Regional Response to the
National Remedy Review Board Comments
on the Site Information Package for
the General Electric (GE)-Pittsfield/
Housatonic River Project, Rest of River

DCN HR-080212-AARX
SDMS 518898

August 3, 2012

Prepared by

^DST«\
v,

iSEh

PRO^

U.S. Environmental
Protection Agency

New England Region
Boston, Massachusetts

11P-0432-3


-------
TABLE OF CONTENTS
REGIONAL RESPONSE TO THE NATIONAL REMEDY REVIEW BOARD
COMMENTS ON THE SITE INFORMATION PACKAGE FOR THE GENERAL
ELECTRIC (GE)-PITTSFIELD/HOUSATONIC RIVER PROJECT, REST OF RIVER

REGIONAL RESPONSE TO NATIONAL REMEDY REVIEW BOARD

RECOMMENDATIONS FOR THE GE-PITTSFIELD/HOUSATONIC RIVER,

REST OF RIVER

APPENDIX A PRELIMINARY DRAFT OF DRAFT OUTLINE OF
POTENTIAL PERFORMANCE STANDARDS FOR
ALTERNATIVE SED 9/FP 4 MOD

Attachment A-l Massachusetts Division of Fisheries and Wildlife Core Habitat Area
Maps

APPENDIX B REVISED COMPARATIVE ANALYSIS OF ALTERNATIVES

Attachment B-l Use of Channel Realignment Along the Housatonic River for
Restoration and Remediation of PCB Contamination

Attachment B-2 Channel Dynamics and Ecological Conditions in the Housatonic
River Primary Study Area

Attachment B-3 Activated Carbon Summary

Attachment B-4 Massachusetts Division of Fisheries and Wildlife Core Habitat
Area Maps

Attachment B-5 Cap Cross Section Refinement - Layer Sizing, Rest of River -
Reach 5 A

Attachment B-6 Comparison Metrics

Attachment B-7 Post-East Branch Remediation Boundary Conditions

Attachment B-8 Food Chain Model Output

Attachment B-9 Preliminary Draft ARAR Tables for SED9/FP 4 MOD

Attachment B-10 Cost Assumptions Memorandum for SED9/FP 4 MOD

APPENDIX C GENERAL ATTACHMENTS

Attachment C-l National Remedy Review Board Recommendations for the
Housatonic River, Rest of River Site

Attachment C-2 Assessment of Recent Tissue PCB Chemistry and Lipid Data

Attachment C-3 Screening Evaluation for Potential Improvement in Sediment
Trapping in Woods Pond

Attachment C-4 Housatonic River Status Report: Potential Remediation

Approaches to the GE-Pittsfield/Housatonic River Site "Rest of
River" PCB Contamination

Attachment C-5 Bank Erosion/Restoration, Housatonic River, Massachusetts

L:\20502169.095\NRRB_RESPONSBMASTERTOC_REV.DOCX

111

8/1/2012


-------
ATTACHMENT B-4 MASSACHUSETTS DIVISION OF FISHERIES AND
WILDLIFE CORE HABITAT AREA MAPS


-------
Commonwealth of Massachusetts

Division off
Fisheries & Wildlife

MassWildlife

Wayne F. MacCallum, Director

July 31, 2012

Robert G. Cianciarulo, Chief
Massachusetts Superfund Section
Office of Site Remediation and Restoration
EPA New England (OSRR-07-01)

5 Post Office Square
Boston, MA 02109-3912

Re: Housatonic River, Core Habitat Areas in the Primary Study Area
Dear Mr. Cianciarulo:

As you are aware, the states of Massachusetts and Connecticut have been working cooperatively
for the last several months to discuss potential approaches to clean up the Rest of River portion of
the GE Housatonic site. These discussions have focused, in part, on the need to address the risks
from polychlorinated biphenyls (PCBs) to humans, fish, and wildlife while avoiding, mitigating
or minimizing the impacts of the cleanup on the unique ecological character of the Housatonic
River. Minimizing impacts to habitat and, in particular, species listed pursuant to the
Massachusetts Endangered Species Act, M.G.L. c. 131A ("MESA"), and 321 CMR 10.00 (the
"MESA Regulations") presents unique challenges as almost the entire Primary Study Area (PSA)
is mapped as Priority Habitat for state-listed species (for a description of Priority Habitat and its
regulatory function please see:

http://www.mass.gov/dfwele/dfw/nhesp/regulatory re view/priority habitat/priority habita
t home .htm. Therefore, in order to help identify the most important areas for habitat protection,
as well as habitats and species that might be particularly sensitive to impacts from PCB
remediation activities, the Massachusetts Division of Fisheries and Wildlife ("DFW") developed
maps of "Core Habitat Areas." The purpose of this letter is to provide an overview of the
approach we used to identify the Core Areas.

As part of our Priority Habitat mapping process, taxonomic experts from DFW's Natural
Heritage & Endangered Species Program ("NHESP") routinely delineate habitat for each state-
listed species, based on actual field-documented records, or "occurrences." There are four types
of Housatonic Core Areas. Core Areas 1, 2, and 3 represent subsets of the delineated state-listed
species habitat found in the PSA. Core Area 4 represents a subset of the documented and
potential vernal pool habitat in the PSA. Please refer to the enclosed maps dated May 21, 2012
which depict the locations of these Core Areas, entitled "Core Habitat Areas, Housatonic River
Primary Study Area (PSA)", "Core Habitat Areas (Core Area 2), Housatonic River Primary Study
Area (PSA)", and "Part of the Housatonic River Showing Primary Study Area, High Species
Richness, and Vernal Pools".

Core Area 1 includes the highest quality habitat for species that are most likely to be adversely
impacted by PCB remediation activities (Table 1). As can be seen in Table 1, most of these species
are plants that are not mobile, and are very sensitive to the expected effects of soil remediation

	www. mcissw ildlife. org

Division of Fisheries and Wildlife

Field Headquarters, North Drive, Westborough, MA 01581 (508) 389-6300 Fax (508) 389-7891

An Agency of the Department ofFish and Game


-------
Housatonic Core Habitat Areas, 7/31/2012, page 2 of 3

activities. Core Area 1 also includes habitat for one state-listed moth species that inhabits mature
floodplain forest, one habitat area for the Jefferson's Salamander, and Triangle Floater mussel
beds. Some of the plant species found in Core Area 1 are located in floodplain forest, which is
not readily restorable and would take decades to return to its current state, if ever. Finally, Core
1 includes areas that are excellent examples of two rare natural communities—High Terrace
Floodplain Forest and Black Ash Bur Oak Hemlock Swamp.

Core Area 2 includes the highest quality habitat for more mobile species that may be less
vulnerable to remediation impacts, species where the habitat is likely to be somewhat more easily
restored, and listed species that may be of a somewhat lower conservation concern, given their
state-wide distribution (e.g. American Bittern; see Table 2). For example, the Mustard White is a
Threatened butterfly species of significant conservation concern that uses a mix of natural areas
along the river and old field habitat. It may be possible to remediate its habitat in phases,
restoring and replacing host plants as the work is completed.

Core Area 3 includes those areas with dense concentrations of state-listed species. Specifically,
Core Area 3 includes areas where Division biologists have delineated overlapping habitat for
eight (8) or more state-listed species.

Core Area 4 includes all certified vernal pools in the PSA as well as additional potential vernal
pool habitat areas which, based on information provided by GE and EPA, are likely to meet the
Massachusetts criteria for vernal pool certification based on the presence of "obligate" vernal
pool breeding amphibians see:

http://www.mass.gov/dfwele/dfw/nhesp/vernal pools/vernal pool cert.htm.

These Core Areas played an important role during recent discussions between the EPA and the
states of Massachusetts and Connecticut regarding potential remediation approaches to Rest of
River. Consistent with the requirements of MESA and the MESA Regulations, the Core Areas are
helping to guide efforts to avoid, minimize and mitigate impacts to state-listed species. Although
a final MESA evaluation will not be completed until the remedy design phase, by focusing on the
Core Areas, EPA and the Commonwealth believe that a framework has been established to
achieve MESA permitting standards of assessing alternatives to both temporary and permanent
impacts to state-listed species, and of limiting the impact to an insignificant portion of the local
populations of affected species. See 321 CMR 10.23. For example, the parties focused on
avoidance of some of the most important and sensitive rare species habitats in Core Area 1.
Similarly, in Core Areas 2 and 3, avoidance of impacts when practical, careful consideration of
PCB remediation methods and the sequence and timing of remediation activities, as well as after-
the-fact habitat mitigation are all approaches that will assist in achieving the substantive
requirements of MESA. Although the Core Areas play an important role in guiding avoidance
and minimization of impacts to state-listed species, in some cases the "take" of state-listed species
is likely to be unavoidable. In those cases, consistent with MESA's status as a location-specific
applicable or relevant and appropriate requirement ("ARAR"), the Commonwealth will work
with GE and the EPA to minimize impacts and to ensure that an adequate long-term net-benefit
mitigation plan for the affected state-listed species is designed and implemented, as required by
321 CMR 10.23(2)(c).

If you have any questions about this letter, please don't hesitate to contact me.


-------
Housatonic Core Habitat Areas, 7/31/2012, page 3 of 3

Sincerely,

^ v. k-

Jon Regosin, Ph.D.

Chief of Conservation Science

Natural Heritage & Endangered Species Program

End.: Table 1. Species and Natural Communities Included in Core Area 1 Delineation
Table 2. Species and Natural Communities Included in Core Area 2 Delineation

cc: Mark Tisa, MA Division of Fisheries & Wildlife
Richard Lehan, MA Department of Fish & Game
Mike Gorski, MA Dept. of Environmental Protection
Eva Tor, MA Dept. of Environmental Protection
Trad Iott, CT Dept. of Energy & Environmental Protection


-------
Jefferson Salamander

Reach 5A

Ostrich Fern Borer.

Narrow Leaved Spring Beauty

l-e

[CR'OF.UJtSTREETd

PITTS El ELD

Legend

Primary Study Area (PSA)
Core Area 1-

These areas show the most
i m porta nt/d istu rba nce-sensitive
habitat areas for state-listed species

\w00D~sJz
-------
Reach 5A

Mustard White

Wood Turtle

Reach 5C

lMOREWO'dDjtlftKEl

BITiTiSRIELFB

American Bittern

lETA^TvgjrSJl

LENOX



Legend

Primary Study Area (PSA)

Core Area 2

These areas show the most
important/disturbance-sensitive
habitat areas for the more mobile
state-listed species

[waeBs'agivB]

Date of Aerial Photo- 2009

Map produced by DFW GIS, 5/21/2012

Core Habitat Areas (Core Area 2)
Housatonic River Primary Study Area (PSA)


-------
'fS&SOvV



wsz*** f l

pwg5>: ' J >"¦

Burr Oa/c andjRed Maple

Mustard White

Wood Turtle

¦\ Reach 5C

fMQREmmm^El

IgilTiSf^lElD

iDBlfliOMI

I RoarinQlBfppk'l

iilEWm

WMf.way^ond}

teEliWtillttKE]

Legend

C3 Primary Study Area (PSA)

Core Areas- High Species Richness
( 8 or more state-listed species)

Vernal Pool Core Areas

l w/Sisp'siSSim

Date of Aerial Photo- 2009

Map produced by DFW GIS, 5/21/2012

Showing Primary Study Area, High Species Richness, and Vernal Pools


-------
ATTACHMENT B-5 CAP CROSS SECTION REFINEMENT - LAYER
SIZING, REST OF RIVER - REACH 5A


-------
Memo

Stanfee

To:

Scott Campbell
Weston Solutions, Inc.
10 Lyman Street, Suite 2

From: Nicholas D'Agostino, PE &

Daniel Nein, CWB
Stantec Consulting
5 LAN Drive
Westford, MA 01886

File:

Pittsfield, MA 01201
195600459

Date: February 24, 2012

Reference: Cap Cross Section Refinement - Layer Sizing
Rest of River - Reach 5A

Housatonic River PCB Remediation, Pittsfield, MA

As requested, Stantec Consulting (Stantec) has completed preliminary calculations to size the habitat,
armor, filter, and isolation layers for the referenced project. It should be noted that the calculations and
recommendations presented herein are preliminary in nature and are not intended to be used for a final
design without further verification. In determining the sizing for the various layers of the cap, it was
assumed that the underlying soil after sediment removal along the river bottom would be granular
(sand) for both riffles and pools.

The Rest of River, from the confluence at the East and West Branches of the Housatonic River to
Woods Pond Dam, is primarily a low-gradient stream and part of an overall riverine system recognized
as having a long history of landscape alteration, settlement patterns, and structural control of the river
(e.g., dams, channelization, bank armoring) (Massachusetts Natural Heritage and Endangered Species
Program 2010). The sediment bed consists of coarse to fine sands with approximately ten percent silts
and clays in Reaches 5A and 5B and fine sands and silts in Reach 5C (Woodlot Alternatives, Inc.
2002), which is consistent with the general trend observed through extensive bed sediment core
sampling that the average percentage of sediment becomes finer from upstream to downstream
(Weston Solutions, 2004).

The sizing of the armor layer of the cap was based upon a 2-year storm event with a peak annual flow
of 1,725 cubic feet per second (cfs) for cross section XS105. A Mannings 'n' value of 0.03 and channel
slope of 0.001 was used in estimating flow and velocity. The flood level in the channel would rise to
elevation (El) 958.9, which corresponds to a velocity of 4.67 feet per second. The river bottom for cross
Section XS105 was estimated to be at El 949± at its deepest point, resulting in a water depth of 9.9
feet.

Several velocity-based methods were utilized to determine the D50 size of the riprap and included the
US Army Corps of Engineers (USACE, 1994), US Bureau of Reclamation (USBR, 1967), US
Geological Survey (USGS, 1966), and Isbash, (1936) methods. The D50 size for each of these
methods was calculated to be 1.71 inches, 3.5 inches, 5.16 inches, and 1.82 inches, respectively. The
attached figure provides an example of D50 size versus velocity for a variety of methods, which
indicates the USGS method as being too conservative when compared to the other methods. A D50

One Team, infinite Solutions.


-------
Stamtec

February 24, 2012
Scott Campbell
Page 2 of 5

Reference: Cap Cross Section Refinement - Layer Sizing
Rest of River - Reach 5A

Housatonic River PCB Remediation, Pittsfield, MA

size of 3.5 inches is recommended for this channel. Angular rock is recommended for the riprap armor
layer with the following gradation and corresponding weight:

D10o	5.25 inches 14 pounds

D50	3.5 inches 4 pounds

D10	1.8 inches 0.6 pounds

The thickness of the riprap armor layer was determined by multiplying the D10o stone size by 2.5 times,
resulting in a layer thickness of 13 inches. The D85 was estimated to be 4.5 inches and D15 to be 2
inches

The habitat layer was sized based on critical dimensionless shear stress calculations using a value of
0.035, and is intended to replicate existing surficial bed habitat conditions and where possible improve
surficial bed habitat conditions for aquatic species present. Applying a factor of safety of 2.0 and using
an empirical data set for gravel bed streams (Rosgen, 2006) a D84 value of 3 inches was determined.
The following gradation of the habitat layer is recommended at an overall thickness of 6 inches.

D10o	3.5 inches

D84	3.0 inches

D50	1.0 inches

D16	0.25 inches

Aquatic species such as dragonflies (i.e. aquatic larvae), mussels, and wood turtle, all found in the Rest
of River, are capable of inhabiting a range of riverine sand and gravel bed compositions, which was a
consideration during cap design. The triangle floater (Alasmidonta undulata) occupies a wide range of
substrate and flow conditions, but like most mussel species prefers habitat of low-gradient river reaches
with sand and gravel substrates (Nedeau et al 2000). The wood turtle (Glyptemys insculpta), also a
state-listed species of Special Concern, prefers slower moving streams and rivers with sandy bottoms
and vegetated banks (Ernst and Lovich 2009). Similarly, the larvae of dragonflies inhabit a bed
sediment gradient of sandy substrates found in flowing waterways (Nikula et al 2003). Aquatic habitat
suitability was also a consideration during earlier phases of remediation design. A post-remediation
aquatic community assessment of the 11/4-mile reach conducted in 2007 found macroinvertebrate taxa
richness, relative abundance, and biomass increased when compared to 2000 data and; fish sampling
identified a diverse and abundant post-remediation fish population consistent with the expected fish
community composition, with noticeably greater fish presence in the vicinity of stone structures
provided as part of habitat restoration in the river channel (Weston Solutions, Inc. 2007).

Once the habitat layer is placed at the site, sands and other small-sized sediments from upstream
should fill the interstitial void space relatively soon thereafter, which is anticipated to offer similar
existing habitat conditions for aquatic species present. Smaller sands can also be added during
construction, and it is recommended that sand comprise approximately 30 percent of the layer, with
gravel as specified above comprising 70 percent of the layer.


-------
Stamtec

February 24, 2012
Scott Campbell
Page 3 of 5

Reference: Cap Cross Section Refinement - Layer Sizing
Rest of River - Reach 5A

Housatonic River PCB Remediation, Pittsfield, MA

The filter layer beneath the armor layer should consist of a gravel (coarse) layer underlain by an
isolation sand (finer) layer. These layers were sized based upon the following equation (Brown, 1989):

D15 (coarse layer)	D15 (coarse layer)

	< 5 <	< 40

D85 (finer layer)	D15 (finer layer)

Utilizing this relationship, it was determined that a gravel filter layer beneath the armor layer
corresponding to an AASHTO No. 57 stone would need to be 6 inches thick. The recommended
gradation for this gravel filter layer is as follows:

Sieve Size	Percent Passing by Weight

1-1/4 inch	100

1	inch	95 to 100
1/4 inch	25 to 60
No. 4 0 to 10
No. 8 0 to 5

An average value for D85 and D15 for this gravel filter layer was estimated to be 0.75 inch and 0.25 inch,
respectively.

For the riprap to gravel filter interface, the size of the gravel was determined from the following equation:

Dis (riprap)	Di5 (riprap)

	= 2.63 < 5 <	= 8.3 < 40 OK

D85 (gravel filter layer)	D15 (gravel filter layer)

Therefore the gravel portion of the filter will provide an adequate transition between the riprap and
underlying isolation layer.

This six inch thick gravel layer would then be underlain by a six inch thick sand isolation layer
corresponding to a Massachusetts Department of Transportation (MassDOT) specification M1.03.0
Type c. The purpose of the isolation layer is to minimize the advective and diffusive flux of PCB
migration from the underlying sediments up through the cap. This sand layer should be blended with
organic material at the source to achieve a total organic content of 0.5 percent by weight. The
recommended gradation for the sand/isolation layer is as follows:

Sieve Size	Percent Passing by Weight

2	inch 100
1/4inch	50 to 85
No. 4	40 to 75
No. 50	8 to 28
No. 200	0 to 10


-------
Stamtec

February 24, 2012
Scott Campbell
Page 4 of 5

Reference: Cap Cross Section Refinement - Layer Sizing
Rest of River - Reach 5A

Housatonic River PCB Remediation, Pittsfield, MA

An average value for D85 and D15 for this sand isolation layer was estimated to be 1.00 inch and 0.009
inch, respectively. The gravel filter layer to sand layer interface was estimated as follows:

D15 (gravel filter layer)	D15 (gravel filter layer)

	= 0.24 < 5 < 	= 27.3< 40 OK

D85 (sand layer)	D15 (sand layer)

The gradation of the gravel filter layer will provide adequate protection for the underlying sand isolation
layer.

The river bed sediment gradations were obtained for Reach 5B from Table 4-6 from the RFI Report
(General Electric, 2003) which provided an arithmetic mean for the percentages of gravel (1.4%), sand
(76%), silt (18%), and clay (4.1%). For the purposes of this analysis, it was assumed that the gradation
of the sediment at the proposed bed excavation limit would be similar to the river bed sediment
gradation noted above. From this information D85 and D15 values were estimated to be 0.02 inch and
0.001 inch, respectively. The sand filter to sediment interface was estimated as follows:

D15 (sand layer)	D15 (sand layer)

	= 0.45 < 5 <	=9 < 40 OK

D85 (river bed sediment)	D15 (river bed sediment)

The gradation of the sand isolation filter will provide adequate protection of the underlying sediments.

A 3-inch thick mixing layer should be placed over the bottom of the excavation and should consist of
gravel to prevent infiltration of the underlying soil/sediment into the isolation layer during construction.

REFERENCES:

Brown, S.; Clyde, E. 1989. Design of Riprap Revetment. Hydraulic Engineering Circular 11 (HEC-11).
FHWA IP-89-016. Washington, DC: U.S. Department of Transportation, Federal Highway
Administration. Section 4.4, Filter Design. http://isddc.dot.qov/OLPFiles/FHWA/OQ9881.pdf

Ernst, C. and J. Lovich. 2009. Turtles of the United States and Canada. Second Edition. The Johns
Hopkins University Press, Baltimore, Maryland.

General Electric Company, 2003 "Housatonic River - Rest of River RCRA Facility Investigation
Report" Volume 1 September 2003

http://www.epa.gov/reqion1/qe/thesite/restofriver/reports/rcra fir/RFI Report Executive%20Summarv
and Section 1.pdf

Isbash, S. V. 1936."Construction of Dams by Depositing Rock in Running Water," Transaction, Second
Congress on Large Dams, Vol 5, pp 123-126.

Massachusetts Natural Heritage and Endangered Species Program (NHESP). 2010. Rare Species and
Natural Community Surveys in the Housatonic River Watershed of Western Massachusetts.
Massachusetts Division of Fisheries and Wildlife. Westborough, Massachusetts.


-------
Stamtec

February 24, 2012
Scott Campbell
Page 5 of 5

Reference: Cap Cross Section Refinement - Layer Sizing
Rest of River - Reach 5A

Housatonic River PCB Remediation, Pittsfield, MA

Nedeau, E., M. McCollough, B. Swartz. 2000. The Freshwater Mussels of Maine. Maine Department
of Inland Fisheries and Wildlife, Augusta, Maine.

Nikula, B., J. Loose, M. Burne. 2003. A Field Guide to the DragonfUes and Damselflies of
Massachusetts. Massachusetts Division of Fisheries and Wildlife, Westborough, Massachusetts.

Rosgen, D.L. 2006. A Watershed Assessment for River Stability and Sediment Supply (WARSSS).
Wildland Hydrology Books, Fort Collins, CO. http://www.epa.gov/warsss/.

USGS 1966, Simons, D. B., and Richardson, E. V. "Resistance to Flow in Alluvial Channels," US
Geological Survey Professional Paper 422-J, US Government Printing Office, Washington, DC.

US Army Corps of Engineers, 1994, "Engineering and Design - Hydraulic Design of Flood Control
Channels (EM1110-2-1601)"

US Bureau of Reclamation. 1967. "General Design Information for Structures," Chapter 2, Canals and
Related Structures, Design Standards No. 3, US Department of the Interior, Denver, CO.

Weston Solutions, Inc. 2004. Model Calibration: Modeling Study of PCB Contamination in the
Housatonic River. Volume 3 Appendix B Hydrodynamic and Sediment/PCB Fate and Transport Model
Calibration. Report prepared for U.S. Environmental Protection Agency Region 1, Boston,
Massachusetts.

Weston Solutions, Inc., 2007. Post-Remediation Aquatic Community Assessment 1 1A Mile Removal
Reach. Report prepared for U.S. Environmental Protection Agency Region 1, Boston, Massachusetts.

Williams, David T 2009. Power Point Presentation "Streambed Stabilization: Resistive Methods",
Montana Department of Natural Resources and Conservation (DNRC) Technical Training Course
"Stream Bank Stabilization and Erosion Control Design" December 3-4, 2009
http://www.dnrc.mt.qov/wrd/water op/floodplain/streambank course/stream stabilization resistive.pdf

Woodlot Alternatives, Inc. 2002. Ecological Characterization of the Housatonic River. Report
prepared for U.S. Environmental Protection Agency Region 1, Boston, Massachusetts.


-------
NARROW CHANNEL, 2H:1V SS, VELOCITY^™ 2009,

4.67 5

6	7

VELOCITY (FT/S)


-------
ATTACHMENT B-6 COMPARISON METRICS


-------
^ ^ ^ ^

i Passing Woods Pond (kg)

1T0 Reach 5 & 6 Floodplain (kg)

Passing Rising Pond Dam (kg)

Figure 1 - Average Annual Mass of PCBs Passing Woods Pond and Rising
Pond and Transported to Reach 5 & 6 During the Model Period for
Combination Alternatives (averaged over last 5 years of simulation)


-------
Reach 5A Reach 5B Reach 5C Reach 5D Reach 6 Reach 8

¦ SED2+FP1 I SED3+FP3 ¦SED5+FP4 ¦SED6+FP4 I SED8+FP7 I SED9+FP8 HSED10+FP9 ¦SED9+FP4MOD
Reach 7 and CT Reaches are not presented on this figure. Concentrations for these reaches are presented as ranges in the CMS

Figure 2 - Model Predicted Average Surface Sediment (0-6") PCB Concentration
at End of Projection Period for Combination Alternatives


-------
Reach 5A Reach 5B Reach 5C	Reach 6	Reach 8

¦ SED2+FP1 BSED3+FP3 ¦SED5+FP4 ¦SED6+FP4	SED8+FP7 aSEDg+FPS HSEDIO+FPg ¦SED9+FP4MOD

Reach 7 and CT Reaches are not presented on this figure. Concentrations for these reaches are presented as ranges in the CMS.

Figure 3 - Model Predicted Average Surface Water PCB Concentration at End of
Projection Period for Combination Alternatives


-------
600

w
S>
o
<

500

474

474 Acres equal to 100% IMPG Attainment
474	474 474—474 474	474 474

474

* &*** &>v*?k sto^p1

si Lower Bound	"Upper Bound

Figure 4 - Benthic Invertebrates IMPG Attainment in Acres for Combination
Alternatives


-------


160



140



120



100

w



o



o

80

<





60



40



20



0





in

45-





119 Acres equal to 100% IMPG Attainment

119

119 119 119 119 119

Se®s*



24

i

S

Lower Bound

,to^p4 sto^p1 sto«*fVa

Upper Bound

Figure 5 - Amphibian IMPG Attainment in Acres for Combination
Alternatives


-------
"O

o
n

0J

o

o

3

(0
(0

m

45
40

35

_30
u>

f25
Ł

r-20

CD
O
Q-15

X 10

o

ro

5

Ł
L.

CO

o

IMPG is 55 mg/kg





&

^ ^	^ ^	w^G	r*

%





^ ^ ^ ^	^ ^ ^ yjF ^ ^ ^ ^ <#

SED2+FP1

SED3+FP3

SED5+FP4

SED6+FP4

SED8+FP7

SED9+FP8

SED10+FP9

SED9+FP4 MOD

Tissue concentrations reported in GE's RCMS were noted to be "autumn average" but appear to be annual
averages. To facilitate comparisons, values shown for SED 9/FP 4 MOD are annual averages

Figure 6 - Projected Warmwater Fish Tissue (whole body) PCB
Concentration at the End of Model Projection Period for Combination
Alternatives


-------
©

0

1	~

^ O)
© ^

M 1>
to Ł

-D

IE

5

o
o

60

50

40

30

20

10

IMPG is 14 ma/ka

0

Reach Reach Reach Reach Reach Reach Reach Reach 7H
7 A 7B 7C 7D 7E 7F 7G

SED2+FP1 ¦SED3+FP3 «iSED5+FP4 I.SED6+FP4 «SED8+FP7 SED9+FP8 BSED10+FP9 HSED9+FP4 MOD

Figure 7 - Projected Coldwater Fish Tissue (whole body) PCB Concentration
at the End of Model Projection Period for Combination Alternatives


-------
900

800

700

600

U)

0) 500
o
<

400

300

200

100

s^pA s^f?1	seo^f"

Figure 8 - Insectivorous Birds (Wood Duck) IMPG Attainment in Acres for
Combination Alternatives

573

265

381


-------
w

&
o
<

600

500

400

300

200

100

221

100% attainment equals approximately 486 acres
474	486	486	486

405

0 I 0	0

S€.o^ &*** se»6rfpA sep»rfp1

Figure 9 - Piscivorous Bird IMPG Attainment for Combination Alternatives


-------
800
700
600

 500

d)

o
<

400
300
200
100

726 Acres equals
100% IMPG

726

726

726726

726

Attainment

0 0

0 0

0

U \J	u u	u	u

0 A	.	.	™—!	¦ , 1

seo^fpA seo6*fpA sep»*fp1 sto9"fpa

0 0



Lower Bound

Upper Bound

Figure 10 - Piscivorous Mammals (Mink) IMPG Attainment in Acres
for Combination Alternatives


-------
595 Acres equal to 100% IMPG Attainment

595 595 595 595 595 595 595 595 595

595

s€!o»*f?1

Lower Bound IMPG

Upper Bound IMPG

Figure 11 - Omnivorous and Carnivorous Mammals IMPG Attainment in Acres for
Combination Alternatives


-------
>*

o

c

0
O

LU

U)

c
"5.

Q.

rc

(0

o

-------
3,000,000

2,500,000
2,000,000

CO
T3

k.

> 1,500,000

o ' '

!q

3

o

1,000,000
500,000

0

~ 22	121,000

121,000

556,000

921,000	14 915,000

2^0
241,700



Sediment Volume (cy)

i Soil Volume (cy)

PCB Mass Removed

Figure 13 - Sediment and Soil Removal Volumes and PCB Mass Removed
for Combination Alternatives


-------
100,000

~o

d)
>

g 400,000

C

c
o

H 300,000

80,000 O

§ «r

Ł. ¦g

60,000 $ §
g S
m

40,000 ^

1 ? qnn

190,000 n 171270

20,000

0	— — _	40,000

I firppnhmiqp fiat; Fmk<;inn<; ~ PCB Mass Removed

o



Greenhouse Gas Emissions
Estimate for SED9+FP4 MOD is based on ratios of similar alternatives.

Figure 14 - Greenhouse Gas Emissions with PCB Mass Removed for
Combination Alternatives


-------
16,000	60

U)

¦i" 14,000	R 52 13|5Q0

¦m	- 50

O 12,000 -|	|	Ht260

-	40
30
20

-	10

¦ Annualized Truck Trips	~Estimated Years of Truck Traffic

SED9+FP4 MOD based on excavation volumes, backfill volumes, truck capacities and
information presented in the Revised CMS.

Figure 15 - Estimated Annual Truck Trips and Total Years of Truck Traffic for
Removal of Excavated Material and Delivery of Capping/Backfill Material for
Combination Alternatives


-------
ATTACHMENT B-7 POST-EAST BRANCH REMEDIATION BOUNDARY

CONDITIONS


-------
l~~OR^ HydroQual

Memorandum

To: Susan Svirsicy
Scott Campbell
Dick McGrath

FROM: Edward Garland

Date: July 23,2012

Re: Post-East Branch Remediation
Boundary Conditions

CC:

FILE: WSTN - 153624

As part of the PCB fate and transport model calibration and validation, relationships were developed
describing total suspended solids (TSS) and PCBs entering the upstream boundary of Reach 5A
from the East Branch of the Housatonic River as a function of river flow conditions (Appendix B.2
of the FMDR (EPA, 2006)). These estimates were based on data collected prior to remediation of
the East Branch. In the early stages of the Corrective Measures Study (CMS), GE developed
estimates of East Branch PCB boundary conditions for the post-East Branch remediation period
(Arcadis, et.al., 2007) for use in simulations of future conditions. For the purpose of the CMS, GE
assumed that the relationship between TSS and East Branch river flow was representative of future
conditions.

GE has an ongoing monthly monitoring program, and in the fall of 2010, an analysis was performed
to compare the post-East Branch remediation boundary condition estimates to post-remediation
data available at that time. Differences between the 2006-2010 data and earlier boundary condition
estimates were noted and model simulations were performed to evaluate if these differences affect
the model projections, particularly the relative effectiveness of the different alternatives. The results
of these analyses were presented at a technical team meeting in December 2010 and are described
below to document the analysis and conclusions

East Branch TSS and PCB Concentrations

TSS and PCB boundary concentrations are highly dependent on river flow conditions, which
complicates comparison of data collected in different time periods, because flow conditions at the
time of sampling can vary substantially between two different periods (i.e. pre- and post-
remediation). Data collected between April 27, 2006 and August 26, 2010 are plotted versus East
Branch flow and superimposed on boundary conditions used in the CMS (Figure 1). Both TSS and

HDR | HydroQual
HDR Engineering, Inc.

