Chapter 6 of Housatonic River Pcb Sediment Management Study Program for Monitoring the Natural Recovery of the River

GENERAL ELECTRIC COMPANY Fairfield, Connecticut

CHAPTER 6 OF HOUSATONIC RIVER PCB SEDIMENT MANAGEMENT STUDY PROGRAM FOR MONITORING THE NATURAL RECOVERY OF THE RIVER

April 1988

CO D (ft

LMSE-88/0171&337/017 o o o LAWLER, MATUSKY & SKELLY ENGINEERS NJ Environmental Science & Engineering Consultants One Blue Hill Plaza Pearl River, New York 10965

PGE00047117 I I I TABLE OF CONTENTS Page No. I LIST OF FIGURES iv LIST OF TABLES vii I SUMMARY S-l 5.1 Introduction S-l 5.2 Perspective on PCB Concentrations in the S-l I Housatonic River 5.3 Temporal and Spatial Trends in the S-2 Housatonic River I 5.4 PCB Fate and Transport Model S-3 5.5 Projection of PCB Concentrations S-3 5.6 Socioeconomic Impact on the Recreational 5-4 Resource I 5.7 Future Development and the No Action Plan S-4 5.8 Monitoring Program S-5 I 6 PROGRAM FOR MONITORING THE NATURAL RECOVERY 6-1 OF THE RIVER I 6.1 Overview of No Intervening Action Plan 6-1 6.1.1 Background 6-1 I 6.1.2 Chapter Organization 6-4 6.2 PCBs in Water, Sediment, and Fish: Housatonic 6-6 I River in Perspective 6.2.1 Water 6-8 6.2.2 Sediment 6-10 I 6.2.3 Fish 6-15 6.2.3.1 PCB Trends in Housatonic River 6-22 I Fishes 6.3 PC BFate and Transport Model 6-36 6.3.1 Description of WASTOX Model 6-36 I 6.3.2 Parameter Evaluation 5-40 6.3.2.1 Segmentation and Hydrology 6-40 I 6.3.2.2 Bed Sediment Characteristics 6-41 6.3.2.3 Settling, Resuspension, and Burial 6-43 6.3.2.4 Sediment-Water Partitioning 6-48 I 6.3.2.5 Bed Sediment-Water Column 6-50 I Diffusion

I Lawlor Mntusky PGE00047118 I I I TABLE OF CONTENTS (Continued) I Page No. 6.3.2.6 Volatilization 6-50 I 6.3.2.7 Biodegradation 6-50 6.3.2.8 Upstream and Tributary Suspended 6-51 Solids and PCBs 6.3.3 Model Calibration 6-53 6.3.3.1 Solids 6-54 I 6.3.3.2 PCBS 6-56 6.3.4 Model Projections of PCS 6-59 I 6.3.4.1 Long-Term Hydrological Period 6-59 6.3.4.2 Upstream and Tributary Inflows 6-60 I of PCBs ™ 6.3.4.3 Projections of PCBs in 6-61 Sediment and Water • 6.3.4.4 Extrapolation of PCS Concentra- 6-63 • tions in Fish 6.4 Environmental Impacts of No Action Plan 6-68 | 6.4.1 Socioeconomic Impacts 6-68 _ 6.4.1.1 Recreational Impacts 6-68 ™ 6.4.1.2 Angler Surveys 6-78 6.4.1.3 Resource Evaluation 6-90 • 6.4.2 Future Development 6-96 6.4.2.1 Regulatory Process 6-97 • 6.4.2.2 Evaluation of Potential PCB 6-100 Losses From Dredging and Filling Projects • 6.4.2.3 Future Development and the No 6-103 • Action Plan 6.5 Proposed Monitoring Program 6-106 | 6.5.1 Objectives and General Principles 6-106 _ 6.5.2 Sediments 6-108 • 6.5.3 Water 6-109 • 6.5.4 Fish 6-112 6.5.5 Data Review and Assessment 6-114 •

I Luu-lor Matuskx SkrIU l'i PGE00047119 I I

I TABLE OF CONTENTS (Continued)

Page No. I REFERENCES CITED R-l APPENDICES I A - Scatter Plots of Fish PCB Concentration (ug/g) Versus Total Length (mm) B - PCB Analyses of One Inch Increments of Sediment From Core Samples of Housatonic River in 1986 C - Analytical Framework for the Evaluation of I PCB Release and its Application to a Hypothetical Dredging Project with Subaquatic Disposal in Lake I Lillinonah I I I I I I I I I iii

I Lawlor, Malusky '^Skellv Engineers PGE00047120 I I I LIST OF FIGURES I Figure No. Title Page No. 6.1-1 Housatonic River and Watershed Area 6-1A 6.2-1 PCB Concentrations (ug/kg) in Connecticut 6-14B I Bottom Sediments (USGS Data Base) • 6.2-2 Housatonic River, American Eel 6-29B 6.2-3 Housatonic River, Black Crappie 6-29C | 6.2-4 Housatonic River, Brown Bullhead 6-29D _ 6.2-5 Housatonic River, Brown Trout 6-29E • 6.2-6 Housatonic River, Carp 6-29F • 6.2-7 Housatonic River, Chain Pickerel 6-29G 6.2-8 Housatonic River, Largemouth Bass 6-29H I 6.2-9 Housatonic River, Lepomis spp. 6-291 6.2-10 Housatonic River, Rock Bass 6-29J I 6.2-11 Housatonic River, Smallmouth Bass 6-29K • 6.2-12 Housatonic River, White Catfish 6-29L 6.2-13 Housatonic River, White Crappie 6-29M • 6.2-14 Housatonic River, White Perch 6-29N 6.2-15 Housatonic River, White Sucker 6-290 I 6.2-1 6 Housatonic River, Yellow Perch 6-29P 6.2-17 Total PCB in Insect Tissue, Housatonic 6-34A I River, Cornwall, CT 6.3-1 Fluxes of PCB Associated With the Bed 6-40A I 6.3-2 Schematic of Model Segments and River 6-40C Flows of Housatonic River

6.3-3 Sediment PCB vs Mile Point 6-42A

6.3-4 Sediment PCB and TOC vs Depth, Falls 6-43A I Village and Bulls Bridge

I Lawlcr, Malusky •"'''Skolly Engineers PGE00047121 I I

LIST OF FIGURES • (Continued)

Figure No. Title Page No. 6.3-5 Sediment PCB and TOC vs Depth, Lake 6-43B I Lillinonah-Rt. 133 Bridge, Lake Lillinonah-Shepaug Dam 6.3-6 Sediment PCB and TOC vs Depth, Lake Zoar- 6-43C • Rt. 84 Bridge, Lake Zoar-Stevenson Dam 6.3-7 Procedure for Evaluating Sediment Burial 6-45B | Rate and Depth of Active Sediment Based on Cesium . 6.3-8 Sediment Cesium vs Depth Without Mixing 6-46A ™ 6.3-9 Flow Dependent Resuspension Relationship 6-48A • 6.3-10 Suspended Solids Calibration, Segments 1-4 6-55A 6.3-11 Suspended Solids Calibration, Segments 5-7 6-55B | 6.3-12 PCB Calibration Results in Water Column 6-57A and Sediment, Segments 1, 2 (Total), 1 & 8, • 2 & 9 (Particulate) • 6.3-13 PCB Calibration Results in Water Column 6-57B • and Sediment, Segments 3, 4 (Total), 3 & 10, | 4&11 (Particulate) 6.3-14 PCB Calibration Results in Water Column 6-57C I and Sediment, Segments 5, 6 (Total, 5 & 12, m 6 & 13 (Particulate) 6.3-15 PCB Calibration Results in Water Column 6-57D • and Sediment, Segment 7 (Total), 7 i 14 (Particulate) • 6.3-16 Housatonic River at Falls Village, CT 6-59A 6.3-17 Decaying PCB Boundary and Tributary Con- 6-60A I centrations • I I

Mnlusky Skolly K I PGE00047122 I I

• LIST OF FIGURES (Continued) I Figure No. Title Page No. I 6.3-18 PCB Projection Under No Action Plan 6-61A • Segments 2/9, 3/10 6.3-19 PCB Projection Under No Action Plan 6-613 1 Segments 4/10, 5/12 6.3-20 PCB Projection Under No Action Plan 6-61C 1 Segments 6/13, 7/14 6.3-21 PCB Projections Under No Action Plan 6-64A • 6.4-1 Recreational Areas on Housatonic River 6-77A and Its Adjacent Water Bodies | 6.4-2 Total PCB Release From Lake Llllinoah With 6-103A and Without Hypothetical Marina Dredging _ and Subaquatic Disposal

I I I I I I I I vi

I Lawler, Mntusky : ^Skelly E PGE00047123 I I I LIST OF TABLES Table No. Title Page No. 6.2-1 PCB Concentration Data for Whole Water 6-8A I Samples From Housatonic River 6.2-2 PCB Concentrations in Whole Water Samples 6-8B1 I From a Variety of Land Use Areas 6.2-3 PCB Concentrations in Whole Water Samples 6-8C I of the North Atlantic Slope 6.2-4 PCB Concentrations in Sediments of Remote 6-11A I Habitats Where PCBs Have Been Located 6.2-5 PCB Concentrations in Sediments of Lakes 6-12A I and Streams From a Variety of Land Use Areas 6.2-6 State PCB Concentration Data for Bottom 6-13A1 Sediments from a Variety of Land Use Areas I 6.2-7 PCB Concentrations in Bottom Sediments of 6-13B Connecticut Rivers and Lakes I 6.2-8 Sediment PCB Concentrations in the State 6-14A1 of Connecticut I 6.2-9 PCB Concentrations in Sediment of Four 6-15A Housatonic River Impoundments 6.2-10 Geometric Mean, Maximum Wet-Weight, and 6-16A I Lipid-Weight Concentrations, and Percentage of Stations Showing Detectable Concentrations I in at Least One Sample 6.2-11 Mean PCB Concentrations in Lake Trout 6-19A From Various Regions of Lake Superior and I Siskiwlt Lake, Isle Royale 6.2-12 Range in Mean PCB Concentrations in Species 6-19B of Fish From Big Cypress Swamp Collected I During Winter and Spring 1972 6.2-13 Range and Arithmetic Mean of PCB Concentrations 6-20A I in Whole Fish Composite Samples of Two Species of Fish From Upper and Lower Reaches of the I Apalachicola River, Florida, 1978 I

I jjr, Matusky 'Skelly PGE00047124 I I

LIST OF TABLES I (Continued)

Table No. Title Pace No.

