20070812 Ver 2_EPA Review Comments for Badin Lake_20100224

EPA Review Comments for Badin Lake Documents
Submitted for North Carolina Department of the Environment and Natural Resources
(NCDENR) and Alcoa
Alcoa-Badin Works
Badin, Stanly County, North Carolina
February 24, 2010
Introduction
In 1917, Alcoa-Badin Works (Alcoa) constructed the Narrows (Badin) Dam across the
Yadkin River creating Badin Lake, which has three geographic arms (NE, NW, and SW).
Alcoa is on the west bank of the SW arm near a swimming cove and boat ramp. (Please
refer to Figure 1, Sediment Sampling and Transect Locations, combined U.S.G.S.
Topographic Quadrangle Maps, Figure 4-32, Badin Lake Surface Water and Sediment
Sample Locations (Plant Area).
In the 1960s and 1970s, Alcoa spills and leaks occurred as a result of plant operations.
Due to these operations and storm water runoff, Alcoa discharged PCB contaminated soil
from the outfalls into Badin Lake. Due to their operations, Alcoa had contaminated on-
site soil at several solid water management units (SWMUs). In the 1990s, Alcoa
excavated the PCB contaminated soil and improved the caps at the SWMUs to prevent
PCB contaminated soil from impacting Badin Lake through surface water runoff.
EPA was asked by NCDENR in fall of 2009 to provide comments to four studies and
related documents conducted on the PCB contamination of Badin Lake and respond to
several questions on these studies.
1. Portions of the following reports were reviewed by EPA:
a. Technical Sediment Sampling Reports:
b. RCRA Facility Investigation Report, Alcoa, Badin, NC, dated 3/2001
c. E-mail from Alcoa: Badin Lake Sediment Investigation Summary, dated
2/12/2009
d. Badin Lake Swim/Picnic Area and Badin Boat Access Sediment
Assessment, Stanly Co., NC, Prepared by Environmental Science Inc.,
dated 1/2008
e. E-mail from John Rodgers: New Badin Lake Report, dated 8/5/2009
2. Toxicology Reports:
a. Proposed Exposure Investigation Protocol for PCB and PAH Fish Tissue
Sampling for Badin Lake, Stanly and Montgomery Counties, North
Carolina, Prepared by North Carolina Department of Health and Human
Services, dated 9/22/2008
b. Health Consultation Badin Lake Fish Tissue Stanly and Montgomery
Counties, NC Prepared by NC Health and Human Services, dated
9/18/2009
EPA's General Conclusions:
Based on the review of the above reports and documents, Alcoa's PCB contamination
appears to remain within the SW Arm of Badin Lake. However, unknown PCB sources
and the railroad may be impacting the NW Arm of Badin Lake. The NE Arm of Badin
Lake, probably the least impacted area of the Lake, could possibly receive PCBs from
other sources. No PCB sediment data has been collected in the NW and NE Arms of
Badin Lake to determine other sources of PCBs. Also, the currents within the Lake are
unknown and it is not know how these sediments may become mixed together within the
Lake's center.
Toxicologic evaluations, conducted by the North Carolina Health and Human Services,
determined that the present advisory of 1 meal/week is protective of human health. This
was based on sediment samples, collected from the SW Arm of Badin Lake and the
discharge points of the NW and NE Arms of the Lake, and fish collected throughout the
Lake. Nevertheless, the fish with the greatest PCB contamination (0.11 ug/kg) was
collected in the NW Arm of the Lake. This indicates that PCB contaminated sediment in
the NW Arm of the Lake may be impacting fish. However, fish migration studies have
not been conducted to verify this hypothesis.
EPA's comments and questions to these studies are presented in Appendix A to this
document. EPA's Response to NCDENR's questions follows.
EPA's Response to NC's Questions:
• Is it possible to identify the sources of the PCBs in Badin Lake? Can the
sources even be determined with any degree of certainty?
It is possible to identify the sources of PCBs in Badin Lake. Aroclor results from grab
sediment samples near Alcoa's outfalls indicate that Alcoa released PCBs into the SW
Arm of Badin Lake at concentrations greater than Industrial Standards (0.057 mg/kg). In
regard to the "degree of certainty", it may be advisable to determine the extent of this
contamination. This can be accomplished by collecting and analyzing discrete sediment
samples at the SW Arm transects shown on the attached USGS topographic map. The
aroclor data from these transects can be compared to the on-site aroclors to determine the
extent of PCB contamination. Note: Alcoa states that they used only the following
aroclors 1260, 1254, and 1248 and that studies of the extent of contamination should
include only these aroclors. However, Alcoa's RFI detected the following additional
PCB Aroclors: 1242, 1232, 1221, and 1016. Therefore, any additional investigations
determining the extent of Alcoa's contamination should, at a minimum, include Aroclors:
1260, 1254, 1248, 1242, 1232, 1221, and 1016.