1200 MacArthur Blvd
Mahwah, NJ 07430-2322

Phone: (201)529-5151
Fax: (201)529-5728
www.hdrinc.com


-------
Susan Svirsky
Scott Campbell
Dick McGrath

July 23, 2012

Page 2

PCB concentrations from the post-East Branch remediation period tend to fall below the original
boundary condition estimates used in the CMS. Lower TSS concentrations may be the result of
changes in the cross sectional shape of the East Branch introduced as part of the remediation to
help stabilize the riverbanks. Armor stone placed on the riverbed of the East Branch could also
contribute to a temporary reduction in TSS boundary conditions, due to infilling of the voids
between the armor stone and a reduction in resuspension of bed sediment. In each of these cases,
the change in solids loading from the East Branch may be a transient condition as a new equilibrium
morphology develops. The effect of the reduction in TSS on PCB boundary conditions was
investigated by performing equilibrium partitioning calculations to estimate particulate PCB
concentrations (i.e. mg PCB/Kg TSS).

Relationships between particulate PCB concentrations and East Branch flow, for different time
periods are shown on Figure 2 (Arcadis, et.al. Supplement to Model Input Addendum, Figure 4-1).
The upper dotted line represents the function derived from pre-East Branch remediation data,
which was used in the calibration and validation modeling. The solid line represents the function
derived by Arcadis, et.al. in 2007, using post-East Branch remediation data. Additional post-East
Branch remediation data (June 14, 2007- August 26, 2010), which supplement Arcadis, et.al. 2007
analysis (Figure 2), were used to calculate particulate PCB concentrations and are shown on Figure
3. These more recent data are in general agreement with Arcadis, et.al. 2007 relationship, suggesting
that the mass of PCBs per mass of solids entering Reach 5A from the East Branch has not changed
relative to Arcadis, et.al. 2007 estimates. Rather, the reduction in total PCB concentrations
expressed on a volumetric basis (i.e. ug/1) is primarily related to the reduction in TSS concentrations
from the East Branch.

Projections with Boundary Concentrations based on 2006-2010 Data

The conclusions of the data analysis that:

•	2006-2010 East Branch TSS concentrations were lower than historical levels,

•	the lower TSS carry sorbed PCB per mass of TSS similar to Arcadis' 2007 estimate, and

•	lower TSS concentrations and similar sorbed PCBs result in lower total PCB loading

led to the question, "How would the model projections, particularly the relative effectiveness of the
different alternatives be affected by replacing the previous estimated boundary concentrations with
TSS and PCB concentrations based on the 2006-2010 data?". The lower total PCB concentrations
(relative to the original CMS estimates) would not necessarily result in an acceleration of natural
recovery simply because TSS inputs from the East Branch represent a source of solids with lower
PCB concentrations than the existing sediment in Reaches 5 and 6. The TSS mass loading is also
reduced relative to the original estimates, which reduces the dilution effects of sedimentation.

HDR | HydroQual
HDR Engineering, Inc.

1200 MacArthur Blvd
Mahwah, NJ 07430-2322

Phone: (201)529-5151
Fax: (201)529-5728
www.hdrinc.com


-------
Susan Svirsky
Scott Campbell
Dick McGrath

July 23, 2012

Page 3

Model simulations were performed with TSS and PCB boundary conditions based on the 2006-2010
data to determine the affect on simulated future concentrations, relative to those simulated with
higher TSS and PCB concentrations estimated in 2007 for the CMS. Two alternatives, SED1/2 and
SED5, were used for this evaluation because they include a wide range in remedial activities, from
no active remediation in SED1/2 to bank stabilization and removal and/or capping in all of the sub-
reaches from Reach 5A through Reach 6 in SED5.

Pre-remediation data in the East Branch were collected in both routine monthly and storm-event
monitoring programs, which allowed the data to be separated into groups based on flow conditions
(rising, falling, and neutral). The same type approach could not be applied to the more-limited post-
remediation data. Instead, log-log regressions were developed for the TSS concentrations above and
below 250 cfs (Figure 4) and used to derive an alternate TSS time series input to Reach 5A from the
East Branch. An alternate PCB boundary concentration time series was developed from GE's post-
remediation particulate PCB versus flow relationship (Figures 2 and 3), the alternate TSS time series,
and equilibrium partitioning calculations. Note that all of the other assumptions incorporated in the
CMS (effect of remediation in Silver Lake, Unkamet Brook, GE plant site, half life) were retained in
developing the alternate PCB boundary conditions.

Reach-average sediment PCB concentration results from simulations performed with the alternate
TSS and PCB boundary conditions are compared to original CMS results on Figures 5-9. In all
reaches, a comparison of the SED1/2 results with the original and revised boundary conditions
shows a small reduction in the rate of natural recovery, which is due to a reduction in the loading of
solids from the East Branch with PCB concentrations lower than the in-place sediments. Results
from the SED5 simulations show little sensitivity in Reaches 5A and 5C (Figures 5 and 7). In these
reaches, the relatively gradual change in sediment PCB concentrations for both the original and
revised boundary condition simulation indicates that deposition of upstream boundary solids is less
than in other reaches, and therefore, the change in boundary conditions has less of an impact. In
reaches 5B and 5D (Figure 6 and 8) the more-rapid increase in sediment PCB concentrations
following remediation in in the original SED5 results indicates that deposition of boundary solids
and PCBs is more significant in these reaches. The result of the reduction in boundary solids
concentrations slows the rate of recontamination; however, the resulting sediment concentrations
differ by less than a factor of two.

In Reach 6, sediment concentrations at the beginning of remediation in SED5 are higher in the
simulation with the reduction in East Branch solids concentrations, compared to the original SED5
simulation, because of the slower rate of natural recovery caused by the lower solids inputs from the
East Branch. As a result, the post-remediation concentration in the revised simulation is slightly
higher than the original SED5 simulation. The slower rate of natural recovery in the revised

HDR | HydroQual
HDR Engineering, Inc.

1200 MacArthur Blvd
Mahwah, NJ 07430-2322

Phone: (201)529-5151
Fax: (201)529-5728
www.hdrinc.com


-------
Susan Svirsky
Scott Campbell
Dick McGrath

July 23, 2012

Page 4

simulation results continues following remediation because of the reduced source of solids from the
East Branch. The post-remediation Reach 6 sediment concentrations differ by less than 20 percent
in the two SED5 simulations.

Given the relatively small effect of the alternate East Branch boundary conditions on Reach 5A-6
sediment concentrations, and the uncertainty in whether the change in the East Branch TSS
concentrations represent a short-term transient or permanent change, it is concluded that application
of the revised boundary conditions to the remaining SED alternatives is not warranted.
Comparisons among the alternatives should not be affected by the uncertainty in the estimates of
East Branch boundary conditions.

References:

ARCADIS BBL and QEA. 2007d. Housatonic Rest of River Model Input Addendum Supplement,
August 2007.

EPA. 2006. Final Model Documentation Report: Modeling Study of PCB Contamination in the
Housatonic River. Prepared by Weston Solutions, Inc., West Chester, PA, for the U.S. Army
Corps of Engineers, New England District, and the U.S. Environmental Protection Agency,
New England Region, November 2006.

HDR | HydroQual
HDR Engineering, Inc.

1200 MacArthur Blvd
Mahwah, NJ 07430-2322

Phone: (201)529-5151
Fax: (201)529-5728
www.hdrinc.com


-------
afj

r-
Ł

1/3
H

Flow(cfs)

Figure 1. Comparison of original (•) TSS and PCB estimates and post-East Branch
remediation period data (~) versus river flow ( - non-detect plotted at detection limit)

Ea^t Branch

>» fin	i		

10

~3j
_>

e

u

n.

10

10


-------
100.00

10.00

QQ

o_,

fl- 01]
qjjxZ

ra"3tj 1.00
3 E

o ^
u
Cl.

0.10

0.01

10	100	1000	10000

Pomerov Flow

(cfs)

Figure 2. East Branch boundary particulate PCBs versus River Flow (	

Calibration/Validation, 	 Arcadis, et.al. 2007 estimate of post-East Branch Remediation

period) (Figure 4-1 of Arcadis, et.al., 2007 Supplement to Model Input Addendum) (•
Routine Monitoring, * GE 2007 Supplemental Storm Event Monitoring; filled symbol =

detected, open=n on-detect).


-------
Pomerov Flow

(cfs)

Figure 3. Additional post-remediation data (•) added to East Branch boundary particulate

PCBs versus River Flow (	Calibration/Validation, Arcadis, et.al. 2007 estimate of

post-East Branch Remediation period) (Figure 4-1 of Arcadis, et.al., 2007 Supplement to
Model Input Addendum) (• Routine Monitoring, • 2007 Storm Event - GE 2007
Supplemental Monitoring Program; filled symbol = detected, open=non-detect).


-------
Post Remediation

1000

100


c
o
IS

*5

C

o 2T

0	d

¦5.

1	E ,0

u- CO
T3 JO
0) l-
c
!5
E
o
O

0.1





•



• •



0

• ^

•

# ^ . #.• •• .

•. •% •.

Jrm
•

... ;.\t. •

• • •• # • y*

•

_





. Intercept=2.84
Slope=-0.06

Intercepts.49E-03

Slope=1.55

10

100

1000

10000

Flow (cfs)

Figure 4. East Branch total suspend solids versus flow - Post-East Branch remediation period data


-------
10

Reach 5A

—p—i—i—|—i—i—i—|—i—i—r—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—]—i—!—i—|—i—i—i—|—i—i—<—|—i—i—i—|—i—i—i—j—i—i—i—|—i—i—r-

— SED1/2 (MNR) - Original Boundary Conditions	5ED5 - Original Boundary Conditions

SED1/2 (MNR} - Revised Boundary Conditions

" SED5 — Revised Boundary Conditio!

10

¦O
ai
X.

o>

Ł

CO 10

O

0.


-------
Figure 6. Reach 5B average sediment PCB concentrations from original and revised simulations


-------
10

Reach 5C

-i—r t i—i—i r ; i—i—i—|—»—i—i—[ i i—i—|—i—r t | i i i |—i—i—i—|—i—i i y i i i—|—i i i |—i r i—|—i—i—r ; i—r—i—f t i—T—

SEDl/2 (MNR) - Original Boundary Conditions	SED5 - Original Boundary Conditions

SED1/2 (MNR) - Revised Boundary Conditions

SED5 - Rev

T3
OJ
*

D>

E,

CO 10

o

CL
*—»
c

0)

E

-a

a>

CO

10

10 	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	l	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	I	i	i	i	l	i	i	i	

0	4	8	12	16	20	24	28	32	36	40	44	48	52	56	60

Year

Figure 7. Reach 5C average sediment PCB concentrations from original and revised simulations


-------
Figure 8. Reach 5D average sediment PCB concentrations from original and revised simulations


-------
Year

Figure 9. Reach 6 average sediment PCB concentrations from original and revised simulations


-------
ATTACHMENT B-8 FOOD CHAIN MODEL OUTPUT


-------
6/12/2012

FCM Runs

June 12, 2012

• Graphs show "Combination 1" to "Combination 9"
(with options 1 and 2 for Combination 9)

Further Notes:

X axis set to 60 years

Contents

IMPG Assumptions within plots:	3

Largemouth Bass, Reach 5A	4

Largemouth Bass, Reach 5B	5

Largemouth Bass, Reach 5C	6

Largemouth Bass, Reach 5D	7

Largemouth Bass, Reach 6	8

Largemouth Bass, Reach 7A	9

Largemouth Bass, Reach 7B	10

Largemouth Bass, Reach 7C	11

Largemouth Bass, Reach 7D	12

Largemouth Bass, Reach 7E	13

Largemouth Bass, Reach 7F	14

Largemouth Bass, Reach 7G	15

Largemouth Bass, Reach 7H	16

Largemouth Bass, Reach 8	17

Largemouth Bass, Bulls Bridge	18

Largemouth Bass, Lake Lillinonah	19

Largemouth Bass, Lake Zoar	20

Largemouth Bass, Lake Housatonic	21

Largemouth Bass, Reach 5A, Probabilistic IMPGs	22

Largemouth Bass, Reach 5B, Probabilistic IMPGs	23

Largemouth Bass, Reach 5C, Probabilistic IMPGs	24

Largemouth Bass, Reach 5D, Probabilistic IMPGs	25

Largemouth Bass, Reach 6, Probabilistic IMPGs	26

Largemouth Bass, Reach 7A, Probabilistic IMPGs	27

Largemouth Bass, Reach 7B, Probabilistic IMPGs	28

Largemouth Bass, Reach 7C, Probabilistic IMPGs	29

Largemouth Bass, Reach 7D, Probabilistic IMPGs	30

Largemouth Bass, Reach 7E, Probabilistic IMPGs	31

Warren Pinnacle Consulting, Inc.

Page 1 of 42


-------
6/12/2012

Largemouth Bass, Reach 7F, Probabilistic IMPGs	32

Largemouth Bass, Reach 7G, Probabilistic IMPGs	33

Largemouth Bass, Reach 7H, Probabilistic IMPGs	34

Largemouth Bass, Reach 8, Probabilistic IMPGs	35

Largemouth Bass, Bulls Bridge, Probabilistic IMPGs	36

Largemouth Bass, Lake Lillinonah, Probabilistic IMPGs	37

Largemouth Bass, Lake Zoar, Probabilistic IMPGs	38

Largemouth Bass, Lake Housatonic, Probabilistic IMPGs	39

Warm water Metric	40

C ol d water Metri c	41

Modeled Subreach Average Fish (Fillet) PCB Concentrations at End of Project Period	42

warren
pinnacle
consulting, inc.

W

Warren Pinnacle Consulting, Inc.

Page 2 of 42


-------
June 12, 2012
IMPG Assumptions within plots:

Receptor

Protection

Type

Deterministic
tPCB IMPG
(mg/kg skinless
fillet)

Probabilistic
(mg/kg skinless
fillet)

IMPG Tissue
Type

FCM Surrogate

Human health

Cancer 10E-06

RME

0.0019

0.0064

fillet ww

"Blended" Fish Fillet

Human health

Cancer 10E-06

CTE

0.049

0.057

fillet ww

"Blended" Fish Fillet

Human health

Cancer 10E-05

RME

0.019

0.064

fillet ww

"Blended" Fish Fillet

Human health

Cancer 10E-05

CTE

0.49

0.57

fillet ww

"Blended" Fish Fillet

Human health

Cancer 10E-04

RME

0.19

0.64

fillet ww

"Blended" Fish Fillet

Human health

Cancer 10E-04

CTE

4.9

5.7

fillet ww

"Blended" Fish Fillet

Human health

Non-cancer (adult)

RME

0.062

0.12



"Blended" Fish Fillet

Human health

Non-cancer (adult)

CTE

0.43

1.5



"Blended" Fish Fillet

Human health

Non-cancer (child)

RME

0.026

0.059

fillet ww

"Blended" Fish Fillet

Human health

Non-cancer (child)

CTE

0.19

0.71

fillet ww

"Blended" Fish Fillet

Human Health

CT Advisory Unlimited
Consum.



0.046



Converted to fillet
ww from fillet
with skin ww

"Blended" Fish Fillet

Human Health

MA Advisory

-

1



fillet ww

"Blended" Fish Fillet

IMPG Notes: Source Table 2-2— http://www.epa.gov/regionl/ge/thesite/restofriver/reports/impg/248143.pdf

•	FCM Results are converted to skinless fillet concentrations using the 5:1 conversion method per CMS dispute resolution letter.

•	For the CT Advisory line, to convert "fillet with skin" to "fillet without skin" the concentration is multiplied by 2.3 to achieve whole body
concentrations (Bevelheimer et al. 1997) and then divided by 5 to achieve skinless fillet concentrations (per CMS dispute resolution letter).

Warren Pinnacle Consulting, Inc.

Page 3 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

10

Combination 1
• Combination 6

20

¦ Combination 2
¦Combination 7

30

Years of Simulation

• Combination 3
•Combination 8

40

Combination 4
— Combination 9

50

60

Combination 5

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 5A
Page 4 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE
TIon"-"Can"cer Adult CTE;

r in i n irr ii rfnir

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

Combination 1
• Combination 6

20

¦ Combination 2
¦Combination 7

30

Years of Simulation

• Combination 3
•Combination 8

40

Combination 4
— Combination 9

50

60

Combination 5

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 5B
Page 5 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE
TIon-"Can"cer Adult CTE;
Cancer 10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory

Non-CancerChild RME
CanceFT0A-5~RME

Cancer 10A-6 RME

10

Combination 1
¦ Combination 6

20

¦Combination 2
¦Combination 7

30

Years of Simulation

¦Combination 3
¦Combination 8

40

Combination 4
— Combination 9

50

60

Combination 5

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 5C
Page 6 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

H5535Se6iAdullU!t

Cancer 10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

Combination 1
¦ Combination 6

20

• Combination 2
•Combination 7

30

Years of Simulation

• Combination 3
•Combination 8

40

Combination 4
— Combination 9

50

60

Combination 5

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 5D
Page 7 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

Combination 1
¦ Combination 6

20

• Combination 2
•Combination 7

30

Years of Simulation

• Combination 3
•Combination 8

40

Combination 4
— Combination 9

50

60

Combination 5

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 6
Page 8 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory
Cancer 10A-5 CTE

TIon-CaHcfcM Auuli CTE
Cancer 10A-4 RME

Non-Cancer Adult RME

	Cancer 10A-6 CTE

CT Advisory

Non-CancerChild RME
CanceFT0A-5~RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7A

Warren Pinnacle Consulting, Inc.

Page 9 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE
p n-"Ca n ce r Ad u 11 CTE'
ancer10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7B

Warren Pinnacle Consulting, Inc.

Page 10 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE
an"c"er Adult CTE;
r10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7C

Warren Pinnacle Consulting, Inc.

Page 11 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

Cancer 10A-5 CTE

"TIon-"Cancer Adult CTE;
Cancer 10A-4 RME

Non-Cancer Adult RME

	Cancer 10A-6 CTE

CT Advisory

Non-CancerChild RME
CanceFT0A-5~RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7D

Warren Pinnacle Consulting, Inc.

Page 12 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE
Son'^nber Adult CTE;
CancfiTI 0" 4 RME

Non-Cancer Adult RME

	Cancer 10A-6 CTE

CT Advisory

Non-CancerChild RME
CanceFT0A-5~RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7E

Warren Pinnacle Consulting, Inc.

Page 13 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

TIon-"Cance"r"

Cancer 10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7F

Warren Pinnacle Consulting, Inc.

Page 14 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
	Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7G

Warren Pinnacle Consulting, Inc.

Page 15 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

TforvUan

Cancer 10A-4 RME

Non-Cancer Adult RME

	Cancer 10A-6 CTE

CT Advisory

Non-CancerChild RME
CanceFT0A-5~RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
	Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 7H

Warren Pinnacle Consulting, Inc.

Page 16 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

^"n-"C"ance"r"AcfLirf"C'
Cancer 10A-4 RME

Non-Cancer Adult RME
	Cancer 10A-6 CTE

CT Advisory
Non-CancerChild RME

Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦ Combination 6

¦Combination 2
¦Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
¦Combination 9 Opt. 1

50	60

Combination 5
	Combination 9 Opt. 2

Largemouth Bass, Reach 8

Warren Pinnacle Consulting, Inc.

Page 17 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE

Non-Cancer Adult CTE
Cancer 10A-4 RME

Non-CancerChild RME
Cancer 10A-5 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦Combination 6

• Combination 2
¦ Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
	Combination 9 Opt. 1

50

60

Combination 5
— Combination 9 Opt. 2

Largemouth Bass, Bulls Bridge

Warren Pinnacle Consulting, Inc.

Page 18 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Years of Simulation

Combination 1		Combination 2		Combinations		Combination 4		Combinations

Combination 6		Combination 7		Combination 8		Combination 9 Opt. 1	Combination 9 Opt. 2

Largemouth Bass, Lake Lillinonah

Warren Pinnacle Consulting, Inc.

Page 19 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE

Non-Cancer Adult CTE
Cancer 10A-4 RME

Non-Cancer Adult RME
6 CTE

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦Combination 6

• Combination 2
¦ Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
	Combination 9 Opt. 1

50

60

Combination 5
— Combination 9 Opt. 2

Largemouth Bass, Lake Zoar

Warren Pinnacle Consulting, Inc.

Page 20 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Cancer 10A-4 CTE

MA Advisory

Cancer 10A-5 CTE

Non-Cancer Adult CTE
Cancer 10A-4 RME

Cancer 10A-6 RME

10

20

30

Years of Simulation

Combination 1
¦Combination 6

• Combination 2
¦ Combination 7

¦Combination 3
¦Combination 8

40

Combination 4
	Combination 9 Opt. 1

50

60

Combination 5
— Combination 9 Opt. 2

Largemouth Bass, Lake Housatonic

Warren Pinnacle Consulting, Inc.

Page 21 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

Combination 1
¦Combination 6

20

¦Combination 2
¦Combination 7

30

Years of Simulation

	Combination 3

40

¦Combination 8

50

Combination 4
¦Combination 9

60

Combination 5

Largemouth Bass, Reach 5A, Probabilistic IMPGs

Non-Cancer Adult CTE

1 MA Advisory

Non-CancerChildCTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

10

Cancer 10A-4 CTE

Warren Pinnacle Consulting, Inc.

Page 22 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

0.01

Cancer 10A-6 RME

0.001

10

Combination 1
¦Combination 6

20

¦Combination 2
¦Combination 7

30

Years of Simulation

	Combination 3

40

¦Combination 8

50

Combination 4
¦Combination 9

60

Combination 5

Largemouth Bass, Reach 5B, Probabilistic IMPGs

Non-Cancer Adult CTE

1 MA Advisory

Non-CancerChildCTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

- Probabilistic IMPGs

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

10

Cancer 10A-4 CTE

Warren Pinnacle Consulting, Inc.

Page 23 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

Combination 1
¦Combination 6

10	20

	Combination 2

	Combination 7

30

Years of Simulation

40

•Combination 3
•Combination 8

50

Combination 4
¦Combination 9

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE

Non-Cancer Adult CTE

MA Advisory
Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

0.01

Cancer 10A-6 RME

0.001

60

Combination 5

Largemouth Bass, Reach 5C, Probabilistic IMPGs

Warren Pinnacle Consulting, Inc.

Page 24 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

Combination 1
¦Combination 6

10	20

	Combination 2

	Combination 7

30

Years of Simulation

40

•Combination 3
•Combination 8

50

Combination 4
¦Combination 9

60

Combination 5

Largemouth Bass, Reach 5D, Probabilistic IMPGs

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

10

Cancer 10A-4 CTE

Warren Pinnacle Consulting, Inc.

Page 25 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

Combination 1
¦Combination 6

20

¦Combination 2
¦Combination 7

30

Years of Simulation

	Combination 3

40

¦Combination 8

50

Combination 4
¦Combination 9

60

Combination 5

Largemouth Bass, Reach 6, Probabilistic IMPGs

Non-Cancer Adult CTE

1 MA Advisory

Non-CancerChildCTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

10

Cancer 10A-4 CTE

Warren Pinnacle Consulting, Inc.

Page 26 of 42


-------
June 12, 2012

10

0.1

0.01

0.001

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

Combination 1
Combination 5
	Combination 9 Opt. 1

—	Combination 2

—	Combination 6

—	Combination 9 Opt. 2

40

¦Combination 3
¦Combination 7

50

60

Combination 4
¦Combination 8

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 7 A, Probabilistic IMPGs
Page 27 of 42

Cancer 10A-6 CTE
CT Advisory

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

- Probabilistic IMPGs

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Cancer 10A-4 CTE

——Non-CancerAdult RME
°-1 Cancer 10A-5 RME

Non-CancerChild RME


-------
June 12, 2012

10

0.1

0.01

0.001

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

Combination 1
Combination 5
	Combination 9 Opt. 1

—	Combination 2

—	Combination 6

—	Combination 9 Opt. 2

40

¦Combination 3
¦Combination 7

50

60

Combination 4
¦Combination 8

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 7B, Probabilistic IMPGs
Page 28 of 42

Cancer 10A-6 CTE
CT Advisory

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

- Probabilistic IMPGs

Cancer 10A-4 CTE
Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

__—Non-CancerAdult RME
°-1 Cancer 10A-5 RME

Non-CancerChild RME


-------
June 12, 2012

10

0.1

0.01

0.001

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

Combination 1
Combination 5
	Combination 9 Opt. 1

—	Combination 2

—	Combination 6

—	Combination 9 Opt. 2

40

¦Combination 3
¦Combination 7

50

60

Combination 4
¦Combination 8

Warren Pinnacle Consulting, Inc.

Largemouth Bass, Reach 7C, Probabilistic IMPGs
Page 29 of 42

Cancer 10A-6 CTE
CT Advisory

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

- Probabilistic IMPGs

Cancer 10A-4 CTE
Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

__—Non-CancerAdult RME
°-1 Cancer 10A-5 RME

Non-CancerChild RME


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

10

Combination 1
Combination 5
	Combination 9 Opt. 1

20	30

Years of Simulation

	Combination 2

	Combination 6

40

¦Combination 3
¦Combination 7

50

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE
Non-Cancer Adult CTE

1

MA Advisory
- Non-CancerChildCTE
¦ Cancer 10A-4 RME
"• Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

0.01

Cancer 10A-6 RME

0.001

60

Combination 4
¦Combination 8

	Combination 9 Opt. 2

Largemouth Bass, Reach 7D, Probabilistic IMPGs

Warren Pinnacle Consulting, Inc.

Page 30 of 42


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

40

50

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 3
¦Combination 7

Warren Pinnacle Consulting, Inc.

	Combination 2	—

	Combination 6	—

	Combination 9 Opt. 2

Largemouth Bass, Reach 7E, Probabilistic IMPGs
Page 31 of 42

60

Combination 4
¦Combination 8

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

: 10

3 Cancer 10A-4 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

40

50

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 3
¦Combination 7

Warren Pinnacle Consulting, Inc.

	Combination 2	—

	Combination 6	—

	Combination 9 Opt. 2

Largemouth Bass, Reach 7F, Probabilistic IMPGs
Page 32 of 42

60

Combination 4
¦Combination 8

: 10

3 Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

40

50

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 3
¦Combination 7

Warren Pinnacle Consulting, Inc.

	Combination 2	—

	Combination 6	—

	Combination 9 Opt. 2

Largemouth Bass, Reach 7G, Probabilistic IMPGs
Page 33 of 42

60

Combination 4
¦Combination 8

: 10

3 Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

40

50

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 3
¦Combination 7

Warren Pinnacle Consulting, Inc.

	Combination 2	—

	Combination 6	—

	Combination 9 Opt. 2

Largemouth Bass, Reach 7H, Probabilistic IMPGs
Page 34 of 42

60

Combination 4
¦Combination 8

: 10

3 Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs

0.01

Cancer 10A-6 RME

0.001

10

20	30

Years of Simulation

40

50

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 3
¦Combination 7

Warren Pinnacle Consulting, Inc.

	Combination 2	—

	Combination 6	—

	Combination 9 Opt. 2

Largemouth Bass, Reach 8, Probabilistic IMPGs
Page 35 of 42

60

Combination 4
¦Combination 8

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Child CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

: 10

3 Cancer 10A-4 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

100

CO
O)

w
CD

C

o

S

-i—•

c
CD
O
c
o
O
m
o

CL

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

10

Combination 1
Combination 5
	Combination 9 Opt. 1

20	30

Years of Simulation

¦Combination 2
¦Combination 6

40

¦Combination 3
¦Combination 7

50

	Combination 9 Opt. 2

Largemouth Bass, Bulls Bridge, Probabilistic IMPGs

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE

0.001

60

Combination 4
¦Combination 8

Warren Pinnacle Consulting, Inc.

Page 36 of 42

Non-Cancer Adult CTE

1 MA Advisory

	Non-CancerChildCTE

Cancer 10A-4 RME
Cancer 10A-5 CTE

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

0.01

Cancer 10A-6 RME


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Chi Id CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

0.001

10

Combination 1
Combination 5
	Combination 9 Opt. 1

20	30

Years of Simulation

¦Combination 2
¦Combination 6

40

¦Combination 3
¦Combination 7

50

60

Combination 4
¦Combination 8

	Combination 9 Opt. 2

Largemouth Bass, Lake Lillinonah, Probabilistic IMPGs

: 0.01

-¦ Cancer 10A-6 RME

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

Warren Pinnacle Consulting, Inc.

Page 37 of 42


-------
June 12, 2012

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

¦ Non-CancerChildCTE
Cancer 10A-4 RME
"Cancer 10A-5 CTE

0.001

10

Combination 1
Combination 5
	Combination 9 Opt. 1

20	30

Years of Simulation

¦Combination 2
¦Combination 6

40

¦Combination 3
¦Combination 7

50

60

Combination 4
¦Combination 8

	Combination 9 Opt. 2

Largemouth Bass, Lake Zoar, Probabilistic IMPGs

: 0.01

-¦ Cancer 10A-6 RME

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory

Warren Pinnacle Consulting, Inc.

Page 38 of 42


-------
June 12, 2012

100

10

0.1

0.01

0.001

Average Fillet PCB concentrations in Largemouth Bass (Average for fish ages 5 to 9)

Compared to Probabilistic IMPGs

10

20	30

Years of Simulation

Combination 1
Combination 5
	Combination 9 Opt. 1

¦Combination 2
¦Combination 6

40

¦Combination 3
¦Combination 7

50

100

Probabilistic IMPGs
10

Cancer 10A-4 CTE

Non-Cancer Adult CTE

1 MA Advisory

Non-Cancer Chi Id CTE
Cancer 10A-4 RME
Cancer 10A-5 CTE

0.001

60

Combination 4
¦Combination 8

	Combination 9 Opt. 2

Largemouth Bass, Lake Housatonic, Probabilistic IMPGs

Warren Pinnacle Consulting, Inc.

Page 39 of 42

: 0.01

-¦ Cancer 10A-6 RME

Non-Cancer Adult RME
Cancer 10A-5 RME
Non-CancerChild RME
Cancer 10A-6 CTE
CT Advisory


-------
June 12, 2012

T3
O
_Q

0
O

¦D

W
W

40

35

30

25

O)

0

cn
E

20

sz m
w O
il CL

iB
ro

15

10

5A

5B

5C

5D

7A

7B

7C

7D

7E

7F

7G

7H

Projected Warmwater Fish Tissue (whole body) PCB
Concentration at the End of Model Projection

i Combination 1
Combination 7

Combination 2
Combination 8

i Combination 3
i Combination 9

Combination 4
i Combination 9 Opt. 1

i Combination 5
Combination 9 Opt. 2

Combination 6

Warmwater Metric

Fish tissues are averaged LMB age class 1 to 10 as specified on page 3-55 of the revised CMS.

Warren Pinnacle Consulting, Inc.

Page 40 of 42


-------
June 12, 2012

50

>s
T3
O
JD

0)
O

45

40

35

Projected Coldwater Fish Tissue (whole body) PCB
Concentration at the End of Model Projection

i Combination 1
Combination 6

Combination 2
Combination 7

i Combination 3
Combination 8

Combination 4 ¦ Combination 5
i Combination 9 Opt. 1 Combination 9 Opt. 2

Coldwater Metric

Fish tissues are averaged LMB age class 1 to 10 multiplied by 2 as specified on page 3-55 of the revised CMS.

Warren Pinnacle Consulting, Inc.