6,2-14 1975 PCS Concentrations in Edible Flesh 6-20B of Multiple Composite Samples for Species of Fish From Nonindustrialized orLightly • Industrialized Watersheds ofNew York State • 6.2-15 Within- and Among-Group Correlation forIn 6-26A • Total PCB Concentration, In Total Length, In | Total Weight, Age, and Percentage Lipid for Selected Fish • 6.2-16 Results From Analysis of Covariance for 6-29A ™ Housatonic River Fishes, 1979-1986 6.2-17 Estimated Rate of Change 1n PCBConcentra- 6-30A • tion Over Years and River Kilometer for Housatonic River, 1979-1986 • 6.2-18 Average Length, Percentage Lipid, and 6-33A Weight of Seven Species of Housatonic _ River Fishes Examined for PCBs During • 1979-1986 • 6.3-1 Drainage Areas andLong-Term Average Flows 6-408 • 6.3-2 Physical Characteristics of Model Segments 6-41A 6.3.3 Bed Sediment Characteristics of Model Segment 6-41B • 6.3-4 Cesium 137 Activity (pC1/gm dry) for Select 6-45A Increments of Housatonic River Cores I 6.3-5 Solids Settling, Resuspension, and Burial 6-47A Rates of Model Segments • 6.3-6 Model Input Flow Data for 18 Month Calibra- 6-54A tion Period 6.3-7 PCB Concentrations Measured in Water Column 6-56A * of Housatonic River in Connecticut 6.3-8 Partitioning-Based Bloconcentration Esti- 6-63A | mates of Direct PCB Uptake by Fish 6.3-9 Percentage of Fish Above FDA Action Level of 6-65A I 27 ug/nn/gn ™ vi ii • Lawlor Matuskv "8k«lly rs • PGE00047125 I I

I LIST OF TABLES (Continued)

I Table No. Title Page No.

I 6.4-1 Existing Recreational Facilities Adjacent to 6-70A Lake Zoar I 6.4-2 Existing Recreational Facilities Adjacent to 6-70B Lake Lillinonah I 6.4-3 CDEP Trout Stocking on the Housatonic River 6-74A 6.4-4 Percentage of Total Catch by Species, River 6-86A I Section, and Season 6.4-5 1986 Fishing Pressure 6-89A I 6.4-6 Approved and Ongoing Development in Areas 6-97A1 of PCB-Contaminated Sediments • 6.4-7 Information Required for COE and CDEP Permit 6-98A ™ Application 6.4-8 Permits Applicable to Sediment Management 6-99A 1 Plan 6.4-9 Hypothetical Distribution of PCBs in Dredged 6-102A 1 Sediments Before and During Dredging 6.4-10 Long-Term PCB Release With and Without 6-102B • Hypothetical Dredging Project in Lake • Lillinonah 6.5-1 Housatonic River PCB Monitoring Program, 6-114A I Summary of Recommended Sampling I I I

I IX

I Lawlor Malusky r *!>kelly K PGE00047126 I I

I SUMMARY I 5.1 INTRODUCTION The 1985 interim report by Lawler, Matusky & Skelly Engineers (IMS) I on the Housatonic PCB Sediment Management Study presented a screen ing of remedial alternatives, such as capping and dredging of Lakes I Lillinonah and Zoar and dredging the Falls Village and Bulls Bridge impoundments. In light of the decrease in PCB concentrations in I the fish between 1979 and 1984 as well as comments on the interim report by various interested organizations, the Connecticut Depart I ment of Environmental Protection (CDEP) recommended no further stu dies of active remediation in favor of a study to develop a program for monitoring the natural recovery of the river, i.e., the no I intervening action plan. Chapter 6, the final chapter of the Sedi ment Management Study, addresses the scope of work defined by I General Electric (GE) and CDEP. I The objectives of this study of the no action plan are to (1) evaluate the Housatonic River's natural recovery from present PCB I levels; (2) assess the socioeconomic impact associated with the PCB situation, e.g., the fish eating advisory; (3) incorporate the administrative and technical considerations of future development I involving sediment and PCB movement, e.g., dredging for a marina; I and (4) recommend a long-term monitoring program. I 5.2 PERSPECTIVE ON PCB CONCENTRATIONS IN THE HOUSATONIC RIVER In order to evaluate the degree to which PCB concentrations in the I Housatonic River are elevated at present and to gain a better understanding of the river's future recovery, an extensive litera- I ture search to quantify background PCB levels was undertaken. The I S-l I Lawler, Matusky Sf Skelly Engineers PGE00047127 I I emphasis was placed on data collected from waterbody locations hav- ™ ing no known point source discharges of PCBs in the area. A com parison of background PCB concentrations in water, sediment, and I fish to levels in the Housatonic revealed the following:

• Water sampling in the late 1970s showed that the incidence of detectable PCB concentrations (>.0.1 ug/1) in the Housatonic River was approximately 10 • times that of other water bodies across the coun- B try. • PC Bconcentrations of bottom sediments in the Housatonic are approximately five to 10 times greater than background levels in other parts of _ Connecticut. • • Although differences in fish species, age, size, and lipid content complicate a direct comparison, • approximately 95% of the fish sampled nationally • have PCB concentrations above the detection limit (from 0.1 to 0.5 ug/g) and background levels m approach the FDA action level of 2 ug/g. PCB con- • centrations in Housatonic River fish vary with species, size, lipid content, location, and year of sampling; the median concentration of the 1979 • through 1986 data is 1.36 ug/g. Approximately 647, • of the fish sampled are below the FDA action level. •

Overall, background levels of PCBs appear to be decreasing follow ing the nationwide ban on PCB prodjction in 1977. • S.3 TEMPORAL AND SPATIAL TRENDS IN THE HOUSATONIC RIVER I According to Housatonic River data collected during the last eight years, PCB concentrations in river water, insect tissue, and fish I appear to be decreasing. PCB concentrations in fish, the most sam pled environmental compartment, generally show a significant de- I crease, both temporally (from 1979 through 1986) and spatially (from Cornwall to Lake Zoar). These trends are a manifestation of • I

Lawler, Matusky Z? Skelly Engineers PGE00047128 I I I the dispersion of PC3s throughout the study area from the once- active major point sou.ce in Massachusetts. The apparent leveling I off of the PCB concentrations in fish from 1984 to 1986 suggests lower reductions with greater time elapsed following the abrupt reduction in GE's industrial discharge in Pittsfield, Massachu I setts.

I 5.4 PCB FATE AND TRANSPORT MODEL I IMS modeled the transport, transformations, and reactions of PCBs in the study area from the Connecticut-Massachusetts border to the I Stevenson Dam in Lake Zoar. A key model parameter is the rate of sedimentation and burial of PCB-laden sediments. Measurements of cesium (a radioactive element from the atmospheric fallout primar I ily in the 1960s) in sediments on the bottom of Lakes Lillinonah and Zoar provided data to evaluate the burial rate and the depth of I the active sediment layer. PCB analyses of bottom sediments in these lakes and in the Falls Village and Bulls Bridge impoundments I showed evidence of limited biodegradation within the study area. The model was calibrated to total suspended solids measurements at I several locations in the Housatonic River over an 18-month period. Although the relatively high detection limit for PCB analyses of water samples does not allow for precise quantification of PCB con I centration in river water, the preponderance of below-detection- limit results supports, at least qualitatively, the model's com I puted PCB concentrations, which are in the subdetection limit I range. I 5.5 PROJECTIONS OF PCB CONCENTRATIONS The calibrated model was applied to project the concentrations of I PCBs in the water and bottom sediments over a 50-year period. A I S-3

I Lawler, Matusky 5^ Skelly Engineers PGE00047129 I I diminishing source of PCBs is assumed for the upstream boundary and * a minimum background FCB concentration is set for the upstream and tributary inflows. A 50% reduction in the PCB concentration of I water and sediment is projected for approximately 20 years from 1986 and an 80% reduction for the year 2036. The extrapolation of | PCB concentrations in fish is based on a direct linear relationship between the response in the fish and the change in PCB exposure I levels in the water and sediment. While the median PCB concentra tions for all species sampled are projected to decrease well below • the FDA action level, approximately 4 to 5% of certain species • (brown trout, white catfish, and carp) are projected to be above ^ the FDA action level 50 years from now. |

S.6 SOCIOECONOMIC IMPACT ON THE RECREATIONAL RESOURCE I

The recreational value of the Housatonic River and its environs is fl well recognized. The river accounts for approximately 1% of the statewide freshwater fishing pressure. The impact of the fish eating advisory on the recreational resource cannot be assessed for I the entire study area due to the lack of pre-advisory data or regional data that would show any potential transfer of recrea- | tional usage to other rivers. However, pre- and post-advisory data for the trout management area show an increase in fishing pressure • attributable to the zero creel policy (return of captured fish to the river) instituted approximately four years after the advisory. •

S.7 FUTURE DEVELOPMENT AND THE NO ACTION PLAN •

The scope of this study was expanded to incorporate the considera tion of future development involving sediment and PCB movement in the Housatonic River, such as dredging for a marina and expansion I I S-4 I Lawler, Matusky ¥ Skclly F.ii^ineers PGE00047130 I I I of a power canal. The existing state and Federal regulatory pro cesses for permitting dredging and filling were reviewed and found I to be comprehensive in addressing relevant issues such as monitoring and mitigation. Of the various construction activities, dredg ing and disposal are considered to have the highest potential im I pact on PCS release. However, the impact can be minimized through proper design, mitigation, and monitoring. There are two basic I alternatives for the disposal of sediments dredged from the Housa honic River study area: upland encapsulation and subaquatic dis I posal. The salient difference is that the upland alternative in volves the transfer of PCBs to an uncontaminated terrestrial en I vironment and poses greater environmental risk than subaquatic disposal, which keeps the PCBs within the aquatic environment. An analytical framework for evaluating the release of PCBs during and I after dredging and subaquatic disposal is presented for a hypothetical dredging project within Lake Lillinonah and shows a neglig- I ible increase in PCB release. Hence, in cases where development is deemed beneficial and dredging of Housatonic River sediments is I necessary, subaquatic disposal is recommended to be consistent with I the no action plan. S.8 MONITORING PROGRAM