2
The NE Arm of Badin Lake is surrounded by the Uwharrie National Forest, with some
campground and housing development. This portion of Badin Lake may have the least
PCB impacts. To determine whether other land sources of PCB contamination are
impacting this portion of Badin Lake, Alcoa should collect grab sediment samples at the
transects shown on the attached topographic map. The sediment samples should be
analyzed for Total PCBs and Aroclors.
It is possible to identify whether there are upstream land sources of PCB contamination in
Badin Lake's NW arm, and, if any PCB contamination is originating from the railroad
along the canal. Alcoa can determine if there are other sources of contamination in these
areas by collecting and analyzing grab sediment samples at the transects upstream in
Badin Lake's NW arm, and the canal at the transects on the attached USGS topographic
maps. These sediment samples should be analyzed for Total PCBs and Aroclors
• What proportion of the PCBs originated at Alcoa?
By comparing the PCB sediment concentrations and trends in the SW Arm transects, the
proportion of the PCBs originating at Alcoa can be determined through transect T-4.
• What proportion originates from other sources, including upstream sources?
PCB data from the transects in the NW Arm of Badin Lake will determine if there are
any upstream sources of contamination and the general extent of this contamination. This
PCB data can be compared to the sediment data from the NE and SW Arms of Badin
Lake to determine the proportion originating from the sources in each arm of the Lake.
PCB data will determine if there are any PCB impacts in the canal from the railroad. If
there are impacts, this data can be compared with the PCB data from the SW Arm of
Badin Lake to determine the proportions originating from the railroad and Alcoa.
• How is Alcoa's PCB contaminated sediment affecting the fish vs. PCB
contaminated sediment from other sources?
To begin to answer this question, several actions must be taken. First, the PCB
concentrations in fish tissue should be compared with the fish's migratory habits.
Second, this data then needs to be compared to the sediment concentrations within each
migratory area. Third, other screening tools used by an Ecological Risk Assessor will
need to be applied to definitively answer this question.
Note: The extent of the PCB contaminated sediment potentially from all sources must be
determined by sampling transects downstream from the Narrows (Badin) Dam (Figure 1
and USGS topographic maps). Depending on the level of sediment and fish PCB
concentrations, the current fish consumption advisory may need to be applied
downstream of the Dam.
3
• Based on the data and analysis NC provided to EPA, is further remediation of
PCBs required by Alcoa and to what levels? t
Yes, additional investigation will need to be conducted. PCB screening levels can be used
to evaluate data initially. However, the design of the Sampling and Analysis Plan requires
input from human health risk and ecological risk assessors. Recommended initial
screening levels are presented below:
• The swimming cove beach should be compared to residential risk standards
(0.016 mg/kg).
• To be protective of human health for direct contact, Badin Lake sediment
should be compared to PCB Industrial Standard (0.057 mg/kg).
• For Unlimited Fish Consumption in Badin Lake, PCB sediment
concentrations should be compared to less than or equal to (<) 50 µg/kg. To
choose a remediation level would prevent further bioaccumulation of PCBs in
fish tissue to greater unhealthy human consumption levels (< 50 µg/kg).
• Although the average PCB fish concentration is within the standards for
limited consumption of 1 fish meal/week at current PCB sediment
concentrations, one (1) fish had 0.11 ug/kg PCBs in its tissue. With the
current sediment concentrations at the Alcoa outfall occurring at
concentrations greater than industrial standards, fish may continue to
bioaccumulate PCBs to greater concentrations.
• To be protective of critters and have no ecological impact, Badin Lake
sediment concentrations must be remediated to < 0.021 µg/kg.
As indicated earlier, the current data and information from these four studies and other
documents will need to be submitted to the EPA Human Health and Ecological Risk
assessors for a recommendation for cleanup levels.
Also, any remediation of PCB contaminated sediment should take into consideration the
negative effects of suspending PCB contaminated sediments into the surface water. In
addition, the observations and questions posed in Appendix A will need to be addressed.
For the remediation of Badin Lake, the selection of an appropriate cleanup level will
require further discussion. Traditionally, PCB contaminated lake sediment requires
remediation to a cleanup level < 50 µg/kg. However, there are other options that could be
considered. EPA, Region 5 required the Fox River to be remediated to 0.250 mg/kg and
then used natural sediment capping for final remediation. This remedial action may be
acceptable, if Alcoa conducts further studies to verify whether natural capping is
occurring in Badin Lake.