Page 41 of 42


-------
June 12, 2012

Reach

Combination
1

Combination
2

Combination
3

Combination
4

Combination
5

Combination
6

Combination
7

Combination
8

Combination
9 Opt. 1

Combination
9 Opt. 2

Fish PCB Concentration (mg/kg wet weight)

Reach 5A

7.44

7.44

0.27

0.26

0.27

0.17

0.32

3.53

0.26

0.26

Reach 5B

8.99

8.99

2.92

0.22

0.22

0.14

0.26

4.80

3.48

3.48

Reach 5C

7.37

7.37

1.85

0.16

0.17

0.12

0.18

4.69

0.82

0.82

Reach 5D

10.31

10.31

7.38

0.36

0.38

0.26

0.23

10.64

1.10

1.10

Reach 6

8.64

8.64

0.69

0.17

0.18

0.14

0.15

2.90

0.74

0.74

Reach 7 A

6.03

6.03

1.17

0.40

0.40

0.35

0.56

2.74

1.12

1.12

Reach 7B

5.56

5.56

2.10

1.60

0.39

0.10

0.29

3.22

0.67

0.67

Reach 7C

6.03

6.03

1.69

1.02

0.19

0.12

0.31

3.09

0.81

0.83

Reach 7D

5.60

5.60

1.40

0.81

0.76

0.71

0.88

2.67

1.37

1.38

Reach 7E

4.00

4.00

0.99

0.55

0.29

0.17

0.29

1.90

0.64

0.66

Reach 7F

3.38

3.38

0.82

0.47

0.44

0.37

0.51

1.56

0.82

0.82

Reach 7G

2.66

2.66

1.08

0.88

0.29

0.11

0.21

1.54

0.38

0.37

Reach 7H

2.73

2.73

0.70

0.43

0.39

0.35

0.43

1.27

0.69

0.69

Reach 8

2.82

2.82

1.31

0.27

0.20

0.13

0.23

1.74

0.37

0.37

BBD

0.113

0.113

0.035

0.018

0.016

0.014

0.011

0.020

0.022

0.022

LL

0.080

0.080

0.025

0.013

0.011

0.010

0.008

0.014

0.015

0.015

LZ

0.056

0.056

0.017

0.009

0.008

0.007

0.006

0.010

0.011

0.011

LH

0.054

0.054

0.017

0.009

0.008

0.007

0.005

0.010

0.010

0.010

Modeled Subreach Average Fish (Fillet) PCB Concentrations at End of Project Period

(Compare Revised CMS Table 8-3)

•	Average of largemouth bass fish ages 5-9;

•	Model endpoint concentrations after projection (autumn average);

•	Whole body concentrations divided by a factor of 5.0 to convert to fillet basis.

Warren Pinnacle Consulting, Inc.

Page 42 of 42


-------
ATTACHMENT B-9 PRELIMINARY DRAFT ARAR TABLES FOR

SED9/FP 4 MOD


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Federal ARARs

Clean Water Act,
National Ambient
Water Quality Criteria
for PCBs

National

Recommended Water
Quality Criteria:
2002, EPA-822-R-
02-047, USEPA,
Office of Water,
Office of Science and
Technology
(November 2002)

Freshwater chronic aquatic life criterion (based on
protection of mink): 0.014 ug/L.

Human health criterion based on human
consumption of water and organisms:

0.000064 ug/L.

Relevant and
appropriate; partially
waived.

The freshwater chronic aquatic
life criterion of 0.014 ug/L would
be met by the proposed
alternative. The human health
criterion based on human
consumption of water and
organisms of 0.000064 ug/L may
be waived in Massachusetts and
in portions of Connecticut on the
grounds that achievement of this
ARAR may be technically
impracticable, given that it may
not be achieved by any sediment
alternative in any reach in
Massachusetts or in one or more
of the Connecticut impoundments
under this alternative. Asa
modified Performance Standard
for this waived criterion, the
project would be required to meet
the Near-Term Biota
Performance Standard outlined in
Section II.b.l.a.(2).

State ARARs

Numeric

Massachusetts Water
Quality Criteria for
PCBs

Massachusetts
Surface Water
Quality Standards,
314 CMR 4.05(5)(e)

Same as federal water quality criteria.

Applicable; partially
waived.

Same as federal standard, see
above.

1


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Numeric Connecticut
Water Quality Criteria
for PCBs

Connecticut Water
Quality Standards

Same as federal water quality criteria.

Applicable; partially
waived.

Same as federal standard, see
above.

To Be Considered

Cancer Slope Factors

EPA Integrated Risk
Information System

Guidance values used to evaluate the potential
carcinogenic hazard caused by exposure to PCBs.

To be considered.

CSFs used to compute the
individual cancer risk resulting
from exposure to carcinogens in
site media.

Reference Doses

EPA Integrated Risk
Information System

Guidance values used to evaluate the noncancer
hazards associated with exposure to PCBs.

To be considered.

RfDs used to characterize human
health risks due to
noncarcinogens in site media.

PCBs: Cancer Dose
Response Assessment
and Application in
Environmental
Mixtures (EPA, 1996)

EPA/600/P-96/00 IF
(National Center for
Environmental
Assessment, Office
of Research and
Development,
September 1996)

Guidance describing EPA's reassessment
regarding the carcinogenicity of PCBs.

To be considered.

The guidance has been
considered in characterization of
site risks.

Guidelines for
Carcinogenic Risk
Assessment (EPA,
2005)

EPA/630/P-03/001F
(EPA Risk
Assessment Forum,
March 2005)

Framework and guidelines for assessing potential
cancer risks from exposure to pollutants and other
environmental agents.

To be considered.

Guidelines used to evaluate all
risk assessments on
carcinogenicity.

Supplemental
Guidance for
Assessing
Susceptibility from
Early-Life Exposure to
Carcinogens

EPA/630/R-03/003F
(EPA Risk
Assessment Forum,
March 2005)

Guidance on issues related to assessing cancer
risks associated with early-life exposures,
including an adjustment for carcinogens acting
through a mutagenic mode of action.

To be considered.

Guidelines used to evaluate all
risk assessments on
carcinogenicity in children.

2


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Massachusetts biota
consumption advisory

MA Department of
Public Health
(MDPH), Freshwater
Fish Consumption
Advisory List (2007)

Advises that the public should not consume any
fish from the Housatonic River from Dalton to
Sheffield due to PCBs; also includes frogs and
turtles.

To be considered.

This advisory would be
considered in reference to biota
consumption and actions to
reduce fish consumption risks,
including Institutional Controls.

Massachusetts
waterfowl

consumption advisory

MDPH, Provisional
Waterfowl
Consumption
Advisory (1999)

Advises that the public should avoid eating all
mallards and wood ducks from the Housatonic
River and its impoundments from Pittsfield to
Rising Pond.

To be considered.

This advisory would be
considered in reference to
waterfowl consumption and
actions to reduce waterfowl
consumption risks, including
Institutional Controls.

Connecticut fish
consumption advisory

Connecticut
Department of Public
Health (CDPH), 2006
Advisory for Eating
Fish from
Connecticut
Waterbodies

Establishes advisories on consuming fish from the
Housatonic River in Connecticut (above Derby
Dam), including Lakes Lillinonah, Zoar and
Housatonic, due to PCBs in fish. Advisories vary
by species, location, and group of consumers,
ranging from "do not eat" to "one meal per week."

To be considered.

This advisory would be
considered in reference to fish
consumption and actions to
reduce fish consumption risks,
including Institutional Controls.

Potential Location-Specific ARARs

Federal ARARs

Clean Water Act -
Section 404 and
implementing
regulations

33 USC 1344

33 CFR Parts 320-
323 (ACOE)
40 CFR Part 230
(EPA)

Under this requirement, no activity that adversely
affects a wetland shall be permitted if a practicable
alternative with lesser effects is available. If
activity takes place, impacts must be limited to the
maximum extent. Sets standards for
restoration/mitigation required as a result of
unavoidable impacts to aquatic resources.

Applicable.

Any remedial activities that
would alter wetlands, particularly
the excavation of contaminated
wetland soils and sediments,
would be conducted in
accordance with these standards.

3


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Rivers and Harbors
Act of 1899, Section
10

33 USC 403

U.S. Army Corps of Engineers approval is
generally required to excavate or fill, or in any
manner to alter or modify the course, location,
condition, or capacity of the channel of any
navigable water in the U.S.

Applicable.

This alternative includes work to
be performed in or near wetland
and floodplain areas. Any
remedy activities would comply
with the substantive requirements
of this provision. Remedy would
be coordinated with the U.S.
Army Corps of Engineers.

Fish and Wildlife
Coordination Act

16 USC 662(a)
40 CFR 6.302(g)

Any modification to a body of water requires
consultation with the U.S. Fish and Wildlife
Service and the appropriate state wildlife agency
to develop measures to prevent, mitigate, or
compensate for losses to fish and wildlife.

Applicable.

This alternative includes work to
be performed in or near wetland
and floodplain areas. There
would be consultation with the
U.S. Fish and Wildlife Service
and other federal and state
resource agencies.

National Historic
Preservation Act and
regulations

16 USC 470f
36 CFR Part 800

A federal agency must take into account the
project's effect on properties included or eligible
for inclusion in the National Register of Historic
Places.

Applicable.

A Phase 1A Cultural Resources
Assessment was prepared in
2008. In addition, any features
with potential historic/cultural
significance would be evaluated
during the remedial design phase.
If this alternative should impact
historic properties/structures
determined to be protected by
these standards, activities would
be coordinated with DOI.

4


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Archaeological and
Historic Preservation
Act

16 USC 469

When a Federal agency finds, or is notified, that
its activities in connection with a Federal
construction project may cause irreparable loss or
destruction of significant scientific, prehistorical,
historical, or archeological data, such agency shall
notify DOI. Such agency may request DOI to
undertake the preservation of such data or it may
undertake such activities. If DOI determines that
such data are significant and are being or may be
irrevocably lost or destroyed, it is to conduct a
survey and other investigation of the areas which
are or may be affected and recover and preserve
such data which are not being, but should be,
recovered and preserved in the public interest.

Applicable.

A Phase 1A Cultural Resources
Assessment was prepared in
2008. In addition, if during
remedial design or remedial
action, it is determined that this
alternative may cause irreparable
loss or destruction of significant
scientific, prehistorical,
historical, or archaeological data,
EPA would notify DOI and
comply with the processes in this
statute.

Executive Order
11988 (Floodplain
Management)

Executive Order

Federal agencies are required to avoid impacts
associated with the occupancy and modification of
a floodplain and avoid support of a floodplain
development whenever there is a practicable
alternative.

To be considered.

It is currently anticipated that this
project would not result in
occupancy and modification of
the floodplain. Any activities
would be conducted in
accordance with these standards.

Executive Order
11990 (Protection of
Wetlands)

Executive Order

Federal agencies are required to avoid adversely
impacting wetlands unless there is no practicable
alternative and the proposed action includes all
practicable measures to minimize harm to
wetlands that may result from such use.

To be considered.

Activities in the project would be
conducted in accordance with
these standards.

State ARARs

Massachusetts
Waterways Law and
implementing
regulations

MGL Ch. 91
310 CMR9.00

Regulates activities in waterways below the high
water mark.

Applicable.

This alternative has activities in
the river. Measures undertaken
would meet the substantive
environmental standards and
limit impacts.

5


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Massachusetts Clean
Water Act - water
quality certifications

314 CMR 9.01-9.08

For discharge of dredged or fill material: (a) no
discharge is permitted if there is a practicable
alternative to the proposed discharge that would
have less adverse impact on the aquatic ecosystem,
so long as the alternative does not have other
significant adverse environmental consequences;
(b) no discharge is permitted unless appropriate
and practicable steps have been taken which will
avoid and minimize potential adverse impacts to
bordering or isolated vegetated wetlands or land
under water; (c) no discharge to Outstanding
Resource Waters, other than specified exceptions;
(d) no discharge without a variance to particular
Outstanding Resource Waters listed in 9.06(4),
including certain vernal pools; (e) stormwater to
be controlled with best management practices; (f)
no discharge shall be permitted in rare
circumstances where the activity will result in
substantial adverse impacts to the physical,
chemical, or biological integrity of surface waters.

Applicable.

The proposed alternative has
activities, including soil/sediment
excavation and handling, that
may occur in and around site
wetlands. To the extent that this
alternative includes any discharge
of dredged or fill material, project
activities would comply with the
substantive environmental
standards of this statute and
regulations, including limiting
adverse effects and preserving,
mitigating, and restoring
disturbed areas.

Massachusetts
Wetlands Protection
Act and regulations

MGL c. 131, section
40

310 CMR
10.53(3)(q)

310 CMR 10.54-58
and 10.60

310 CMR 10.59

Assessment, monitoring, containment, mitigation,
and remediation of, or other response to, a release
or threat of release of oil and/or hazardous
material in accordance with 310 CMR 40 is
authorized as a limited project if it (a) has no
practicable alternatives that are consistent with the
MCP and that would be less damaging to resource
areas; (b) is designed and operated to avoid or,
where avoidance is not practicable, to minimize
impacts to resource areas.

Applicable

Any remedial activities that may
alter wetlands, including
excavation of contaminated soils
and sediments, would be
conducted in accordance with
these standards.

6


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Massachusetts Dam
Safety Standards

302 CMR 10.00

Regulations govern design and construction of
new and existing dams, and removal of existing
dams, and inspection of dams.

Potentially applicable

To the extent that these
regulations are applicable to a
Massachusetts dam for which
there is project activity, the
project would comply with these
regulations.

Massachusetts
Historical Commission
Act and Regulations

MGL c. 9, section
27C

950 CMR 71.07

If a project has an area of potential impact that
could cause a change in the historical,
architectural, archaeological, or cultural qualities
of a property on the State Register of Historic
Places, these provisions establish a process for
notification, determination of adverse impact, and
evaluation of alternatives to avoid, minimize or
mitigate such impacts.

Applicable to state
actions; relevant and
appropriate to state-
authorized actions
that have impact on
such properties.

If such properties are present in
the area of the response action in
this alternative, the project would
comply with these provisions.

Connecticut Dam
Safety Regulations

CGS 22a-401 to 22a-
411

Conn. Agencies
Regs. Section 22a-
409-2.

Regulations govern design and construction of
new and existing dams, and removal of existing
dams, and inspection of dams.

Potentially
applicable.

To the extent that these
regulations are applicable to a
Connecticut dam for which there
is project activity, the project
would comply with these
regulations.

Connecticut Inland
Wetlands and
Watercourses Act and
regulations

CGS 22a-36 et seq.
Conn. Agencies
Regs. Sec. 22a-39-4

Permit required for activities that remove material
from inland wetlands or watercourses; CT DEEP
allowed to issue general permit for minor activities
with minimal environmental impacts, defined to
include monitoring and sampling.

Relevant and
appropriate in
Connecticut.

To the extent that work in
Connecticut removes material
from inland wetlands or
watercourses, the project would
comply with this provision.

7


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Potential Action-Specific ARARs

Federal ARARs

Toxic Substances
Control Act (TSCA)
regulations on cleanup
of PCB remediation
waste

40 CFR 761.50
40 CFR 761.61

General requirements (761.50) and specific
options (761.61) for cleanup of PCB remediation
waste, including PCB-containing sediments and
soils. Options include self-implementing
provisions (not applicable to sediments) and risk-
based approval by EPA. Risk-based approval is
pursuant to 40 CFR 761.61(c) and requires
demonstration that cleanup method will not pose
an unreasonable risk of injury to health or the
environment.

Applicable.

Project would comply with these
provisions.

TSCA regulations on
storage of PCB
remediation waste

40 CFR 761.50
40 CFR 761.65
40 CFR 761.61(c)

General and specific requirements for storage of
PCB Remediation Waste. Regulations include
specific provisions for storage of PCB
Remediation waste in piles at the cleanup site or
site of generation for up to 180 days (761.65(c)(9).
Also allows for risk-based approval by EPA of
alternate storage method (j61.61(c), based on
demonstration that it will not pose an unreasonable
risk of injury to health or the environment.

Applicable.

Project would comply with these
provisions.

TSCA regulations on
discharge of PCB-
containing water

40 CFR 761.50(a)(3)

Prohibits discharge of water containing PCBs to
navigable waters unless PCB concentration is <3
mg/L or discharge is in accordance with NPDES
discharge limits.

Applicable.

Any discharge to navigable
waters must comply with this
provision.

TSCA regulations on
decontamination

40 CFR 761.79

Establishes decontamination standards and
procedures for removing PCBs from water,
organic liquids, and various types of surfaces.

Applicable.

To the extent the project involves
decontamination activities, this
provision must be complied with.

8


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Clean Water Act and
NPDES regulations

33 USC 1342
40 CFR 122
including, but not
limited to
122.3(d) and
122.44(a) & (e)

40 CFR 125.1-125.3

These standards include the provision that point
source discharge must meet technology-based
effluent limitations (including those based on best
available technology for toxic and non-
conventional pollutants and those based on best
conventional technology for conventional
pollutants) and effluent limitations and conditions
necessary to meet state water quality standards.

Applicable.

These standards would be
complied with if water from the
remedial action, such as from
dewatering or other processing of
sediment and wetland soils, is
discharged to surface waters.

Clean Water Act -
NPDES regulations
(stormwater
discharges)

40 CFR

122.26(c)(l)(ii)(C)
40 CFR 122.44(k)

Best management practices (BMPs) must be
employed to control pollutants in stormwater
discharges during construction activities.

Applicable.

These standards would be
complied with during
construction activities.

Endangered Species
Act and regulations

16 USC 1536(a) -(d)
40 CFR 6.302(h)
50 CFR Part 402,
Subparts A & B

These provisions state that a federal agency action
is not likely to jeopardize the continued existence
of a listed threatened or endangered (T&E) species
or result in destruction or adverse modification of
critical habitat, unless an exemption is granted.

Applicable.

These provisions would be
complied with in regard to
federally-listed T&E species and
their critical habitat.

State ARARs

Massachusetts Clean
Water Act - water
quality certification
regulations

314 CMR9.01 -9.08

This includes provisions dealing with discharge of
dredged or fill material: (a) no such discharge is
allowed if there is a practicable alternative with
less adverse impact on aquatic ecosystem; (b)
appropriate and practicable steps must be taken to
avoid and minimize adverse effects on land under
water and on bordering or isolated vegetated
wetlands, including 1:1 restoration or replication
of such wetlands (unless waived); (c) there must
be no discharge that would adversely affect
estimated habitat of rare wildlife species under the
Wetlands Protection Act or to certain designated
"Outstanding Resource Waters," including
certified vernal pools, unless a variance is
obtained; (d) stormwater discharges must be

Applicable.

The alternative includes
activities, including soil/sediment
excavation and handling that may
occur in and around site
wetlands. To the extent that this
alternative would include any
discharge of dredged or fill
material, project activities would
comply with the substantive
environmental standards of this
statute and regulations, including
limiting adverse effects and
preserving, mitigating, and
restoring disturbed areas.

9


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs





controlled with BMPs; and (e) there must be no
substantial adverse impacts to physical, chemical,
or biological integrity of surfaced waters.
For dredging and dredged material management:
(a) no dredging is allowed if there is a practicable
alternative with less adverse impact on aquatic
ecosystem; (b) appropriate and practicable steps
must be taken to avoid, minimize, or mitigate
adverse effects on land under water; (c) dredging
must be conducted to meet performance standards
designed to minimize impacts on the aquatic
ecosystem and protect human health; and (d)
placement of dredged material in an intermediate
facility for sediment management (dewatering,
processing, etc.) prior to disposal or reuse must
meet certain requirements, including requirements
governing method of placement/storage of
dredged material and siting criteria.





Massachusetts Clean
Water Act and
Wetlands Protection
Act - stormwater
management standards

310 CMR

10.05(6)(k)

314 CMR 9.06(6)(a)

Projects subject to regulation under the Wetlands
Protection Act or that involve discharge of
dredged or fill material must incorporate
stormwater BMPs to attenuate pollutants in
stormwater discharges, as well as to provide a
setback from receiving waters and wetlands, in
accordance with 10 specified stormwater
management standards.

Applicable.

The project would comply with
stormwater requirements.

10


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Massachusetts
Endangered Species
Act (MESA) and
regulations

MGLc. 131A
321 CMR 10.00,
Parts I, II, and V

A proposed activity in mapped Priority Habitat for
a state-listed rare, threatened, endangered species
or species of special concern, or other area where
such a species has occurred may not result in a
"take" of such species, unless for conservation and
management purposes provided there is a long-
term net benefit to the conservation of the species.

Applicable.

The project would take place in
priority habitat for one or more
state-listed species. In
implementing the project,
impacts to state-listed species
would be avoided wherever
possible. To the extent that
impacts cannot be avoided, any
impacts would be insignificant
and a plan for conservation and
management would be
implemented. The processes
outlined as part of the alternative
for work in Core Habitat areas
were developed in consultation
with the Commonwealth and
would satisfy these requirements.



321 CMR 10.00, Part
IV

Projects that will alter a designated Significant
Habitat must be reviewed to ensure that they will
not reduce the viability of the habitat to sustain an
endangered or threatened species.



There are no designated
Significant Habitats in the project
area.

11


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Massachusetts
hazardous waste
regulations on
identification of
hazardous waste

310 CMR 30.100

Establishes criteria and lists for determining
whether a waste is a hazardous waste under state
law.

Wastes that contain PCBs > 50 mg/kg (which are
listed wastes) are exempt from the state hazardous
waste management regulations so long as they are
managed in compliance with EPA's TSCA
regulations (40 CFR Part 761) (see 310 CMR
30.501(3)(a)).

The state hazardous waste management
regulations also exempt dredged material (even if
it constitutes non-PCB state hazardous waste) that
is temporarily stored at an intermediate facility
(pursuant to 314 CMR 9.07(4)) and managed in
accordance with a state water quality certification
and §404 requirements under the Clean Water Act
(see 310 CMR 30.104(3)(f)).

Applicable.

Wastes that contain PCBs at
levels greater than or equal to 50
mg/kg would be managed in
compliance with EPA's TSCA
regulations (40 CFR Part 761).
Dredged material stored
temporarily would be managed in
accordance with the substantive
state and federal requirements.

Massachusetts air
pollution control
regulations

310 CMR 7.09

Prohibits person engaged in dust-generating
activities from creating condition of air pollution,
defined as air concentrations that would cause a
nuisance, be injurious or potentially injurious to
human or animal life, vegetation, or property, or
unreasonably interfere with comfortable
enjoyment of life and property or conduct of
business.

Applicable.

Project would comply with this
provision during any dust-
generating activities.

Massachusetts
Contingency Plan,
Method I Soil
Standards

310 CMR 40.0970
through 40.0979

Establishes soil cleanup standards, including those
for residential or unrestricted use ("S-l
Standards").

Potentially
Applicable or
Potentially Relevant
and Appropriate.

Performance Standards based
upon unrestricted use or
residential use in Massachusetts
would be based upon the S-l
Method I standard.

12


-------
PRELIMINARY DRAFT

SUMMARY OF APPLICABLE OR RELEVANT AND APPROPRIATE REQUIREMENTS (ARARs) FOR ALTERNATIVE

SED 9/FP 4 MOD

(Preliminary list of ARARs, additional federal or state ARARs may be added prior to any remedy proposal)

Statute/Regulation

Citation

Synopsis of Requirements

Status

Action(s) to be Taken to
Achieve ARARs

Connecticut
Endangered Species
Act

Conn. Gen. Stat. 26-
303 through 26-316

Requires state agency to: (a) ensure that any action
authorized or performed by it does not threaten the
continued existence of a listed endangered or
threatened species or result in destruction or
adverse modification of habitat essential to such
species, unless an exemption is granted; and (b)
take all reasonable measures to mitigate any
adverse impacts of the proposed action on such
species or habitat. Prohibits "taking" of
endangered or threatened species, except where
State determines that a proposed action would not
appreciably reduce likelihood of survival or
recovery of the species.

Potentially
applicable.

To the extent that any project
activity takes place in
Connecticut, the project would
comply with these regulations.

Connecticut fisheries
and game laws

Conn. Gen. Stat. 26-
60

Authorizes CT DEEP to issue permits to properly
accredited persons for sampling of fish,
crustaceans, and wildlife for educational and
scientific purposes, with CT DEEP to determine
number, species, area, and method of collection.

Relevant and
appropriate.

For any sampling performed in
Connecticut, the project would
comply with substantive
provisions of this statute.

Connecticut
Remediation Standards
Regulations, Direct
Exposure Criteria for
Soil

Conn. Gen. Stat. 22a-
133k-l through K-3
Appendix A

Establishes soil cleanup standards, including those
for residential or unrestricted use ("Residential
Criteria").

Potentially
Applicable or
Potentially Relevant
and Appropriate.

Performance Standards based
upon unrestricted use or
residential use in Connecticut
would be based upon the
Residential Direct Exposure
Criteria.

To Be Considered

TSCAPCB Spill
Cleanup Policy

40 CFR Part 761,
Subpart G

Policy used to determine adequacy of cleanup of
spills resulting from the release of materials
containing PCBs at concentration of 50 mg/kg or
greater.

To be considered.

To the extent that such a spill
occurs in the project, this policy
would be considered in the
response.

13


-------
ATTACHMENT B-10 COST ASSUMPTIONS MEMORANDUM FOR

SED9/FP 4 MOD


-------


SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

To: Scott Campbell	Date: 7/25/12

From: Tony Delano

Re: GE/Housatonic River Site - Rest of River, SED 9/FP 4 MOD/TD 1 RR Cost Assumptions
Memorandum

The purpose of this memorandum is to summarize the assumptions used to develop costs for SED 9/FP 4
MOD and TD 1 RR, for the Housatonic Rest of River Project. In general, costs provided by GE supporting
its October 2010 Revised CMS (RCMS) were used as a baseline for all construction costs for the SED 9/FP
4 MOD portion of the alternative. Modifications to those costs resulting from changes in quantities,
materials or methods were then made to provide the best possible basis for comparison to the
alternatives presented in the RCMS.

Table 1 summarizes the quantities assumed in the cost estimate by reach for SED 9/FP 4 MOD.

Table 1 - Summary of Volumes for SED 9/FP 4 MOD



Vo

umes (cubic yard [CY])

Area
(acre)

Reach

Riverbed
cut (feet)

Riverbed

Riverbank

Pond/
Backwater

Floodplain

Total



5A

2.5

168,000

25,000





193,000

42

5B

none,
except for
500 CY

500

500





1,000



5C

2

186,000







186,000

57

Backwaters







95,000



95,000

61.5

Woods Pond







285,000



285,000

60

7

1 to 1.5





84,000



84,000

79

8

1 to 1.5





71,000



71,000



FP

1 or 3







75,000

75,000

45



Totals

354,500

25,500

535,000

75,000

990,000

344.5

The following sections provide additional details regarding the costing assumptions for the remediation
and TD components, respectively, of this alternative:

Page 1


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

SED 9/FP 4 MOD Component

Item numbers (e.g., 1.0) refer to items numbers in the GE detailed cost information, which is the basis
for all costs presented in the RCMS. In some cases, a different item number was used for the SED item
and corresponding FP item for the same category of cost (e.g., Dewatering). To consolidate the
explanation of assumptions, these items have been combined in this section.

Item 1.0 Pre-Design Investigation

•	For all reaches, this cost was estimated at 5% of all on-site construction-related costs.

Item 2.0 Mobilization/Demobilization

•	For all reaches, this cost was estimated at 5% of all on-site construction-related costs.

Item 3.0 Construction of Staging Areas/Access Roads

•	All costs were estimated on a reach-specific basis using a unit cost of $/acre or lump-sum pricing
(Reaches 5D and 6 only) based upon unit costs for GE's SED 9 and FP 8 alternatives as presented in
the RCMS. To determine a total estimated cost, these unit costs were multiplied by the estimated
area of staging areas and access roads developed to optimize staging area sizes and locations.

•	The costs for this line item are shown by reach in Table 2.

Table 2 - Assumptions and Costs for Item 3.0 Construction of Staging Areas/Access Roads









SED 9/FP4





GE SED 9

GE SED 9

GE SED 9

MOD

SED 9/FP 4 MOD

Reach

(acre or LS)

($/acre or LS)

Total

(acre)

Total

5A

40.6

$156,158

$6,339,000

27.5

$4,300,000

5B

15.5

$194,323

$3,012,000

0.5

$100,000

5C

12.5

$231,520

$2,894,000

11

$2,520,000

5D

LS

$247,0001

$247,000

—

$247,000

6

LS

$496,0001

$496,000

—

$496,000

7

7.6

$223,816

$1,701,000

7

$1,500,000

8

3.2

$236,563

$757,000

5

$1,278,000

FP 4 MOD









SED 9/FP 4





GE FP 8 Basis

GE FP 8

GE FP 8

MOD

SED 9/FP 4 MOD

Reach

(acre)

($/acre)

Total

(acre)

Total

5A

17.6

$28,580

$503,000

6.23

$179,000

5B

9.62

$24,324

$234,000

2.45

$60,000

5C

7.59

$34,783

$264,000

1.26

$43,000

6

0.51

$13,725

$7,000

0.81

$12,100

7

0

$0

$0

0

$0

Notes:

1Lump sum pricing
LS = lump sum

Page 2


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

Item 4.0 Sheeting - Not Used
SED Item 5.0/FP Item 4.0 Dewatering

•	Dewatering estimates by reach were based upon the number of days required to complete
excavation or dredging and the unit costs determined from GE's SED 9 and FP 8 alternatives as
presented in the RCMS.

•	The costs for this line item are shown by reach in Table 3.

Table 3 - Assumptions and Costs for Dewatering









SED 9/





GE SED 9

GE SED 9

GE SED 9

FP 4 MOD

SED 9/FP 4 MOD

Reach

(day)

($/day)

Total

(day)

Total

5A

536

$7,015

$1,876,000

386

$1,880,000

5B

320

$25,534

$8,171,000

4

$100,000

5C

378

$23,294

$8,805,000

451

$10,503,000

5D

264

$4,852

$1,281,000

230

$1,110,000

6

385

$5,068

$1,951,000

450

$2,270,000

7

0

0

0

0

0

8

129

$6,341

$818,000

129

$820,000









SED 9/





GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

SED 9/FP 4 MOD

Reach

(day)

($/day)

Total

(day)

Total

5A

342

$395

$135,000

171

$68,000

5B

150

$420

$63,000

70

$30,000

5C

176

$398

$70,000

47

$19,000

6

5

$400

$2,000

4

$1,000

7

36

$360

$13,000

10

$4,000

Page 3


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

SED Item 6.0/FP Item 5.0 Water Treatment

•	Water Treatment estimates were based upon the number of days required to complete
excavation or dredging and the unit rate determined from GE's SED 9 alternative as presented in
the RCMS.

•	No water treatment costs were included for the FP 4 MOD portion of the alternative based on
the assumption in the RCMS for GE's FP 8 that there would be no water treatment costs.

•	The costs for this line item are shown by reach in Table 4.

Table 4 - Assumptions and Costs for Water Treatment









SED 9/





GE SED 9

GE SED 9

GE SED 9

FP 4 MOD

SED 9/FP 4 MOD

Reach

(Days)

($/day)

Total

Days

Total

5A

536

$10,709

$2,875,000

386

$4,140,000

5B

320

$2003

$641,000

4

$10,000

5C

378

$32,373

$12,237,000

451

$14,597,000

5D

264

$2,674

$706,000

230

$615,000

6

385

$3,062

$1,179,000

450

$1,378,000

7

305

$2,669

$814,000

305

$814,000

8

129

$2,658

$345,000

129

$345,000

Item 7.0 Debris Removal

•	Debris Removal estimates were based upon the number of acres requiring remediation and the
unit rate determined from GE's SED 9 total costs for debris removal and number of acres as
presented in the RCMS.

•	The costs for this line item are shown by reach in Table 5.

Table 5 - Assumptions and Costs for Debris Removal









SED 9/





GE SED 9

GE SED 9

GE SED 9

FP 4 MOD

SED 9/FP 4 MOD

Reach

(acre)

($/acre)

Total

(acre)

Total

5A

42

$3,500

$147,000

42

$147,000

5B

27

$7,074

$191,000

1

$7,000

5C

57

$7,000

$399,000

57

$399,000

5D

68

$7,000

$476,000

61.5

$430,000

6

60

$7,000

$420,000

60

$420,000

7

20

$6,600

$132,000

20

$132,000

8

41

$7,073

$290,000

41

$290,000

Page 4


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

SED Item 8.0/FP Item 6.0 Excavation

•	Excavation Cost estimates were based upon the number of cubic yards of material requiring
remediation and the unit rate determined from GE's SED 9 and FP 8 total costs for excavation
and number cubic yards as presented in the RCMS.

•	Because the volume of floodplain excavation has not been defined by reach for FP 4 MOD, the
proportions of material to be excavated within each reach were estimated by using the volumes
in GE's FP 4 as an approximation. The percentage of the total volume in each reach for FP 4 was
applied to the total volume of 75,000 cubic yard (CY) for FP 4 MOD to develop reach-specific
volume estimates.

•	The costs for this line item are shown by reach in Table 6.