I A monitoring program is proposed to track the natural recovery of the river. Sampling and analysis of PCB concentrations in sedi- I ment, water, and fish are specified. A comprehensive reassessment of the recommended monitoring data is scheduled for 1992 to chart I the river's recovery in comparison to the projections. Subsequent monitoring would be contingent upon the findings of the reassess I ment study. I I S-5 I Lawler, Matusky Sf Skelly Engineers PGE00047131 I I I CHAPTER 6 I PROGRAM FOR MONITORING THE NATURAL RECOVERY OF THE RIVER I 6.1 OVERVIEW OF NO INTERVENING ACTION PLAN I 6.1.1 Background This chapter completes the Housatonic River PCB Sediment Management I Study. The study is an outcome of the Housatonic River Agreement (HRA) between General Electric (GE) and the Connecticut Department of Environmental Protection (CDEP) signed on 1 June 1984. Chapters I 1 through 5 constitute the interim report published as a separate volume by Lawler, Matusky & Skelly Engineers (IMS) in June 1985. I The interim report (IMS 1985a) characterized the Connecticut por tion of the Housatonic river (Figure 6.1-1), including Lakes Lilli I nonah and Zoar and the impoundments at Bulls Bridge and Falls Vil lage, and presented data on the sediment and polychlorinated bi- I phenyl (PCB) material based on a 1979-1980 study performed jointly by the Connecticut Agricultural Experiment Station (CAES), the U.S. Geological Survey (USGS), and CDEP (Frink et al. 1982). The I 1985 report also described the preliminary screening of more than 40 individual treatment techniques for destroying PCBs or otherwise I rendering them harmless according to three major criteria: I • Engineering feasibility • Cost effectiveness I • Environmental effects I Of these treatment techniques, only capping with clean sand and dredging/excavation followed by upland disposal or incineration I warranted secondary screening. Costs for excavation or dredging I were estimated at $60 million to $190 million for either Lake Zoar 6-1 I -lnr, Matusky ~'Sk«lly Knjrineers PGE00047132 FIGURE 6.1-1 I HOUSATONIC RIVER AND WATERSHED AREA I I I I I MASSACHUSETTS I NEW YORK I I I I I I I I I SCALE IN MILES 5 T o 5 10 15 2O I Source: Frink et a I. 1982 6-1 A I PGE00047133 I I or Lake Lillinonah and $1 million to $3 million for either the Bulls Bridge or Falls Village impoundments. The report further I stated that: • "These estimates were based on the assumption that I the closest available landfill site is capable of accepting - and willing to accept - the PCB-laden material. Indications are that this 1s unlikely I and that disposal cost will be much higher. • "Environmental effects include destruction of bot tom fauna and consequent disruption of food I chains. Fish spawning sites may also be buried. Of particular concern are the effects of these perturbations on the endangered bald eagles in the I Lake Lillinonah area." I The detailed cost/benefit evaluation of the alternatives that passed secondary screening (dredging and capping for the larger I lakes and dredging for the impoundments) as well as the no action alternative was to be performed in the subsequent phase of the sed I iment management study. The interim report was reviewed by CDEP and organizations identi- I fied in the HRA and comments were submitted by: the Housatonic Valley Association, the Housatonic Fly Fishermen's Association, the I Lake Zoar Authority, Northeast Utilities Service Company, the Connecticut Council Trout Unlimited, and the PCB Watchdog Commit I tee. (As stated in the HRA, the other organizations that also re- ceived copies of the interim report but did not submit comments were: Lake Lillinonah Authority, Candlewood Lake Authority, I Housatonic River Commission, and regional planning agencies.) Based on the aforementioned organizations' comments and data fron I Housatonic River monitoring, CDEP stated on 2 May 1986: 1. Further studies of capping and dredging of Lake I Lillinonah and Lake Zoar, and further studies of I hydraulic dredging of Falls Village and Bulls I 6-2 I La\vlei\ Malusky okelly KntiruMj PGE00047134 I I Bridge, are not recommended. Reconsidering of dredging and capping alternatives may be war- • ranted if future monitoring indicates a reversal p of current trends in fish which can be linked to PCB contamination of bottom sediments in impound- _ merits. • 2. Further study of the no action alternative is recommended as the focus of the final sediment • management report in accordance with Sections ID I and IE of the HRA. 3. "The Housatonic Valley Association and the Housa tonic Fly Fishermen's Association.. .recommended that remedial action in the Connecticut section _ of the river be held in abeyance until positive • results are accomplished in Massachusetts to ™ reduce PCB transport to Connecticut. The DEP concurs with these points of view." •

CDEP also recommended that the no action study address the follow- i ng: I • Assessment of socioeconomic (i.e., recreational ft resource) impacts • Additional collection and analysis of sediment cores to evaluate natural PCB degradation I • Estimation of burial rate of PCB-laden sediment by _ deposition of clean sediments • • Development of a long-term monitoring program to assess trends of PCB levels in the sediment and • biota (mainly fish) •

In its letter of 15 August 1986 to CDEP, GE agreed to study the no I action plan as the recommended course for managing the Housatomc River's PCB situation. The detailed scope of work proposed for B completing the sediment management study included an analytical framework or mathematical model. By integrating the cause and ef- • feet mechanisms, such as sediment transport and burial of PCB-laden • sediment by clean sediment, the model provides a basis for analysis I Lau'Icr Maluskv Sk«ll\ l.ntrineers I PGE00047135 I I of the fate of PCBs in the Housatonic River. PCB concentrations in the sediment, water, and fish can be projected for a number of I possible conditions within the no action scenario, e.g., minimal PCBs crossing the Massachusetts state line. These projections I provide guidelines for designing the monitoring plan. In addition, trends depicted by the model can be validated as new monitoring I data become available, or the model can be adjusted and projections revised, if necessary. Hence, the model is a tool for charting the I natural recovery of the river. I 6.1.2 Chapter Organization This chapter of the study relies on information presented in the I first five chapters (the interim report). Section 6.2 reviews the scientific literature and data on PCB levels in the water, sedi I ment, and fish of a diversity of water bodies to provide a perspective on the present status of the Housatonic River. The emphasis I is on areas having no known PCB sources so as to evaluate back ground levels that would be similar to tributary inputs to the Housatonic River; also trends in background levels would be indica- I tive of the endpoint condition at full recovery. The trends in the PCB body burden of Housatonic River fish are also discussed in this I section. I Section 6.3 presents a generalized toxic contaminant model, WASTOX, to simulate the transport of PCBs in the Housatonic River. It models the particulate and dissolved components of PCBs in the I water and bottom sediments as a function of the transport, transfer, and reactions. Model projections of PCB concentrations in I water and sediment are then used to project PCB levels in fish I based on empirical relationships of bioaccumulation. I I 6-4 I Lnwler Malusky Skelly Kntjineers PGE00047136 I I Section 6.4 discusses the environmental impacts of the no action plan in two parts: (1) the socioeconomic impacts associated with • the recreational resource, and (2) the potential impact from future development involving sediment/PCB movement, e.g., dredging for a • manna.

Section 6.5 puts forth a monitoring program that includes a pro- I posed schedule for data collection and interpretive evaluation. I I I I I I I I I I I I

6-5 Liiwler Matusky Skcllv K

PGE00047137 I I 6.2 PCBS IN WATER, SEDIMENT, AND FISH: HOUSATONIC RIVER IN I PERSPECTIVE As a result of their Industrial application before the mid- to late I 1970s, PCBs today are widely distributed throughout the United States. These persistent compounds, whether released directly into I marine or freshwater systems or deposited initially on land, are eventually carried over and ultimately found in the aquatic envi- I ronment (NAS 1979). Entry of PCBs Into aquatic systems depends not only on the location I of specific water bodies but also on their size, PCBs commonly enter large water bodies through point sources or tributary contri I butions, and other nonpoint sources, including dispersion of contaminated sediments and atmospheric deposition. These nonpoint sources, as well as continued use and disposal of PCBs, are prob I ably responsible for the ubiquity of PCB concentrations in fish throughout the United States (Schmitt et al. 1981). Atmospheric I contributions alone have been cited as the major cause of contami nation of waters in the vicinity of Isle Royale in Lake Superior I (Swain 1978) as well as in the waters of Lake Michigan (Swackhamer and Armstrong 1986). Water bodies with high surface areas, extreme I depths with low hydraulic turnover, and little suspended particulates relative to the total water volume are more susceptible to I PCB bioaccumulation (Swain 1983). Point sources were primarily responsible for the initial entry of I PCBs into the Housatonic River; however, contamination currently depends on the distribution and transport of fine-grained river I sediments (Frink et al. 1982), as discussed in Section 6.3. I To better evaluate the status or condition of the Housatonic River I with regard to PCBs, and to provide a comprehensive understanding I 6-6 I Lawlor, Matusky Gkelly L PGE00047138 I I of the physical and chemical transformations and the resultant bio logical accumulations of PCBs in the river, it is of value to re • view PCB trends in concentrations throughout the country. Of spe cial interest are concentrations reported from areas containing no • known point sources, especially in the northeast.

While analytical capabilities have improved, the quantification of I PCB concentrations is still limited. Total PCB concentration is still the most reliable means for comparison, regardless of which J Aroclors are included (Schmitt et al. 1985). Even so, the data presented later in this section are based on the best estimate of • the Aroclor pattern observed, and may under or overestimate the actual quantity present due to a number of factors (Finlay et al. • 1976), Including the equipment and staff employed. Furthermore, ™ sampling within a state is often not truly representative of the M state or may be biased toward a given basin. Finally, arithmetic I means, medians, and ranges do not adequately describe PCB distribu tion within the aquatic system. Earlier studies (to be cited) made p little or no effort to quantify the detection of PCB concentrations with regard to sediment, water, and fish samples. •

Even so, such a review will help emphasize the ubiquitous nature of • PCBs, and, since the Housatonic is located in the northeastern part * of the country, comparison with data from other waterways in this region will provide background for better understanding of the I relevance of PCBs in sediment and fish tissue. Specifically, by noting ambient levels of PCBs (occurring through nonpoint sources) • in the northeast, we may infer what percentage of PCBs in the Housatonic is attributable to similar nonpoint sources via tribu • taries and make estimates as to the background levels attainable at full recovery. I I Luwlnr Matuskv »>kclh K I PGE00047139 I I Comparisons of riverine systems with well-known PCB discharges, such as the Hudson River or New Bedford Harbor, have purposely been I avoided. Such comparisons would not be consistent with the objec tives of this report, which are to predict what concentrations may I be expected in sediment, water, and fish over a given period of I time and the total time to full recovery. 6.2.1 Water

I Analyses of whole water samples from the Housatonic by USG5 showed PCB levels in both unfiltered (total PCBs) and filtered (dissolved I PCBs) water samples (Table 6.2-1) (Frlnk et al. 1982). The level of total PCBs in the Housatonic River generally tends to be greater I than or equal to that of whole water samples collected by USGS from a variety of land use areas throughout the United States from 1971 I to the partial year 1975 (Table 6.2-2) (Crump-Wlesner et al. 1973; Finlay et al. 1976) and for whole water samples collected from the I north Atlantic slope from 1971 to 1974 (Table 6.2-3) (Dennis 1974). In whole water samples collected from 49 states, detectable concen- I trations ranged from 0.1 to 4.0 ug/1, although they occur only 3% of the time for all states sampled. The range of detectable PCB I values reported for the north Atlantic slope was smaller and the maximum decreased from 2.1 ug/1 in 1971 to 0.1 ug/1 in 1974. The I frequency of PCB occurrence also decreased: from 17-60% in 1971 1972 to 2-4% in 1973-1974. In the Housatonic the range of detect able concentrations (0.1-0.6 ug/1) found in 1979-1980 was similar I to that reported for the north Atlantic slope, but the frequency for all three stations combined is greater than 50%. This indi- I cates that for the Housatonic the median value reported for all samples of detectable PCBs is somewhat representative of actual or I "true" whole water concentrations. The true median of all samples I collected at Falls Village and Gaylordsville, however, may be sub- I 6-8 I Laivlcr Matusky Skelly KMoineors PGE00047140 I H>