4
Enclosures
Figure 1, Sediment Sampling and Transect Locations
Figure 4-32, Badin Lake Surface Water and Sediment Sample Locations (Plant Area)
USGS topographic Quadrangle Maps, Badin Lake, NC
Handy, NC
High Rock, NC
New London, NC
Figure 1, Sediment Samples in Badin Lake
Figure 2, Badin Lake sediment Sampling
Figure, Sediment sampling locations, Stanly County Entrix, April 2, 2009
PCB Excel Spreadsheet
Appendix A
5
Appendix A
EPA's Evaluation of the four (4) PCB Studies
Background
As part of the RFI and other reports, Alcoa and John H. Rodgers, Stanly County's
consultant, submitted polychlorinated biphenyls (PCB) sediment and fish data to the
NCDENR from 2001-2009. Alcoa submitted Aroclor sediment and congener fish data,
while John H. Rodgers submitted congener sediment and fish data.
EPA compared PCB sediment data in these reports to the following standards: 1)
Industrial Standards (0.057 mg/kg), 2) Unlimited Fish Consumption (50 µg/kg), and 3)
No Effect on Critters (21.6 µg/kg). These levels were selected as follows:
(1) Because the sediment in Badin Lake is covered with water, people do not have
direct contact with the sediments. Therefore, comparing the sediments to
Industrial Standards is protective for direct human contact.
(2) For there to be Unlimited Fish Consumption in Badin Lake, PCB sediment
concentrations must be less than or equal to (<) 50 µg/kg to prevent the
bioaccumulation of PCBs in fish tissue to unhealthy human consumption
levels (< 50 µg/kg). For there to be No Effect on Critters in Badin Lake and
for PCBs to pose no Ecological Risk, sediment concentrations must be < 21.6
µg/kg.
Evaluation of Alcoa's Data
In 1996-1997, Alcoa collected and analyzed discrete sediment and surface water
samples from Badin Lake. The sediment samples' analytical results showed that PCB
contaminated sediments at concentrations greater than (>) Industrial Standards extend
from Alcoa's outfall 002 to outfall 013 (Figure 4-32, Badin Lake Surface Water and
Sediment Sample Locations (Plant Area)). The concentration of PCB contamination
decreased just beyond Alcoa's outfall 013 to concentrations less than (<) Industrial
Standards. Four (4) sediment samples were collected at NEP 1-3, just beyond outfall
013. While the NEP-2 sediment sample had concentrations greater than Unlimited Fish
Consumption standards, the remaining sediment samples had such low concentrations
that the sediment would have no effect on critters.
In the Aloca 1996-1997 Surface Water and Sediment Study, shallow and bottom
sediment samples were collected. EPA compared the PCB concentrations of the shallow
sediments with the bottom sediments at the following locations: NEP-3, NEP 5, NEP-10,
and NEP-17. The sediment data show that PCBs at NEP-17 are more concentrated on
the surface than at depth. The other sediment data does not show a definite change in the
PCB concentration between the surface and bottom sediments. Depths were not given for
the shallow and bottom samples.
6
The results of other studies show that Alcoa's PCB contaminated sediments in the
outfall area have decreasing concentrations. Nevertheless, the PCB sediment
concentrations remain > industrial standards near Outfalls 011 and 013. Alcoa's
remediation measures are monitored natural attenuation. Therefore, it is recommended
that Alcoa conduct a more thorough study of Badin Lake within the outfall area to
determine if natural sediment capping is occurring.
In 2009 Alcoa conducted a PCB sediment study of Badin Lake that included ten
(10) transects as shown on Figure 1, Sediment Sampling and Transect Locations. Per
NCDENR's directions, Alcoa collected numerous samples at each transect, that were
composited, along with one discrete sample, as shown on Figure 1, Sediment Samples in
Badin Lake.
• All transect data was non-detect for PCBs in the sediment. These results
could be due to the effects ofAlcoa's 1990s site remediation activities,
sediment spreading, or natural sediment capping. Nevertheless, EPA has
concerns with the study's design. While the detection limits were less
than the Industrial Standards, they were greater than the Unlimited Fish
Consumption Standards. The compositing of the sediment samples could
have diluted some PCB contaminated sediments, causing false non-
detection results.
The discrete sediment samples were collected from the middle of each arm
of Badin Lake with the finer sediments from each sample analyzed for
PCBs. These samples represent the concentration of PCB sediments in the
area of the greatest flow within each of Badin Lake's Arms. These
samples may be non-detect for PCBs due to dilution or being carried by
the greater stream flow. The greatest PCB contaminated sediments may
be in areas where the flow is reduced and sediment deposition is
occurring.