Table 6 - Assumptions and Costs for Excavation



GE SED 9

GE SED 9

GE SED 9

SED 9 MOD

SED 9 MOD

Reach

(CY)

($/CY)

Total

(CY)

Total

5A

159,000

$34

$5,414,000

193,000

$6,700,000

5B

98,000

$120

$11,745,000

1,000

$120,000

5C

156,000

$83

$12,899,000

186,000

$15,370,000

5D

109,000

$65

$7,091,000

95,000

$6,180,000

6

244,000

$45

$10,939,000

285,000

$12,770,000

7

84,000

$73

$6,112,000

84,000

$6,110,000

8

71,000

$63

$4,499,000

71,000

$4,500,000



GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

FP 4 MOD

Reach

(CY)

($/CY)

Total

(CY)

Total

5A

85,500

$58

$4,973,000

42,500

$2,474,000

5B

37,500

$62

$2,318,000

17,400

$1,074,000

5C

43,800

$59

$2,576,000

11,600

$682,000

6

1,202

$72

$86,000

972

$70,000

7

9,039

$53

$475,000

2,519

$133,000

SED Item 9.0/FP Item 7.0 Backfill Material Placement

•	Backfill Material Placement estimates were based upon the number of cubic yards of backfill
material requiring placement and the unit rate determined from GE's SED 9 total costs for
backfill and number of cubic yards as presented in the RCMS.

•	Where a habitat layer is specified as part of the backfill cross-section, a premium of $1 was
added to the $/CY estimated price to account for additional costs associated with meeting the
specification of the grain sizes required for the habitat layer. The habitat layer is included in
Reaches 5A, 5B, 5C, 5D, and 6.

Page 5


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

•	For Reach 5B, a 3-inch layer of activated carbon was assumed to be broadcast over the entire
surface of the riverbed. It was assumed that activated carbon cost is $1.50/pound and would be
placed at a rate of 30,000 pounds per acre over the 27 acres of Reach 5B. Because this item was
not included in GE's estimates, a new method of installation for this material had to be
developed. A vortex spreader would be used from a boat, with support on land for handling the
activated carbon. Estimated productivity for this operation is 1 acre per week, with a total of 27
acres requiring coverage. The work crew is assumed to consist of seven workers with three on
the boat, two hauling material, and two managing the stockpile. Application of the activated
carbon in this manner would require approximately 27 weeks to complete.

•	The costs for this line item are shown by reach in Table 7.

Table 7 - Assumptions and Costs for Backfill Material Placement

Reach

GE SED 9
(CY)

GE SED 9
($/ CY)

GE SED 9
Total

SED 9/
FP 4 MOD
(CY)

SED 9
MOD
($/CY)

SED 9/FP 4 MOD
Total

5A

134,000

$64

$8,563,000

193,000

$65

$12,530,000

5B

88,000

$94

$8,265,000

1,000

$95

$95,000

5B (activated carbon only)

1,500

$1,179

$1,769,000

5C

156,000

$89

$13,960,000

186,000

$90

$16,830,000

5D

114,000

$84

$9,537,000

95,000

$85

$8,100,000

6

96,800

$102

$9,832,000

96,800

$103

$9,930,000

7

83,800

$90

$7,581,000

83,800

$90

$7,580,000

8

71,700

$84

$5,956,000

71,700

$84

$5,950,000

Reach

GE FP 8
(CY)

GE FP 8
($/ CY)

GE FP 8
Total

FP 4 MOD
(CY)

FP 4 MOD
Total

5A

85,500

$53

$4,506,000

42,500

$2,474,000

5B

37,500

$55

$2,049,000

17,400

$949,000

5C

43,800

$53

$2,339,000

11,600

$620,000

6

1,202

$70

$84,000

972

$68,000

7

9,039

$51

$462,000

2,519

$128,000

Page 6


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

Item 10.0 Bank Stabilization

Reach 5A

•	Unit rates for bank stabilization presented in the RCMS were used to develop a generic cost per
linear foot of riverbank for SED 9/FP 4 MOD. In terms of estimating cost, it was assumed that
Reach 5A bank stabilization would consist of 32% bioengineering (with a riprap toe), 1% riprap,
with the remaining 67% having no action taken. These assumptions yield a unit rate of $218.97
per linear foot for an assumed 33% of the riverbanks, or approximately 17,400 feet (ft) (3.35
miles). Based on these unit rates and assumptions, the total cost for riverbank stabilization is
$3,820,000.

Reach 5B

•	Bank stabilization was assumed to be required to restore the area of bank where 1,000 cubic
yards (CY) of material is proposed to be removed as part of SED 9/FP 4 MOD. To estimate a cost
for this restoration, the unit rate for stabilization in Reach 5A was converted to a per square foot
basis by assuming the average width of bank stabilization is 10 ft, yielding a unit rate of
$13/square foot (sq ft). This was then applied to the area required to excavate 1,000 CY of
material at an excavation depth of 1 ft, which is approximately 9,000 sq ft, resulting in a bank
stabilization cost of $110,000.

SED 9 Item 11.0/FP 4 MOD Item 8.0 Site Restoration

•	Site restoration estimates were based on the number of acres requiring restoration and the unit
rate as determined from GE's SED 9 total costs for site restoration and the number of acres
restored as presented in the RCMS.

•	Determination of the unit rate for site restoration assumes a certain proportion of the following
types of habitats: forested wetland habitat, shrub and shallow emergent habitat, backwater and
deep emergent marsh, vernal pool, grassy upland, and forested upland. Each of these habitats
has a different restoration cost per acre. The blended rate used in the RCMS is applicable to the
mix of habitats determined by GE for FP 8. The actual mix of habitats is unknown because both
the road and staging area network and the removal areas for the floodplains is currently
unknown. Therefore, the costs shown in Table 8 are also based on the assumption that the
proportions of the various habitat types for FP 4 MOD are the same as determined by GE for FP
8 in the RCMS.

•	The costs for this line item are shown by reach in Table 8.

Page 7


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

Table 8 - Assumptions and Costs for Site Restoration









SED 9/FP 4





GE SED 9

GE SED 9

GE SED 9

MOD

SED 9/FP 4 MOD

Reach

(acre)

($/acre)

Total

(acre)

Total

5A

34.74

$32,326

$1,120,000

33.7

$1,090,000

5B

19.46

$31,449

$612,000

1

$30,000

5C

16.48

$35,133

$579,000

13.2

$463,000

5D

0

0

0

0

0

6

0

0

0

0

0

7

7.43

$29,610

$220,000

7

$208,000

8

5.44

$29,596

$161,000

5.4

$160,000

Reach

GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

FP 4 MOD



(acre)

($/acre)

Total

(acre)

Total

5A

52.0

$37,858

$1,969,000

32.6

$1,233,000

5B

22.9

$42,582

$973,000

16.1

$685,000

5C

26.6

$38,924

$1,035,000

11.3

$441,000

6

0.71

$40,845

$29,000

0.12

$5,000

7

5.6

$27,143

$152,000

2.8

$76,000

SED 9 Item 12.0/FP 4 MOD Item 9.0 Transportation and Disposal

•	Transportation and Disposal costs as presented in this section are related to the disposal of
staging area materials, which are assumed to be non-Toxic Substances Control Act (TSCA), and
are directly related to the size of the total area of staging areas and access roads. Unit rates are
therefore based upon the GE's SED 9 total costs as presented in the RCMS for site restoration
and the number of acres restored.

•	No additional costs were included for FP 4 MOD, except those for Reach 5B, because all roads
and staging areas would be removed under the SED portion of this alternative, except those for
Reach 5B, for which very little sediment removal is proposed.

•	The cost added for FP 4 MOD in Reach 5B is $450,000, based upon a unit rate of $183,000 per
acre for approximately 2.5 acres of roads and staging areas.

•	The costs for this line item are shown by reach in Table 9.

Table 9 - Assumptions and Costs for Transportation and Disposal









SED 9/FP 4





GE SED 9

GE SED 9

GE SED 9

MOD

SED 9/FP 4 MOD

Reach

(acre)

($/acre)

Total

(acre)

Total

5A

34.74

$304,499

$14,888,000

33.7

$10,260,000

5B

19.46

$192,240

$3,741,000

1

$190,000

5C

16.48

$414,684

$6,834,000

13.2

$5,470,000

5D

0

0

0

0



6

0

0

0

0



7

7.43

$298,789

$2,220,000

7

$2,100,000

8

5.44

$320,404

$1,743,000

5.4

$1,730,000

Page 8


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

SED 9 Item 13.0/FP 4 MOD Item 10.0 Topographic Surveys

•	GE originally based surveying costs upon the number of months to complete remediation of a
reach or a lump-sum cost per survey. Unit rates were therefore based on the duration of SED 9
or number of surveys required for SED 9, as specified in GE's RCMS. SED 9 MOD costs were
based on the same units, whether duration or lump sum.

•	The duration for Reach 5B was adjusted downward for SED 9/FP 4 MOD to reflect the limited
excavation and backfill work proposed.

•	For FP 4 MOD costs, surveying costs were based upon duration of construction.

•	The costs for this line item are shown by reach in Table 10.

Table 10 - Assumptions and Costs for Topographic Surveys



GE SED 9

GE SED 9



SED 9/FP 4 MOD

SED 9/FP 4



(duration

($/month or

GE SED 9

(duration or

MOD

Reach

or surveys)

each)

Total

surveys)

Total

5A

34 months

$28,163/month

$958,000

42 months

$1,180,000

5B

2 each

$35,000 each

$70,000

2 each @ $5,000

$10,000

5C

2 each

$35,000 each

$70,000

2 each

$70,000

5D

2 each

$35,000 each

$70,000

2 each

$70,000

6

2 each

$35,000 each

$70,000

2 each

$70,000

7

2 each

$35,000 each

$70,000

2 each

$70,000

8

2 each

$35,000 each

$70,000

2 each

$70,000



GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

FP 4 MOD

Reach

(month)

($/month)

Total

(month)

Total

5A

31

$28,065

$870,000

15

$433,000

5B

15

$27,067

$406,000

7

$188,000

5C

16

$28,438

$455,000

4

$120,000

6

0.41

$31,342

$13,000

0.34

$11,000

7

2.9

$28,651

$84,000

0.82

$23,000

Page 9


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

SED 9 Item 14.0/FP 4 MOD Item 11.0 Environmental Monitoring

•	In the RCMS, environmental monitoring costs were based upon the number of months to
complete remediation of a reach. The SED 9 MOD unit rates were therefore based upon the GE
SED 9 duration as presented in the RCMS.

•	For SED 9 MOD, the duration for Reach 5B was adjusted downward to reflect the limited
excavation and backfill work proposed.

•	The costs for this line item are shown by reach in Table 11.

Table 11 - Assumptions and Costs for Environmental Monitoring











SED 9/FP4



GE SED 9

GE SED 9

GE SED 9

SED 9/FP 4 MOD

MOD

Reach

(month)

($/month)

Total

(month)

Total

5A

34

$41,336

$1,405,000

42

$1,740,000

5B

24

$42,696

$1,025,000

1

$50,000

5C

28

$42,731

$1,196,000

36

$1,538,000

5D

28

$41,786

$1,170,000

24

$984,000

6

28

$41,518

$1,163,000

30

$1,264,000

7

31

$42,16 0

$1,307,000

31

$1,307,000

8

13

$46,154

$600,000

13

$46,154



GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

FP 4 MOD

Reach

(month)

($/month)

Total

(month)

Total

5A

31

$27,806

$862,000

15

$429,000

5B

15

$27,267

$409,000

7

$189,000

5C

16

$28,000

$448,000

4

$119,000

6

0.41

$45,894

$19,000

0.34

$16,000

7

3

$27667

$83,000

0.8

$23,000

SED 9 Item 15.0 Annual O&M/Monitored Natural Recovery; Item 12.0 FP 4 MOD Annual O&M

•	For SED 9, no changes were made to GE's assumptions and costs are therefore identical.

•	For FP 4 MOD, GE's FP 8 costs as presented in the RCMS were prorated by the total estimated
restoration area. As was the case in the RCMS, annual O&M was assumed to continue for 5
years following completion of construction.

•	The costs for this line item are shown by reach in Table 12.

Table 12 - Assumptions and Costs for Annual O&M



GE FP 8

GE FP 8

GE FP 8

FP 4 MOD

FP 4 MOD

FP 4 MOD

Reach

(acre)

($/year)

Total

(acre)

($/year)

Total

5A

52

$177,000

$885,000

33

$111,000

$555,000

5B

23

$78,000

$390,000

16

$55,000

$275,000

5C

27

$92,000

$460,000

11

$39,000

$195,000

6

0.71

$15,000

$75,000

0.12

$3,000

$15,000

7

5.6

$20,000

$100,000

2.8

$10,000

$50,000

Page 10


-------
ygpcaKa

SED 9/FP 4 MOD/TD 1 RR Cost Assumptions Memorandum

TD 1 RR Component

Costs for railroad infrastructure, railroad transport, and disposal at rail-ready facilities were developed in
2011 and re-confirmed in 2012.

Railroad Infrastructure-Approximately $300,000 in railroad infrastructure upgrades would be needed
to construct several spurs to provide access to a transfer and loading station. The costs associated with
construction of rail spurs to access the rail staging areas from the main railroad line are included in the
TD 1 RR capital costs. The construction cost associated with the staging areas necessary to support the
rail spurs and loading areas is included in the SED 9 Reach 5A staging area costs. GE confirmed in the
RCMS that the existing track from Housatonic, MA, to Pittsfield, MA, was of sufficient design to handle
rail cars loaded with up to approximately 110 tons of material. No costs for upgrade of the existing track
were included based on GE's analysis.

Transportation and Disposal Pricing-Two rail-ready facilities, one for TSCA and one for non-TSCA, were
considered. Table 13 summarizes the costs for both transportation to and disposal at these facilities:

Table 13 - Summary of Transportation and Disposal Costs forTD-1 RR



Non-TSCA

TSCA

TD-1 RR

Facility

Model City

EQ

Type

Non-TSCA

TSCA

Location

Niagara Falls

Michigan

Rail Transport Price ($/ton)

$43

$110

Disposal Price

$55

$85

Total T&D

$98

$195

CMS Truck Transport

GE TD 1 Transport

$56

$130

GE TD 1 Disposal

$44

$90

Summary of Overall Costs

Table 14 summarizes the cost by reach for SED 9/FP 4 MOD and TD 1 RR.

Table 14 - Summary of Overall Costs by Reach

Reach

Total Costs

Reach 5

$223,660,000

Reach 6 - Wood's Pond

$42,805,000

Reach 7

$29,860,000

Reach 8 - Rising Pond

$23,200,000

Long-Term Monitoring

$8,733,000

Total Cost of Alternative before T&D

$328,000,000

Transportation and Disposal (TD 1 RR)

$228,000,000

Total cost of Alternative SED 9/FP 4 MOD/TD 1 RR

$556,000,000

The total present-worth of the alternative is $412,000,000, which includes $229,000,000 for non-TD
costs and $183,000,000 for the TD portion of the alternative.

Page 11


-------
APPENDIX C
GENERAL ATTACHMENTS


-------
ATTACHMENT C-1 NATIONAL REMEDY REVIEW BOARD
RECOMMENDATIONS FOR THE HOUSATONIC RIVER, REST OF

RIVER SITE


-------


UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

WASHINGTON, D C. 20460



OFFICE OF
SOLID WASTE AND
EMERGENCY RESPONSE

October 20, 2011

MEMORANDUM

SUBJECT: National Remedy Review Board Recommendations for the Housatonic River, Rest of
River Site

The National Remedy Review Board (the Board) and the Contaminated Sediments Technical Advisory
Group (CSTAG) have completed their review of the proposed cleanup action for the Housatonic River,
Rest of River site, in Pittsfield, MA. This memorandum documents the Board's and CSTAG's advisory
recommendations.

Context for Review

The Administrator established the Board as one of the October, 1995 Superfund Administrative Reforms
to help control response costs and promote consistent and cost-effective remedy decisions. The Board
furthers these goals by providing a cross-regional, management-level, "real time" review of high cost
proposed response actions prior to their being issued for public comment. The Board reviews all
proposed cleanup actions that exceed its cost-based review criteria.

The Board review is intended to help control remedy costs and to promote both consistent and cost-
effective decisions. Consistent with CERCLA and the National Oil and Hazardous Substances Pollution
Contingency Plan (NCP), in addition to being protective, all remedies are to be cost-effective. The
Board considers the nature of the site; risks posed by the site; regional, state, tribal and potentially
responsible party (PRP) opinions on proposed actions; the quality and reasonableness of the cost
estimates; and any other relevant factors or program guidance in making our advisory recommendations.

FROM: Amy R. Legare, Chair
National Remedy Revi

Contaminated Sedimei

Stephen J. Ells, Chair

TO:

James T. Owens III, Director

Office of Site Remediation and Restoration

U.S. EPA Region 1

Purpose

Internet Address (URL) • http;//www.epa.gcv
Recycled/Recyclable • Printed with Vegetable Oil Based Inks on 100% Postconsumer. Process CWonne Free Recycled Paper


-------
I he overall goal of the review is to ensure sound decision making consistent with current law,

regulations- and guidance.

Generally, the Board makes the advisory recommendations to the appropriate regional division director.
Then, the region will include these recommendations in the administrative record for the site, hpicnllv
before it issues the proposed cleanup plan Tor public comment. Whi.e the region is expected to give the

Board's recommendations substantial weight, other important factors, such as subsequent public
comment or technical analyses of response options, may influence the region's final remedy decision.
The IBoard expects the regional division director to respond in writing to its recommendations within a
reasonable period of time, noting in particular how the recommendations influenced the proposed
cleanup decision, including any effect on the estimated cost of the action. Although the Board's
recommendations are to be given substantial weight, the Board does not change the Agency's current
delegations or alter the public's role in site decisions; the region has the final decision-making authority.

Office of Solid Waste and Emergency Response (GSWER) Directive 92X5.6-08, February 2002.
Principles for Managing Contaminated Sediment Risks at Hazardous Waste Sites, established the
CSTAG as a technical advisory group to -...monitor the progress of and provide advice regarding a
small number of large, complex, or controversial contaminated sediment Superlund sites...." One main
purpose of the CSTAG is to guide Regional site project managers on how to appropriate!} manage then-
sites throughout the cleanup process in accordance with the 1 1 risk management principles set forth in
the OSWKR Directive. HP A decided not to have a separate technical review by the C'S'l AG per OSWER
Directive No. 9285.6-2U. September 2009. Changes io the Roles and Responsibilities of the
Contaminated Sediments Technical Advisory Group fCSTAGi. but instead elected to have a combined
NRRB CSTAG meeting for this site, litis joint meeting format is the approach FPA plans to take in the
future at CSTAG sites.

Overview of the Proposed Action

The preferred alternative outlined by the Region for the Board and CS 1 AG included the following

general actions:

•	Sediment removal followed by capping in Reach 5, Reach 6 (Woods Pond). Reach 7
impoundments, and Reach 8 (Rising Pond) to meet human health and ecological risk reduction
goals; monitored natural recovery (MNR) in the remaining portions of Reach 7 as well as in
Reaches 9 through 16.

•	Excavation and backfill of contaminated floodplain soils exceeding human health and
environmental risk-based thresholds.

•	Disposal of excavated/dredged river sediment, bank and floodplain soil off-site in properly
permitted and maintained TSCA and notv'l SCA landfills

•	Operation, monitoring, and maintenance: five-vear reviews; and, institutional controls (ICs).

Housatonic ROR

Final 10/20/11


-------
National Remedy Review Board and Contaminated Sediments Technical Advisory Group
Advisory Recommendations

The Board and CSTAG (hereafter referred to as the Boards) reviewed the information package
describing this proposal and discussed related issues with Region i staff and management (Susan
Svirsky. Boh Cianciarulo. Tim Conway, and Jim Owens) on July 27 and 28, 2011. Region 1 OiTice of
Environmental Stewardship {Susan Studlien. Joanna Jerison, and Audrey Zueker) and Lam Brill,

Region 1 Office of Site Remediation and Restoration, participated via web conference. Massachusetts
Department of Environmental Protection (Ken Kimmell. Mike (iorski, Paul Locke, and Eva lor);
Massachusetts Department of lish and Game (Rich Lehan. Mark Lisa, and Mary Griffin); Curt
Spalding. Region 1 Regional Administrator. Ira Leighton. Region 1 Deputy Regional Administrator; and
Jim Murphy, Region I Office of the Regional Administrator participated by web conference for the
presentation b\ the Commonwealth of Massachusetts ( the Commonwealth). I he Boards reviewed this
site generally as if it were a ChRCLA proposed remedial action, recognizing that at this stage of the
cleanup process, the application of certain aspects oi the CLRCl.A/NCP remedy selection process is
complicated and unique. Based on this: review and discussion, the Boards oiler the following
comments:

Site Characterization

In the package presented to the Boards, modeling results played an important role in evaluating MNR as
a remedial option. The Boards recommend that additional adult largcmouth bass fish tissue data be
collected and analyzed in the context of historical data and model output. If the apparent discrepancy
between the 2008 data (mean of about 5 ppm PCB in fillet) and model output (about 18 ppm) remains,
the modeling should be updated to provide risk projections that more appropriately reflect current
conditions. In addition, the updated sampling results may be used to evaluate the eifectivcncss and
benefits of the upstream remediation.

Related to the above recommendation, but from broader perspective, the Boards recommend that the
Region expand the adult fish tissue collection efforts to pro\ idc an adequate baseline database for
evaluating the effectiveness of completed, ongoing and planned remedial actions.

Based on the model predictions described in Appendix F ol the package, the Region concluded that
Woods Pond, even if modified by deepening and changing the flow direction of the input channel, could
not be an effective sediment trap. Based upon a brief analysis of the empirical data for the site, however,
it appears to the Boards that the model predictions for trapping efficiency may not be consistent with
some of the historical sedimentation data for the site. The Boards belie ve that a modified Woods Pond,
acting as a sediment trap, could reduce the amount oi PCBs released over the darn in addition to the
reductions that would result from other proposed active remedial measures. There lore, the Boards
recommend that the Region further evaluate the potential incremental improvement in sediment trapping
of a modified Wood Ponds and recommends that the Region ask engineers from the US Army Corps ol
hngineers to assist in this evaluation.

Housatonic ROR

Final 10/20/11


-------
Human Health/Ecological Risk

During the presentation, the Region stated it is conducting a risk-based PCB cleanup as described in 40
Code of Federal Regulations (CFR) 761,6'(c). The Boards recommend that, since, for example, the
Region plans to leave soils with PCB contamination in excess of 50 parts per million (ppm). the
Superfund program closely coordinate with the Region's i oxie Substances Control Act program to
ensure the reined) meets the requirements of 40 CFR 761.61(c).

From the presentations by the Commonwealth and the Region to ilie Board, it. appears that there is a
fundamental disagreement concerning the interpretation and application of some of the criteria for
remedy selection. Particularly noteworthy are the differences in perspective on the balancing of short-
term and potential long-term environmental impacts from remedy implementation and the reduction of
long-term risks predicted to be achieved by a protective remedy. fhe presentation by the
Commonwealth indicated that it sees the impacts to Commonwealth-listed species resulting from the
need to control stream meandering as a long-term impact whereas the Region contends that habitat
restoration and other impact reduction measures wit! be effective in meeting the requirements of the
Commonwealth's endangered species law and therefore any impacts will be only short-term. 1 lie
Commonwealth's presentation also indicated that it believes the long-term ecological risks (e.g. adverse
effects to mink and wood duck) w ere acceptable when balanced against the impacts of remediation on
habitat loss. Alternately. hPA sees these long-term ecological risks as requiring remediation to meet the
threshold criteria for selecting a remedy that is protective. The Boards recommend thai the Region
consolidate the discussion on the documented ecological impacts at the site and compare them to the
Agency's requirements under CHRC1.A and the. RCRA Permit to select a remedy protective of all
identiiied receptors (assessment endpoints). This consolidated presentation will allow for a direct
comparison of short -term and long-term risks and impacts and how these risks are balanced, justified
and consistent with remedy selection criteria in am decision documents.

The Boards note that CkRCLA and the RCRA Permit identify protcctiveness of human health and the
em ironment as a threshold criterion that ail remedies must achieve. Furthermore, the NCP states that
the use ol institutional controls should supplement (not substitute for) active response measures (e.g.,
K's should not substitute for active response measures as the sole remedy unless such active measures
are determined not. to be practicable). The remedy supported by the Commonwealth appears to rely
solely on institutional controls (ICs) to protect human health through consumption offish by restricting
all consumption, whereas the remedy preferred by the Region would achieve a measure of risk reduction
that results in risks from iish consumption within the acceptable risk range and at a hazard quotient of 1
under a central tendency exposure scenario in virtually all reaches. The Board recommends that the
Region emphasize in the decision document (through both deterministic and probabilistic risk methods;
that the remedy allows for some degree offish consumption and, consistent with the NCP, docs not rely
solely on ICs to achieve a level of protcctiveness for this exposure.

Principal Threat Waste

The package presented to the Boards included a discussion of principal threat waste (P'l \Y). While the
discussion addressed contaminant mobility , it did nor specifically address toxicity and why the high
concentrations of PCBs (some locations at greater than 800 ppm) in floodplain soils would not be
Housatonic ROR	,	Final 10/20/11


-------
considered PTW materials subject to Comprehensive Environmental Response. Compensation and
I iability Act's (CERCL./Vs) and the NCP's preference for treatment to the maximum extent
practicable. Consistent with A Guide to Principal Ihreat and Low Level Uveal Wastes (OSWER
Directive No. 9380.3-06KS) which addresses the preference fc-r treatment of highly toxic materials, and
in light of A Guide on Remedial Actions ai Superfwui Sites it iih PCB Contamination i OSWER Directive
No. 9355.4-01 FS) which states that P'l W will generally include soils contaminated at concentrations
greater than 100 pprn PCBs, the Boards recommend that in its decision documents, flic Region more
thoroughh explain how its reading of Agency guidance and its approach to treatment at this site are
consistent with the statute and NCP.

Remedial Action Objective

The review package states that RAOs will address human and ecological risks as well as downstream
migration of PCBs. The Boards recommend that any decision documents for an engineering
performance-based (dredging to a depth to allow placement of a 2-2.5 foot cap) remedy that isolates
PCBs in the sediments through a bank-to-bank design should clearly explain why a numeric remediation
goal (known as interim media protection goals |IMPGsj in the review package) for sediments that is
protective of human health will not be developed. The decision documents should also better explain
where the IMPGs/cleanup standards will be applied (i.e.. in which exposure area.) in the floodplain and
how meeting these levels will be met and how the RAO will be achieved.

Remedy Performance

Based on the information presented, the Boards believe that the proposed cleanup at this site would
leave large quantities of PCBs in floodplain soils. In the future, EPA may determine that leaving this
remaining waste on site is not protective of human health and the environment. fherefore. the Boards
recommend that the Region consider including a contingency remedy (e.g. pursuing other response
actions in an adaptive framework) in the decision documents that would describe a cleanup- approach
resulting in more risk reduction through additional floodplain soil source removal or other active
remediation allernatives.

The Region's presentation included a discussion on implementing an adaptive management approach to
the remedial action. The Hoard and CS 1 AG recommend that the decision document better describe that
die selected remedy is based on the current understanding and knowledge of ihe site and that its
implementation will be phased and conducted within the adaptive management framework, l or
example, the first phase of implementation could begin wiih remediation ( or a demonstration project") ol
Reach 5A and Woods Pond {pending the results of further analysis of Woods Pond being a potential
sediment trap) that includes habitat replacement and reconstruction. Additionally, the Region should
describe the various implementation contingency approaches ;e.g., remediation and habitat
mitiiiation/replacement/reconstruction methods) that will be developed to provide implementation
options within the adaptive framework. 1 his description should also include provisions to pilot test
amendments to the cap, such as active amendments and 'or granular activated carbon, to reduce the
bioavailabilitv of PCBs Recent pilot projects for in-situ amendments at I iunter's Point (CA) and Grasse
River (NY) have demonstrated reduction in PCB bioavailability.

Housatonic ROR

Final 10/20/11


-------
The Region siaied ihat ihere are a number of dams (including the ones at Woods Pond and Rising Pond)
lhai must be maintained, in order for the remedy to be protective. The Boards note that dams are being
removed in a number of places across the country lo improve the env ironmental conditions of rivers.
Therefore, the Boards recommend that the remedy include requirements for addressing contaminated
sediments stored behind the dams as part of any future dam maintenance and/or dam removal activities.
Costs tor dam maintenance (to the extent necessarx to ensure that sediments remain contained) and/or
sediment removal activities should be included in the cost estimates.

A critical component potentially affecting the success of the Region's preferred remedy is the prevention
of the future releases of PCBs from the eroding banks in the upper seven miles or so of Ihe ri\er The
Commonwealth and many of the stakeholders acknowledge that the banks are eroding significant
amounts of PCBs but are strongly opposed to the type of hard bank stabilization techniques that were
used in the upper two miles. The Boards recommend that the Region provide additional information in
the decision documents supporting the effectiveness of softer bioengineering techniques in this part of
the river with its low gradient, locations with steep banks, and high flow rates during storm events. The
Region also should explain the key uncertainties that were considered in evaluating the long-term
effectiveness of these biocnginccring techniques. In its presentation to the Boards, the Commonwealth
was confident that the extensive bank stair iization pro-posed in the preferred remedy would prevent the
river trom meandering and the subsequent formation ol'new oxbow lakes. The Commonwealth belie\es
that containment of the river within its c invent banks would have long-lasting detrimental and
irrcA. ocablc impacts on the floodplain wetlands, vernal pools, and mam of the Commonwealth-listed
wildlife and plant species that depend on these habitats. I he Boards recommend that in the decision
documents the Region expand its rationale on why bank stabilization will not result in the long-term
adverse impacts to the ecosystem suggested by the Commonwealth. The rationale should address the
relative importance of oxbow lake formation versus periodic flooding on the long-term continued
existence of wetlands, vernal pools, and the Commonwealth-listed species that rely on a wetland
ecosystem. The Boards also recommend that in the decision documents., the Region directly address the
Commonwealth's position that channel migration is critical to "maintain]ing] a diverse mosaic oi'
wetlands and habitats that support species diversity over time." The Boards believe it would be useful
for purposes of evaluating alternatives and ensuring meaningful public participation for the Region to
estimate how many of the 6(> vernal pools and how mam acres of wetlands would disappear or be
ecologically non-functional if the river stops meandering.

W	*r	A. '	xr ...... W .. ..Ł2

Stakeholders

The Boards appreciate all of Ihe time and effort taken by the stakeholders to provide their thoughts on

the future actions to be taken at this portion of the site.

The package provided to the Board outlines the complexity of the remedy components as selected
through the RCRA permit process yet implemented as a Supertund remedial action, li may be
challenging to stakeholders to understand the logic/basis oi the reined} option components, how they fit
into the overall reined). and how the remedy as a whole meets and is consistent with Supcrfund rcmed\
selection criteria and guidance. The Board recommends that the Region develop a communication plan
for the stakeholders to concisely and clearly convey how the individual components of the remedx lit
together to achieve the remedial action objectives and meet the criteria for remedy selection.

Housatonic ROK	Final 10/20/11


-------
Early Action

In the presentation the Region identified three residential area? above Superfund residential PCB action
levels (i.e. 1 ppm per OSWER Directive No. 9<5>.4-Ul FS. ,1 Guide on Remedial Actions at Superfund
Sites fl'ith Pi B ('ontamination) and high u>e recreational areas (river access, camping, etc) above PCB
action levels. Since ihe Rest of River will be implemented as a Superfund remedial action, the Board
recommends that the Region consider conducting ail earK action (e.g.. removal or earl) interim action)
in parallel with the other Rest of River activity to address the exposure as soon as practical.

Conclusion

We commend the Region's collaborative efforts in working with the Board, C8TAG and stakeholder
groups at this site. We request that a draft response to these recommendations be included with the draft
proposed plan when it is forwarded to the Office of Superfund Remediation and Technology
Innovation's Site Assessment and Remedy Decisions (SARD) branch for review. The SARI") branch will
work with both your staff and the Boards to resolve any remaining issues prior to your release of the
proposed cleanup plan and subsequent remedy decision. This memo will be posted to the Board's
website!	¦	i and CS1 AG's website

i	) wi Ihin 30 calendar days of our

signatures. Once your response is final and made part of the site's administrative record your response
will also be posted on the Boards websites.

Thank you for your support and the support of your managers and staff in preparing for this review.
Please call Amy Legare at (703) 347-01 24 or Sieve FHs at (705) 603-8822 should you have any

questions.

ec: J. Wool ford {OS RTI)

k. Southerland (OSR 11 )
h. Gilberg (OSR!'.)