• • o CD CD UJ I

OJ to I 0 1— CM •>»• ^—Ji JJ £2 0 0 19 1 — 1 1 ^ OC JJ CD -I .-l I z |t | LXJ _J o 0 O CD oc 0 o o I o o 1— to ^c LU to CJ I o LU T o: LO rx *~H o± o o c* o u. o I to LU CJ z ^ CM •-! ^-l c < t__l I ^ • * • c o to O O CD o •• -LJ -r— •»->

2 LU U 1- «-• 1— 03 J-> 1 <£ j l_ C I CM 3 (/> ^y • u CD UJ ™ •a cr O _ a> o LJJ O IO CM i-l > CJ 1 ^T" _ JJ • • • co 3: cc ^Q^£ O O O O CO I ^c LU «- 1 1 i/> O — ~^ h- Q£ __ r—1 <—1 »—* I/I Q 0 LL * • » .r^> U_ *—• O O O O T3 O)

»•—• 2: O (U l/l f* r— ^3 CJ p^ aJ rrj I *»—. a. c *-> C7 £ >^ ^—• CO I/I OJ 4J t— LU i_ ai c ^ O l_ CLT3 0 h- ~z. O) OJ I < LU 4-> \— ^2 tfl 0 C£ *r to CM 10 4-> a) ci; •—* S a) ••— o O c •DO . 01 »—« CJ ^•™^ *— F^ »/> (_ I i— o CM o re) a) L. < oo -cr ui — 3 Ctf O S Q. E o O LU u |_ -M I/) CJ « CL. *^" ^~ OO • C Lf) l_ r-( ?—1 »—t I ^ o s: i— O) ^* M O o 3c < (« 10 0 S o CO OJ l_ 4-> L_ <1> L_ flj CO HI Q--<-> (U < o QJ QJ ^^ •> ro •— i*— "O "^^ CJ "O L_ »^ -r— a> o c t_ t— > ~u i- T3 c: • TD •r— 4->> L_ L_ I/I o L_ <» .— QJ 1_ OJ I (—1 *J to O •r- -r- 4-> 3 CT> }— (TJ F— ~ TJ t- r U C c 0) r- >, O C -i U rO 1— Z 3 u_ o a: (/• O U_ C3 ro ja o T3 cu I

6-8A

PGE00047141 1 IABL t b.L-e. irage i or jj • PCB CONCENTRATIONS (ug/11 IN WHOLE WATER SAMPLES FROM A VARIETY OF LAND USE AREAS3 1 USGS Data 1971 to Partial Year 1975 1 m No. OF STATE SAMPLES OCCURRENCES13 CONCENTRATIONS0 MEDIANC 1 Alabama 6 0 1 Alaska 3 0 Arizona 72 0 1 Arkansas 91 1 0.2 Cal ifornia 439 4 0.1-0.1 0.1 1 Colorado 50 1 0.3 I Connecticut 74 13 0.1-0.2 0.1 Florida 204 4 0.1-1.0 0.1 1 Georgia 30 0 Hawaii 8 0 1 Idaho 8 0 11 1 inois 2 0 Indiana 1 0 1 Iowa 54 1 0.1 Kansas 41 1 0.2 1 Kentucky 7 0 Louisiana 373 2 0.1-0.2 0.2 1 Maine 10 0 1 Maryland 6 1 0.1 Massachusetts 21 4 0.1-0.2 0.1 I • aSamples include surface and groundwater samples without distinction, bOccurrences are number of samples with detectable PCBs. 1 cRanges and medians are for occurrences only. 1 1 6-8B1 PGE00047142 I TABLE 6.2-2 (Page 2 of 3) PCB CONCENTRATIONS (ug/1) IN WHOLE WATER SAMPLES FROM A ™ VARIETY OF LAND USE AREAS3

No. OF STATE SAMPLES OCCURRENCES13 CONCENTRATION0 MEDIANC I Michigan 51 1 0.1 Minnesota 5 2 0.1, 0.3 I Mississippi 121 0 I Mi ssourl 47 0 Montana 96 0 I Nebraska 90 0 Nevada 11 0 I New Hampshire 6 0 New Jersey 74 6 0.1-<0.5 0.1 I New Mexico 56 0 I New York 373 53 0.1-4.0 0.3 North Carolina 5 0 I North Dakota 50 0 Ohio 18 2 0.1, 0.2d I Oklahoma 149 9 0.1-3.0 O.ld I Oregon 34 0 Pennsylvania 48 1 0.2 I Rhode Island 3 0 South Carol ina 16 0 I South Dakota 25 0 I aSamples include surface and groundwater samples without distinction. • bOccurrences are number of samples with detectable PCBs. • cRanges and medians are for occurrences only. ^Estimated value.

6-8B2 |

PGE00047143 I TABLE 6.2-2 (Page 3 of 3) PCB CONCENTRATIONS (ug/1) IN WHOLE WATER SAMPLES FROM A I a I VARIETY OF LAND USE AREAS No. OF I STATE SAMPLES OCCURRENCES'3 CONCENTRATION MEDIAN0 Tennessee 3 0 I Texas 1424 23 0.1-0.7 0.3d Utah 25 0 I Vermont 1 0 I Virginia 22 1 0.1 Washington 38 1 0.1 I West Virginia 5 0 Wisconsin 48 0 I Wyomi ng 29 0 I Source: Crump-Wiesner et al. (1973), updated by Finley et al. (1976). aSamples include surface and groundwater samples without distinction. bOccurrences are number of samples with detectable PCBs. I cRanges and medians are for occurrences only. I Estimated value. I I I I I I I 6-8B3 PGE00047144 I I I TABLE 6.2-3 PCB CONCENTRATIONS IN WHOLE WATER SAMPLES OF THE NORTH ATLANTIC SLOPE I I WHOLE WATER NO. OF SAMPLES OCCURRENCES3 CONCENTRATION13 MEDIANb I 1971 12 7 0.1-2.1 0.2 1972 112 19 0.1-0.3 0.1 I 1973 109 4 0.1 0.1 I 1974 51 1 0.1 0.1

Source: Dennis 1976. Data gathered by USGS. Occurrences are samples with detectable PCBs. • bRanges and medians are for occurrences only. • I I I I I I I I

6-8C I PGE00047145 I I stantially below the median reported (0.1 ug/1). How much below depends on the actual PCB concentrations of all samples, including I those below detection limits (to be discussed further in Section 6.3.2.8). In contrast, the preponderance of zero readings from I scattered whole water measurements (Table 6.2-2) appears to indi cate that actual whole water background levels are relatively lower I than the medians reported. This suggests that background PCB concentrations are somewhat lower than PCB concentrations in Housa I tonic River water, although it is difficult to say by how much. Even so, due to the analytical procedures used at the time, which limited detection to 0.1 ug/1 (Finlay et al. 1976), zero readings I may actually mask widespread contamination.

I It is well documented that concentrations in water are often nondetectable or, if detectable, many times lower than the levels I found in the immediate underlying sediments (Bruner and Hill 1977; Dexter et al. 1978). This is not surprising considering the physiochemical properties of PCBs, which impart a high specific I gravity, low water solubility, and high partition coefficient to I these organic compounds (Crump-Wiesner et al. 1973; Dennis 1974). The bulk of PCB concentrations In fresh water is therefore bound to I suspended particulates, bottom sediment, or other solid surfaces (Crump-Wiesner et al. 1973; Horn et al. 1979; Pavlou and Dexter I 1979), and stream transport of PCBs is primarily by means of waterborne particulates (Nisbet and Sarofim 1972; Sawhney et al. 1981). The Housatonic River generally has low levels of suspended sedi I ments, as demonstrated by suspended sediment discharge studies (Frink et al. 1982), and this may explain the relatively low PCB I concentrations in whole water samples. I Considering the above, it is apparent that water samples alone give I only a poor indication of whether PCBs are present, and that a more I 6-9 I Lawler Matusky GkelJv Ku

6.2.2 Sediment •

In the Housatonic River, Sawhney et al. (1981) found PCB concentra- | tions to be highly correlated with clay and organic matter. They suggested that both retention and transport of PCBs in the river • were dependent on the distribution of fine particles on clay organic-matter complexes. This accounts not only for relatively • low levels of PCBs in free-flowing sections of the river between * impoundments, where Frink et al. (1982) noted fine sediment to be • scarce, but also for higher PCB concentrations within the lakes, • where organic sediments are more abundant. In the impoundments of _ Falls Village and Bulls Bridge seismic profiles show bottom sedi- | ments to be about 6 in. deep, evenly distributed, and consisting mostly of sand and some silt with little organic content. Sediment • along the shore near the dam of Falls Village consists mostly of clay and silt with little organic matter (Frink et al. 1982), and • PCB concentrations do not vary with distance from the dam (Frink et * al. 1982). By contrast, in both Lake Zoar and Lake Lillinonah re- « cent sediment and associated organic material are thickest at the • respective dams and gradually thin out with increasing distance up

6-10 • Lawlor Malusky okclh Knqiruicrs I PGE00047147 I

I stream. In these large water bodies organic material Is thicker where bottom topography Is flat or gently sloping or where wave i action is minimal, allowing for accumulation. PCB concentrations are thus not evenly distributed in either lake, but decrease with I distance from the dam. For this reason it was more difficult to determine PCB content in these water bodies than 1n the smaller I impoundments of Falls Village and Bulls Bridge, and regression analysis was necessary to estimate the PCB mass.

I Since the bulk of PCB residues are associated with sediments (either suspended or settled), concentrations depend upon the quan- I tity and composition of sediment present. Though sediment data provide a more discernible Indicator of PCB contamination than I water data alone, meaningful interpretation of these data in terms of the extent above background levels is difficult. Nevertheless, I the relevance of PCB contamination in the Housatonic River in terms of sediment and fish tissue (as discussed in Section 6.2.3) can be better understood if compared with that of other rivers and lakes I within the immediate and outlying regions.

I On a broad scale, nationwide sampling of freshwater bottom sedi ments indicates widespread occurrence of PCB residues (Finlay et I al. 1976; Nisbet 1976). Monitoring of lakes, streams, and wetlands by USG5 (STORE!) has disclosed isolated incidences of PCBs in what I were otherwise considered pristine environments as well as in other remote areas (Table 6.2-4).