• The transect locations in this study do not identify whether other point or
non-point sources of PCB contamination exist within the watershed. Did
Alcoa attempt to determine other sources of PCB contamination by
studying historical air photos? More transects should have been sampled
within each arm of the Badin Lake. The railroad along the NW Arm of
Badin Lake and the canal may be a source of PCB contamination. While
composite samples are useful for a toxicology study, grab samples should
have been collected along each transect to determine other potential
sources of contamination.
Evaluation of John H. Rodgers Sediment Studies
In response to Alcoa's data, John H. Rodgers, Stanly County's consultant,
conducted two limited sediment studies in 2007 and 2009. The 2007 Study involved
collecting and analyzing discrete sediment samples from Badin Lake near Alcoa's
7
outfalls 002, 011, and 013 (Figure 4-23, Badin Lake Surface Water and Sediment Sample
Locations (Plant Area) and Figure 2, Badin Lake sediment Sampling, and PCB Excel
Spreadsheet). The data shows that PCB contaminated sediment at concentrations greater
than Industrial Standards extends from Outfall 002 to Outfall 013 with one sample at
concentrations greater than Industrial Standards and two samples non-detect for PCBs.
The background sample, located in a cove of the NE arm of Badin Lake, was non-detect.
The location of this background sample is acceptable. However, the detection limits are
unknown and no laboratory quality control data was submitted with this study.
In the 2009 Study (Figure Sediment Sampling Locations, Stanly County Entrix,
4/2/2009, and PCB Excel Spreadsheet) discrete sediment samples were collected and
analyzed near Alcoa's outfalls (SW Arm), the NW Arm of Badin Lake, and upgradient of
the Narrows Dam. This data shows PCB contaminated sediment at concentrations
greater than Industrial Standards extends from Alcoa's outfall 011 and 013. No samples
were collected near Alcoa's outfall 002. The background sample site in Badin Lake's
NW arm had 0.022 mg/kg PCBs. The samples from upstream of the Narrows Dam had
0.011-0.022 mg/kg PCBs.
Based on this data John H. Rodgers, Stanly County's consultant concludes that
the PCB contaminated sediment in Badin Lake is due primarily to Alcoa, it has impacted
the fish, and the contamination extends to the Narrows Dam.
EPA has problems with John H. Rodgers conclusions from this study. The
background sediment sample, collected at a location within the NW Arm
of Badin Lake, may be diluted with flow from Garr Creek. His conclusion
that Alcoa's contamination is 10 x the contamination in the NW Arm may
be erroneous.
There are significant problems with this study's data. The laboratory
assigned their sample identification numbers to the specific sample results.
John Rodgers in the final report and sample location map assigned
different sample identification numbers to the sample results. No table
was submitted showing how the laboratory's sample identification
numbers relate to the report and location map's sample identification
numbers. EPA needs this information to verify there is no confusion with
the samples. The PCB Congeners table shows that most of the sediment
data is laboratory qualified data with the following codes:
B Analyte was detected in the Lab Method Blank (LMB) at a
concentration greater than 20pg/L or 5pg/g, and the concentration in
the associated sample is less than 10 times the LMB concentration,
K Data is being reported with a failing ratio and should be considered as
an estimate value,
J Amount detected is between the Method Detection Limit and the
Lower Calibration Limit,
U Identifies a compound as not being detected.
8
John Rodgers did not submit a legend explaining these laboratory
qualification codes. The above code definitions were based on a report
from another lab.
• John Rodgers in his report compares unqualified lab data (data that met all
laboratory QA/QC standards without reservation) with the fish tissue
results. He determined that there was a relationship between PCB
congeners in sediments from the southwest arm of the lake and PCBs in
fish from Badin Lake.
• John Rodgers later in this report compares all qualified (data that did not
meet all laboratory QA/QC standards) and unqualified data with the fish
tissue results. He states that all congener data was used then states that
actual and estimated concentrations were use. It is not clear what samples
with QA/QC labels (B, K, J and U) were used in this analysis. He later
compares the qualified data, only, with the fish tissue results. John
Rodgers concludes that the congeners associated with Alcoa and the SW
arm sediments co-occur with congeners detected in fish.
• There are problems with the John Rodgers data quality and conclusions.
His study is based on correlating sediment and fish tissue data.
Correlation does not mean causation. To establish causation, a fish
migration study must be conducted in Badin Lake.
Other Questions that should be considered during further investigations of Badin
Lake:
1. Have PCBs contaminated the sediments of Badin Lake's NW arm, and are they
continuing to contaminate the Badin Lake's sediments?
2. Are PCB contaminated sediments from various sources commingling due to
currents within Badin Lake?
a. Limnological studies should be conducted to determine the currents within
Badin Lake.
3. Has the PCB sediment contamination throughout the Lake reached asymptotic
conditions?
4. Upon further studies, will the current fish advisory continue to be protective of
human health?
9