R. Cheat ham (fTRRO)

D. A mm on (OSR Tl)

D. Cooper (OSRTf)

NRRB members
CSTAG members

Housatonic ROR

Final 10/20/11


-------
ATTACHMENT C-2 ASSESSMENT OF RECENT TISSUE PCB
CHEMISTRY AND LIPID DATA


-------
W Golder
Associates

TECHNICAL MEMORANDUM

DATE July 23, 2012

REFERENCE No. 0814210104-500-TM-Rev1

TO Scott Campbell
Weston Solutions

CC Dick McGrath (Isosceles Group)

Jonathan Clough (Warren Pinnacle)

Susan Svirsky (USEPA)

FROM Gary Lawrence	EMAIL giawrence@golder.com

Barb Wernick	bwernick@gotder.com

HOUSATONIC RIVER FOOD CHAIN MODEL TECHNICAL SUPPORT -
ASSESSMENT OF RECENT TISSUE PCB CHEMISTRY AND LIPID DATA

1.0 INTRODUCTION

Golder Associates Ltd. (Golder) was retained by Weston Solutions to provide technical support related to
assessment of recent Housatonic River fish tissue chemistry. The scope is summarized in the document
"Technical Support Associated With GE/Housatonic River Project, Pittsfield, MA, Task Order 2, 08/08/2011
[Task 0233 - Remedy Review Board - Food Chain Model Technical Support]" prepared by Scott Campbell.
Warren Pinnacle Consulting (WPC) was separately retained by Weston Solutions to assist with additional
Food Chain Model (FCM) simulations in support of model-data comparisons.

This memorandum summarizes the findings of an updated model validation analysis for FCM and can be used in
response to comments made by the National Remedy Review Board (NRRB) and Contaminated Sediments
Technical Advisory Group (CSTAG)1. The model-data comparisons are intended to address comments from
NRRB that relate the role of modeling results in evaluating monitored natural recovery (MNR) as a remedial
option.

This memorandum is limited to the discussion of fish tissue chemistry up to and including 2010 sampling results.
Recent fish tissue collections (2006 to 2010 data) have been included in a multi-year trend analysis and may be
used to evaluate the effectiveness and benefits of the upstream remediation. The NRRB and CSTAG boards
also recommended that additional adult largemouth fish tissue data be collected and analyzed in the context of
historical data and model output. The latter sampling and statistical comparisons have been conducted (using
Fall 2011 sampling data), but are reported under separate cover.

1 The NRRB final comments were submitted to EPA Region I on October 20, 2011.

Golder Associates Ltd.

500 - 4260 Still Creek Drive, Burnaby, British Columbia, Canada V5C 6C6
Tel: +1 (604) 296 4200 Fax: +1 (604) 298 5253 www.golder.com

Golder Associates: Operations in Africa, Asia, Australasia, Europe, North America and South America

Golder, Golder Associates and the GA globe design are trademarks of Golder Associates Corporation.


-------
Scott Campbell
Weston Solutions

0614210104-500-TM-Rev1
July 23, 2011

2.0	METHODS

The general approach used to assess FCM performance was to conduct updated model simulations and
compare these simulations against the findings of recent tissue monitoring studies. Because the model was
calibrated and validated using historical data sets, comparisons made to recent tissue collections (2006, 2008,
and 2010 sampling events} provide an independent test of model performance.

Other considerations for assessing model performance included:

Simulation of MNR - The updated model runs were designed to simulate the exposure conditions

encountered in the watershed in the years leading up to the recent tissue collections;

i Lipid Content - The FCM, as currently constructed, does not simulate PCB concentrations in the fillets of
fish, but instead estimates whole-body concentrations on a wet-weight basis. Fillet concentrations may
then be estimated using a scaling factor. The scaling factor that is appropriate for application to estimation
of fillet concentrations is very sensitive to the lipid content in the fillet portion and remaining carcass portion
of sampled fish;

Appropriate Spatially Matched Results - The FCM was not developed to estimate concentrations in
individual sampled fish, but rather to simulate the "typical" or average concentrations observed over a
sampling unit of biological relevance to fish. For largemouth bass and many other fish species, the
appropriate unit of spatial averaging is the river sub reach, using subreach designations developed in
conjunction with the Housatonic River modeling study and ecological risk assessment;

¦ Full Monitoring Data Set for Largemouth Bass - In assessing model performance, it is important to

consider the full record of monitoring data, rather than single sampling events, and to assess the
consistency of trends across spatial units. There are several sources of sampling variability and uncertainty
that limit the conclusions that can be drawn from a single collection; and,

Trends for Other Species - In evaluating the presence and strength of temporal trends, it is important to

consider patterns for one species in the context of other species' patterns. The interconnected nature of
the biota tissue types sampled from the river (via food web linkages and similar sources of PCBs) means
that inferences regarding trends should be made using a "weight of evidence" of available data.

2.1	New Validation Data

The FCM model was formally validated using a tissue PCB data set that included measurements from 1979 to
2004. Since 2004, three significant collections have increased the tissue data set available for model validation,
including:

2006 Young of Year (YOY) Sampling - GE collected whole-body tissue samples for YOY specimens of
three resident species (bass, perch, sunfish). The sampling locations were consistent with previous
monitoring events and included Reach 5B (GE Sampling ID of HR2), Reach 6 (Woods Pond; GE Sampling
ID ofWP), Reach 7G (Giendale Impoundment; GE Sampling ID of GD), and Reach 9 (GE Sampling ID of

HR6) south of Rising Pond. Of these data, all but the Reach 9 samples can be compared to FCM outputs,
which were simulated for Reaches 5 through 8;




-------
Scott Campbell
Weston Solutions

0614210104-500-TM-Rev1
July 23, 2011

2008 Adult and YOY Sampling - In the autumn of 2008, GE collected another round of YOY sampling
using similar methods to those applied in 2006. In addition, an adult largemouth bass tissue collection (total
of 40 adult specimens greater than 12 inches length) was conducted, with sampling in each of Reach 5B/C,
Reach 6, and Reach 8, Each specimen was separately analyzed for fillet and carcass portions, and the
results were combined in the calculation of a "reconstituted" whole-body concentration for both PCB and
lipid; and,

2010 Young of Year Sampling - GE collected whole-body tissue samples for YOY specimens of three

resident species, similar to sampling in 2006 and 2008.

Lipid-normalized concentrations of PCBs in fish tissues were calculated using sample-specific lipid content data.
The wet-weight concentration was divided by the fraction lipid (expressed on a decimal basis) to yield a
PCB concentration in the units of mg/kg lipid.

A refinement was made to the assignments of Sampling Reach for the 2008 adult largemouth bass data. The
global positioning system (GPS) coordinates were used to assign each specimen from "Reach 5B/C" to the most
appropriate subreach (<.e., Reach 5B or Reach 5C). These assignments facilitated direct model-data
comparisons because the FCM simulates each subreach. Some of the adult specimens were collected in
proximity to a subreach boundary (Reach 5B/5C) or backwater areas adjacent to the main channel
(Reach 5C/5D). In these cases, Reach 5C was used as the exposure area assumed to be relevant, as it
represents a large area of suitable bass habitat with PCB exposure conditions intermediate between other
subreaches.

2.2 New FCM Simulation

The exposure conditions used in the original FCM calibration and validation runs were not considered to be
appropriate for comparison to recent tissue collections. The recent tissue data reflect an exposure condition that
includes a modified boundary condition at the upstream end of the model's spatial domain. In order to better
reflect the recent conditions, it was desirable to incorporate modeling assumptions that take into account the
influence of the now completed East Branch Housatomc River remediation. The following changes were made
to the FCM simulations:

r The fate and transport model simulations for the "Monitored Natural Recovery" (MNR) scenario used in the
Corrective Measures Study were used as a baseline for the model linkages to FCM (sediment and water
concentrations used as exposure to aquatic organisms);

v The model was "spun up" for four years using the first year of MNR exposures - this simulation period is
used to provide an appropriate initial condition for the tissue compartments in the model, prior to the
simulation of the time period of interest;

is The first calendar day of the new simulation period (post spin-up) was 1/1/2005, and the model was run for
several years to allow for comparisons against tissue collections from 2006, 2008, and 2010; and,

!• Outputs were captured from every 30 days, which provides sufficient temporal detail {i.e.. maximum of
2-week offset between simulation date and actual collection date).

3/18

v©ras«


-------
Scott Campbell
Weston Solutions

0614210104-500-TM-Rev1
July 23, 2011

2.3 Pairing of Data

The FCM simulations provided an updated time series of concentration estimates for each biota compartment

simulated by the model. This included separate estimates for age classes of the fish species. Each field
measurement was first assigned an estimated fish age based on consideration of the length and weight
measurements for the individual specimen. The allocations were made using procedures outlined in the
EPA/Weston Model Calibration Study and Final Model Documentation reports. All YOY specimens corresponded
to age class 0+ in the bioaccumulation model. Adult bass specimens ranged from age 5+ to age 9+ (the latter
was used to simulate the largest age class of largemouth bass).

Tissue measurements were assigned to the most relevant model simulation using the species, reach, date, and
age class designations.

The FCM estimates whole-body wet-weight concentrations for all species of interest. It is possible to derive
lipid-normalized simulations using the model-specified lipid content, which is a controlling parameter in the model
code. These lipid contents were assigned based on long-term averages in regional data sets, and therefore do
not reflect individual variations in lipid content that occur naturally among individual fish and over time. The
purpose of FCM is to simulate the long-term trends in PCB concentrations, and the model cannot predict
small-scale variations (i.e., variability among individual fish) or seasonal fluctuations attributable to variation in
lipid content. For this reason, it is important to consider the following when performing model-data comparisons:

e Emphasis on central tendencies within each subreach/species combination rather than individual
measurements; and,

ii! Investigation of measured lipid contents for temporal consistency and degree of similarity to
FCM parameters.

The lipid contents observed over the course of Housatonic River sampling events in the last two decades are
discussed in Sections 3.1 and 3.2, below.

The results provided herein emphasize largemouth bass, due to volume of recent tissue PCB data available for
this species. The statistics for model fit based on lipid-normalized data are summarized in Table 1; these
statistical metrics are the same ones applied during the Model Calibration and Model Validation studies
conducted by EPA. The model bias statistic (MB*; Arnot and Gobas, 20042) is a useful performance metric
because it tracks the central tendency of the ability of the model to simulate PCB concentrations within each
subgroup of fish tissues. MB* is the geometric mean of the ratio of simulated and measured concentrations for
tPCBs, and is a measure of the systematic overprediction {MB* > 1) or underprediction (MB* < 1) of the model.

•' Arnot, J-A. and F A P C. Gobas. 2004, A food web bioaccumulation model for organic chemicals in aquatic ecosystems, Environ. Toxicol,
Chem. 23:2343-2355,

3.0 RESULTS

4/18




-------
Scott Campbell
Weston Solutions

0614210104-500-TM-Rev1
July 23, 2011

3.1 PCBs in Adult Bass

Figure 1 is a scatter plot comparing the measured and simulated PCB concentrations for iargemouth bass (LMB)

sampled from Massachusetts reaches of the Housatonic River in 2008, The model performance can be
assessed by comparing the centroid of the data for each subreach to the optimal fit (1:1 slope) line. There is an
apparent discrepancy of approximately two-fold between the mean measured concentrations of PCBs in whole
body 1MB samples from the PSA reaches. The simulated concentrations are, on average, greater than the
measured concentrations in these reaches. In contrast, the Rising Pond (Reach 8) data show a small
under-prediction.

The FCM model was parameterized such that whole-body 1MB lipid contents increase with the age of the fish.
For the largest age-classes (of relevance to the adult fish specimens), the lipid content was assumed to range
from 4% to 5%, with the higher value applicable to the largest and oldest fish. Evaluation of the 2008 lipid data,
however, indicate that only Reach 5B specimens had these assumed lipid levels. The mean whole-body lipid
contents (reconstituted) from Reaches 5C, Reach 6, and Reach 8 were 3.8%, 2.5%, and 2.8%, respectively.
Therefore, the discrepancies in predicted concentrations for these subreaches result in part from lipid differences
relative to the long-term data set.

Figure 2 presents the same LMB scatter plot, with the data standardized to lipid content (lipid normalized). For
PSA reaches, there is improvement in model fit, with no systematic over- or under-prediction for either Reach 5C
or Reach 6. A systematic discrepancy is still observed for Reaches 5B and Reach 8; however, these differences
are within the model performance criteria. Overall, the model bias statistic for the 2008 adult bass samples is
1.13 and the mean percent error is -11% (Table 1), indicating only a small tendency for model over-prediction.

Both wet-weight and lipid-normalized data exhibit pronounced variability (i.e., factor of 5 or more between
maximum and minimum concentrations for each species/reach). This is expected due to the heterogeneous
distribution of PCBs in the river sediments, causing individual bass to be exposed to different exposure levels
even within the same sampling reach. It is not possible for FCM to simulate these micro-scale variations
because the model is run using a central tendency estimate of PCB concentrations in sediment within each
reach.

Overall, the comparison of model simulations to 2008 adult LMB data indicate acceptable model performance.
Although the concentrations of PCBs in bass in 2008 are reduced relative to other recent sampling events, and
indicate a reduction in concentration prior to lipid-normalization, this finding can partly be explained on the basis
of lower lipid content in 2008 (a transient phenomenon not consistent with long-term trends). Appropriate
matching of model simulations to data indicates that the FCM does not over-predict concentrations.

Figure 3 shows the 2008 adult LMB PCB concentration data in the context of historical measurements of large
LMB specimens (i.e., aged 5+ or older) in the Primary Study Area (Reaches 5 and 6). The data show a small
overall decrease in PCB concentrations. However, there is substantial variability among sampling programs,
and among individual specimens within each program.

Evaluation of temporal trends in LMB fillets is complicated by the relatively large interannual variability in fillet
lipid content (Figure 4). Although the variability in individual fillets is substantial, systematic differences in lipid
content are apparent among Housatonic River sampling programs. The dark black bars in Figure 4 depict the
mean fillet lipid content for each of the four major tissue collections. The 2002 sampling event produced the
largest average fillet lipid (approximately 2%), which contributed to elevated fillet PCB concentrations in
that sampling year. In contrast, the 2008 sampling event yielded the lowest average fillet lipid content

5/18




-------
Scott Campbell
Weston Solutions

0614210104-500-TM-Rev1
July 23, 2011

(mean of 0,5%), arid numerous individual measurements of fillet lipid were 0.2% or lower. General Electric has
previously questioned the biological plausibility of these low values, yet they are observed in both EPA and
GE fillet collections. Because PCB concentrations are strongly controlled by lipid content, these variations and
uncertainties in lipid analyses render the fillet-based trends highly uncertain relative to whole-body analyses,
With such large fluctuations in lipid content among sampling events, it is not possible to discern the presence or
slope of a temporal trend with confidence. Although lipid-normalization of fillet results can control for some of
this variability, uncertainty remains in the use of a fillet measurement for tracking long-term bioaccumulation
trends. As the majority of the PCB mass in fish is associated with the carcass (offal) portion of fish, whole-body
tissue concentrations in largemouth provide a more reliable basis for conducting trend assessments.

In summary, the apparent large drop in fillet PCB concentrations between 2002 and 2008 is largely attributable
to the difference in fillet lipid contents between these sampling events. When whole-body results (and
lipid-normalized results) are evaluated, the trend in adult LMB concentrations is much flatter. A statistically
significant decrease over time is apparent in the long-term data set for adult bass, but the slope of the trend is
lower than suggested by the NRRB/CSTAG comments.

3.2	PCBs in YOY Bass

Figure 5 is a scatter plot comparing FCM simulations to data for the young-of-year LMB specimens collected in
2008, 2008, and 2010. Overall, there is a small systematic over-prediction of measured PCB concentrations.
The magnitude of over-prediction (based on species-reach averages) is not large, with most reach/year
combinations yielding a mean absolute percent error of less than 50%.

Similar to the large adult LMB, the evaluation of model performance is improved through lipid-normalization
(Figure 6). Most model-data comparisons are significantly improved through lipid-normalization, and overall the
2008 data exhibit low bias statistics once data are lipid-normalized. Once again, this indicates that the 2008
specimens contained lower lipid contents (on average) than have been observed in the historical data set used
to parameterize the model. The mean percent error statistics for 2006, 2008, and 2010 were -23%, -6%, and
-39%, respectively (Table 1), indicating a minor tendency for over-prediction. The corresponding model bias
statistics are 1.27, 1.07, and 1.50, all of which are well within performance criteria for the model. The best model
fit for YOY fish was observed during the 2008 sampling event, and both adult and YOY samples exhibited overall
bias lower than 15% for that event.

The LMB YOY predictions are within FCM model performance criteria for validation and indicate no cause for
concern in terms of long-term simulation of remedial alternatives.

3.3	Other Species

Further insight regarding the long-term trends on PCB bioaccumulation is provided by analysis of other fish
species collected as part of fish tissue monitoring. The linked bioaccumulation model predicts a gradual decline
in PCB concentrations of all fish species over time, with multiple decades required for average fish tissue
concentrations to be reduced by 50%) (i.e., half life for natural depuration processes). The above analysis
indicates that the reduction in adult largemouth bass PCBs observed between 2002 and 2008 is at least partially
explained by lipid variations and other natural stochastic factors, as opposed to being an indication of rapid
progression toward sustained lower PCB bioaccumulation. To help evaluate the broader trend in fish tissue PCB
burden, examination of other monitored species is helpful.

8/18

(2>taoitar,«


-------
Scott Campbell
Weston Solutions

0814210104-500 TM-Rcv1
July 23,2011

Figures 7 and 8 depict the PCB concentration trends from lorvg-ten-n monitoring of multiple YOY species in the
PSA (wet-weight and lipid-normalized, respectively). These figures are suggestive of a long-term gradual
decline in PCB concentrations in biota. In terms of temporal trends, statistically significant decreases were
observed for most species and sampling locations over the period of record, and similar trends were observed
for both wet-weight and lipid-normalized concentrations.

Figures 9 and 10 depict the PCB concentration trends from long-term monitoring of multiple YOY species in the
downstream reaches (wet weight and lipid-normalized, respectively). Again, there are suggestions of slow
declines in PCB concentrations for some species and reaches, although the variability among sampling events is
relatively large.

None of the species evaluated above indicate a rapid decline in PCB concentrations over the period of record.

In terms of model validation, it is important to recognize that the linked bioaccumulation model is unable to
predict the micro-scale changes in bioaccumulation that occur over small time and spatial scales. Nor does the
model simulate the transient variations in bioaccumulation, such as those influenced by anomalous lipid
conditions within sampled fish. Rather, the model simulates the long-term trends in bioaccumulation under the
assumption that controlling parameters such as lipid content and sediment PCB exposure will vary about a
central tendency over the long-term.

When the expanded historical data set for PCBs in adult largemouth bass is evaluated, a decline in
PCB concentrations is observed for both wet-weight and lipid-normalized whole-body concentrations (Figure 3),
and these decreases are statistically significant. However, the slope of the trend for most model-data
comparisons is gradual, and the decline in largemouth bass concentrations between 2002 and 2008 appears to
be mainly attributable to natural lipid variations among sampling events. The effect of the atypical lipid observed
in 2008 is that the apparent rate of decline is exaggerated when evaluated on a wet-weight basis; analysis
based on lipid-normalized data corrects for this problem.

Stochasticity in PCB concentrations over time is evident for all species of interest, and confounds the precise
evaluation of temporal trends. Although it is possible that the linked bioaccumulation model for the
MNR scenario predicts a rate of decline in PCB concentrations that is lower (or greater) than the actual rate,
there is no evidence that the model is improperly calibrated based on the available information.

5,0 CLOSURE

We trust that the enclosed information is sufficient to meet your current needs. If you have any questions, feel
free to contact the undersigned at (604) 297-2003.

4.0 SUMMARY

Gary Lawrence, M.R.M, R P. Bio,
Associate, Senior Environmental Scientist

Barb Wernick, M.Sc., R.P. Bio.

Associate, Senior Environmental Scientist

GSL/BGW/mrw

t-0t04\0Bl42 IQiCK-SDQ-lrr-rovi-pcti 'issue Irtish	tfecx


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 1. Scatter plot comparing FCM simulations of PCBs in whole-body adult largemouth bass to
field specimens from General Electric's 2008 fish sampling program in the Housatonic
River. Dashed line represents equal concentrations between simulated and measured
PCBs.

120

Wet-Weight Whole-Body (GE 2008 Adult LMB)

5

3 100

bo

M

U)

JL so

CO
U

6	60

rtj

u>

>

< 40
~u

4-»

_ro
3

E

in

20

o

-B—cs>-

~m

— o

A	s'0 a

-~—O-,-'

^c?' O o
X O °o o

20	40	60	SO 100

Measured tPCB (mg/kg ww)

O Reach 8 ~ Reach 6 A Reach5C

O Reach SB— — 1:1 Slope

120

8/18

Cjty Colder
Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 2. Scatter plot comparing FCM simulations of PCBs in the lipid portion of adult largemouth
bass to field specimens from General Electric's 2008 fish sampling program in the
Housatonic River. Dashed line represents equal concentrations between simulated and
measured PCBs.

Lipid-Normalized Whole-Body (GE 2008 Adult LMB)

4500

2 4000

Q.

&P 3S00

M

E,

CO

u

Q.

Q)
on

TO

i-


<
•O
o
AAA

A AA ~A A-^

A





1000

2000

3000

4000

Measured tPCB (mg/kg lipid)

O Reach 8 ~ Reach 6 A Reach 5C

O Reach SB'

1:1 Slope

9/18

Cjty Colder
Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 3. Temporal trend of PCB concentrations (wet weight [top panel] and lipid-normalized
[bottom panel]) in large adult LMB over the period of record, for all samples collected
within the PSA.

600
¦5 500
o> 400
1 300
g 200

CL

~ 100
0

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Sampling Date

14000
_12000

J 10000

f 8000

j= 6000

qq 4000
o

9b 2000
0

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Sampling Date

Large Bass (n=122) 9

y=-0.0071x +353.96	•

°8 8 $

i	t-fr-l	1

Large Bass (n=122)

y = -0.2417x + 11624

08 a ®

i ¦ 1 1 1 1 1 1 			 1 1 1 1 1 1 1 ™i

10/18

Colder
Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 4. Temporal analysis of lipid content in Housatonic River largemouth bass specimens,
subdivided by subreach.

_aj

a.
E

ID
t/)

7%

6%

5%

•- 4%

¦o

'a

c

ai
u

ai
a.
a.

3%

2%

1%

0%

O

Jan-90

Fillet Lipid Contents in Massachusetts Largemouth Bass

Jan-94

A
A

-A-

O

O
O

O

7T

Jan-98	Jan-02

Sample Collection Date

O

Jan-06

Jan-10

O Reach 5A 0 Reach5B AReachSBC ~ Reach 5C O Reach 6 O Reach 8

¦ Program Mean

11/18

Golder
>Ł~ Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 5. Scatter plot comparing FCM simulations of PCBs in whole-body juvenile largemouth bass
(YOY) to field specimens from General Electric's biennial fish sampling program in the
Housatonic River (2006 to 2010). Dashed line represents equal concentrations between
simulated and measured PCBs.

Wet-Weight Whole-Body (GEYOYLMB)

so

— 45

5
3

<-)
Q.

Ofl
CO

k-

01
>
<
T3
0J

4->

JO
3

E

CO

40

m 35

30

Si. 25

20

15

10

CD CO- CO	x'

~



A ACCpC?

O
O
O

0	10

Reach 5B- 2006
Reach6-2006
Reach 7G -2006
1:1 Slope

20	30	40

Measured tPCB (mg/kg ww)

A Reach SB-2008 O
A Reach 6 -2008 O
A Reach7G-2008 O

50

Reach SB-2010
Reach 6-2010
Reach 7G - 2010

12/18

Colder
Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM- Rev1
July 23, 2011

Figure 6. Scatter plot comparing FCM simulations of PCBs in the lipid portion of juvenile

largemouth bass (YOY) to field specimens from General Electric's biennial fish sampling
program in the Housatonic River (2006 to 2010). Dashed line represents equal
concentrations between simulated and measured PCBs.

Liptd-Normalized Whole-Body (GE YOY LMB)

2000

— 1800
12

= 1600

"55
Ł

CD
U

CL

QJ
0l0
(0

k.

Of
>
<
TJ
0J
*->
J5
3

E

tj)

1400
1200

1000

800

600

400

200

O CO ,6 COO

	~	

^	a

<3CS\YA

500

1000

1500

2000

O
O
O

Reach7G -2006
Reach SB - 2006
Reach 6-2006
1:1 Slope

Measured tPCB (mg/kg lipid)

A Reach7G-2008 O
A Reach 5B - 2008 O
A Reach 6-2008 O

Reach7G-2010
Reach SB-2010
Reach 6

13/18

Colder
Associates


-------
Scott Campbell
Weston Solutions

O01421O1O4-5OQ-TM-Rev1
July 23, 2011

Figure 7.

Wet-Weight PCB Concentrations over Time in YOY Fish from Biennial Sampling (PSA Reaches).

60
50
40
30
20
10
0

~ Median (with maximum and minimum)
-75th percentile
"•25 th percentile

Reach 5B

fi

8G

BG

BG

BG

BG

BG

BG

8G



LB

LB

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS

PS



YP

YP

YP

YP

YP

YP

YP

YP

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

06

10



94

96

98

00

02

04

06

08

Species/Year

60
50
40 '
30 "
20 ¦
10
0

4,

~ Median (with maximum and minimum)
-75th percentile
_25th percentile

A fi

Reach 6

U



I

BG

BG

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB

LB



YP

YP

YP

YP

YP

YP

YP

YP

YP

94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10

Species/Year

14/18

Golder

Associates


-------
itions

2500

2000

1500

1000

500

0

2500

2000

1500

1000

500

0

O01421O1O4-5OO-TM-Rev1
July 23, 2011

Lipid-Normalized PCB Concentrations over Time in YOY Fish from Biennial Sampling (PSA Reaches).

~ Median (with maximum and minimum)

—	75th percentile

—	25th percentile

Reach 5B

J1





BG

94

BG

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS

PS



YP

YP

YP

YP

YP

YP

YP

YP

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

YP

10

Species/Year

~ Median (with maximum and minimum)

—	75th percentile

—	25th percentile

r*i i

Reach 6

*

I

BG
94

BG

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB

LB



YP

YP

YP

YP

YP

YP

YP

YP

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

Species/Year

YF
10

15/18

Golder

Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM-Rev1
July 23, 2011

Figure 9, Wet-Weight PCB Concentrations over Time in YOY Fish from Biennial Sampling (Downstream Reaches in
Massachusetts).

30
25
20
15

10 1

5
0

Q Median (with maximum and minimum)
™75th percentile
— 25th percentile

Reach 7G (Glendale Dam)

I

n

ifUO



a

EL

n

S

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS



YP

YP

YP

YP

YP

YP

YP

98

00

02

04

06

08

10



96

98

00

02

04

06

08

10



96

98

00

02

04

06

08

10



96

98

00

02

04

06

08

Species/Year

VP
10

10

6 *

n Median (with maximum and minimum)
• 75th percentile
-25th percentile

Reach 9 (Downstream of Rising Pond)



A

BG

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS

PS



YP

YP

YP

YP

YP

YP

YP

YP

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

YP
10

Species/Year

16/18

Golder

Associates


-------
Scott Campbell
Weston Solutions

0814210104-500-TM-Rev1
July 23, 2011

Figure 10. Lipid-Normalized PCB Concentrations over Time in YOY Fish from Biennial Sampling (Downstream Reaches in
Massachusetts).

700
600
500
400
300
200
100
0

BG
96

n Median (with maximum and minimum)
"75th percentile
- 25th percentile

Reach 7G (Glendale Dam)

i

n

Q



n_fl.

BG

BG

BG

BG

BG

BG

BG



L8

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS



VP

YP

YP

YP

YP

YP

YP

98

00

02

04

06

08

10



96

98

00

02

04

06

08

10



96

98

00

02

04

06

08

10



96

98

00

02

04

06

08

Species/Year

YP
10

250

200

150

100

50

I

° Median (with maximum and minimum)
" 75th percentile
-25th percentile

Ti i

Reach 9 (Downstream of Rising Pond)

I



BG
94

BG

BG

BG

BG

BG

BG

BG

BG



LB

LB

LB

LB

LB

LB

LB

LB

LB



PS

PS

PS

PS

PS

PS

PS

PS

PS



YP

YP

YP

YP

YP

YP

YP

YP

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

10



94

96

98

00

02

04

06

08

YP
10

Species/Y ear

17/18

Golder

Associates


-------
Scott Campbell
Weston Solutions

0814210104-50Q-TM-Rev1
July 23, 2011

Tabie 1:	Statistics for Mode! Fit Based on Lipid-Normalized PCB Concentrations in Largemouth Bass Tissues (2006 to 2010}

Description

n

ME

MAE

MPE

MAPE

MB*

All Samples

183

-38

520

-15%

46%

1.18

By Sampling Event:













2006 GE Young-of-Year Fish Sampling

21

-116

185

-23%

28%

127

2008 GE Adult Fish Sampling

120

-5

705

-11 %

54%

1.13

2008 GE Young-of-Year Fish Sampling

21

18

126

-6%

20%

1.07

2010 GE Young-of-Year Fish Sampling

21

-190

190

-39%

39%

1.49

By Sample Type:













Adult Fish Sampling

120

-5

705

-11%

54%

113

Young-of-Year Fish Sampling

83

-96

167

-23%

29%

1.26

By Reach:













Reach 5B

39

-396

436

-39%

42%

1.53

Reach SBC

2

-807

807

-84%

84%

2.55

Reach 5C

22

-462

661

-53%

65%

1.92

Reach 5CD

3

-218

480

-26%

42%

1.33

Reach 6

86

-23

503

-12%

35%

1.13

Reach 7G

21

-90

90

-37%

37%

1.48

Reach 8

30

822

850

58%

84%

0.51

By Reach (Adult Fish Only):













Reach 5B Adult

18

-699

759

-61%

64%

1.97

Reach 5BC Adult

2

-807

807

-84%

84%

2.55

Reach 5C Adult

22

-462

661

-53%

65%

192

Reach 5CD Adult

3

-218

480

-26%

42%

1,33

Reach 6 Adult

45

-6

619

-12%

39%

1 14

Reach 8 Adult

30

822

850

58%

64%

0.51

Notes: n = sample size; ME = mean error; MAE = mean absolute error; MPE = mean percent error; MAPE = mean absolute percent error; MB* = model bias statistic
In this table, separate statistics are presented for Reach SBC (n - 2) and 5CD (n=3) to assess model fit lor tissue collections close to the subreach boundaries.

18/18


-------
ATTACHMENT C-3 SCREENING EVALUATION FOR POTENTIAL
IMPROVEMENT IN SEDIMENT TRAPPING IN WOODS POND


-------
Screening Evaluation of Hydraulic Efficiency of Woods Pond

SUMMARY

Below is an initial pond design for increasing sediment retention in Woods Pond. The
effectiveness of the design is supported by a screening evaluation based on hydraulic efficiency.
As hydraulic efficiency increases, so will sediment settling and trapping efficiency. The
proposed design increases effective retention time by nearly four times and is sufficient to
eliminate most of dead zones (about 90% of maximum as compared to 22% in the existing
configuration). In this design, the inlet and outlets are maximally spaced to eliminate short-
circuiting. Increasing depth in the pond would reduce resuspension of sediments and further
improve trapping efficiency. Finally, the proposed design breaks the fetch lengths in Woods
Pond, which may reduce resuspension induced by wind. Well-established design principles
(criteria for detention times, length-to-width and length-to-depth ratios) known to achieve 70
to 90% retention of suspended solids would be incorporated, to the extent possible, in the
design. As a reference, the current trapping efficiency is estimated at 17%.

To quantitatively assess the sediment trapping efficiency of this and other designs, additional
analyses could be performed using reservoir trapping efficiency analysis (Heinemann 1984),
flocculent settling analysis, or mass balance analysis of other reservoirs on the Housatonic River
system. Such evaluations would support an optimal design of Woods Pond to maximize
sediment trapping and immediately reduce downstream transport of contaminants through
Woods Pond before and during Reach 5 remediation actions.

ANALYSIS

Hydraulic efficiency is a measure of the effective area of the pond or active flow zones (one
minus the hydraulic efficiency is the fraction of the pond acting as dead zones).

Hydraulic Efficiency = 0.9 [ 1 - exp (-0.3 L/W) ] (Shields et al. 1987)

Existing Effective Length (L) to Width (W) Ratio: 0.84

Estimated Hydraulic Efficiency Based on L/W Ratio: 20.0%

Effective L/W Ratio with Proposed Training Dikes (see attached figure): 7.54
Estimated Hydraulic Efficiency Based on L/W Ratio: 80.6%


-------
The hydraulic efficiency with training dikes is four times larger than the existing hydraulic
efficiency; therefore, the effective retention time is nearly four times greater than the existing
pond (slightly less than the ratio of hydraulic efficiencies due to loss of volume occupied by the
training dikes). Adding more dikes would likely have diminishing returns since the dikes would
consume as much or more area and volume than one would gain in eliminating dead zones. In
this design, the inlet and outlets are maximally spaced to eliminate short-circuiting. The
increased retention time provided by this design will permit greater settling/deposition.