I Crump-Wiesner et al. (1973) collected sediment data on PCB occurrence in streams included in the national hydrologic benchmark net- I work. This network, established by USGS in 1958, used 57 stream basins located throughout the United States to provide basic hydro- I logic data. The benchmark sites were expected to remain in their I natural condition and were not expected to be significantly altered I 6-11 I La\vler Malusk Skolly I'. PGE00047148 I I TABLE 6.2-4 PCB CONCENTRATIONS (ug/kg) IN SEDIMENTS OF | REMOTE HABITATS WHERE PCBs HAVE BEEN DETECTED I No. OF CONCEN PRISTINE WATERSHEDS SAMPLES OCCURRENCES2 TRATIONb MEDIANb I South Fork Rocky Creek near 1 1 5.2 Briggs, Texas0 Upper Twin Creek, McGraw, Ohioc 1 1 8.8 I Big Cypress Swamp, Big Cypress 28 10 1.5-130 5.5 National Preserve, Floridad I Everglades Conservation Area, 17 10 12-8500 96 National Park, Florida6 I

Occurrences are samples with detectable PCB concentrations. • ^Ranges and medians are for occurrences only. cFrom Crump-Wiesner et al. (1973). Data gathered from 46 of 57 national I hydrologic benchmark sites, originally selected by USGS to represent nat ural stream basin conditions throughout the United States where altera- • tions by man were expected to be minimal. Of 46 sites sampled in 1972 for | PCB analysis of sediment, only two showed detectable PCB levels and are included in this table. _ dFrom Bruner and Hill (1977). Occurrence of 130 ug/kg value from one of • four samplings at one of seven stations deviated greatly from other values. Point source or electric transformer along Florida Interstate 84 • were suspected of contaminating this one sample, which is not considered • representative of true background levels. eFrom Bruner and Hill (1977). Data from STORET water quality file. Val- • ues represent annual means, and the range and median of annual means from all stations where total PCB concentrations were equal to or greater than 100 ug/kg in sediments in one or more samples. Point sources along U.S. I Route 41 suspect for high value of 8500 ug/kg. m I I 6-11A I

PGE00047149 I

I by man. To ensure minimal interference by man, many sites were located in national parks, wilderness areas, state parks, national I forests, and areas set aside for scientific study. Despite careful screening for pristine locations of benchmark sites, two out of 46 I bottom sediment samples analyzed in 1972 contained detectable PCB I concentrations. Bruner and H111 (1977) monitored sites in Big Cypress Swamp of the Big Cypress National Preserve, where human activity is considered I minimal, to establish background levels for the southwestern United States. They reported background levels as high as 30 ug/kg in I sediment and noted several sites within the Everglades conservation area with annual means greater than 100 ug/kg and one site as high I as 8500 ug/kg. (The latter was located along U.S. Route 41, and I point sources were suspect.) Crump-Wiesner et al. (1973) also summarized sediment data collected from lakes and streams known to drain a variety of land use areas, I generally located away from industrialized centers (Table 6.2-5). They reported sediment concentrations taken at random from 16 I states across the country to range from 5.0 to 2400 ug/kg, and noted that, overall, one of every five sediment samples examined I contained measurable PCBs. Because the PCB detection limit of sediment, approximately 10 ug/kg, is relatively low compared to the I range of detectable PCB concentrations reported, sediment zero readings do not obscure the actual mean PCB concentration to the I extent noted for water concentrations. While USGS (1971-1972) whole water and sediment data suffer from a I lack of representative sampling within states and the inclusion of multiple samples from the same locations (Finlay et al. 1976), I these data still demonstrate significant PCB concentrations to be I widespread in the water resources of the nation (Crump-Wiesner et I 6-12 I La\\ ler. Maluskv okelly Knirinc PGE00047150 I I TABLE 6.2-5 PCB CONCENTRATIONS (ug/kg) IN SEDIMENTS OF LAKES AND I STREAMS FROM A VARIETY OF LAND USE AREAS I STATE No. OF SAMPLES OCCURRENCES3 CONCENTRATION0 MEDIAN0 Alaska 3 0 I Arkansas 23 4 20-2,400 60 I California 13 3 20-190 85 Connecticut 1 1 40 I Hawaii 4 0 Georgia 12 10 10-1,300 300 I Maryland 11 5 10-1,200 30 I Mi ssi ssippi 8 2 50;170 New Jersey 12 10 8-250 20 I Oregon 4 2 15;140 Pennsylvania 16 11 10-50 20 I South Carol ina 11 8 30-200 50 Texas 293 23 7.9-290 80 I Vi rgi nia 10 8 5-80 40 I Washington 10 0 West Virginia 2 1 10 I

Source: Crump-Wiesner et al. (1973). Samples taken at random (by USGS Jan I 1971 - Jun 1972) from lakes and streams draining a variety of land use areas, generally located away from industrial centers. I Occurrences are samples with detectable PCB concentrations. bRanges and medians are for occurrences only. I I 6-1 2A

PGE00047151 I I al. 1973). Finlay et al. (1976) also reported STORE! bottom sediment data collected from a variety of land use areas in states I throughout the country (Table 6.2-6). While the majority of states reported zero occurrences, sediment concentrations higher than 40 I ppb were common in states that monitored sediment PCBs to any I appreciable extent during the year 1972 to partial year 1975. In Connecticut itself (excluding the Housatonic River) PCBs appear to be ubiquitous in river and lake sediments at levels commonly I approaching and sometimes surpassing 40 ug/kg. Based on sampling in 1974, Finlay et al. (1976) reported PCB contamination to be both I widespread and at levels up to 350 ug/kg throughout the state. Greatest concentrations were noted around highly developed areas of I Hartford, New Haven, Stamford-Greenwich, and the Housatonic River from Massachusetts through Connecticut. The mean value for all of I the 1974 state-reported sediment PCB concentrations was 57.0 ug/kg (Table 6.2-7).

I Higher sediment PCB levels were reported 1n more limited surveys. Sediment samples collected and analyzed by USGS (1973-1977) from I the Connecticut and Thames River basins ranged 1n concentration from 0 to 1000 ug/kg and had a mean of 43 ug/kg (Frink et al. I 1982). Analyses of sediment from the Still River during 1974 and 1976 also showed high PCB concentrations, up to 2400 ug/kg, with a I mean of 433 ug/kg (Frink et al. 1982). Limited sampling of seven Connecticut lakes and ponds by the Connecticut Agricultural Experi ment Station showed Ball Pond in New Fairfield, Bantam Lake in I Litchfield, and Linsley Pond in Branford and North Branford with sediment PCB levels of 360, 30, and 50 ug/kg, respectively (Frink 1 et al. 1982). These three freshwater bodies were reported by Nor vell et al. (1979) to be highly eutrophic, and urban runoff is a 1 likely source of contamination. While these surveys show that sed

6-13 Lawler Malusky JJkelly Kn^inoer.s

PGE00047152 ID S C71 1 CM CVJ 1 T

u 4S y £ § 4 1

JE" t/ii>i 5 a CM O ^* 1 3 Q Set

UJ *-i LO 1 S r~t •^

ao to ir> c»i 10 o 1 *X S^ S F^ 5 % m ^

O> U"> Lrt rO ^-« i—< o CM tc> ro t/) in tjO uS i T i i i T T 1 rtz <; \s \s \s \? \s \s \/ on B^

i 2? UJ p u 1 1 o o^cMO^-tior*^ o rsj o o co rn r~< 5 to Sr5 £• gg r*J *-4ir>CO*-iOOOcn ^-t ^ CNJOO rn 1 o« »-«*-» ^r 2) ^H ^-» «- U- UJ o ^

u ^ CD CD« -iO)< W i/i 1 a.^ —Z3 _J 1 t-« ro

<£> t— \ CM O K il s ? * 1 LU •—'

-» 1 t/) § 1| K U O ^5 O O O O^ *~* *V CD C3 O ^^ O> ^5 O O 1 o Hg 8 I/ i" I9 1£— I/) gi *-» •-< ^ £.° <4- a _ : I ^ si reu BV $™ d. cr (» 88^ •

U) l/> CVJ x^ -^ xx lit ^ 1 CO O G> 0) fO i^> 3 ?NI L_ -»~> D •— I— -*-J -^ 8 B "a> i- C .— t/l O TO • l|t « 1 *1 ^ CM *"* """ *" "^ ""* a> d> c *— •J3 10 *° ^ "o f oUJ ) TI • C TO 3 40 3 ^ d-g ^ r3 OJ 4-> Q n gSE-8+Jigio -în « j?ra > i_ iô fjj T5 U O £ j: 5'«Sra— oc'CE1 'joj?-H-2|n!Ji"c:5S in 3 en ^- c: <55ls§Sc^l3EEl|îi£ sj^jjpj ^

6-13,"

PGE00047153 1 u i 3 1 4% li_J Ct OH 0 o en X i-4 1* 1 sy_> •?* Oooc: 1 b5 si u> ^ C\J OO 1 s<• Pi JS s i| S 1 z3* —•< «? y OP 3* t-t s;l-e^, oo^ooor*vo • !ins 2w3^ o o to —» or^-oooo J u_ 5 t=« in SS 1 """" ce sa r-,

s ri 1^H liOzu OC. ESb <-> gi M 8 in &^ a CM <-t cr> r-l in CM •-« S 1 t/i? z to •^ 1 a 1 *i o f- cr> oi y w 3 VJ ii ^-« L. d> CU P— T> rH at.•f UJt c £' 5 OOOCMO OOt^OO *J (O -^- U 1 4J > U -,-*-> ^ « (U L en ^— i^t O &y 1 cr> CM •—• °l H- 0> QJ d -r i_ k. rp T) fO OJ O +J * E Q •r— V) OJ ~Q * ••- 3 u C "D to 1 & "O •— C fO _ UJ CJ•• irsn. 3u QJtD) — c 1 1 6-13A2 PGE00047154 I I TABLE 6.2-7 PCB CONCENTRATIONS (ug/kg) IN BOTTOM | SEDIMENTS OF CONNECTICUT RIVERS AND LAKES I No. OF FRESHWATER BODY SAMPLES OCCURRENCES11 CONCENTRATION6 MEAN5 I Various rivers and lakes3 36 27 0-350 57.0 Connecticut and Thames riversb 79 0-1000 42.8 I Still River0 13 12 0-2400 433.4 Ball Pond in Fairfieldc 1 1 360 360 I 0 Bantam Lake in Litchfield 4 - 30 I Linsley Pond in Branfordc and 3 - 50 North Branford I - Data not available. aFrom Finlay et al. (1976). Data collected by the State of Connecticut, • October 1974. bl)SGS, Water Resources Data for Connecticut (1974-1978). m cFrom Frlnk et al. (1982). Ponds and lakes cited had detectable PCB con- • centrations. A total of five other sediment samples from four other ponds/lakes showed nondetectable PCB concentration. Occurrences are samples with detectable PCB concentrations. • eRanges and means are for occurrences only. • I I I I I 6-13B I

