The Baytown Township Groundwater Contamination site (the site) was first discovered in 1987; since that time, investigation and response actions by a number of state and local entities have been ongoing. The site is the result of the disposal or spillage of a large quantity of trichloroethylene (a.k.a. trichloroethene, or TCE) at an unknown source or sources in the vicinity of the Lake Elmo Airport, and the use or spillage of carbon tetrachloride (CCl 4) at a nearby former grain storage facility. The site consists of an area of groundwater contamination that is in excess of six square miles, and affects four major groundwater aquifers.
Several hundred private water supply wells, and one of three existing municipal water supply wells in the city of Bayport have been impacted by the TCE contamination. The highest current concentration of TCE in a private well (approximately 100 micrograms per liter (µg/L)) has been found at the Lake Elmo Airport.
In early 2002, the Minnesota Department of Health (MDH) issued a new interim recommended exposure limit of 5µg/L for TCE in drinking water from private wells. This precipitated a series of response actions and new investigations relative to the groundwater contamination. Further response actions by the Minnesota Pollution Control Agency (MPCA) are under consideration.
Exposure to TCE above health-based criteria is currently being prevented by recommending property owners have new private water supply wells constructed to deeper, clean aquifers where possible, and by providing granular activated carbon (GAC) filtration units for existing private wells and on new private wells where a clean aquifer is not available or feasible and concentrations of TCE exceed health-based criteria. However, the Metropolitan Airports Commission (MAC), and in most instances, the MPCA, have indicated that for homes platted after April 9, 2002, GAC filters will not be provided. Baytown Township, with the assistance of MDH, enacted township ordinance number 36 in part to ensure that newer homes not provided GAC filters by the MAC or MPCA will have appropriate GAC systems installed and maintained by the homeowners. West Lakeland Township has adopted a similar ordinance.
To date, TCE concentrations in the one impacted municipal well and the Bayport municipal water supply system have not exceeded regulatory or health-based criteria. Although exposures have exceeded the current health-based criterion for TCE, no evidence exists of an unusual incidence of adverse human health effects as a result of exposure to TCE in groundwater at the site.
Since the date of the previous Health Consultation for the Baytown Township Groundwater Contamination Site (MDH 1999), MDH has established an interim recommended exposure limit for TCE that is substantially lower than the previous standard. As a result, a significant number of additional private water supply wells (over 125) exceeded the new interim recommended exposure limit. Additional site investigation, water testing, and installation of whole-house GAC filter systems, and expansion of the Special Well Construction Area (SWCA; see "Background" section below) have occurred in response to the new, lower exposure limit.
The MPCA, local governments, and area citizens have requested that MDH staff review the results of water well sampling at the site to determine whether levels of volatile organic compounds (VOCs), primarily TCE in groundwater pose an unacceptable health risk to area residents. MDH staff also have participated in several community meetings to discuss the site, have coordinated several mass mailings, and have participated in other meetings with state, county, city, and townships officials and consultants to discuss long-term options for a safe water supply.
The purpose of this document is to summarize the most recent information about the site, primarily the activities that have taken place since the 1999 MDH Health Consultation, and to provide information to the community regarding exposure to TCE and its potential health risks.
Site Description and History
Much of the information in this section has been presented in previous MDH documents (MDH 1996, MDH 1999), but is included for completeness. The groundwater contamination at the site was initially discovered in June 1987, when MDH sampled ten private wells in the vicinity of the Bayport (a.k.a. Stillwater Prison) Dump under a program designed to assess groundwater quality near old dumpsites in the Twin Cities area. One VOC was detected in one well at that time. As a result, additional groundwater sampling was initiated by the Washington County Department of Public Health and Environment, the MPCA, and MDH. Those agencies oversaw the sampling of monitoring wells, community water supply wells, and residential drinking water supply wells in the area of the site.
These investigations revealed a sizeable area of VOC contamination in groundwater. The contaminants detected consisted of TCE, CCl4, or both. Historically, TCE was found in more wells and at higher concentrations than CCl 4. Because of the elevated levels of VOCs, the MDH first issued a well drilling advisory in 1988, which was modified in 2001 to encompass the approximately nine-square-mile area shown in Figure 1. The advisory, which is now called a Special Well Construction Area (SWCA) requires:
- Well contractors and well owners to submit well construction plans and to receive written approval from MDH prior to constructing, reconstructing, or sealing wells in the Baytown Special Well Construction Area; and
- Well owners to test water from new private water supply wells for VOCs, prior to hookup, to verify that the groundwater meets MDH health-based criteria.