Channel Width with Training Dikes

as a Fraction of Existing Effective Width: 0.33

Effective Channel Width with Training Dikes as a Fraction
of Existing Effective Width Considering Hydraulic Efficiency: 1.34

Effective width with training dikes is about 34% greater than the existing effective width;
therefore, the velocities in the active flow areas with training dikes are about 75% of the
existing velocities in the active flow areas. The reduced velocities also reduce the shear
stresses and the potential resuspension or erosion of the sedimentation in Woods Pond.
Increasing the depth of Woods Pond would further decrease velocities and shear stress, but the
design depth would be limited by the effective length of a reconfigured Woods Pond and the
detention time to optimize sedimentation. Finally, the training dikes also break the fetch
lengths in Woods Pond, which may reduce resuspension induced by wind.

Sediment retention basin design procedures recommend 24 to 40 hours of hydraulic detention
time in a properly designed basin to achieve 70 to 90% retention of suspended solids. Proper
design would provide a Length-to-Width ratio of at least 4:1 and a Length-to-Depth ratio of
about 200:1; however, the depth should be sufficient to provide storage without scouring,
erosion or significant resuspension, typically at least 3 to 6 ft (Caltran 2000), but design depth is
a function of the maximum design velocity and storage capacity.

To quantitatively assess the sediment trapping efficiency of an optimized design, additional
analyses could be performed using reservoir trap efficiency methods that consider the grain
size distribution of the suspended solids, the settling velocity distribution of the particles, and
specific flow characteristics.

CONCLUSION

It is expected that the increased hydraulic efficiency and effective retention time, coupled to
the decreased velocities, shear stresses, and reduction in wind-induced resuspension will
enhance sedimentation comparably to the performance of a sediment retention basin, which


-------
would typically achieve 70 to 90% trapping efficiencies. However, a more detailed analysis of
sediment deposition and trapping efficiency could be conducted on this and other designs using
approaches such as flocculent settling analysis, mass balance analysis of other reservoirs on the
Housatonic River system, and reservoir trapping efficiency analysis (Heinemann 1984). Such
evaluations would support an optimal design of Woods Pond to maximize sediment trapping
and immediately reduce downstream transport of contaminants through Woods Pond before
and during Reach 5 remediation actions.

REFERENCES

Shields, F.D.; Thackston, E.L.; Schroeder, P.R.; Bach, D.P. 1987. Design and Management of
Dredged Material Containment Areas to Improve Hydraulic Performance. Technical Report D-
87-2, NTIS No. AD-A187 235. Prepared by U.S. Army Engineer Waterways, Experiment Station.

State of California Department of Transportation (Caltran), November 2000. Stormwater
Quality Handbooks - Construction Site Best Management Practices (BMPs) Manual.

Heinemann, 1984. "Reservoir trap efficiency," Agricultural Research Service, in Erosion and
Sediment Yield: Some Methods of Measurement and Modeling, edited by R.F. Hadley and D.E.
Walling, Geo Books, Norwich, England.


-------

-------
DRAFT

PRIVILEGED

9 January 2012

Screening Evaluation of Trapping Efficiency for Woods Pond Alternatives

by Paul R. Schroeder, PhD, PE, Research Civil Engineer
US Army Engineer Research and Development Center

SUMMARY

Below is a screening evaluation of potential effectiveness of various Woods Pond alternatives
and conditions on the trapping efficiency of TSS and PCB. The alternatives include the present
condition; Woods Pond dredged to a depth of 6 ft, 10 ft and 15 ft; Woods Pond with USACE
spur dike design and dredged to 6 ft, 10 ft and 15 ft; State of Massachusetts Concept 1 with and
without deepening; and State of Massachusetts Backwater Concept with and without
deepening. In the State of Massachusetts Concept 1, flow passes through the backwater area
only during periods with higher flow rates (above 300 cfs).

ANALYSIS

Assumptions and approach are given below:

Log Kow for PCBs = 6.4
foe = 0.01

Kd = 0.617 * foe * Kow = 15,500 L/kg (assuming equilibrium partitioning)

TSS data from 1981 (low flow year), 1988 (medium flow year) and 1990 (high flow year)

PCB Fraction associated with particulates, Fp = (Kd * TSS) / [1 + (Kd * TSS)] (computed daily)

Area, Length and Volume estimated from drawings of each alternative (given in Table 1)

Width = Area / Length

Hydraulic Efficiency = 0.9 [ 1 - exp (-0.3 L/W) ] (Shields et al. 1987)

Flow data from 1981 (low flow year), 1988 (medium flow year) and 1990 (high flow year)

Residence Time, t = Volume * Hyd. Eff. / Flow Rate (computed daily)

Settling Decay Rate, k = 0.05/hour/(mg/L) (adjusted as function of TSS concentration daily,
estimated from settling test on freshwater upstate New York sediments and calibrated to
trapping efficiencies estimated from measured deposition rates and predicted by EFDC)

TSS Removal, RTSs = exp (-k * TSS * t) (computed daily)

PCB Removal, RPCB = Fp * RTSs (computed daily)

Annual results of trapping efficiencies are given in Table 2.


-------
DRAFT	PRIVILEGED	9 January 2012

Table 1. Physical Characteristics of the Woods Pond Alternatives

Alternative

Area

Volume

Length

Depth

Width

Hyd Eff

sq ft

cu ft

ft

ft

ft

Woods Pond No Dredging

2,429,000

9,720,000

1510

4.0

1609

0.221

Woods Pond 6 ft

2,429,000

14,570,000

1510

6.0

1609

0.221

Woods Pond 10 ft

2,429,000

24,290,000

1510

10.0

1609

0.221

Woods Pond 15 ft

2,429,000

36,440,000

1510

15.0

1609

0.221

Woods Pond w/USACE dikes 6 ft

2,429,000

14,570,000

3890

6.0

624

0.761

Woods Pond w/USACE dikes 10 ft

2,429,000

24,290,000

3890

10.0

624

0.761

Woods Pond w/USACE dikes 15 ft

2,429,000

36,440,000

3890

15.0

624

0.761

MA Concept 1 >300 cfs* No Dredging

5,919,000

19,180,000

5910

3.2

1001

0.747

MA Concept 1 <300 cfs* No Dredging

2,429,000

14,570,000

1510

6.0

1609

0.221

MA Concept 1 >300 cfs* w/Deepening**

5,919,000

25,470,000

5910

4.3

1001

0.747

MA Concept 1 <300 cfs* w/Deepening**

2,429,000

14,570,000

1510

6.0

1609

0.221

MA Backwater Concept No Dredging

5,572,000

17,810,000

5680

3.2

981

0.742

MA Backwater Concept w/Deepening**

5,572,000

24,480,000

5680

4.4

981

0.742

* Stage is assumed to be high enough to induce flow through backwater areas and over spillways when flow is greater
than 300 cfs.

** Upper backwater areas are assumed to be deepened from 2 ft to 3 ft and Woods Pond is assumed to be deepened
from 4 ft to 6 ft. Lower backwater areas are assumed to be 3 ft deep.


-------
DRAFT	PRIVILEGED	9 January 2012

Table 2. Trapping Efficiencies for Various Woods Pond Alternatives

Alternative

Trapping Efficiency (percent)

Low Flow (1981)

Medium Flow (1988)

High Flow

1990)

TSS

PCB

TSS

PCB

TSS

PCB

Woods Pond No Dredging

13.5%

8.7%

14.3%

8.6%

12.3%

8.3%

Woods Pond 6 ft

19.5%

12.4%

20.6%

12.3%

17.8%

11.9%

Woods Pond 10 ft

30.0%

19.1%

31.6%

18.9%

27.6%

18.6%

Woods Pond 15 ft

41.1%

26.0%

43.0%

25.6%

38.0%

25.7%

Woods Pond w/USACE dikes 6 ft

51.3%

32.2%

53.4%

31.8%

47.8%

32.4%

Woods Pond w/USACE dikes 10 ft

68.8%

42.8%

70.9%

42.0%

65.3%

44.5%

Woods Pond w/USACE dikes 15 ft

81.6%

50.4%

83.4%

49.2%

78.7%

53.9%

MA Concept 1 300 cfs* No Dredging

49.1%

28.2%

50.6%

27.3%

52.7%

35.4%

MA Concept 1 300 cfs* w/Deepening**

58.5%

33.9%

59.6%

32.6%

60.5%

41.0%

MA Backwater Concept No Dredging

57.3%

35.8%

59.4%

35.3%

53.6%

36.4%

MA Backwater Concept w/Deepening**

68.2%

42.4%

70.3%

41.6%

64.6%

44.1%

* Stage is assumed to be high enough to induce flow through backwater areas and over spillways when flow
is greater than 300 cfs.

** Upper backwater areas are assumed to be deepened from 2 ft to 3 ft and Woods Pond is assumed to be
deepened from 4 ft to 6 ft. Lower backwater areas are assumed to be 3 ft deep.


-------
DRAFT

PRIVILEGED

9 January 2012

The performance of State of Massachusetts alternatives could be significantly improved by
adding a spur dike in Woods Pond and by additional deepening. The trapping efficiency of PCB
may be greater than predicted (although still less than the trapping efficiency of TSS) because
equilibrium partitioning would not be expected to occur during high flow conditions, which is
also when the PCB and solids concentrations are greater. The analysis did not consider the
effect of flow on stage or water depth and residence time; therefore, the analysis would
underestimate the trapping efficiency. However, the settling rate was adjusted for TSS
concentration, which would offset the underestimate in residence time and trapping efficiency.

CONCLUSION

Analysis of the trapping efficiencies of the various alternatives indicate that adaptation of the
flow characteristics in Reach 5C and Woods Pond can significantly improved the trapping
efficiency for both TSS and PCB. The proposed alternatives are predicted to remove about 50%
of the TSS and 30% of the PCB without deepening. Moderate deepening with flow alterations
can improved the trapping efficiency to about 70% for TSS and 45% for PCB. Adding a spur dike
to the State of Massachusetts alternative would likely improve the trapping efficiency to more
than 80% for TSS and 50% for PCB. The maximum trapping efficiency for PCB is estimated to be
a little more than 60% (assuming equilibrium partitioning and 100% removal of TSS).

REFERENCE

Shields, F.D.; Thackston, E.L.; Schroeder, P.R.; Bach, D.P. 1987. Design and Management of
Dredged Material Containment Areas to Improve Hydraulic Performance. Technical Report D-
87-2, NTIS No. AD-A187 235. Prepared by U.S. Army Engineer Waterways, Experiment Station.


-------
Concept 1

Low Levee

Potential
Dredge Area

Potential
Dredge Area

6 Meter
Remediation Buffer

Submerged Weir

*• —-r»L- •

Potential
Dredge Area

' LimnoTech

6 Meter

Remediation Buffer

Potential
Dredge Area

Broad


-------
^~11111016^

Water | Environment | Scientists | Engineers

MEMORANDUM

FROM: Richard McCulloch and Dan Lautenbach

TO:

Eva Tor, Massachusetts DEP

DATE: 1/12/2012
PROJECT: HRPCB

CC: Bruce Haskell, CDM Smith

SUBJECT: Woods Pond Alternatives Conceptual Trapping Analysis

DRAFT - FOR REVIEW AND DISCUSSION ONLY-NOT FOR DISTRIBUTION
Summary of Findings

This memorandum presents a summary of the methods applied and results of LimnoTech's
spreadsheet analyses of Woods Pond trapping efficiency. Our analyses were aimed at evaluating
whether or not higher trapping efficiencies than those reported in recent EPA / HydroQual, Inc.
investigations could be attained in Woods Pond. Findings of our analyses are summarized below.

Effective trapping of solids requires the ability to trap solids delivered under moderate to high
flow conditions

•	90% of the solids load to Woods Pond is delivered by flows in excess of the median flow

•	65% of the solids load to Woods Pond is delivered by flows in excess of the 90th
percentile flow

Preventing short circuiting of Woods Pond significantly increases trapping efficiency

•	Under short-circuiting conditions, the effective surface area is less than 15% of the total
surface area of the pond

•	Installation of an outlet weir in combination with dredging may be effective in reducing
short-circuiting

•	If the full surface area upstream of the proposed weir could be utilized for settling, the
long-term solids trapping efficiency would be approximately 50%, which is higher than
the estimated trapping efficiency for all scenarios reported in recent EPA / HydroQual,
Inc. investigations

Elevated bed shear stresses, as predicted by the Housatonic River EFDC model daring event
flows, appear to prevent particles from depositing to the bed and limit solids trapping in Woods
Pond under the current pond configuration

•	The proposed weir, along with potential dredging of the pond, would reduce velocities
and shear stresses, allowing for settled solids to deposit and remain in the sediment bed

Background

Several rounds of analysis of Woods Pond trapping efficiency have been performed by
HydroQual, Inc. (HQI) as part of past hydrodynamic and sediment transport investigations.

501 Avis Drive
Ann Arbor, Ml 48108
734-332-1200

Fax: 734-332-1212
www.limno.com


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 2

Short circuiting of flows through the pond has been identified as a major limitation for in-pond
solids settling, with river flows modeled as traveling directly from the north center of the pond,
following the existing thalweg, and leaving through the pond spillway during high flows. The
most recent investigation performed by HQI in the spring and summer of 2011 analyzed several
alternatives for the trapping of solids and associated PCBs using the EFDC mode 1 for the site,
with the goal of increasing Woods Pond trapping efficiency. Results of this investigation are
reported in a September 13, 2011 draft document titled "Analysis of Woods Pond Solids andPCB
Trapping Efficiencies" (HQI trapping analysis).

Trapping efficiency results for the HQI-evaluated model alternatives ranged from a low of 17%
for the MNR scenario to a high of 36% solids trapping when Woods Pond was significantly
deepened by up to 15 feet. LimnoTech was subsequently asked to evaluate additional conceptual
alternatives to see if these alternatives could potentially provide a higher level of trapping
efficiency. The evaluations performed build upon the HQI analysis, with a focus on reducing
pond short circuiting, lowering in-pond velocities and promoting a longer residence time during
higher flows. LimnoTech has developed a spreadsheet based approach to evaluate solids loading
to Woods Pond, settling and deposition within the pond, and an estimate of trapping efficiency to
assist in determining whether these alternatives should be explored further.

The short circuiting of flow through Woods Pond during higher flow events that is limiting
trapping efficiency could be mitigated in concept through the installation of a long low head weir
stretching north to south across the western side of the pond in combination with in-pond
dredging. This weir would be approximately 960 feet in length and would be set at the same
height as the spillway elevation at the downstream end of the pond. It is assumed that the
installation of this weir with dredging would reduce the short circuiting currently occurring by
distributing flows along the western edge of the pond, allowing the full area of the pond
upstream of the weir to be used for settling. A schematic of this concept is shown in Figure 1.

The effectiveness of this weir concept is enhanced by the proposed in-pond dredging on the
upstream side of the weir, further reducing velocities and increasing solids residence time during
high flow by eliminating the preferential flow path currently present. The trapping efficiency of
the weir option alone was not estimated, because the spreadsheet level analysis could not
estimate the extent of short circuiting that would still exist if the thalweg present upstream the
weir remains intact. During high flow events, river flows and associated solids could potentially
be routed preferentially through the thalweg to the weir, with flows then spreading across the
length of the weir before continuing downstream. This routing could utilize a percentage of the
effective surface area similar to or greater than existing conditions, but the degree of change
caused by the weir could not be determined using the spreadsheet level analysis. These concepts,
however, can and should be vetted through the use of a model or other higher level
hydrodynamic computational analyses to ensure that the proposed weir would be effective in
preventing pond short circuiting, with or without dredging.

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 3

mmmwstmma

l

¦ N..-.. ' !4 i

¦b#.. mmM

5 ' Kmrnw

¦KmvA ^kAWKvv

/sK-f . :• >4.	(WAVA

V	v - ; V, ¦ jfe \

wKw^KwK^i
s /

Mh*KvXvX*XvX%%v«

CKvK'KsmwKvd
/^r je ¦

W

/»W»W»V\ • /

	—

wXwK "7
ftwwKwKvJ

?JAV»V»V*V«V*J «•
f . ~: _ v

	-— ¦ /

LimnoTech

Figure 1: Woods Pond Trapping Alternative Utilizing Long Low-Head Weir

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 4

Solids loading to Woods Pond

Flow and load estimates for Woods Pond were developed using relationships to gaged flows
outlined in Attachment B.3 of Appendix B of the December 2004 Model Calibration Report
(Weston 2004). Flow and load relationships were developed for several locations within the
project study area, including New Lenox Road and the exit of Woods Pond. For purposes of this
analysis the relationships developed for New Lenox Road were used to develop a time series of
flows and solids loads delivered to the Woods Pond entrance.

Flow time series estimates used were based on measured flow rates at the USGS Coltsville gage
and staff gage measurements and rating curves developed by Weston and HQI for each site of
interest. To develop this series, available hourly USGS Coltsville gage data from October 1,

1990	through September 30, 2009 were obtained from the USGS gage website. Interpolated flow
values were created for any gaps in the hourly data. A complete hourly flow series for this time
period was then created for New Lenox Road incorporating the regressions created for rising,
steady, and falling flows at Coltsville as detailed in Attachment B.3 of the Model Calibration
Report. These regressions are labeled as equations 5, 6, and 7 in the document.

The New Lenox Road flow time series was used to develop a TSS load time series based on
regressions developed by Weston and HQI and described in Attachment B.3. These regressions
were developed using the predicted flow for New Lenox Road and locally collected water
column TSS data. The complete flow data set for New Lenox Road was smoothed as described
in Attachment B.3 for use in calculating TSS, and TSS time series for rising, steady, and falling
flows were then compiled using equations 25, 26 and 27, with one exception. Using the
concentration coefficient of 1.78 x 101 as stated in equation 26, the steady flow TSS regression,
gave TSS values orders of magnitude higher than the TSS loads predicted for the rising and
falling limb and caused the yearly TSS load to be orders of magnitude higher than the yearly
average load shown in Table 2 of Attachment B.3. Changing the coefficient value to 1.78 x 10"1
for the steady flow regression corrected these issues.

The flow and load time series calculated using these regressions were then compared to the
available outputs shown in Attachment B.3. The calculated flow time series visually compared
well with the calculated New Lenox Road time series shown in Figure 5 of the attachment that
used the developed regression relationships, with a similar peak flow rate and similar rise and
fall times calculated for the September 1999 event shown. Table 2 within Attachment B.3 states
the annual TSS load at New Lenox Road as 4,767 MT/yr between 1988 and 2001. A direct
comparison with this value is not possible because complete years for 1988 and 2001 were not
available on the USGS website. However, the average yearly TSS load calculated for the years

1991	through 2001 using the regressions provided was calculated to be 5,015 MT/yr, a difference
of 5 percent. Visual comparison of the average flow rates and yearly TSS loads graphed in
Figure 15 of Attachment B.3 for New Lenox Road gave good agreement with calculated values
as well. Overall, the comparisons of flows and solids developed for this analysis to those
reported in the Model Calibration Report demonstrated that the methodology applied for this
analysis is consistent with that used for the modeling.

The Housatonic River flows, as calculated above at New Lenox Road, were used directly in this
Woods Pond trapping efficiency analysis. The solids load estimates, however, were modified to
reflect deposition occurring between New Lenox Road and the entrance to Woods Pond. A
reduction of 28% was used in order to reduce the annual average solids load computed at New

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 5

Lenox Road (5,463 MT/yr for 1990 - 2009) to the annual average solids load at the entrance of
Woods Pond (4,000 MT/yr), as predicted for the 52-year MNR model simulation (Table 1 of
HQI trapping analysis). This reduction also compares favorably with visual comparisons of
modeling and data shown in the Model Validation Report (Weston 2006) at New Lenox Road
(Fig. 6-2.20) and at Woods Pond Headwaters (Fig. 6-2.21).

The resulting October 1990 through September 2009 flow and solids loads at the Woods Pond
entrance are shown in Figure 2. The flows are plotted as a cumulative frequency and the solids
loads are plotted as a cumulative contribution to the total solids load. This figure highlights the
importance of high flows in delivering solids to Woods Pond. For example, flows less than or
equal to the median flow deliver only 10% of the solids load to Woods Pond, with the remaining
90% coming from flows in excess of the median. Flows less than or equal to the 90th percentile
flow deliver only 35% of the solids load, with the remaining 65% of the solids loads coming
from the upper 10th percentile of the flow distribution. To trap a large percentage of the solids
load delivered to Woods Pond, any pond modifications need to target the higher flows that occur
more infrequently.

100%

T3

2 90%

LO

to

2	80%

C

.o

5 70%

c

0

V	60%

.2

Ł 50%

3

u

« 40%

V
3

Ł 30%

S
_o

t 20%

_>

_ra

1	10%


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 6

Trapping efficiency description

The portion of solids mass that deposits to the sediment bed relative to the solids mass that enters
a water body is termed the "trapping efficiency". It expresses the fraction or percentage of solids
that are retained (trapped) within the water body. The HQI trapping analysis presents an
expression for trapping efficiency as a function of particle settling velocity, surface area of the
water body, and flow rate:

TE = 1 - e~wA^Q	(1)

Where:

TE = trapping efficiency [dimensionless]
w = particle settling velocity [L/T]

As = surface area of water body [L2]

Q = flow rate [L3/T]

Equation 1 is used for this spreadsheet analysis of Woods Pond trapping efficiency. It describes
the one-dimensional gradually-varied flow case for computing trapping efficiency. It assumes
complete mixing of solids in the water column, no erosion of solids from the sediment bed, and
no "non-deposition" (i.e., solids that settle to the bed are assumed to deposit there, rather than
being entrained back into the water column).

As described by Equation 1, the following items improve the trapping efficiency: faster settling
particles, larger surface area, and lower flow rates. Relatively little can be done to alter flow rates
in the system or to alter the particles transported to Woods Pond and their associated settling
velocities. The primary focus of this analysis is to describe the expected improvement in trapping
efficiency attainable by maximizing the surface area of Woods Pond that is available for
deposition. Currently, short-circuiting of flows prevents the pond from attaining its full trapping
potential. This analysis also considers the validity of the no erosion and no non-deposition
assumptions by examining expected shear stresses within Woods Pond.

The settling velocity used in applying Equation 1 is based on the settling velocities used to
calibrate the Housatonic River sediment transport model. The model uses a settling velocity that
varies based on TSS concentration, calculated by applying a velocity of 0.1 m/day to the first 5
mg/L of TSS (intended to represent washload particles) and a velocity of 30 m/day to the
cohesive TSS in excess of 5 mg/L. It is a way to model two different types of cohesive particles,
while actually including only one cohesive particle in the model. For this spreadsheet analysis,
the same model-calibrated settling velocities are used. However, Equation 1 is applicable only
for a constant settling velocity; therefore, the equation is applied separately for each of the two
settling velocities. The total trapping efficiency is then computed by combining the trapping of
washload solids with the trapping of non-washload cohesive solids, as shown in Equation 2:

rrr?	_ 1 cOUT _ 1 CWL(1-TEWL)+ cCOh0-~TeCOh)	/-0\

I tT0TAL — I - — I	-	(2)

lin	lin

Where:

TEtotal = combined trapping efficiency for all cohesive solids [dimensionless]

TEwl = trapping efficiency for washload solids (from Equation l) [dimensionless]

TEcoh = trapping efficiency for non-washload cohesive solids (from Equation l)
[dimensionless]

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 7

Corn = total suspended solids concentration out of Woods Pond [M/L3]

Cin = total suspended solids concentration into Woods Pond [M/L3]

Cwl = incoming suspended solids attributed to washload (up to 5 mg/L) [M/L3]

Ccoh = incoming suspended solids attributed to non-washload cohesive solids (CiN-CWl)
[M/L3]

Trapping efficiencies computed from Equations 1 and 2 are plotted in Figure 3 as a function of
flow rate. The surface area used for these calculations is 48.4 acres, equal to the Woods Pond
surface area upstream of the weir shown in Figure 1. The TECoh and TEWL results are computed
from Equation 1. The TEtotal results are computed from Equation 2 for three example Cm
concentrations (10 mg/L, 20 mg/L, and 50 mg/L). The trapping efficiency for washload solids is
low, even at low flow rates, due to the slow settling velocity of these particles. The non-
washload solids trapping efficiency is nearly 100% at low flow and declines with increasing flow
rate.

As shown in Figure 3, the total trapping efficiency depends on flow rate as well as the incoming
suspended solids concentration. As the incoming concentration increases beyond the 5 mg/L
washload concentration, TETOtal increases from TEWL towards TECoh• This is the result of faster
settling solids becoming more dominant as the suspended solids concentration increases.

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

pagei

100% »~»~».

S

>
u
c

.Si

"u

Ł

ClD
c

a

Q.


"O

o
o

~ TE, COH
¦ TE, WL

ATE, TOTAL (lOmg/L)
XTE, TOTAL (20mg/L)
X TE, TOTAL (50mg/L)

:^55sb&»

1,000

2,000

3,000

4,000
Flow (cfs)

5,000

6,000

7,000

8,000

Figure 3: Potential Woods Pond Trapping Efficiencies with Weir as a Function of Flow and TSS

Results of TE analysis

The equations shown above describe the trapping efficiency for a given flow and load condition.
By applying these equations to a long-term record of flows and loads, an estimate of the long-
term trapping efficiency of Woods Pond can be developed. Equations 1 and 2 were applied on an
hourly basis to the flow and solids load time series developed from October 1990 through
September 2009 to determine the overall trapping efficiency for this 19-year period.

Two different surface areas were used in computing the long-term trapping efficiency - one to
represent the proposed weir scenario and one to represent the current Woods Pond condition.

Trapping efficiency with the proposed weir

The surface area used for the weir scenario is the pond surface area that is upstream of the
proposed weir shown in Figure 1 (48.4 acres, or 2,110,000 ft2). The weir is intended to prevent
short-circuiting in the pond, and the computed trapping efficiency for this scenario assumes the
flow spreads throughout the pond upstream of the weir, utilizing this large surface area for
deposition.

For this scenario, the estimated long-term trapping efficiency is 50%. This estimated trapping
efficiency is three times larger than the predicted trapping efficiency under MNR and larger than
any of the alternative scenarios evaluated in the HQI trapping analysis.

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 9

Figure 4 shows the estimated trapping efficiencies for the October 1990 through September 2009
time period as a function of flow. The figure shows how the competing factors of settling
velocity and residence time play out over the various flow and solids load conditions in Woods
Pond. Trapping efficiency is very low at the lowest flows due to the slow settling velocity of
washload particles, which dominate the TSS load at low flow. As the flow rate increases from
low-flow conditions, there is an increase in non-washload particles with faster settling velocities,
which causes the trapping efficiency to increase. Trapping efficiency peaks at approximately
70%, when flows are near 1,000 cfs. For flow rates in excess of 1,000 cfs, no further increases in
settling velocity are expected as the TSS load is already dominated by non-washload particles.
The trapping efficiency declines above 1,000 cfs due to the shorter residence time associated
with increasing flow rate.

100%
90%
80%

u

Q 20%

8,000

10%

0%

1,000

2,000

3,000

4,000
Flow (cfs)

5,000

6,000

7,000

Figure 4: Estimated Woods Pond Trapping Efficiencies with Weir for Oct 1990 - Sept 2009 Flow
and TSS Load Time Series

Increasing the surface area of the pond through dredging of selected areas immediately upstream
of the pond would increase the trapping efficiency at all flow rates. For example, the additional
dredging of 16.5 acres upstream of the pond, a distance of roughly 850 feet upstream, would
increase the trapping efficiency to approximately 80% of all particles at a flow of 1,000 cfs and
corresponds to a long term trapping efficiency of 55%. To obtain this additional trapping

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis	page 10

efficiency would require large amounts of initial dredging, and any benefit would have to be
weighed against the cost of this dredging and accompanying habitat loss.

Trapping efficiency for the current condition

A second long-term trapping calculation was done by varying the pond surface area until the
computed long-term trapping efficiency matched the 17% reported for the MNR model scenario
in the HQI trapping analysis. This derived surface area represents the current "effective surface
area" of Woods Pond. The computed effective surface area is 310,000 ft2, which is roughly 15%
of the total pond surface area. The computed effective surface area appears consistent with the
primary flow pathway in Woods Pond during high flow conditions (Figure 3 of HQI trapping
analysis), which supports the notion that short-circuiting is a major reason for the relatively low
trapping efficiency currently observed in Woods Pond.

Deposition of solids

The above analysis of trapping efficiency assumes that near-bed shear stresses are low enough to
allow particles settling near the bed to deposit. If near bed shear stresses are too high, particles
settling near the bed may be entrained in the flow, rather than deposit to the bed, reducing the
trapping efficiency. This non-deposition may be contributing to the relatively low trapping
efficiencies predicted by the Housatonic River sediment transport model. Non-deposition is
represented in EFDC and other models by a probability of deposition function. The probability
of deposition is used compute an effective particle settling velocity, which is a reduction of the
quiescent settling velocity:

wse = Pk*ws	(3)

Where:

wse = effective settling velocity [L/T],

Pk = probability of deposition [dimensionless], and

Ws= quiescent settling (fall) velocity [L/T],

The probability of deposition has a value between 0 and 1, defined in the Housatonic River
model application as:

Pk =

fX_Jb\f°rTb < Tcd

V Tcd J

(4)

0	: M rh > Tcd

Where:

rb = total bed shear stress [M/L/T2], and

Tcd = critical shear stress for deposition [M/L/T2].

The Housatonic River model was calibrated with a value of rcd equal to 2 dynes/cm2. Under
quiescent conditions (zero bed shear stress), Pk= 1, and particles settle to the bed and deposit
based on their quiescent settling velocity (ws). Under flowing conditions, as the bed shear stress

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 11

approaches 2 dynes/cm2, Pk approaches zero, the effective settling velocity (wse) approaches
zero, and no particles deposit to the bed.

EFDC model results shown in the HQI trapping analysis suggest that solids are, at times,
prevented from depositing to the bed. Specifically, the "Flocculant Addition" scenario, which
doubles the cohesive settling velocity, does not result in the expected increase in trapping
efficiency. In this model scenario, the trapping efficiency increases from 17% to 21%. Based on
the one-dimensional trapping efficiency analysis described above for the current condition, an
increase to approximately 28% is expected when settling velocities are doubled. The less-than-
expected increase may be due to elevated shear stresses preventing deposition. Results of other
model scenarios also support the hypothesis that non-deposition due to elevated shear stresses is
limiting the predicted trapping efficiency. The model scenarios with the highest predicted
trapping efficiencies all involved some deepening of Woods Pond ("Woods Pond Deepening",
"Massachusetts Alternative", and "Baffles" scenarios). While, in theory, deepening does not
directly increase trapping efficiency, deepening of the pond would result in lower velocities and
lower shear stresses within the pond, increasing the probability of deposition.

With the proposed modifications to Woods Pond (weir and in-pond dredging), shear stresses
would typically be low enough to prevent non-deposition from limiting the pond's trapping
efficiency. As an example, with a weir length of 960 ft and a water depth of 9 ft, a high flow of
2,000 cfs results in an average velocity of 0.23 ft/s. Using the bed roughness of 0.02 m from the
Housatonic River model, the bed shear stress would be 0.55 dynes/cm2 and the probability of
deposition would be 0.721. Under the current pond configuration, with shallower depths and
short circuiting, the average velocity within the preferential flow path for this 2,000 cfs flow
could easily be twice as high (0.5 ft/s or greater), resulting in a shear stress in excess of 2
dynes/cm2 and no deposition (probability of deposition = 0).

This trapping efficiency analysis, based on Equation 1, also assumes no erosion of the sediment
bed in Woods Pond. The critical shear stress for erosion is higher than the critical shear for
deposition. Since shear stresses with the proposed weir and in-pond dredging are expected to be
low enough to largely avoid non-deposition, erosion of the bed is also not expected. Even with
the current pond configuration, results of the EFDC model show only a small contribution of
erosion to the total solids load in Woods Pond (Table 1 of HQI trapping analysis).

