hand protection as follows:

8.0 Preventive Maintenance Procedure and Schedule

9.0 Investigation-Derived Waste

9.1 Types of Investigation-Derived Waste

• Personnel protective equipment (PPE). This includes disposable coveralls, gloves, booties, respirator canisters, etc. It is expected that normal work clothes will be used by samplers with disposable gloves and booties where appropriate.
• Disposable equipment (DE). This includes plastic ground and equipment covers, aluminum foil, Teflon tubing, broken or unused sample containers, sample container boxes, tape, etc.
• Groundwater obtained through well development or well purging.
• Cleaning fluids such as spent solvent and wash water.

9.2 Management of Investigation-Derived Waste

10.0 Well Abandonment

REFERENCES

1. Statement of Work (SOW) for the Source Control Remedial Design and Remedial Action at Neal's Landfill, Monroe County, Indiana, March 5, 1999

2. Remedial Design / Remedial Action Work Plan, Neal's Landfill, Monroe County, Indiana, CBS Corp., April 23, 1999

3. Inspection at Neal's Landfill in Monroe County near Bloomington, IN, USEPA December2,1980.

4. Stipulation and Order, Civic Action IP 83-9-C, U.S. District Court for the Southern District of Indiana, January 4, 1983.

5. Consent Decree, Civil Action No. IP 83-9-C, USA and the State of Indiana v. Westinghouse Electric Corporation and Monsanto Company, and Civil Action No. IP 81~4~C, The City of Bloomington and Monroe County, Indiana v. Westinghouse Electric Corporation and Monsanto Co., entered August 1985.

6. Final Report- Remediation of Neal's Landfill, Bloomington, Indiana, CBS Corp, December 2000.

7. Powell, Richard L., Geosciences Research Associates, Inc., "Geology and Hydrology of Neal's Landfill, Monroe County, Indiana", August 8, 1983.

8. Ecology and Environment, Inc., "Field Investigations of Uncontrolled Hazardous Waste Sites, FIT Project, Interim Report of the Neal's Landfill Facility near Bloomington, Indiana," November 1982, for USEPA

9. O'Brien & Gere Engineers, "Neal's Dump and Neal's Landfill Investigation," December 1982, for Jones, Day, Reavis and Pogue.

10. Memorandum to File from R. Bloese "Drilling at Neal's Landfill", Ecology and Environment, Inc., November 16, 1982.

11. Bloese, R., Ecology and Environment, Inc., "Indiana/TDD No.R5-8109-3E, Bloomington/Neal's Landfill," Memorandum to File, July 7, 1983. (Wells 2SS, 11, 5SS, 10SS drilling info).

12. Bloese, R., Ecology and Environment, Inc., "Indiana/TDD#5 8109-3E, Bloomington Neal's Landfill," Memorandum to File, January 31, 1983. (Replacement well 1A - drilling info).

13. Lithologic Logs, Slug Test data and field notes; Lee Watson and J.M. Tanner, U.S.G.S. and PELA, November 1983.

14. Tetra Tech EM Inc., Revised Current Status Report for Groundwater, Surface Water, Sediment, and Fish Data, Neal's Landfill Site, March 17, 2000, for USEPA

15. Onsite Groundwater Monitoring Plan, Neal's Landfill, Blasland & Bouck Engineers P.C., August 1987.

16. CH2M HILL, "Neal's Landfill Stability Evaluation", prepared for CBS Corp., March 1999.

17. Ground-Water Monitoring in Karst Terranes: Recommended Protocols and Implicit Assumptions, EPA 600/X-891050, Feb 1989

18. RCRA Ground-Water Monitoring Technical Enforcement Guidance Document, OSWER-9950. 1 November 1992

19. Hopkins, D., USEPA, ~Conard's Branch Sampling - USEPA Water Analysis Results," fax to J. P. Patrick, Westinghouse Electric Corporation, January 7, 1993.

20. Sparks, D. W., U.S. Fish & Wildlife Service, PCB's in Richland Creek Downstream from Neal's Landfill, April 1991," November 1991

21. Indiana University School of Public and Environmental Affairs, Well Water User Survey around Four PCB Contaminated Sites," ESAC-85-03, December 1985, for Westinghouse Electric Corporation.

22. Bloomington Project Quality Assurance Project Plan (QAPjP), Revision 4, Viacom, September 23, 1998

23. Blasland & Bouck, Engineers, P.C., "On-Site Ground-Water Monitoring Plan, Neal's Landfill, 2 Volumes," August 1987, for Westinghouse Electric Corporation.

24. Viacom, Inc., FULCRA Cap Inspection and Maintenance Plan for Neal's Landfill, Bloomington, Indiana, March 2001.

25. USEPA, UFinal Report Neal's Landfill Fish Assessment", USEPA ERTC, USFWS, Lockheed Martin REAC, February 2002

26. Hebert, C.E. and Keenleyside, K.A., Environment Canada, "To Normalize or Not to Normalize? Fat is the Question", February 1994.