The MPCA listed the site on Minnesota's Permanent List of Priorities (PLP), the state Superfund site list, in 1988. The site was added to the U.S. Environmental Protection Agency's (EPA's) National Priorities List (NPL), the federal Superfund list, in January 1995.
The MPCA originally identified the Lake Elmo Airport as a suspected source for the TCE contamination because of the known use of TCE as a degreaser and parts cleaner at other airports. The shape of the plume, and the fact that the highest concentrations of TCE in groundwater were found in private wells and monitoring wells at the airport were also contributing factors. In May 1988, the MPCA issued a Request for Information to the Metropolitan Airports Commission (MAC). The MAC owns and operates the Lake Elmo Airport, which is located near the western end of the Baytown SWCA. Subsequent investigations, conducted by the MPCA in 2003, suggest that similar TCE concentrations are present in an area west of the airport, which may indicate the presence of a source area other than, or in addition to, the airport.
The MPCA identified a former grain storage facility northeast of the Lake Elmo Airport as the suspected source of the CCl4 contamination, because CCl4 had been widely used as a grain fumigant. The locations of the Lake Elmo Airport, the former grain storage facility, and other site features are shown in Figure 2.
The MAC is a public agency charged with managing public airports in the Twin Cities area, and it opened the Lake Elmo Airport in 1951. MAC maintains the small airfield and leases space to tenants who store and operate general aviation type aircraft. By 1972, the airport had approximately 35 hangars in what is known as the main (or south) hangar area and accommodated small, recreational airplanes, which require little maintenance beyond routine oil changes. The airport also included several maintenance facilities known as fixed-base operators. The north hangar area was not built until the late 1970s. The current airport has approximately 132 hangars and associated facilities, two runways under 4,000 feet long, and is designed to accommodate primarily private and recreational flyers operating single- and twin-engine propeller aircraft. The period of 1950-1970 is important, as this is the most likely time that the release of TCE occurred, as discussed on page 12.
In 1988, the MAC's consultant (Wenck Associates Inc. [Wenck]) began an investigation of the Lake Elmo Airport and the immediate surrounding area. The investigators attempted to characterize the site and the TCE contamination, as well as identify potential sources of TCE in the area. Wenck submitted an "Interim Evaluation Report" to the MPCA in September 1988, concluding that there was no residual TCE contamination in the surficial soil or glacial drift (shallow) groundwater in the immediate vicinity of the Lake Elmo Airport. The report further concluded, based on hangar inspections and tenant interviews, that there was no evidence of use or disposal of TCE at the airport.
In April 1994, the MDH completed a Public Health Assessment, concluding that the site posed a public health hazard because, 1) groundwater quality data showed a limited number of well users may have been exposed to levels of TCE and/or CCl4 that were above the existing MDH Health Risk Limits (HRLs) for groundwater at that time (30 µg/L for TCE and 3 µg/L for CCl4); 2) other well users in the site area were likely still being exposed to VOCs at variable concentrations; and 3) the source(s) of both the CCl4 and TCE contamination, which were not yet known, were likely still present.
Geology / Hydrogeology
The geology of the region in which the site is located consists of glacial drift (stratified sand, silt, and clay deposited by glaciers) overlying a thick sequence of Paleozoic sedimentary rock formations made up of sandstone, limestone, dolomite, and shale. These, in turn, overlay pre-Cambrian volcanic rock formations composed primarily of basalt. The bedrock formations tilt and thicken slightly to the west, forming the eastern rim of a large geologic structure known as the Twin Cities Basin. A generalized view of the geology and hydrogeology of the site is shown in Figure 3.
On the eastern edge of the site, the bedrock formations are bent into an upward convex fold, known as the Hudson-Afton anticline, which has been deeply eroded to form the St. Croix River valley, as shown in Figure 4. Recent work by the Minnesota Geologic Survey (MGS, 2004) indicates that a large fault trends northeast-southwest near the western shore of the St. Croix River (Figure 5). Within the river valley, the upper bedrock formations (Prairie du Chien and Jordan) have been removed by erosion and buried by valley fill sediments (Mossler and Bloomgren 1990).