Conclusions

This analysis demonstrates that the use of a weir on the western side of the pond in conjunction
with in-pond dredging could produce higher solids trapping efficiencies than those reported in
recent EPA / HydroQual, Inc. investigations. The results reported in the HQI trapping analysis
are based on EFDC modeling, which is more computationally sophisticated than the spreadsheet
analysis presented here. However, despite the relatively simple computational framework for this
analysis, it is based on sound theory and reasonable assumptions. In fact, a spreadsheet analysis
is often needed to better understand the results generated by more complex modeling tools. A

1 The velocity estimate assumes flow is evenly distributed throughout the pond. The model-calibrated bed roughness
of 0.02m is relatively high for Woods Pond. A high roughness value results in high model predictions of total
bottom shear stress. This estimate of shear stress does not consider the role of macrophytes, which would tend to
reduce velocities and bottom shear stress when present during the year. If the pond is dredged, macrophyte growth
will be limited.

LimnoTech


-------
Woods Pond Alternatives Conceptual Trapping Analysis

page 12

detailed understanding of the model results is essential for making an assessment of the model's
predictive capability.

If solids trapping efficiencies of around 50% for Woods Pond are in the range of what is needed
to make better use of the pond within a remediation alternative, then the results of this analysis
and the potential modifications to Woods Pond warrant further investigation. Several inputs to
this analysis, such as settling velocities and assumed washload concentration, could be adjusted
within reasonable ranges to assess the impact each has on predicted trapping efficiency.
Alternative assumptions could also be assessed, such as the assumption of complete mixing of
solids in the water column that is implicit in the one-dimensional trapping efficiency expression.
The primary consideration, however, would be to assess the effectiveness of the proposed weir
(with and without in-pond dredging) in preventing short circuiting, reducing velocities, and
increasing residence time in the pond. These concepts should be vetted through the use of a
model or other higher level hydrodynamic computational analyses. The existing EFDC model
could be used for this purpose.

The existing EFDC model could also be used to translate improvements in solids trapping
efficiency to improvements in PCB trapping efficiency and reductions in downstream PCB
transport. Increased solids trapping in Woods Pond would increase PCB trapping in the pond,
reducing downstream transport. If in-pond dredging were implemented, downstream transport of
PCBs from Woods Pond would also be reduced through PCB removal during dredging to attain
and maintain a deeper pond. Due to higher organic carbon content, washload particles will
typically transport relatively more PCB than lower organic carbon particles delivered during high
flows. These considerations, as well as non-erosion based fluxes of PCB from the sediment bed
play a role in the total trapping efficiency for PCBs. The EFDC model is an effective tool for
quantifying the benefits of hydrodynamic and solids trapping improvements on PCB transport.

References

HydroQual, Inc. 2011. Analysis of Woods Pond Solids and PCB Trapping Efficiencies - Draft.
September.

Weston. 2004. Model Calibration: Modeling Study of PCB Contamination in the Housatonic
River, Appendix B, Attachment B.3. Prepared for U.S.Army Corps ofEngineers and U.S.
Environmental Protection Agency. DCN: GE-122304-AMCG. December.

Weston. 2006. Model Validation: Modeling Study of PCB Contamination in the Housatonic
River. Prepared for U.S. Army Corps ofEngineers and U.S. Environmental Protection Agency.
DCN: GE-030706-ADBR. March.

LimnoTech


-------
DRAFT

PRIVILEGED

CEERD-EP-E	14 February 2012

Memorandum for: Mr. Dean Tagliaferro, Project Manager for GE-Pittsfield/Housatonic River Site, U.S.
Environmental Protection Agency Region 1

By: Paul R. Schroeder, PhD, PE, Environmental Laboratory, USACE ERDC

SUBJECT: Review of Limnotech's Woods Pond Alternatives Conceptual Trapping Analysis Memo dated
1/12/2012

1.	I have reviewed the subject memo and agree in general terms with the findings of the analysis
as summarized in the memo. In particular, effective trapping of solids requires the ability to trap solids
delivered under moderate to high flow conditions; preventing short circuiting of Woods Pond
significantly increases trapping efficiency; and elevated bed shear stresses, as predicted by the
Housatonic River EFDC model during event flows, appear to prevent particles from depositing to the bed
and limit solids trapping in Woods Pond under the current pond configuration. Addressing these three
issues in the design of Woods Pond and the backwater areas will significantly improve trapping
efficiency of solids and PCBs.

2.	Limnotech proposes dredging and a submerged weir in Woods Pond to address these three
issues. The submerged weir is designed to spread the flow and it will decrease short-circuiting, but it
will not distribute the flow in Woods Pond nearly as effectively as training dikes. The memo does not
give an estimate of the short-circuiting predicted with the weir in place. The memo implies that short-
circuiting would be eliminated. Dredging is proposed to slow the flow, reduce short-circuiting, eliminate
resuspension, and increase trapping under high flow conditions. The depth of the dredging is not
provided in the memo, but it is not needed based on the trapping efficiency formulation used in the
analysis as long as it is sufficient to reduce bottom shear stress sufficiently to limit resuspension. The
analysis predicted a solids trapping efficiency of 50%, which should be achievable.

3.	Greater trapping could be achieved by including the backwater areas in the flow to increase the
surface area for settling, which would increase the solids particularly under high flow conditions.

4.	Limnotech did not provide an estimate for PCB trapping efficiency; however, the efficiency was
predicted to less than that for solids. Based on my previous analysis of PCB and solids trapping
efficiency for Woods Pond and backwater system alternatives, a solids trapping efficiency of 50% would
yield about 32% trapping of PCBs. The results of Limnotech analysis for solids trapping efficiency are
similar to the estimates that I had made for Woods Pond, but perhaps somewhat smaller depending on
the particulars of the reduction in short-circuiting and the deepening.


-------
ATTACHMENT C-4 HOUSATONIC RIVER STATUS REPORT:
POTENTIAL REMEDIATION APPROACHES TO THE GE-
PITTSFIELD/HOUSATONIC RIVER SITE "REST OF RIVER" PCB

CONTAMINATION


-------
U.S. E PA I HOUSATONIC RIVER STATUS REPORT

LEARN MORE AT:www.epa.gov/region 1/ge

Potential Remediation Approaches to the GE-Pittsfield-
Housatonic River Site "Rest of RiverPCB Contamination

THE RIVER The Housatonic River is contaminated
with polychlorinated biphenyls (PCBs) and other hazardous
substances released from the General Electric Company
(GE) facility in Pittsfield, MA. The entire site consists of the
254-acre GE facility; the Housatonic River and its banks and
floodplains from Pittsfield, MA, to Long Island Sound; and
other contaminated areas. Under a federal Consent Decree,
GE is required to address contamination throughout the site,
including in the River.

INTRODUCTION:

EPA and the states of Massachusetts and Connecticut
(collectively, the "Parties") have been working coopera-
tively for the last several months to discuss potential ap-
proaches to clean up the Rest of P,iver portion of the GE
Housatonic site. These discussions have focused, in part,
on the need to address the risks from polychlorinated
biphenyls (PCBs) to humans, fish, wildlife and other
organisms while avoiding, mitigating or minimizing the
impacts of the cleanup on the unique ecological charac-
ter of the Housatonic River. This summary document re-
flects the current status of the Parties' preliminary, good
faith efforts to discuss and identify potential remedial
approaches for the Rest of River in light of the Parties'
shared goals and interests. The Parties recognize that no
remedy decisions have yet been made, EPA will consider
all relevant information, and any remedy proposal and all
other information in the administrative record, including
this status report, will be subject to public comment at the
time that EPA issues a proposed cleanup plan.

In areas of the Housatonic River and its floodplain, PCBs
in the sediment, soil and surface water pose an unac-
ceptable risk to human health and/or the environment.
The governmental agencies are considering a cleanup
plan consisting of a combination of targeted soil and sedi-
ment removal, riverbed capping, and monitored natural
recovery as a potential means of addressing the PCBs
posing the greatest threat and achieving goals such as
the following:

• reduce risks to children and adults from di-
rect contact with soil and sediment;

•	reduce soil contamination in the floodplain to
levels which allow continued recreational use
without unacceptable risk;

•	reduce PCB concentrations in fish to levels
that allow increased consumption of fish
caught from the River in Massachusetts and
Connecticut and reduce impact to affected
communities relying on the fish for economic
considerations or cultural practices;

•	reduce the potential movement of PCBs from
the river onto the floodplain, from the banks
into the River, and from upstream to down-
stream locations, including the downstream
transport into Connecticut;

•	reduce contamination to acceptable risk for
ecological receptors (fish, wildlife, and other
organisms) in the river, floodplain, and vernal
pools;

•	reduce PCB surface water and sediment con-
centrations by addressing PCB sources in sedi-
ment and soil to advance future compliance
with water quality standards in Massachusetts
and Connecticut and attainment of the high-
est possible use of the River consistent with
the Clean Water Act; and,

•	protect and preserve the unique ecological
characteristics of the Upper Housatonic Wa-
tershed in conducting remedial efforts.

Based upon those discussions, and EPA's ongoing analysis
of the nine criteria in the RCRA permit (subject to fur-
ther information and analysis as EPA continues to review

KEY CONTACTS:

JIM MURPHY

U.S. EPA

(617) 918-1028

murphy.jim@epa.gov

DENNIS SCHAIN

CT Dept. of Energy &
Environmental Protection
(860) 424-3110
dennis.schain@ct.gov

ED COLETTA

Mass DEP

(617) 292-5737

edmund.coletta@state.ma.us

BOB GRECO

MA Dept. of Fish & Game
(617) 626-1556
bob.greco@state.ma.us

GENERAL INFO:

EPA NEW ENGLAND

5 Post Office Sq,,

Suite 100

Boston, MA 02109-3912

TOLL-FREE
CUSTOMER SERVICE

1-888-EPA-7341

continued >

Jl United States

Environmental Protection
M * Agency

May 2012


-------
Pitts fidil

Confluence

Reach 5

ReaGh 6 (Woods Pond)

Reach 7
Reach 8 (Rising Pond)

Berkshire

Reach 9

Reach 10

Reach 11

Cornwall
(Bridge

Reach 12

CONNECTICUT

Reach 13

Fairfield

GE-Pittsfield/Housatonic River Site
HOUSATONIC RIVER, REACHES 5 THROUGH 17

Hampshire

NEW YORK

Hampden

r

MASSACHUSETTS

Hartford

¦ Reach 14 (Lake Lillinonah)

New Haven

Reach 15 (Lake Zoar)

Reach 16

(Lake Housatonic)

Long Island Sound

11D-0379


-------
Confluence of East and West Branches

Reach 5A

MsrtontAtlb'

luersCorrer

11 EM OK

lldm-Jtf-j

Reach 5C

^chJSDJ
^waters

iRWmoni

Vtfi. B HI liOTO I

J %, " 1' lir»3xStsfc>n

'Old 'ijtfiyjcfe Pond Dam j
WcCtids Pond Dam 7

Reach 6 (Woods Pond)

Reach 7A

Reach 7B

Columbia Mil/1!
Lee.'Eagle Mills Dai

Reach 7C

:sl SIceMjrUDe Omfcr

EtlMt ~

lendale

Willow Mill Dam \



F a : 'i'f fte'U2 'JG gep ttWn xds t b_ n acle s_ fetfc r_m 3/11 jti x d. 02-11^-11 11:1 T. rtefcic

LEGEND:

o Towi/City
Roads

A/ Readi Division Line

Housatonic River
State P ark
Municipal Boundary1
10-'Year Floodplain

~s d.«s ~ nsKilometera

GE - Pittsfield.1Housatonic River Site
Rest of River

HOUSATONIC RIVER,
PRIMARY STUDY ARE A
(REACHES 5 AND 6) AND
REACHES 7 AND 8


-------
the CMS and other information), a potential framework for a remedy tenta-
tively would include the following elements:

•	Removing PCB-contaminated bed sediment and containing residual
PCB contamination in some reaches in the Housatonic River using a
combination of excavation/dredging and capping.

•	Removing PCB-contaminated soil from some areas in the 10-year
floodplain adjacent to the river.

•	Minimizing bank erosion and PCB transport downstream through
the use of natural channel design, sediment trapping measures, and/
or other engineering controls, while preserving the dynamic River
habitat.

•	Avoiding, minimizing or mitigating impacts to state-listed species of
concern, including targeting of floodplain remediation to avoid cer-
tain critical core state-listed species habitat.

•	Employing an adaptive management approach to ensure that the
cleanup is performed using the best available technologies and meth-
ods.

•	Taking advantage of existing rail infrastructure, transporting of con-
taminated soils and sediment to existing licensed off-site disposal fa-
cilities.

•	Providing for long-term monitoring, maintenance, and institutional con-
trols as well as reviewing the remedy every five years to evaluate the
effectiveness and adequacy of the cleanup.

Rough volumetric estimates for a cleanup plan meeting these goals could range
from 750,000 cubic yards to 1 million cubic yards (yd3) of removal, in addi-
tion to containment of residual PCBs and monitored natural recovery. Actual
volumes would certainly vary based on the performance criteria and design,
which are yet to be decided, and the removal volumes could be reduced below
this range. To put this in perspective, some estimates show that there are over
4 million cubic yards of contaminated soil and sediment exceeding 1 part per
million PCBs in the River and floodplain.

STATE INVOLVEMENT MOVING FORWARD
We expect these discussions to continue as EPA develops the proposed per-
mit modification in consultation with the states. EPA will propose a remedial
approach for the Rest of River for public comment, and the Parties reserve
judgment on the appropriate remedy until after the comment period closes.
The approach set forth in this document is only preliminary and tentative, sub-
ject to further information that will be reviewed as part of the consideration
of the CMS, and intended to provide the public with information about these
joint governmental discussions. The Parties reserve all rights available under
the Consent Decree, CERCLA, RCRA, CWA or other applicable law.

EPA, Massachusetts and Connecticut commit to a long-term effort to
work cooperatively in reviewing comments on the to-be-proposed clean-
up plan, finalizing the cleanup plan, reviewing the details of the Remedial
Design, and overseeing the remedy's implementation and long-term moni-
toring. EPA expects to work closely with the States on the development
of the performance standards, corrective measures and on identification
of ARARs for the draft permit before it is issued for public comment. EPA
sees an organized, on-going and inclusive partnership with the states as
essential to successful implementation of the final remedy.

There will be an opportunity for review and comment by Massachusetts
and Connecticut on key GE deliverables, including the draft Statement of
Work, work plans, design submittals, and other documents submitted dur-
ing design, construction, and long-term monitoring. We look forward to
a cooperative approach to moving this project forward and ensuring that
issues of importance to the states are addressed at each step of the process.

EPA will work with the states to develop a structured process for inter-
agency engagement prior to finalization of the Permit as well as during de-
sign and implementation of the remedy. In addition, the states and EPA will
develop an interagency process as a means to address issues that may arise.

PRELIMINARY THOUGHTS ON POTENTIAL
REMEDY COMPONENTS

The following pages provide a brief analysis of various elements of a poten-
tial proposed cleanup plan. This plan represents the preliminary thinking
at this stage of the technical discussions among the States and EPA

NOTE: The assumptions contained in this document are based on current data
and modeling; current data and monitoring are very useful to compare various
alternatives and actions to assess relative effects of different choices. However,
more deliberations will occur prior to the proposal or selection of corrective mea-
sures for Rest of River. In addition, following selection of the corrective measures
for Rest of River, remedial design activities will include additional investigations
and sampling of the river, floodplains and vernal pools. The data and refined
information gathered during that investigation and sampling will include data on
concentrations of PCBs, the presence of MESA species, and other parameters.
The refned information on appropriate parameters will be used to determine the
final scope, location, and volumes of the final corrective measures.

ACRONYMS

ARARs	Applicable or Relevant and Appropriate Requirements

CERCLA	Comprehensive Environmental Response, Compensation, and Liability Act

CMS	Corrective Measures Study

CWA	Clean Water Act

IMPGs	Interim Media Protection Goals

MassDFG/DFW 	Massachusetts Department of Fish and Game / Division of Fishieries & Wildlife

PCBs	polychlorinated biphenyls

RCRA	Resource Conservation and Recovery Act

page 4


-------
I. REACH 5 RIVER BED AND BANKS
Area of the river

For the purpose of current deliberations, the government agencies have exam-
ined a 10-mile long section of the river which includes Reaches 5A, 5B, and 5C
from the confluence of the East and West Branches of the Housatonic River
to the headwaters of Woods Pond.

A potential approach to cleanup

At this stage of deliberations, the government agencies are discussing remedia-
tion of River Bed in Reaches 5A and 5C. There have been tentative discussions
of removing contaminated sediment from the river bed to a depth sufficient
to place a multi-layered engineered cap to sequester remaining contamination.

Current thinking is that the cap would be expected to consist of, at a mini-
mum, an isolation layer with organic carbon, a protective layer, and a habitat
layer. This may, of course, change as further information is developed, such
as during design.

Remediation of erodibie river banks in Reach 5A. At present, the preliminary
thinking is that any bank remediation/restoration approach would follow a
hierarchy of most preferred to least preferred:

1.	Leave banks intact (no disturbance/excavation)

2.	Reconstruct disturbed banks with bio-engineering "soft" restoration tech-
niques

3.	Reconstruct disturbed banks with a cap layer extending into the river bank
placed under a bio-engineering/"soft" layer

4.	Place rip-rap cap or hard armoring on surface of banks (only for protection
of adjacent infrastructure)

This potential approach being considered by the government agencies would, based on current data, entail the disturbance of one third of the Reach
5A riverbanks (or 3.5 miles), recognizing that minimizing the length of affected bank could require capping into the riverbanks throughout that area
in order to achieve sufficient bank stabilization. This potential remediation approach would follow the hierarchy of bank stabilization options outlined
above, and following this hierarchy may result in the disturbance of less than one-third. This potential approach would minimize impacts to species
listed under MESA and help preserve the dynamic character of the river. The government agencies will continue to evaluate these and the other goals
on page 1 above throughout the design of any response action.

Hot spot remediation of River Bed and Banks in Reach 5B, The government agencies are currently looking at an approach under which a limited
amount of work in Reach 5B would be conducted in order to address surface sediment or riverbank soil exceeding 50 ppm PCBs. Under such a
potential approach, any excavated Reach 5B riverbanks could be restored using bioengineering "soft" restoration techniques. EPA could also explore
the potential addition of activated carbon to surface sediment in the Reach 5B riverbed to further sequester residual contamination in that sub-reach.
If such an option were to be pursued, EPA estimates this hot spot remediation would be limited to less than 5% of the river banks and 5% of the
river bed, likely less than 1,000 cubic yards.

Cleanup goals associated with this potential approach

As part of the agencies' deliberations, EPA modeled the potential approach above, and believe that it would be expected to result in reductions in
fish tissue concentrations to allow increased consumption of fish caught from the river a short time after remediation is completed in these Reaches.
That is, the approach would achieve acceptable cancer and non-cancer residual risks for EPA's "Central Tendency" exposure scenario, representing
the average person's fish consumption rates. Remediation of sediment in Reaches 5A and 5C, erodibie river banks in Reach 5A, and hot spots in
Reach 5B would also reduce the transport of contamination into downstream reaches (thus improving overall water quality). This cleanup approach
would achieve human health benchmarks for children and adults for direct contact with sediment and reduces contamination to acceptable levels for
ecological receptors (fish, wildlife, and other organisms) over the long term in much of this stretch of the river.

Preliminary estimates of this approach indicate that it is expected to greatly reduce the contribution to downstream transport of contamination.
The potential approach could also include long-term monitoring, to evaluate if downstream transport is being minimized. Early analyses indicate that
reducing downstream transport would help continue the reduction of PCB levels in fish tissue in Massachusetts and throughout Connecticut, which,
over time, should allow the consumption of additional fish meals and revision or potential elimination of fish consumption advisories in parts of the
river in the future.as determined by each state through application of its state-specific fish consumption goals.

Estimated volumes

If cap thickness were assumed to have an average excavation depth of 2 Vi feet in Reach 5A and 2 feet in Reach 5C, the agencies estimate that
this component of the above potential remedial approach could entail the excavation of approximately 380,000 yd:l of contaminated bank soil and

page 5


-------
sediment or approximately one third of the total remediation volume estimated for this potential remedial approach. Of this amount, approximately
25,000 yd5 would be expected to be riverbank soil, a small percentage (2 to 3%) of the remediation volume under this potential approach.

Cap thickness could be a major driver of excavation volumes. In any proposed remedial approach, EPA would tend to specify certain cap design prin-
ciples and performance standards, but not a particular material thickness. Thus, during Remedial Design, GE would likely look to optimize cap designs
in order to lower the amount of sediment requiring excavation. For instance, if the designed cap thickness was reduced by one foot in Reaches 5A and
5C, this could result in more than a 40% reduction in volume to approximately 220,000 yd:!.

This potential approach would provide specificity on the options for removal so as to minimize impacts to the many species covered under the Mas-
sachusetts Endangered Species Act (MESA) which would also help preserve the dynamic nature of the river.

II. WOODS POND

Area of the River at issue

Woods Pond is Reach 6 of the Rest of River study area and covers approximately 54 acres
from the end of Reach 5 to the Woods Pond Dam.

A potential approach to cleanup

Excavation and Capping. At this stage of deliberations, the government agencies are dis-
cussing a proposed approach for the removal of contaminated sediment from the pond
prior to placement of an engineered cap to sequester the remaining contamination. The
proposed approach under consideration would involve the removal of sediment that
would result in a final average water depth of at least six feet in the pond after placement
of a one-foot thick cap. Current thinking is that the cap would be expected to consist of,
at a minimum, an isolation layer with organic carbon coupled with a protective layer and
a habitat layer.

EPA, in consultation with the states, is also considering additional deepening and/or place-
ment of structures within the pond to further promote sedimentation as well as the use
of activated carbon before and/or after final cap placement to sequester residual PCBs
and minimize the potential for downstream migration of PCBs. The agencies are looking
at possible designs for the final contours of the pond bathymetry that would support a
functioning littoral zone. Final removal depths, structures, and engineered cap configura-
tions would, of course, be determined during remedial design.

Cleanup goals associated with this potential approach

As part of their deliberations, the government agencies have modeled a variety of potential approaches, and based on that, believe that this approach
would be expected to result in reductions in fish tissue concentrations to allow increased consumption of fish caught from the river shortly after reme-
diation is completed. The agencies are further considering removal of additional mass of contamination from Woods Pond which, combined with cap-
ping, might also heip to reduce the transport of contamination into downstream reaches and further control a potential source of downstream release
in the unlikely event of dam failure. If downstream transport can be reduced, it should help continue the reduction of PCB levels in fish tissue in Mas-
sachusetts and throughout Connecticut, which, over time, should allow the consumption of additional fish meals and revision or potential elimination
of fish consumption advisories in parts of the river in the future, as determined by each State through application of its state-specific fish consumption
goals. Preliminary estimates of this potential cleanup approach indicate that it is expected to achieve human health benchmarks for children and adults
for direct contact with sediment and reduce contamination to acceptable levels for ecological receptors (fish, wildlife, and other organisms) over the
long term in Woods Pond. Any remedy component that is ultimately proposed and selected by EPA would, of course, be implemented to be consistent
with ARARs, including by avoiding, minimizing and mitigating impacts to state-listed species,

This option would potentially increase the sediment retention of Woods Pond and the retention of residual PCB contamination. This retention of
sediment from the river and floodplains may be expected to have the benefit of further decreasing the movement of PCBs downstream. Under any
remedial option, Woods Pond would be monitored over the long term and if substantial PCBs accumulate in the pond, removal of the residual PCBs
would be evaluated. As noted above, EPA would also explore the use of activated carbon before and/or after final cap placement to sequester residual
PCBs and minimize potential for downstream migration of PCBs.

Estimated volumes

The agencies estimate that if the excavation and capping of Woods Pond were carried out as described in the potential approach outlined above, it
would entail the excavation of approximately 285,000 yd3 or 29% of the total remediation volume. Under that scenario, the major factor behind the
excavation volumes would be the deepening of the pond. Pond deepening may be advantageous in some ways, as it can result in (a) removing mass of
PCBs from the Pond and (b) potentially enhancing the Pond's ability to trap solids and PCBs, thus reducing downstream transport.

page 6


-------
III. DOWNSTREAM IMPOUNDMENTS (REACHES 7 AND 8)

Area of the River at issue

For the purpose of current deliberations, the government agencies have examined
a potential remedy that would address the impounded areas behind several dams
located in the River south of Woods Pond. These include four dams in Reach 7 (Co-
lumbia Mill Dam, Eagle Mill Dam, Willow Mill Dam and Glendale Dam (subreaches
7 B, 7C, 7 E, and 7G, respectively) and Rising Pond (Reach 8). Each of the four Reach
7 Impoundments ranges from 8 to 12 acres each, while Rising Pond comprises ap-
proximately 45 acres.

A potential approach to cleanup

Excavation and Capping, At this stage of deliberations, the government agencies
are discussing a proposed approach for the removal of contaminated sediment
from the river bed prior to placement of a multi-layered cap to sequester remaining
contamination. Conceptually, this approach would include excavation of an esti-
mated 1 to 1 Vj foot (at least) of contaminated sediment, followed by placement
of a 1 to 1 % foot cap. Also under active consideration by EPA is the possibility of

establishing a secondary or "contingency" remedy allowing for the excavation of the impoundment behind Columbia Mill Dam or other Reach 7 or
Reach 8 impoundments to a level of 1 ppm PCBs to better dovetail the remedy with any potential dam removal projects by others.

Cleanup goals associated with this potential approach

As part of their deliberations, the government agencies have modeled the potential approach above, and believe that it would be expected to result in
reductions in fish tissue concentrations to allow increased consumption of fish caught from the river after remediation is completed. The agencies are
further considering removal of additional mass of contamination from these impoundments which, combined with capping, might also help to reduce
the transport of contamination into downstream reaches and further control a potential source of downstream release in the event of dam failure. If
downstream transport can be reduced, it should help continue the reduction of PCB levels in fish tissue in Massachusetts and throughout Connecticut,
which, over time, should allow the consumption of additional fish meals and revision or potential elimination of fish consumption advisories in parts of
the river in the future as determined by each State through application of its state-specific fish consumption goals. Preliminary estimates of this potential
cleanup approach indicate that it is expected to also achieve human health benchmarks for children and adults for direct contact with sediment and
reduce contamination to acceptable levels for ecological receptors (fish, wildlife, and other organisms) over the long term in this stretch of the river.

Estimated volumes

The agencies estimate that if excavation and capping of the Reach 7 impoundments and Rising Pond were carried out as described in the potential ap-
proach outlined above, it would entail the excavation of approximately 155,000 yd;" or 15% of the total remediation volume.

Under this type of remedial approach, GE could, as part of the design phase, evaluate whether volume can be reduced in Reach 7 Impoundments by
excavating sufficient sediment to meet a level of 1 ppm PCBs as opposed to excavating and capping the entire area behind each impoundment. Such an
approach may also be considered as an alternative approach in conjunction with a dam removal project. For example, such an approach at Columbia
Mil! Dam would likely entail the excavation of approximately 2 feet of sediment to meet the 1 ppm standard, potentially increasing the volume exca-
vated by approximately 10,000 cubic yards. Institutional controls may need to be part of the remedy for these and other downstream areas, including
those in Connecticut, to ensure that GE remains responsible for any PCB contaminated sediment generated as part of dam maintenance or any other
permitted activities along the River.

IV. FLOODPLAINS, VERNAL POOLS, AND RIVER
BACKWATERS

Area of the River at issue

The floodplain study area in the vicinity of Reaches 5 and 6 (roughly equivalent to
the 10-year flood elevation) consists of approximately 1,000 acres. Within that area
there are 66 vernal pools identified by EPA comprising 18 acres. There are 85 acres
of river backwaters.

A potential approach to cleanup

Excavation and backfill. At this stage of deliberations, the government agencies are
discussing a proposed approach for the excavation of 1 foot of floodplain soil to
generally meet Human Health risk target of 10-5 or Flazard Index =1. in certain
"frequently used subareas", the agencies are discussing the option of excavating to

page 7


-------
a depth of 3 feet rather than 1 foot. The agencies' current thinking, would be to avoid the highest priority habitat areas (Core Area i)
except as needed to meet to meet 10-4 and HI— 1. Such an approach would be expected to avoid, minimize, or mitigate impact to other
priority habitat areas, as long as residual risk is at or below 10-4 and HI-1, and also address other factors such as downstream transport
of PCBs, The option currently under consideration would attempt to remediate vernal pools but avoid those that are within the highest
priority habitat areas (Core Area I), and employ a phased, adaptive management approach to vernal pool remediation. An example of
how this approach might work in the field is provided in the attached Appendix,

As part of the River sediment cleanup, the potential approach under consideration would also (a) call for removal of 1 foot of bed sedi-
ment followed by capping in backwater areas (River Reach 5D), and (b) generally avoid backwaters, or portions thereof, located in highest
priority endangered species habitat areas ("Core Area I"), except in discrete areas with PCB concentrations greater than 50ppm,

Cleanup goals associated with this potential approach

The government agencies believe that the potential approach discussed above could be designed to achieve human health benchmarks
for children and adults for direct contact under a variety of area-specific exposure scenarios. It is further estimated that removal of con-
taminated soil from the floodplain, especially in areas near the river banks, could prevent these materials from becoming a potential future
source of contamination to the River, and might also benefit ecological receptors (fish, wildlife, and other organisms) in the primary study
area. The agencies believe that remediation in the backwater areas is an important component of the overall River sediment remediation,
as it would likely contribute to allowing increased consumption of fish caught from the river after remediation is completed.

Preliminary estimates of the potential remedial approach currently under consideration indicate that over the long term this option would
meet the long term ecological goals in a reasonable time frame, while providing protection for state-listed species.

Estimated volumes

The agencies estimate that if excavation and backfill in the floodplain and vernal pools were carried out as described in the potential reme-
dial approach outlined above, then it would involve the removal of approximately 75,000 yd3 of contaminated materia! in approximately
45 acres of the floodplain (which represents only 5% of the entire floodplain area), and backwater remediation involving an additional
95,000 yd" and 61,5 acres while excluding approximately 8,5 acres in Core Area I (it is estimated that less than 0,5 acres of Core Area I
would be remediated due to concentrations greater than 50 ppm). Together, these two aspects of the potential remedial approach would
account for about 17% of the remedy's total volume.

CONCLUSION

This status report is intended to provide the public with a summary of the status of the governmental discussions prior to a decision on a remedy
proposal. This status report represents a conceptual approach, and is not intended to provide a detailed written analysis of elements considered
in the discussions. Nor does this status report affect or waive any privilege of the governments. Further discussion and analysis will be done prior
to EPA issuing a proposed set of corrective measures, EPA's discussions with the states will be considered in responding to GE's draft corrective
measures study, in any dispute resolution process under the Consent Decree concerning the corrective measures study, and in the issuance of the
proposed set of corrective measures.

page 8


-------
APPENDIX

[Caveat: The approach to remediation set out in this appendix reflects an option that is currently under review by EPA and the State agencies, and
which is believed to exhibit a number of promising attributes. No decision has been made, and EPA will continue to deliberate until it has made a
decision on the CMS and remedy proposal.]

FLOODPLAIN/VERNAL POOL - POTENTIAL REMEDIATION APPROACH UNDER REVIEW WITH
EXAMPLES OF FIELD WORK PROTOCOLS

FLOODPLAINS

Set the Cleanup Levels:	Primary: based on Human Health Benchmarks 10-5/H 1 = 1

Secondary: based on Human Health Benchmarks 10-4/H 1 = 1

Note 1: Any areas exceeding 2 ppm (unrestricted use standard) after remediation may require Institutional Controls to restrict certain future uses.

Note 2: Primary Cleanup levels presumed to be protective of Ecological Receptors in Flood plain.

Approach during Pre-Design/Remedial Design:

1,	GE conducts additional sampling of floodplain soil (as needed).