PGE00047155 I I iment PCBs commonly approach and often surpass 40 ug/kg, they are I mostly restricted to urban areas or major waterways. A more comprehensive data base for both urban and rural areas in I Connecticut consists of sediment data gathered from 1973 to 1980 by USGS and published in their annual (1974-1981) Water Resources Data I for Connecticut (Table 6.2-8). These data, which cover 72 sampling sites throughout Connecticut, were classified as urban or rural on I the basis of proximity to population centers. Although the classi fication of certain sites may not be consistent with land use of the drainage area, the overall data base further attests to the I widespread distribution of PCBs in the state. Figure 6.2-1 shows rural areas interspersed with urban areas, a combination that has I produced detectable sediment PCB concentrations (6 ug/kg, weighted average) even in exclusively rural, non-Housatonic areas of the I state. The much higher concentrations in urban, non-Housatonic areas (99 ug/kg, weighted average) is to be expected, and indicates PCB contribution from point sources. (In Housatonic areas the I higher average PCB levels for rural sites is most likely due to I their location closer to the upstream source.) Overall, this data base appears to agree with the conclusions I reached by Finlay et al. (1976), and background levels for sediment PCBs of approximately 40 to 100 ug/kg do not seem unreasonable for I Connecticut. Connecticut also does not appear to be more heavily contaminated than other areas of the country where extensive PCB I sampling has been conducted. Highest sediment PCB concentrations in the state are found in areas I associated with the Housatonic. Comparing Connecticut data alone, combined rural and urban Housatonic areas have sediment PCB concen I trations about five times greater than concentrations of combined I rural and urban non-Housatonic areas. More recent sampling by I 6-14 I Lawler Matusky okolly lintrincor.s PGE00047156 1 LU t/1

~~0 z~~

d« f*v Irt IO ^3 O ^3 t^J ^ •—t (O l/> ^^ fxj o O <""* rn f^ r^ CNJ ^3 CNI O r^ O tO CNJ *J rNl i—^ (Nl f-i csj CM «r 1 iaJ< o * CJ ^3 o o O O 0 0 CO •— LU 1 CJ H- O CT v~t Ol U *^ CNJ *"* Csj *H ?—4 *^ r^ c\t ^ (^ f) t—> I-H I-H r^ ^± CO •—i r*~) O- < Z 1 1 lOOfOt 1 1 1 1 1 IOOOI 1 1 IOILDIO I 1 1 1 1 1 oc <: moo k-Of^ir»oootf> oooo OOCNJO 101-00000 h- ce tn h- *-*^tcsj CM •—« CM ^^ 1

UJ *— g ;(-3> o i— 8 o UJ to 1 o1 LU fe 1 UJ >— «r 1 t— -— \n tOIOCONIO <0<0<0<0<0<0lD«><0(£><£><010 l£linOIO>O LU 1, —r _ oto>cnao> o>o>aicncnoicr)O>o>cncr>cnc»ar>o>cn CT>CT>cnoia,cncr> o t— CE UJ g.g.o.g.o,^ 5,0,0,0,0,5 0,0,0,0,0,0,0,3,5,0,^ §,0,0,0, a, g,o>o Z a. 0) o> ^, z Illllrlllllllllllllllll^-llllllll^ H3 O _ : Q_ J* S;feS;SS;oS;5 ;S;S;5;S;mfeS;wa;&;S;55;5; SSm^SmJ?;^ S ^ 3 a OO 3 1 1 rg ooozo ooooooooooooooo o ooooooo to to z o UJ t— o oioooooooo -ooinoo - -ooooooooo o 1 ^ o ^»- «* «-*o o o o OOOOOOOOCDOO O coo a. a. 0 >— a. z in mcai/>< M m coi n 5 1 0 UJ (/I

~~C — •*• 0) ^~~

~~U+JO^"c>»* ~ V -D^ -r-EO*-^1 ^ «+J ^ C €> C ^ l_ l^i-eT3C O^~ O •— O GJO3:T!OE — *—~~

~~O ^™ ^C I/) ^- ^— *~ C V- ttl -*~* a) E^ Q> »— O (E C •*•" T3 3E *^ 4^ +* U 31 ~D ~tJ ^X! 1C '«~ 2 ^— I— ^— O -4* E^ C -g" o >• Li. *^" ^O^—C Z I— ^. L, CI T3 "C~~

O ^ O 1 0 L_ ^ o m i- ^ _^ a: j^ ••» •*->•*-» CTicntT) u OO-ÔOO u i- uu ur: uj ucnvracQ } an a — *jû*J^-»-*-> -^-j*£ CD E — — •— —^ 3 ^>£ 1 U U -4* ~? »^ ^> C c C c +^ E C^ *"^ *""" E 3 £ E E ^ C rr> j^ *J jw cr d E c cz 2 -*^ V o> ^~ O Q) *^ *^" ^~ C fO CZ «*^ i— C O" I— L_ t- "U CI t- i— o C C i— d ••— 3 £ z ^55St55S2i E < S^5£û l 2u!1 S51 ,2££SS^S8 i 1

~~' -'.4A1 1 PGE00047157 I LU~~

~~O Z «c I —i~~

1 z LO «, o a, o o o ^ cv to o co . oinin ^« moo fsirotocNjrnoo "a 2 I * X • — « * - * f—• I z ^ LU cr C_> 3 OZ — O Z O O OOO O o O O O OO OO OOO O CO — LU C3 CN* ro ^r pv rn p«* ^ to oo *~* fn m to o CT* to 0 1— O LO i—i O »* IT) "tf-c-tlO OCMtO f^^Hr-i^HfsJ m TT *—< I ( IOIOOOI O 1 1 1 O O O 1 IOU3OI 1 1 t llOOl 1 1 a. «t z ct < I- ce

I (/> LU O Z LU CC ce *-H i t OJ t— = 13 I cj {_> •—• t— 0 LU Z z (/> o LU I o fe CSI l-H OJ C/1 LU ^ I «t ^~ l/> (£> to to o o> a>tocn tooito toootoo r**, <£> <& m O LU a> en 01 en o> c?*o>a» cr>oa> a»o*oocrj a>a>cn i*_ nr o O t— Of * LU c?* o> 01 o* c*> f^ fi +J rn >• u) ^ to to 10 01 u O)CT>O* 01 m 01 CTI en CT> cr> «-.* 01 o> CM Z a. 3333r*.r^rxu fxoôr^rx.rN.ôshs.rN.un^^^r^rv.ua^ I (V rn I^T"|2S2? 2tâ fS2S^^1< fS2< ??^< f'*f^l^l?< f < f en ^- Z PO^^"«* 10 r-s.^*-r*» ^-r>^r ^•o>ornrs s ^r^r^r o_ .v " oiaio>o>uoucn oo>o*o>fo§*§> oio>cnzj30»a>oio>Sooa>a;aI r cr ^- *™* ^*'^^^ O *-H*-»^5^^^H^H^-t^^*^^H^^»-«»-«^-(ZZ'— tt —«^ H oo 3 ^^ ^J +t | * 4«t > ji,.- j^j j^j ^j f\ ^j ^> ^ ^ _ p- tj ^j j^j to OOOO O !3 O O O&ILJ O O O 13 i3 OUU I (NJ OOOO Z '-lOO O CO O OZZ«-j\--> OOO CO ID Z 0 LU *» —1 t— Lr>Lnknu~)OtNjo I CJ OOOO O O O O O CD O O O OO OOO 0. a. O /N O" s\. iO m in o o /N o ^. m m m in **v m /% >— a. z in in in in LU 3E o I LU (/5 . J cue t/i ftj ~ C»>> — T- ro ;> >^-P-^>» ^^-^-^La^-jJTJ^co M ia H x » 1 1 1 5 ^ ^ 1 » * 1 5 1 1 1 1 £ 1 s 5 1 1 1 ^ 1 11 § g I 5 z ^555ui£ fiSSSSSSiiSS^SS^SSSSSSS

~~i_ j^ o o> o JT u p j^ O <3 -'- I T_ a> a i c c «a 01 —» c* o o~~

LJ(J V>J£.^J£r— ^-*J(JO"DQJO.U(U^ — OOfTJ oc iaioc--uioro^ o--»-»-o>sr>>— cni-i-L_i_i_ — — 10^^: JJ 5'^«S'i~J ~ 1 ^ fe J ° 2 *^ Stixtt^ cni'ffl o o o^^ooS.!^^ I cc l|l||||=S|||l=ti55|||||S||s|||l|

~~c I z rofomcommf^im ^•^r^t'^j'^'^'^^'^'^rinininininininLniriLnioto I -H,\2 PGE00047158 I~~

~~o z -eeeee-e-e-e-e-e-e •o .c *-> I a> i_ o inoo>OCMOcnoooo«-» I Ul a> O .a z o O O O o o o o o CO —• O h— CM CM r-l CM CM I 0. < in o i o in o i i i i i IX rt O O O CO I~~

~~ai cc •« o m o o fn •» in CM co~~

~~o L. I O 8 o o o c/> O LLJ 1 o _J Q.~~

t/3 f t— I l/l o oo i^ r^ 8 O) Ot o> oe a.LU O O)O>O)(D I i Q. 3Cno>cr> toot O CL. u I o I oo o o o> o o t^ § o o CD o o o in o o in o o o o o o o - o o - o o o in *« o m CM in o CM in in o CD "VI V O I o o o o o in Q. o *~* o — o o /** en I o u o> ro D ai — c c u I 0} O nj ' jD c - v- o ^J-»— O 1 ^" 3 ^ £ «TJ ca: MCZ1-T3C ^ ^ m r— o a Ji u t- •>-•>- U) >—••-• +J •— O O —1 —I i^ a> w •»- (O (O IO -t~* u- I C E vt o> * ui 6 ai «> <0 I- 3 3 C E ID J< Ji p- 'e Q> «O O «O o o •*-> to o I 6-14A3 I PGE00047159 , o ^ T" r u ff> ' .<. 2! ^f ^> O

~~«r 9> Cj (T~~

~~Q Z UJ~~

~~-t CQ «/ cr O tr: O~~

~~I- UJ o co UJ < I z m 9Z~~

~~7 •* CO S?3 ISP o Z Z u. O U-l p si < Q CE UJ H- CO £ o?o?~~

~~siCO o Q.~~

~~6-14B~~

PGE00047160 I I Prink et al. (1982) shows average bottom sediment concentrations in the Housatonic River lakes and the Falls Village impoundments • (Table 6.2-9) to be approximately 700 ug/kg, or about 10 times greater than background. As the riverine sections and the Bulls M Bridge impoundment have less than half the PCS concentrations of the lakes, approximately 10 to 20% of the overall concentrations • may be attributed to background sources: tributaries and atmos m pheric fallout, for example.