Groundwater flow at the site is through several different geologic units and is generally from west to east, towards the St. Croix River, where it discharges. The upper-most saturated layer is glacial drift at the western portion of the site, the Prairie du Chien dolomite through the central portion of the site, and Quaternary sands and gravel at the eastern end of the site, within the river valley. Layers of sediment or rock that are saturated with water and sufficiently permeable to transmit water to wells or springs are referred to as aquifers.
The glacial drift fills previously eroded bedrock valleys that can channel or change the direction of groundwater flow. A large bedrock valley is located on the western edge of the site, and may be affecting groundwater flow in that area (Figure 6).
Groundwater flow in the Prairie du Chien dolomite is heavily influenced by fractures (cracks and voids) in the formation. The Prairie du Chien is actually considered a "group" composed of two separate dolomite formations, the Shakopee and Oneota. For general purposes, this report will consider the Prairie du Chien Group as a single unit. However, it is useful to note that although the lower Oneota formation tends to be more massive than the sandier Shakopee formation, and the Oneota tends to have more solution cavities. According to the hydrologic atlas for this area (Lindholm, et. al., 1974), the Oneota provides the higher yield of water to wells because of this characteristic.
Below the Prairie du Chien is the Jordan sandstone. The Jordan sandstone is a medium to coarse-grained sandstone and is widely used in the region for residential and municipal water supply wells. Where conditions exist for groundwater to move downward rather than horizontally, and where the Oneota formation at the base of the Prairie du Chien Group has solution cavities and/or fractures, groundwater can readily move from the Prairie du Chien into the Jordan, because there is no low-permeability rock layer (or confining unit) separating them. However, these conditions do not appear to be present across the site, so groundwater does not appear to flow freely between the Prairie du Chien and Jordan throughout the study area. This may account, in part, for the irregular shape of the contaminant plume in the Jordan.
Underlying the Jordan sandstone is the St. Lawrence formation, composed of dolomite and siltstone. This formation is not considered an aquifer but rather a confining unit because it has low vertical permeability to groundwater. Below the St. Lawrence formation, in descending order, is the Franconia-Ironton-Galesville sandstone aquifer, the Eau Claire confining unit, and the Mount Simon sandstone aquifer. Because the majority of wells in the area are completed in the Prairie du Chien and Jordan aquifers, the extent and characteristics of the St. Lawrence and Franconia (and lower) formations are not well understood in Baytown Township and Bayport. At some places near the St. Croix River bluffs, portions of the St. Lawrence may be compromised by faults or fractures associated with the Hudson-Afton anticline.
The MGS, in cooperation with Washington County, MPCA and MDH, evaluated the bedrock geology and faulting in the SWCA area and elsewhere in southern Washington County (MGS, 2004). They identified a large northeast-southwest trending fault (Figure 5). In the area where the plume approaches the St. Croix River, the fault is located beneath the western shoreline of the river, and is unlikely to have any significant influence on the plume migration. However, smaller faults and/or fractures associated with the anticline may be present at a scale too small to have been detected with the available data.
Groundwater flow in the unconsolidated sediments, Prairie du Chien group, and Jordan sandstone is to the east; groundwater discharges to the St. Croix River at Lake St. Croix (Wenck 1999a). At the western end of the site, groundwater is found in the unconsolidated sediments at depths ranging from 35 to 65 feet below ground. To the east of the airport, groundwater is first encountered in the Prairie du Chien at a depth of approximately 100 feet below ground. At the far eastern end of the site, the groundwater surface follows the overlying topography as it drops towards the St. Croix River. The flow characteristics of the different aquifers are very different. Groundwater flow in the Prairie du Chien is primarily through fractures (cracks and voids). Groundwater flow in the unconsolidated sediments and Jordan sandstone is mainly through the pore spaces between the sand particles that make up the formations.
Vertical flow between the layers, as measured by well 'nests' (several wells of varying depth at the same location) is quite variable across the site, being downward in some locations, and upward in others. Vertical flow is likely influenced by the pumping of nearby water supply wells. However, the vertical flow does not appear to have resulted in significant amounts of mixing of the groundwater in the various aquifers at the site.