2,	GE designs excavation plan based on meeting Primary Cleanup Levels,

3,	GE conducts additional field reconnaissance as needed to quantify potential state-listed species impacts of proposed excavation,

4,	Evaluate impacts on state-listed species and their habitats, formulate approach to avoid, minimize, or mitigate any such adverse impacts in accordance
with the substantive requirements of a "Conservation and Management Plan" under MESA, including providing a long-term Net Benefit for the con-
servation of the affected state-listed species. All evaluations to be conducted in consultation with MassDFG/DFW, following the above referenced
MESA regulatory standards. In conducting this evaluation, the following approach would be used:

a,	remediation in Frequently Used Sub-Areas to maintain Primary cleanup goals in these sub-areas

b,	"avoidance" of "Core I" habitat areas, except limited areas to meet Secondary Cleanup Level

c,	"minimization or mitigation" for "Core Area II"

d,	Case-by-case determination for "Core Area III"

5,	Based on #4 above, identify any areas to be avoided, and recalculate exposure point concentration (EPC) to ensure that resultant excavation plan
meets, at a minimum, Secondary Cleanup Levels,

6,	To the extent that Secondary Cleanup Levels are not met, propose additional areas to be excavated in order to meet this threshold (repeating Steps
4 and 5 as needed),

7,	In conjunction with the evaluation of the scope of river bed and bank remediation, evaluate presence of any areas of remaining PCB concentrations
in floodplain soils for erosion potential and the likelihood of future downstream transport at levels that could result in unacceptable downstream
contamination. Reevaluate, as needed, any area of proposed floodplain soil remediation, considering the erosion potential and steps 3 through 6
above, proposing further action as necessary to reduce downstream transport of PCBs.

For example, MassDFG/DFW has identified a portion of Exposure Area (EA) 10 as a Core I area that should be avoided, while portions of Core I areas in EAs
19, 37a, and 62 could be excavated to meet Secondary Cleanup Goals.

VERNAL POOLS

Approach during Pre-Design/Remedial Design:

1.	GE conducts site visit with MassDFG/DFW personnel to confirm presence of vernal poo! (as opposed to backwater or wetland)

2.	GE conducts additional sampling and characterization of vernal pools, as needed, to generate baseline data on the concentrations of PCBs and
health/abundance of animal species, including but not limited to state-listed species, GE also conducts additional field reconnaissance as needed to
quantify the likely effects of potential remediation of the vernal pools on any state-listed species,

3.	GE determines vernal pools requiring cleanup to meet Vernal Pool-specific Cleanup Level,

4.	Consistent with the adaptive management approach described below, for vernal pools identified as requiring cleanup solely to meet ecological
remediation goals, EPA will consult with MassDFG/DFW to make case-specific remedial decisions (including traditional excavation/restoration,
alternative remedial strategies, deferment of remediation, and preservation of existing conditions) weighing field evidence of species health/abun-
dance with applicable ecological IMPGs,

5.	For those vernal pools selected for remediation, evaluate impacts on state-listed species and their vernal poo! habitat, formulate approach to avoid,

page 9


-------
minimize, or mitigate any such adverse impacts in accordance with the substantive requirements of a "Conservation and Management Plan" under
MESA, including providing a long-term Net Benefit for the conservation of the affected state-listed species. All evaluations to be conducted in con-
sultation with MassDFG/DFW, following the above referenced MESA regulatory standards. In conducting this evaluation, the following approach
would be used:

a "avoidance" of Core I Areas",

b,	"minimization or mitigation" for "Core Area II"

c.	Case-by-case determination in "Core Area III"

6,	Develop strategy for vernal pool remediation based upon evaluations conducted above plus considerations of contamination levels and accessibility
and the adaptive management approach outlined below.

7,	In consultation with MassDFG/DFW and MassDEP, EPA identifies performance metrics and evaluation criteria for evaluation of success and potential
adaptive management measures (see below),

8,	Pilot the use of activated carbon in a downstream, easily accessible pool and monitor the performance and any potential negative impacts.

An Adaptive Management Approach for Vernal Pools outside of Core Area I:

1,	Identify first phase of vernal pools to be addressed, in consultation with MassDEP and MassDFG/DFW:

a.	Select initial number (8 to 10) of pools for remediation by traditional means (excavation and reconstruction)

b.	Select additional pools for pilot testing of activated carbon addition in lieu of excavation

c.	Select additional pools for pilot testing by a third remediation method to be proposed by GE for EPA approval in consultation with MassDEP
and MassDFG/DFW and/or additional pools to be monitored concurrently with remediated pools as a "control" group for comparison
purposes.

2,	In consultation with MassDFG/DFW and MassDEP, EPA identifies performance metrics and evaluation criteria for comparison of the various re-
mediation approaches

3,	Complete first round of vernal poo! remediation and conduct evaluation of each method

4,	Determine preferred method/approach to remediation of subsequent vernal pools

Acreage Assumptions for potential approach (based on GE's CHS Alternative FP4):

Initial Acreage (pre MESA evaluation) for Floodplains	57 acres

Area of Overlap with Core Areas I, III, IV	35 acres

Assume net 50% avoidance of Core Areas

from "step 4" processes above	17,5 acres

Net Potential Acreage of Floodplain Remediation	40 acres (approx)

Assuming Vernal Poo! approach outlined above	5 acres (approx)

Total assumed potential Floodplain/Vernal Pool acreage	45 acres (+-10%)

STATE AND LOCAL RESOURCES

MORE INFORMATION

Berkshire Athenaeum Public Library
Reference Department
Pittsfield, MA 01201
(413) 499-9480

Cornwall Public Library
Cornwall, CT 06796
(860) 672-6874

Kent Memorial Library

(Kent Library Association)

Kent, CT 06757
I 927-3761

Massachusetts Department
of Environmental Protection
Springfield, MA 01103
(413)784-1100

Connecticut Department of
Energy & Environmental Protection
Hartford, CT 06106
) 424-3854

EPA contact:

Jim Murphy
(617) 918-1028
murphy.jim@epa.gov

EPA Records Center
Boston, MA 02114
(617) 918-1440
www.epa.gov/region1/ge

Housatonic Valley Association
Cornwall Bridge, CT 06754
I 672-6678

page 10


-------
ATTACHMENT C-5 BANK EROSION/RESTORATION, HOUSATONIC

RIVER, MASSACHUSETTS


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

Bank Erosion/Restoration - Housatonic River, Massachusetts

BANK EROSION/RESTORATION - HOUSATONIC RIVER, MASSACHUSETTS

TABLE OF CONTENTS

1.	INTRODUCTION	1

2.	OVERVIEW OF BANK EROSION PROCESSES	1

3.	BANK EROSION ALONG THE HOUSATONIC RIVER	2

4.	TYPICAL CHANNEL RESTORATION CROSS SECTIONS	3

5.	APPROACHES TO BANK RESTORATION ALONG THE HOUSATONIC RIVER	5

5.1	WOODY DEBRIS TOE PROTECTION	5

5.2	SOIL BIOENGINEERING TECHNIQUES	6

5.3	J-HOOKS/LOG VANES	6

5.4	RIFFLE HABITAT	7

6.	EFFECTIVENESS OF BANK STABILITY TECHNIQUES	8

7.	UNCERTAINTIES IN LONG-TERM EFFECTIVENESS	9

8.	CONCLUSIONS	9

9.	REFERENCES	10

LIST OF FIGURES

Figure 1:	View of Highly Eroded Bank along the Housatonic River	1

Figure 2:	Extreme Erosion along a Section of the Housatonic River	2

Figure 3A:	Bed and Water Surface Slope at Baseflow and Stormflow	4

Figure 3B:	Riffle/Pool Sequence	4

Figure 4:	Riffle Cross Section	4

Figure 5:	Woody Debris Toe Protection Detail	6

Figure 6:	Wood Debris Toe Protection During Installation	6

Figure 7:	J-Hook Log Vane	7

Figure 8:	Example of a Log Vane	7

Figure 9:	Examples of Log/Rock Constructed Riffle	8

i


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

Bank Erosion/Restoration - Housatonic River, Massachusetts

1.	INTRODUCTION

The health of a riverine ecosystem is directly related to the stable and cyclical nature of river
processes, which dictate channel and floodplain form and function (Richards, 1982). Bank
erosion is one such natural process that influences stream ecosystems in both stable and unstable
channels. During flood events, stream banks undergo deformation and erosion as a result of
applied forces. These forces erode sediment from stream banks, and this sediment is then
deposited along downstream reaches of the channel. Although all channels experience erosion,
the erosion rates for stable channels are low. The purpose of this paper is to provide background
information on stream bank erosion processes, discuss stream bank erosion along the Housatonic
River between the confluence of the East and West Branches and Woods Pond, and describe
methods for restoring the stream banks following environmental remediation.

2.	OVERVIEW OF BANK EROSION PROCESSES

River systems are complex and contain many inter-connected parts. Stream banks are just one
component in this system and form the critical boundary between the channel and floodplain.
Bank height and slope determine the ability of the stream to interact with the floodplain, are
important indicators of channel stability, and in healthy systems, provide the foundation on
which native riparian vegetation colonizes, grows, and thrives. The near-channel vegetation that
grows on stream banks and the materials from it drive healthy ecological processes by being the
source of organic matter in the form of leaves and woody debris, by shading the stream and
providing cover for aquatic species, and by increasing the strength of soil through the soil-
binding ability of the roots (FISRWG, 1998).

Banks can both build through deposition and
retreat or deform through erosion. Erosion is
defined as the detachment and removal of
particles or aggregates from the stream bank
surface. Bank erosion occurs when shear
stress, the force applied to the bank by flowing
water, is greater than the ability of the bank to
resist deformation or failure (Leopold, 1992).
Critical shear stress and applied shear stress are
important factors in bank erosion. Critical
shear stress is the minimum amount of force
necessary to initiate erosion. Critical shear
stress is based on the boundary characteristics
of the channel, which include vegetation
density and rooting depth, substrate
composition, soil cohesion, and channel
armoring.

Figure 1: View of Highly Eroded Bank
along the Housatonic River

Critical shear stress is most influenced by the hydraulic radius of the channel (typically equal to
the mean depth) and water surface slope. As mean depth and slope increase, the applied shear
stress created by flow in the river also increases. If the applied shear stress produced by the flow
in the river exceeds a critical shear stress, then erosion will occur. Natural stable rivers exhibit
bank erosion, although in small quantities (less than approximately 0.005 feet per year [ft/yr])
(Rosgen, 2006). In unstable rivers, accelerated bank erosion often occurs, and it is not

1


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

Bank Erosion/Restoration - Housatonic River, Massachusetts

uncommon for banks to migrate several feet in a single storm event (Leopold, 1992). Although
natural erosion in a stable stream system can be a healthy process for a river system, accelerated
bank erosion decreases water quality, can cause channels to over-widen, and can be detrimental
to stream side vegetation.

3. BANK EROSION ALONG THE HOUSATONIC RIVER

Over the past 200 years, the Housatonic
River ecosystem has undergone a long
history of channel disturbances and
channel relocations, and in some cases
has adapted to these channel and
watershed disturbances through changes
to planform and dimension. As a result
of these past disturbances, significant
evidence of bank erosion is present
throughout the Housatonic River. These
disturbed banks are often nearly vertical,
contain sparse vegetation, and contribute
significant amounts of sediments to the
river system. The Housatonic River is
currently recovering from these past
disturbances and over time, the
ecosystem will continue to adapt until the
river reaches a sustainable dynamic
equilibrium.

Although the current stream bank and
Housatonic River, this ecosystem is not sustainable in its current state. Over time, the
Housatonic River will move toward a state of uniform energy dissipation that will result in
reduced bank erosion, a reduction in bar formation, and fewer channel processes that form and
maintain the oxbows.

To better quantify the instabilities on the Housatonic River, a Meander Survey and Soil Bank
Loss study (WESTON, 2006) and a Bank Erosion Hazard Index (BEHI) and Near Bank Stress
(NBS) evaluation (Stantec, 2009) were performed. The BEHI/NBS methodology quantified
sediment loading from bank sources, and identified areas that may require restoration efforts and
management controls during any remediation activities. For a detailed explanation on BEHI and
NBS methodology, refer to Watershed Assessment of River Stability and Sediment Supply
(WARSSS) (Rosgen, 2006).

During the Meander Survey and Soil Bank Loss study, aerial photographs from 1952 to 2000
were used to document the movement of the river and estimate the amount of bank migration.
Additionally, short term changes in the volume of bank loss were measured following a bankfull
flow event. Based on this study, the estimated range of erosion rates in Reach 5A was
determined to be 0 to 0.9 ft/yr with an average value of 0.3 ft/yr. Likewise, the erosion rates for
Reach 5B were estimated to by 0.1 to 0.8 ft/yr with an average rate of 0.5 ft/yr. During the study
period, two meander cut-offs occurred resulting in a net loss of river surface area (Woodlot,
2002). The results of the Meander Survey and Soil Bank Loss study were used for bank erosion

Figure 2: Extreme Erosion along a
Section of the Housatonic River

floodplain processes define the ecosystem of the

2


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

Bank Erosion/Restoration - Housatonic River, Massachusetts

rates in the EFDC Monitored Natural Recovery (MNR) simulation for the Housatonic River.
During the MNR simulation, a value of 1,328 MT/yr (1,464 tons/yr) of eroding solids from
riverbanks was used, which resulted in the delivery of 14 kilograms (kg) (30.8 pounds [lbs]) of
polychlorinated biphenyls (PCBs) to the water column and an additional 11 kg (24.3 lbs) of
PCBs to the riverbed on an average annual basis. Based on this, PCBs from eroding riverbanks
represent 45 percent of the overall mass of PCBs entering the river (EPA, 2011).

As part of the BEHI/NBS analysis, the banks were divided and inventoried according to changes
of physical bank characteristics (e.g., bank angle, rooting depth, bank stratification) and the
applied shear stresses. BEHI/NBS assessments obtained along a reach were converted to
estimated sediment load in tons/yr. The bank migration rates were predicted based on published
bank erosion rates as related to the BEHI/NBS ratings from North Carolina and Colorado
(Rosgen, 2006).

The total bank erosion predicted from the 41,000 linear feet (ft) of the Housatonic River
evaluated (in Reaches 5A and 5B) was estimated to be on the order of 7.300 tons/yr. This
equates to an average bank erosion rate of 0.16 tons/ft/yr or 0.32 ft/yr in these reaches (Stantec,
2009). A reference geomorphic bank erosion rate for most stable alluvial reference reaches is
less than approximately 0.005 ft/yr (Rosgen, 2006). Based on this reference rate, these reaches
are considered to be in a state of accelerated bank erosion. One important finding of this study is
that the areas of high bank erosion are generally out of phase with the planform of the river,
which is an indicator of channel instability. In alluvial systems, areas of highest erosion are
related to lateral scour pools on the outside and lower third of the meander bend (Leopold, 1992).
On the reaches studied on the Housatonic River, many of the extreme and very high bank erosion
rates are located upstream of point bars on the inside banks, which is indicative of channel
migration and horizontal instability (Stantec, 2009).

The Housatonic River is currently recovering from historical impacts and modifications.
Although the River will eventually reach a stable state through natural changes over time, such
change will necessarily include accelerated erosion of the floodplain and stream banks, which are
contaminated with PCBs.

4. TYPICAL CHANNEL RESTORATION CROSS SECTIONS

The goals of channel restoration for the Housatonic River include maintaining the natural
geomorphic function of the river, as well as the natural beauty and biological function of the
Housatonic ecosystem. It is possible to design the remediation/restoration in a manner that
meets the restoration goals while improving the geomorphic function of the river. As noted
above, significant portions of the Housatonic River are out of phase with the channel planform,
indicating channel instability. In a natural river, riffles are located within the straighter crossover
section between two bends, and pools are located on the outside of bends in the river (Harman
and Jennings, 1999).

3


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1

Figure 3: (A) Bed and Water Surface Slope at Baseflow and Stormflow;
^	(B) Riffle/Pool Sequence

3	Remediation and subsequent restoration should consider the channel's geomorphic function.

4	Additionally, modifying planform instabilities, including very tight radii of curvature (typically

5	less than two times the bankfull width of the channel), should be considered and evaluated in the

6	restoration plan. Figure 4 below depicts a typical riffle cross section that can be constructed over

7	a capped area following removal of contaminants. In the illustrated example, a deformable soil

8	layer composed of clean fill is placed over the isolation cap along the banks. An appropriate

9	channel substrate is placed on top of the cap over the channel bed.

sediment

10

Figure 4: Riffle Cross Section

4


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

Bank Erosion/Restoration - Housatonic River, Massachusetts

5. APPROACHES TO BANK RESTORATION ALONG THE
HOUSATONIC RIVER

Bank restoration can be achieved through the use of natural materials such as woody debris, soil
bioengineering, and log and rock structures, as well as by adjusting the slope of stream banks and
revegetating the riparian zone (USACE, 2003). Stream bank stabilization should take into
consideration the unique conditions that will be present after contaminant removal, as well as
reference conditions from a stable stream channel (i.e., reference reach), and often involves
restoring stream dimension and profile to improve channel stability. This can be accomplished
by (1) constructing a channel of proper dimension, (2) adding grade control structures, and (3)
regrading the floodplain (Rosgen, 1997). To meet the restoration objectives of this project, it is
important that any bank restoration methods employ, where appropriate, the use of living
systems to enhance the ecosystem and provide for natural ecologic functions.

Regrading a floodplain involves lowering bank heights by excavating a bankfull bench adjacent
to the channel. A bankfull bench is a graded terrace at the bankfull elevation. The bankfull
bench allows flood flows to access the adjacent floodplain, thereby reducing in-channel shear
stresses. In general, the Housatonic River is an incising river system, meaning that the river has
moderate access to its floodplain. One method to reduce future bank erosion is to excavate a
bankfull bench along the Housatonic River and reduce bank heights by approximately 2 to 3 ft,
thus improving floodplain access. The use of riparian plantings would enhance stream bank
stability while providing important habitat.

Bank stabilization should be examined from the engineering, geomorphic, and biological
perspectives. Engineering considerations include the ability of the stream banks to resist erosion,
hydraulic conveyance of the channel, scour, and deflection of erosive forces to other locations
along the reach. Geomorphic considerations include location of the proposed structures,
channel-floodplain interaction, sediment competence and capacity, bankfull cross-section, width-
to-depth ratio, sediment supply, location of depositional areas, bar formations, and locations of
scour. Biological considerations include selection and survivability of planted riparian species,
growing seasons, and fish and macro-invertebrate habitat.

Examples of some of the techniques used to provide bank stability are illustrated below.
5.1 WOODY DEBRIS TOE PROTECTION

Woody debris toe protection is an innovative structure that incorporates readily available on-site
materials that would otherwise be sent off-site for disposal. Woody debris toe protection can be
used for both temporary and long-term bank stabilization on the outside of stream meanders.
The woody debris structure is planted with live stakes, bare roots, and transplants, as well as sod
if available. Large woody debris is placed at an elevation such that the wood remains
submerged, providing important fish habitat and significantly reducing the decay time of the
wood.

5


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

5 5.2 SOIL BIOENGINEERING TECHNIQUES

6	Live cuttings and other soil bioengineering techniques can readily be used to restore and stabilize

7	stream banks (USDA, 1995). Live cuttings consist of cut branches from appropriate tree and

8	shrub species. These cuttings are typically obtained while the plants are dormant. Typical soil

9	bioengineering techniques include live staking, live branch layering, and brush mattresses.

10 5.3 J-HOOKS/LOG VANES

11	J-hooks and log vanes are used for energy dissipation, flow redirection, and creation of

12	downstream scour. These structures help create a large range of velocity and depth combinations

13	throughout the project site, thus increasing biodiversity (Rosgen, 2006). J-hook vanes are

14	composed primarily of large boulders, whereas log vanes are composed of logs typically

Figure 6: Woody Debris Toe Protection During Installation

6


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1	removed from the site to be restored. A schematic of a j-hook/'log vane, as well a photograph of

2	a typical installation, are shown below.

3

4

7	5.4 RIFFLE HABITAT

8	Riffles serve a very important role for both the geomorphic and ecologic functions within a river

9	system. A riffle is the hydraulic control for a river, helping to maintain sediment transport

10	functions. If a riffle cross-section is under-sized for the sediment being delivered to the system,

11	the stream can experience down-cutting. Likewise, if a riffle cross-section is over-sized, the

12	stream can be subject to aggradation. From an ecological function perspective, riffles provide

13	bed diversity and important habitat for macro-invertebrates.

Figure 7: J-Hook Log Vane (courtesy of Wiidland Hydrology)

Cut off Sill

FLOW

POINT BAR

Buried 10-15 feet

GLIDE

7


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1	Typically, riffles can be constructed of rock, wood or a combination of each. Examples of a

2	log/rock constructed riffle (pictures taken immediately after construction and several years after

3	construction) are included in Figure 9 below.

4

5

6

7

8

9

10

11

12	Figure 9: Examples of Log/Rock Constructed Riffle

13	6. EFFECTIVENESS OF BANK STABILITY TECHNIQUES

14	There are many examples of sites where these bank stabilization techniques have been

15	implemented successfully (EPA, 2011), and numerous publications on the use of bioengineering

16	techniques for bank erosion control and habitat enhancement (e.g., USACE, 1997; Sotir and

17	Fischenich, 2001; Sylte and Fischenich, 2000; Allen and Fischenich, 2000; Allen and Fischenich,

18	1999; Li and Eddleman, 2002; and VDCR, 2004).

19	On the Connecticut River in Massachusetts, the Franklin Regional Council of Governments

20	implemented the successful stabilization of more than 10,000 linear feet of river bank using

21	several techniques, including fascines, live planting and seeding, hard toe structures, and coir

22	rolls (FRCOG, 1999, 2003, 2007). On Town Branch Creek in Russellville, Kentucky, the

23	Kentucky Department of Environmental Protection oversaw the removal and restoration of 3.5

24	miles of stream bank soils in three phases between 1997 and 2001 (Land and Water, 2009). For

25	Phases II and III, several techniques, such as j-hook rock vanes, tree crowns, and submerged

26	wooden shelters, were successfully used to stabilize banks and promote habitat restoration.

27	A combination of stabilization techniques was used successfully at the Army Research

28	Laboratoiy Site in Watertown, MA. These stabilization techniques included coir fascines for toe

29	stabilization and brush layers and live stakes for the upper slope treatment (Bioengineering,

30	2012a). On the Manhan River in Easthampton, MA, 600 linear feet of banks were stabilized for

31	the emergency protection of a natural gas pipeline. Both vegetation and structural materials were

32	used to stabilize the bank and re-direct flows toward the channel center (Bioengineering, 2012b).

33	In 1998, General Electric conducted a remedial action to restore portions of the upper riverbank

34	along the West Branch of the Housatonic River in Pittsfield, Massachusetts. The restoration

35	included placement of topsoil, a layer of biodegradable erosion control blanket, coconut fascines

36	and various seed mixtures, tree, shrubs, and herbaceous species. General Electric completed a

37	second remedial action in 2008/2009 that stabilized and restored sections of the lower riverbank

8


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

Bank Erosion/Restoration - Housatonic River, Massachusetts

and channel in the West Branch using aquatic structures, such as current deflectors, boulders,
boulder clusters, large woody debris, and root wads. In addition, coir logs and plant plugs were
used on the toe of the slope as bank stabilization features. Post-construction monitoring reports
indicate that the restoration and stabilization techniques are performing successfully with
minimal maintenance requirements (GE, 2010 and 2011).

7.	UNCERTAINTIES IN LONG-TERM EFFECTIVENESS

Bank stabilization techniques are generally categorized into traditional methods, such as hard
armoring and bioengineering (sometimes also referred to as biotechnical engineering) techniques
(Li and Eddleman, 2002). Each technique has advantages and disadvantages in terms of
applicability, cost, and effectiveness, each of which must be considered on a project-by-project
basis. In addition, each technique will have limitations based on numerous site factors. For
these reasons, and to reduce the potential for failure, it is necessary to implement an inter-
disciplinary (engineering, geomorphic, and biological) approach to design and construction of a
long-term effective bank stabilization solution. The inter-disciplinary approach can be effective
at reducing uncertainties by designing the appropriate stabilization techniques for the project in
consideration of both current and anticipated future conditions, e.g., a 100-year flow event.
Moreover, establishing an effective post-construction monitoring and maintenance program can
further prevent stabilization failures and potentially more severe impacts resulting from such
failures (USACE, 1997).

Changes in watershed use or responses may impact the long-term effectiveness of any bank
stabilization technique. Commonly observed responses include extensive hillslope erosion that
leads to floodplain and channel aggradation during deforestation, followed by channel incision
and bank erosion upon reforestation and/or the implementation of upland erosion control
measures. The downstream movement of sediment created by aggradational and degradational
processes occurring over long periods of time can lead to significant local post-construction
channel instabilities (Miller and Kochel, 2009).

Reducing uncertainty in the long-term effectiveness of bank stabilization can be achieved with
proper planning in selection of the stabilization technique and materials, incorporating site
considerations (e.g., hydrological regime and regional watershed uses) with design
considerations and appropriate construction techniques. Uncertainties associated with the
various materials, design, and construction methods used can result in a range of positive and
adverse environmental impacts. Through proper planning and design, negative impacts can be
minimized and positive impacts maximized. A robust operation and maintenance program
implemented early in a project will further reduce uncertainties in long-term effectiveness (Sylte
and Fischenich, 2000; Fischenich, 2001).

8.	CONCLUSIONS

The Housatonic River has been highly impacted over the past two centuries and currently
exhibits accelerated bank erosion and other signs of instability, including a profile that is out of
phase with the channel planform. Based on data collected from the River, the stream is eroding
at a rate on the order of 0.3 to 0.5 ft/yr, which is significantly higher than stable reference
streams. This erosion is contributing 45% of the PCB load. Accelerated bank erosion decreases
water quality, can cause channels to over-widen, and can be detrimental to aquatic habitat and
stream-side vegetation.

9


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1	Restoration of rivers and stream banks is a common practice used throughout the United States

2	and has evolved significantly over the past 50 years. In the past, many bank stabilization

3	techniques focused on the use of hard armoring with concrete, gabion baskets, or riprap to

4	achieve bank stabilization. Effective long-term bank stabilization can be readily achieved

5	through the use of vegetation and other natural materials as evidenced from the bank restoration

6	techniques presented in this paper. Advantages of these techniques over more traditional hard

7	armoring approaches include increased water quality, temperature reduction, increased biological

8	function, and aesthetics.

9	9. REFERENCES

10	Allen, H.H. and J.C. Fischenich 1999. Coir geotextile roll and wetland plants for streambank

11	erosion control. EMRRP Technical Notes Collection (ERDC TN-EMRRP-SR-04), U.S. Army

12	Engineer Research and Development Center, Vicksburg, MS.

13	Allen, H.H. and J.C. Fischenich 2000. Brush mattresses for streambank erosion control. EMRRP

14	Technical Notes Collection (ERDC TN-EMRRP-SR-23), U.S. Army Engineer Research and

15	Development Center, Vicksburg, MS.

16	Bioengineering Group. Emergency Riverbank Stabilization for Natural Gas Pipeline Protection,

17	Manhan River, Easthampton, MA. Project Summary downloaded July 2012.

18	http://www.bioengineering.com/proi ects/view.php?id=29

19	Bioengineering Group. Bank Stabilization and Wetland Terrace Creation at the Charles River

20	Park, Army Research Laboratory, Watertown, MA. Project Summary downloaded July 2012.

21	http://www.bioengineering.com/projects/view.php?id=20

22	Doody, J.P. A.N. Esposito and D. Weeks Jr. 2009. A Comeback Story: Restoring Town Branch

23	Creek in Russellville, Kentucky. Land and Water, Volume 53, Issue 5, pp 14-18.

24	September/October 2009.

25	Dunne, T. and L.B. Leopold. 1978. Water in Environmental Planning. W.H. Freeman and

26	Company. New York, NY.

27	EPA (U.S. Environmental Protection Agency), 2011. National Remedy Review Board Site

28	Information Package for the Housatonic River, Rest of River, SDMS 487318, June 2011.

29	Privileged, Attorney-Client Work Product, Deliberative Document, Do Not Release.

30	Fischenich, J.C. 2001. Impacts of stabilization measures. EMRRP Technical Notes Collection

31	(ERDC TN-EMRRP-SR-32), U.S. Army Engineer Research and Development Center,

32	Vicksburg, MS.

33	Franklin Regional Council of Governments (FRCOG). 1999. Connecticut River Watershed

34	Restoration 319 Project, 1996-1998.

35	Franklin Regional Council of Governments (FRCOG). Connecticut River Watershed Restoration

36	- Phase II Project 00-04/319. 2000-2003.

37	Franklin Regional Council of Governments (FRCOG). Project Final Report, Connecticut River

38	Watershed Restoration - Phase III Project No. 03-07-319. 2004-2007.

10


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1	FISRWG. 1998. Stream Corridor Restoration: Principles, Processes, and Practices. By the

2	Federal Interagency Stream Restoration Working Group (FISRWG) ISBN-0-934213-59-3

3	http://www.nrcs.usda.gov/wps/portal/nrcs/detail.

4	General Electric Company. 2010. Letter from Richard Gates (GE) to John Ziegler (MDEP),

5	Regarding West Branch of the Housatonic River (GESD02), Summary of October 2010 Post-

6	Remediation Inspection Activities. November 10, 2010.

7	General Electric Company. 2011. Final Completion Report for Parcel G9-8-1 and the Adjacent

8	Portion of the West Branch of the Housatonic River. MDEP Site Nos. GEACO250 and

9	GESD02. RTN: 1-11953. April 22, 2011.

10	Harman, W.A. and G.D. Jennings. 1999. River Course Fact Sheet Number 1: Natural Stream

11	Processes. North Carolina Cooperative Extension Service.

12	Li, M. and Eddleman, K.E. 2002. Biotechnical Engineering as an Alternative to Traditional

13	Engineering Methods, A Biotechnical Streambank Stabilization Design Approach. Landscape

14	and Urban Planning 60 (2002) pp 225-242.

15	Miller, J.R. and Kochel. R.C. 2010. Assessment of Channel Dynamics, In-Stream Structures and

16	Post-Project Channel Adjustment in North Carolina and its Implications to Effective Stream

17	Restoration. Environmental Earth Sciences. February 2010, Volume 59, Issue 8. Pp 1681-1692.

18	Richards, K.S. 1982. Rivers: Forms and Processes in Alluvial Channels.

19	Riley, A.L. 1998. Restoring streams in cities: a guide for planners, policymakers, and citizens.

20	Island Press, Washington, D.C.

21	Rosgen, D.L. 1997. "A Geomorphological Approach to Restoration of Incised Rivers."

22	Proceedings of the Conference on Management of Landscapes Disturbed by Channel Incision.

23	S.S.Y. Wang, E.J. Langendoen, and F.F. Shields, Jr. (editors).

24	Rosgen, D.L. 2006. A Watershed Assessment for River Stability and Sediment Supply (WARSSS).

25	Wildland Hydrology Books, Fort Collins, CO. http://www.epa.gov/warsss/.

26	Rosgen, D.L. Updated 2006. "The Cross-Vane, W-Weir, and J-Hook Structures." Wildland

27	Hydrology, Ft. Collins, CO.

28	Stantec Consulting Services Inc. 2009. Housatonic River BEHI Report.

29	Sotir, R.B. and Fischenich, J.C. (2001). Line and Inert Fascine Streambank Erosion Control.

30	EMRRP Technical Notes Collection (ERDC TN-EMRRP-SR-31), U.S. Army Engineer Research

31	and Development Center, Vicksburg, MS.

32	Sylte, T.L. and Fischenich, J.C. (2000). Rootwad composites for streambank stabilization and

33	habitat enhancement. EMRRP Technical Notes Collection (ERDC TN-EMRRP-SR-21), U.S.

34	Army Engineer Research and Development Center, Vicksburg, MS.

35	USACE (U.S. Army Corps of Engineers). 1997. Bioengineering for Streambank Erosion Control

36	Report 1 Guidelines. Technical Report EL-97-8. U.S. Army Corps of Engineers, Waterways

37	Experiment Station.

38	USACE (U.S. Army Corps of Engineers). 2003. Stream Mitigation Guidelines. U.S. Army Corps

39	of Engineers, Wilmington District; North Carolina Division of Water Quality; U.S.

40	Environmental Protection Agency, Region 4; Natural Resource Conservation Service; and North

41	Carolina Wildlife Resources Commission, Raleigh, NC.

11


-------
Bank Erosion/Restoration - Housatonic River, Massachusetts

1	United States Department of Agriculture, 1995. Engineering Field Handbook Chapter 18: Soil

2	Bioengineering for Upland Slope Protection and Erosion Reduction; Natural Resource

3	Conservation Service.

4	Virginia Department of Conservation and Recreation (VDCR). 2004. The Virginia Stream

5	Restoration and Stabilization Best Management Practices Guide.

6	WESTON (Weston Solutions, Inc) 2006. Final Model Document Report: Modeling Study of

7	PCB Contamination in the Housatonic River, Appendix A.4: Erosion and Accretion Mapping

8	Meander Study. Prepared for the U.S. Army Corps of Engineers and the U.S. Environmental

9	Protection Agency. DCN GE-111006-ADII.

10

12


-------