6.2.3 Fish 1 PCBs have been discussed in terms of the abiotic components of the Housatonic River, but a complete evaluation must consider the bio- • logical elements as well. Of ultimate concern are PCB levels in ™ fish tissue since fish constitute a large and growing proportion of m the human diet (18.7 g/day per capita consumption of fish and I shellfish) (Cordle et al. 1978). Uptake of PCBs by fish primarily follows two pathways. Biomagnification may occur when organisms in | the food chain are exposed to PCB-contaminated water and sediment, and bioconcentration is likely to occur via water passing across • the gills. In the Housatonic River system, both pathways probably function simultaneously. A detailed discussion of the PCB concen- • trations in Housatonic River fish is presented in Section 6.2.3.1 ' and the perspective on backgound levels found in fish is documented in this section. All fish data discussed in this section (6.2.3) I represent wet-weight whole fish composite data unless otherwise stated.

Finlay et al. (1976) summarized data collected by the National Fish I and Wildlife Monitoring Program and Henderson et al. (1971). Upon compiling data collected from 1969 to 1973 and ignoring species differences and geographic location, Finlay et al. (1976) reported I marked declines in both the occurrence of PCBs in fish and the num- _

6-15 I L;i\vlcr Matuskv Ukellv K I PGE00047161 I I TABLE 6.2-9 I PCB CONCENTRATIONS (ug/kg) IN SEDIMENT I OF FOUR HOUSATONIC RIVER IMPOUNDMENTS No. OF ADJUSTED I SAMPLES CONCENTRATION MEAN MEAN3 MEDIAN Falls Village 9 190-1220 700 730 I Bulls Bridge 5 30-230 90 40 Lake Lillinonah (excluding Shepaug Arm)b I Surficial sediments 34 200-3160 1085 985 Core sediments 53 0-2700 533 380 I Combined sediments 87 0-3160 840 740 550 Lake Zoar Surficial sediments 23 10-2200 668 690 Core sediments 40 0-2600 861 700 I Combined sediments 63 0-2600 800 580 690 I Source: Modified from Frink et al.1982. aAdjusted mean (total mass PCBs divided by total mass sediments) repre sents expected PCB concentration if sediments were thoroughly mixed to I eliminate differences in texture, bulk density, and PCB concentrations. bShepaug Arm treated as a separate impoundment due to limited excursion of Housatonic River water into this region and significantly lower PCB levels I compared to main section of Lake Lillinonah. I I I I I I I 6-15A PGE00047162 I I ber of fish containing over 5 ppm PCS during the years 1969 through 1973. However, values vary by at least two orders of magnitude and B no clear trends are evident. Furthermore, PCS concentrations reported prior to 1973 consist of "estimated" values only, and • analytical protocols by which these analyses were made are suspect ™ and may have greatly underestimated actual levels (Finlay et al. _ 1976). As a result, comparisons of these early data with data col- | lected after 1973 may actually underestimate the decreasing trend.

Fish surveys conducted by the Food and Drug Administration (FDA) from 1973 through 1975 also depict a decrease in the incidence of I PCBs in fish but, unlike Finlay et al. (1976), they show an in crease in the fraction of fish containing over 5 ug/g PCBs (EPA • 1978). Again, the reliability of conclusions drawn from these B data is suspect since information was collected from many sources _ with a number of different objectives. |

Of greater consistency and reliability is trend analysis reported • by the National Pesticide Monitoring Program (NPMP) conducted by the U.S. Fish and Wildlife Service, which analyzed residues of • organochlorine chemicals in three to five fish collected from each of approximately 100 stations nationwide from 1970 to 1974. An- • other 620 samples were collected from 109 stations during the two • periods 1976-1977 and 1978-1979, and an additional 315 samples were collected from 107 stations during the period 1980-1981. Documen- | tation of the most recent survey (1982-1984) is currently pending (Schmitt, pers. commun.). •

A summary of geometric mean and maximum wet-weight and lipid-weight • PCB concentrations given in Table 6.2-10 shows a large degree of ™ variability, but overall suggests a gradual nationwide decrease. The same table, however, shows no substantial decrease in the per- I centage of stations from which PCBs were detected. The lack of a

6-16 • Lauler Matu.sky SkHh IHomccrs I PGE00047163 in it C in in Qu. TJ oc m O Cft »—-• CM _J *r ^ ( O O O m *&• CL 00 m in o> E en »— *-« LU ~ o la 3 f-4 >> M «J rtJ J -D 4-J J> (0 O Q. rn cr> y> oo * c ^ •—* to i to rd _ D IX in co . » IV 00 O> y> = * .i UJ T3 LU CM O I —I i- o 3: o> 0 ^ to Q «-> 4J u! I re a> 10 E E= UJ O oo t*i —1 UJ o «a o z Q. rv cd L. U I O J> I*- «.—• •3. _J rv in •f-> ^. 1 1 » C/ ) D oo •-« u» ^ CO OO O1 a. n) ^ LU 7> l^— ^z ^c: O> —J LU re .3 7 •N_^ |— g ° i_ < « t/> c co a CT> o • O oo rv eg ^i^ I— to OL oo en u o in CM II l o — co •* > en *—4 ro !~ (-4 z H J> en ct) LU f^ VI *J a. o «—i CD *—i UJ t— o rv ul TJ _i >— UJ UJ Q) Q a 2 in o v) z CXJ t/1 z ^!D uL. >> 0) i— 3 O (V •-* rv .0 -o a. -D T> 1 1 ^-4 r* OO r* 4-> fO 3 CO rv O i z: o> 00 a> o> t— o «—« o «— hx LU •-« t— O CM r- O) 3 t UJ o •-< — m Z> t/i jC .E to en •M Z3 X 0 S E O JZ §.2. uL U rv en 4-> C3 Q_ in m to Z h- to o t—• oo .-i CO f^ rx CO CO en o> LU LU o> en S O fr> CC >— 0 0 E 0 LU UJ m o —• CL Z£ t-t r~t 03 L. cr <*t en (* h- Q LU z: a> 03? < LU »-l CD * O rv. co CO 3 «J O O* -M (X to r^ o OO *-« ID *~* o t3 rx »— o cn en " g ra r**! f—t 0 t/i cn >— CO CM -*-J f^J «-H L 01 X

~~CM •H -M ^O) 4-J ^— 3 — E O C E — L o ^ .— £: •^ u -*-> 4-> t/i t u u o •* s •• a> nj -— rv Q_ l_ 4-» *-> +J C7> __ -4J C CO -^ 4) U *—< o> m a E u~~

~~6-16A~~

PGE00047164 I I decrease in fish testing positive may be attributable to an im proved analytical method for the 1974 through 1979 data and a de- • crease in the detection limit for the 1980-1961 data. Schmitt et al. (1981) reported positive PCB concentrations (X).05 ppm) in fish • collected from every station located near a municipal or industrial • center in 1974. I Trend analysis shows the geometric mean PCB concentrations for both lipid weight and wet-weight concentrations to have decreased during • the total sampling period (1970 to 1980-1981). Periods of oscilla tion (Table 6.2-10) are more profound during the early to mid- I 1970s, possibly due to changes in sampling and analytical techni que, or differences in fish size, eating habits, lipid content, • age, movement or sampling and measurement error. Maximum lipid- * and wet-weight concentrations of PCBs show a less consistent mm trend. Maximum lipid PCB concentrations declined steadily from | 1974 on, but maximum wet-weight PCB concentrations actually rose until 1978-1979, after which they dropped sharply in the last sam- I pi ing period. Maximum values, however, are more dependent on ran dom sampling of certain fish and are therefore less reliable for I trend analysis.

On a station-by-station basis, total PCB concentrations changed • very little from 1974 to 1979 and it was not until 1980-1981 that _ decreases were noted at stations where fish were previously found B to have high PCB levels (Schmitt et al. 1985). Stations where fish had lower levels of PCBs showed little difference in PCB concentra- I tion, however, and no appreciable loss throughout the entire period (Schmitt et al. 1985). This suggests that, where PCBs are highest • in fish, exposure levels of PCBs are decreasing in response to the * 1977 ban and there exists the greatest potential for depuration mm (excretion) or mortality of larger, older fish. At lower tissue • I

6-17 • Luuhtr Matuskx Skolh r I PGE00047165 I I concentrations the change in PCS source is minimal and there I appears to be more of an equilbrium between uptake and depuration. Over the entire monitoring period (1974 to 1981) mean PCB concen- I tration, wet weight, decreased from 1.20 ug/g to 0.53 ug/g (Schmitt et al. 1985). Even so, the geographic distribution of PCBs re I mained fairly stable, and the percentage of stations where at least one fish was found to contain PCBs was as high as 98.1% as late In I the program as 1978-1979 (Schmitt et al. 1983). This high value shows the ubiquity of PCBs in fish tissue through I out the country. That the percentage of PCB occurrence in fish is greater than the previously reported occurrence in sediments (Sec- I tion 6.2.2) suggests that fish can accumulate detectable quantities of PCBs even when sediment concentrations are below current levels I of detection. At such minimal concentrations biomagnification through the food chain would seem to play a more important role I than direct uptake (bioconcentration), as found for Lake Michigan (Thomann and Connolly 1984).

I Also indicating widespread occurrence of PCBs is work by Veith et al. (1979a) in which multiple composite samples were collected from 1 approximately 45 major watersheds. Results show over 93% of the composites analyzed to contain PCBs. More than half contained I total PCB concentrations greater than 5 ug/g, and only 14% contained less than the FDA action level of 2 ug/g.