The results of well sampling conducted by MDH in 1995 suggested a clear stratification of groundwater between the rock layers based on the results of tritium analyses (MDH 1996). Tritium is a radioactive isotope of hydrogen (H3) generated by atmospheric nuclear weapons testing that began in the late 1940s and early 1950s. Its presence can be used to estimate the age of groundwater. Groundwater samples collected by MDH in 1995 from deeper wells at the site generally had lower levels of tritium, indicating the water entered the aquifer system before the early 1950s, while samples from shallower wells showed higher levels of tritium indicating relatively younger water. This indicates that groundwater between the upper sands, the Prairie du Chien limestone, and the lower Jordan sandstone is not mixing to a large degree. The tritium levels were also roughly correlated with VOC levels in the respective aquifers, indicating the contamination likely first reached the groundwater after the early 1950s.
The 1995 tritium testing results are consistent with the TCE distribution observed in private wells at the site. In the Prairie du Chien limestone, a fairly large area of relatively high TCE concentrations (greater than 20 µg/L) is present (see Figure 7a). In contrast, the Jordan formation has a fairly small area of these relatively high TCE concentrations (as shown in Figure 7b). This suggests a localized area(s) where mixing occurred between the two aquifers, rather than widespread downward transport of TCE from the Prairie du Chien to the Jordan. This localized effect could have been an area of greater vertical permeability at the contact between the two aquifers, an improperly constructed or unsealed well, or other geologic feature(s), such as a fault, that allowed for greater downward flow in the area near 40th Street North and Neal Avenue. An irrigation well located in this area, subsequently sealed by the MAC in 2001, was open across both aquifers and may have contributed to TCE reaching the Jordan at higher than normal rates in that area. However, the large area of TCE contamination in the Jordan cannot be attributed to only one well.
Tritium samples collected in 2002 from the city of Bayport's three municipal wells have yielded less conclusive information regarding groundwater mixing in the Franconia. All three of the wells are completed in the Franconia and deeper aquifers. Tritium concentrations in the samples ranged from 0.8 to 1.3 TU (tritium units). These low concentrations suggest very little or no mixing with shallow or post-1950s groundwater, which is what would be expected. However, in 2003 municipal well #2 was found to contain increasing concentrations of TCE.
Additional tritium samples were collected in 2003 from private Franconia wells in the area by MDH, after TCE was detected in some of the wells. Relatively low tritium concentrations (<0.8 to 1.0 TU) were detected in the western portion of the SWCA and along the northern and southern edges of the TCE plume. Generally higher concentrations (1.1 to 2.7 TU) were detected in wells nearer the river bluff, with the highest concentrations (5.0 to 11.4 TU) being detected in the wells where TCE was detected (Figure 8). The exception to this trend was Bayport well #2, which yielded a low tritium concentration (0.8 TU), yet has had consistent detections of TCE since May 2003.
Extent of the Contamination
The TCE plume at the site covers an area of approximately six square miles, and affects primarily two bedrock aquifers, the Prairie du Chien limestone and the Jordan sandstone. Low concentrations of TCE have also been detected in Quaternary sands in the westernmost portion of the site, beneath the Lake Elmo airport, and in the easternmost portion of the site, beneath the city of Bayport, where both the Prairie du Chien and Jordan aquifers are absent due to erosion. In addition, TCE has been found in concentrations ranging from 1.0 to 11 µg/L in the Franconia formation near the St. Croix River bluff.
Site investigations to date have not detected significant contamination in soils above the water table, with the possible exception of boring TB-4 (see Figure 9) located near the railroad tracks in the fields west of the airport. Low organic vapor levels were measured during the drilling of this boring, but analysis of a soil sample collected from the boring did not detect any VOCs (the depth at which the sample was collected is not included in the report). The absence of conclusive evidence of contamination in the unsaturated soils above the water table either means that any shallow contamination at the source area has degraded, the location of the original release has not been identified, or the TCE was introduced directly to the groundwater via a well or other conduit.
Similarly, significant contamination has not been detected in the saturated sands and gravels overlying the bedrock at the Lake Elmo Airport, except in a small area in the north hangar area, where two private wells completed in the Quaternary aquifer, labeled wells A and B on Figure 10, were found to have 55 and 23 µg/L of TCE, respectively, in 1995. Well A, the deeper and more contaminated well, was sealed, so no information is available regarding concentration trends in that well. Well B has shown a steadily decreasing trend in concentrations, the most recent sample in 1999 containing 9.9 µg/L TCE. Both wells were completed in an area where there is a depression in the surface of the bedrock (see Figure 11).
Well A was completed at an elevation similar to nearby Prairie du Chien wells with nearly identical contaminant concentration levels. There is likely free exchange of groundwater between the bedrock and the Quaternary sediments filling the bedrock "low," so no conclusion about source areas may be drawn from the presence of contamination in this particular Quaternary well. The shallower, less contaminated well B, however, was completed at an elevation well above the bedrock and is screened immediately below the water table (Figure 11). It is possible that contamination in this well may mark the area (or near to it) where some of the TCE entered the groundwater system. There are no wells completed in the Quaternary sediments upgradient of this area to confirm this hypothesis.
The highest concentrations of TCE in the Prairie du Chien have been detected under the eastern portion of the Lake Elmo Airport (Figure 7a). Sampling of wells there in 2002 detected TCE concentrations as high as 110 µg/L at Hangar 13C. The highest concentration detected beneath the airport, 210 µg/L TCE in 1989, was also in the Hangar 13C well. Lower TCE concentrations (14 to 98 µg/L) historically have been found in the Prairie du Chien in monitoring well MW-10B, located approximately 1,000 feet west of the airport. However, a sample collected in August 2003 detected 130 µg/L TCE in MW-10B.
Additional investigation conducted during the summer of 2003 indicates that TCE concentrations west of the airport may be higher than previously detected (see page 17). While the map of the TCE plume in the Prairie du Chien appears to show two "fingers" of the plume migrating to the northeast and the southeast, this is partly due to the near absence of Prairie du Chien wells southeast of Omaha Avenue and 30th Street North. It is also likely that fracture and/or solution cavity patterns in the Prairie du Chien may be influencing the distribution of the TCE plume. However, the higher concentration portion of the plume (where TCE is greater than 10 µg/L) does appear to trend to the east-northeast.
The TCE plume in the Jordan Sandstone shows a similar distribution, with the main portion of the plume trending to the east-northeast and a lower concentration "finger" in the southeastern portion of the site (Figure 7b). It is not clear if the latter is the result of pumping by recently installed Jordan wells "dragging" part of the plume to the southeast, or if the plume is following a geologic feature in the Jordan that directs the contamination to the southeast. Additional monitoring may help to clarify this.
The TCE plume in the Jordan differs from that in the Prairie du Chien in that the concentrations generally are lower, the highest concentrations in the plume are present east of the airport property, and the high concentration "core" of the plume (TCE greater than 20 µg/L) is much more compact. The shape of the Jordan plume suggests that lower concentrations of TCE are migrating downward into the Jordan from the Prairie du Chien as the plume in the upper aquifer migrates away from the airport and, as discussed above, that there may be a localized feature near Neal Avenue North and 40th Street North that has allowed higher concentrations of TCE to reach the Jordan in that area.
TCE has also been detected, recently, in some Franconia wells near the St. Croix River (Figure 12). As discussed in the Geology/Hydrogeology section of this report, it is possible that in this area portions of the St. Lawrence confining unit have been compromised by faults or fractures that may be allowing TCE to enter the Franconia in some areas near the river. Franconia wells installed in the western half of the site, and north of 47th Avenue North and south of 30th Avenue North do not have any detections of TCE, suggesting that the St. Lawrence in these areas is competent and is preventing downward spread of the contamination. As noted on page 8, tritium sampling in 2003 provided additional evidence that groundwater from the upper aquifers is not entering the Franconia in the western portion of the site.
Groundwater modeling (Wenck 2001a) suggests a travel time of approximately 15 years from the airport to the St. Croix River. Contamination was detected near Bayport in 1987. This suggests that the contamination may have originated near the western end of the site no later than the early 1970s. The original release (or releases) of TCE may have occurred sometime between the 1950s and the 1970s.