I While surveys by Schmitt et al. (1981, 1983, 1985) and Veith et al. (1979) are valuable for depicting trend analysis and widespread I occurrence of PCBs, they monitored sites more likely to have point I sources or to be close to large populations. I I 6-18 I Lawlcr Malusky Skelly Kngim:(;rs PGE00047166 I I Widespread occurrence of PCB concentrations has also been reported for nonagricultural or uninhabitated areas, and atmospheric contri I butions alone have accounted for the regular occurrence of trace quantities approaching or surpassing 0.1 ug/g in fish from pristine • environments (Bruner and Hill 1977; Swain 1978; Haines 1983) or ™ relatively remote, sparsely populated areas (Spagnoli and Skinner • 1977; Schmitt et al. 1981; Winger et al. 1984). R

Swain (1978) sampled fish from Lake Superior and Siskiwit Lake, | Isle Royale. His work at Siskiwit Lake (Table 6.2-11), a remote area, showed a mean value of 1.2 ug/g for total PCBs in lake trout I (Salvelinus namaycush) where no known industrial or domestic source has ever existed. The same species collected from various loca- • tions in the Isle Royale area showed a wide range of PCB levels, with the mean for total PCBs ranging from 0.3 to 1.72 ug/g. The • major factor affecting uptake was whole body lipid content of the I fish being grouped for composites. Since human activity is mini mal in this area, with the island uninhabitated in winter and traf- | fie limited to hikers and backpackers in summer, the source of PCBs was believed to be atmospheric. Examination of fresh snowfall from I Isle Royale confirmed high atmospheric levels of PCBs, indicating the importance of atmospheric precipitation 1n the distribution of • PCBs (Swain 1978). •

Bruner and Hill (1977) analyzed individual and multiple composites I of 17 fish species (Table 6.2-12) from stations in the Big Cypress Swamp area of the Big Cypress National Preserve in Florida, an area • where human impact is considered minimal. They reported PCBs in 13 of the 17 species analyzed in concentrations up to 0.12 ug/g as I representative of background levels for the southern United States. This is indicative of the detectable levels of PCBs found • in undeveloped areas. • I

6-19 I L;j\vlt:r Matusky iikolly Knt>in<:ers I PGE00047167 I TABLE 6.2-11 MEAN PCB CONCENTRATIONS (ug/g WET WEIGHT AND LIPID, WHOLE FISH) I IN LAKE TROUT (SALVELINAS NAHAYCUSH) FROM VARIOUS I REGIONS OF LAKE SUPERIOR AND SISKIWIT LAKE. ISLE ROYALE No. OF No. OF PCB (ug/g) I FISH COMPOSITES % LIPID WET LIPID Lake Superior, Isle 4 3 5.0 0.3 6.9 Royale area: lean3 lake I trout (Salvelinus namaycush) Lake Superior, Isle 11 5 12.5 1.72 15.9a I a Royale area: fat lake trout (Salvelinus I namaycush siscowet) A11 Lake Trout For: I Lake Superior, all 24 16 6.19 0.9 14.54 stationsb I Lake Superior, exclusive 10 8 2.68 0.6 22.39 of Isle Royale area Lake Superior, Isle 15 8 9.71 1.2 12.36 I Royale area

~~I Siskiwit Lake, Isle 2 1 3.5 1.2 34.29 Royale~~

I Modified from Swain 1978. aFish groupings based on body burdens, with fat lake trout having consis I tently higher contaminant levels than lean lake trout. Inconsistency in the number of trout from Lake Superior: all stations I does not equal Isle Royale plus non-Isle Royale. Swain (1978) presents I numbers as listed above. I I I 6-19A I PGE00047168 I I TABLE 6.2-12 RANGE IN MEAN PCB CONCENTRATIONS (ug/g, WET WEIGHT, I WHOLE FISH) IN SPECIES OF FISH FROM BIG CYPRESS SWAMP ^ COLLECTED DURING WINTER AND SPRING 1972 _

RANGE Largemouth bass 0.035a-0.104a I Brook silversides 0.012- 0.022 Spotted sunfish 0.012- 0.023 Florida gar 0.034a-0.110a Golden shiner 0.009a«b I Redear sunfish ND-0.022 Gambusia 0.011- 0.042 Flagf ish 0.024b I Marsh killifish 0.052a»b Snook 0.085a-0.120a Bowfin NDa Yellow bullhead NDa I Bluegill 0.016- 0.026 Warmouth NDa Sailfin molly 0.014 b I Striped mullet NDa a Redfin needlefish ND-0.076 I Modified from Bruner and Hill (1977). Pooled data on fish col- « lected 1n January and April 1972 from eight sampling stations I are summarized by range. ND - Not detected. • a Value derived from single sample. bSpecies sampled during one season only. •

I I I I 6-19B I I PGE00047169 I I In a second baseline study in the southeast conducted by Winger et al. (1984) individual and multiple composite samples of fish spe- I cies and other biota from the sparsely populated Apalachicola drainage system of the Florida panhandle were analyzed for concen I trations of metals, organochlorine insecticides, and PCBs. PCBs were detected in 96°/. of the biological samples tested. Mean con I centrations of total PCBs in upstream male channel catfish (Icta- 1urus punctatus) (0.7 ug/g) and largemouth bass (Micropterus sal moides) (1.40 ug/g) exceeded median levels for 1974 NPMP fish of I the southeastern and Gulf coastal stations and approached (and in the case of largemouth bass surpassed) the national geometric mean I of 0.95 ug/g for 1974 (Table 6.2-13). These elevated levels of PCB concentrations were concluded to have entered the Apalachicola 1 drainage system from tributary drainage systems receiving point and I nonpoint source contaminants from metropolitan areas to the north. Sampling in uninhabitated areas of the northeastern United States, Haines (1983) reported PCB concentrations approaching 0.1 ppm in I samples from a single two-year-old brook trout (Salvelinus fonti- nalis) collected from one of six remote (forested) lakes in Maine, I New Hampshire, and Vermont. The singular occurrence of PCBs in I only one lake of the study is unexplained. Analysis of fishes from 41 stations throughout New York State (Spagnoli and Skinner 1977) showed PCB concentrations present in B edible flesh of nearly all fish examined. Sites sampled included rural (sparsely populated and nonagricultural) areas of the Black I River, St. Lawrence River, Genesee River, and the watersheds of the Seneca, Oneida, and Oswego rivers (Table 6.2-14). Most sites sam- I pled were nonindustrial, although some lightly industrialized areas I were included. I 6-20 I Lawlcr Matusky

~~I PGE00047170 I~~

~~(~^ f^l~~

~~to •C^ *—1 f —. LU • * uo 1 f7^ •>p™^" c^ I »—t 1 CD 1 '-^ I 4_> ~5*~ CD ro CD O [ LU CM O 1-1 1-1~~

~~CD O CD O r = a I 0 0 1 o ro~~

~~_ CO I-H CD ;j LU 1 .—. 1 .—• r— —1 CD CD O LO IB Tf (> CM CM I «=c j 00 O 0~~

^^ CO 0 0 ~ I o —i 0 »i»4 ^0 0^ t— 3: j. 1 *""** 1 «^~N < 1- «. to o o o o o: LU i C3 CM ro CM ro ri^ LXi,_ J—— »— Z O 1— LL. CD O O 0 III _ I o to u- LU Z LU O 2 0 0 O _J ' CO 0 '-I O (X CO \ LL • • 5 1 ,—. 1 , . CO < 3C CJ I o to o co ^£ 00 00 0 < 1^ LU CM LO LO CO LU LU CT> U_ • • » • u- K- o: •-« CD CD CD CD O •-! *—' « ' co o: x—. O LU

i t—I Z LL. *n 0 CM Z tO < t/ LO I > co LU •-• a; LU • CO OC ti **v Q£ j _j 0 0 «i LU LL •—< LU Q- LU Q- > > z CD CD • CO _J _I CL, •-. : LU m n o o m z o ;D cat: . • i • ; —i I < —• i: i^ CD CD O CD e ~— 3 z <: ui •—_ s * •?* >^* 0 _l Z Z O£ O t/ t < . u. o 3 CD CD LU i—i CD CM IE "•"-• ^Z ^ c J CO • • I t— CO O 4J LU CM O CJ Z >-i < 1 _) 1 ^"* ) •*"•* »—1 L3 1 1 5 -" -« t/ 1 O LO » • (/ I .LU <: co CM O • Q end. a3 03 1 •—. 1 -—. ^> O O CD IO CO a z •^ to CJ

~~TJ C at o~~

~~«_ 1- IJI 0) OJ •i i i O. 2 T3 Q 0 O CD Z Z i 6-20A i PGE00047171 CM CM CO CO CD "^ CO T LO IT) I O or • CM • O --H O O —' k— O CD CD —H I Oi l •—1 O u_ o~~

CO I LU CD I—1 o o o o r^* oo r^. cr> n r*^ CJ i LO UO <-l O «*• in rv. • CM •—< LU CJ » • • • or .. . in > • O. •—1 10 LO CM rt f— O CD O f^x CM t—1 O O •—I CO Lf> LO ^ "S" I or LU O 1— t— CO CO LU CM I 1 XX z co in o co CT> CO O> ^ CO CO 0. or c^ CO Tf CM l-v •* LO CD . •-( ro X O LU • » • • or ... co <^ >— z: it- n .-H o 1— O OO COCM OO O CO CM CO I LU LU i—i CO U_ 0 0 sQ:- co O Q I O LU E ^ LU CO = OJ _i or ^ 0 x: r— o O. LU E =i "o £1. i/> in cn t— i—i »— n in o> •• in 3 3 •r— ~n 1— < QJ 3 -r Q 3 a o T O I _l i t- in E QJ o c >• ID •MO O — >• QJ ra z o •r- p— r— O re •— t- E o "O . LU > 3 O «- E L ra ra -p QJ ra U_ fvl XI T3 CJ re QJ x: c ca o •— c:oi n} c c re o i_ -U -4-> _j o c toE TD -U T3 m ra •• c I in L. 4J 3 3 LJ ra QJ ar «-i -a in t_ ra 3 -i— 3 OJ C CO or a> 3 ai E c ra t_ m T m ex : o LU t— 4-> L_ -*-> O •r D1 U oj •— 3 E m a. CM —I CO 10 3 O -»-> •— 01 C U. =3 O i— o m QJ o o ra > •• , », y) I Q M ra i- O > E E in •— t_ O -r- LU "Z. •i- •*-• U X p— i— •— o ra a. U t- 1 P—4 -4-> O -P O 1 re ra re p^ LO > O >] CO co i™" ^" y CO CO CO «f C_3 p-1 •> •r~ (-* . » Q _J 4-J in LU I— CO QJ 4-> cr> 3 I 3: LU in ra c • T3 ^^ CJ L_ ^ JC QJ r— C p—* i—i O X) in o -4~> •— ra 'i~ LU TJ m jc CJ -^-. ' XX CJ 3 3 3 ra o a. ra z o JC t_ O O -4-> O -t-J -C >> CJ 1 I O •— x: 4-> C_ 4~> L ^_J 33 4_J t ^ jj -p-> Ol -t O >-i Q p— •»-> T3 34->34->3O34> O3 era ,— LU LU O O O L. o t_ o L -i— 2:M Ol XI O O) 1-5 1-2 l_4-> L.20J 4-> 1- QJ TD QJ >i r—t >> EX: 4-JO4->O-4-> 4->Ol4— U 5 t= i- xi 1— —1 OJ C r— -t-> XI XI C XI T- OJ O QJ QJ LU S QJ cr 5 j^ QJ o. p— Q ^ VI I J£ •«— JC -i— JC O JC *r— QJ IO ^^ t ^ »—* .— o re c o QJ ca ^v re c_ E ~ rere rere rac_ rare p^ i ra ra r^ QJ T3 i- cj "t— 3: oo co 10 i or i or i co —j or «^ i cj r^*. 3 QJ n cnco CT> •— TZJ -*-* *^^ i r-1 ra QJ ra 14- O1O «» > s_ s o I 33 Z cj ra 1- QJ •^-^ p—i QJ C OJ Qj O <4 ji: o cn c c; 14- >r- QJ CO O L ra XJ --3 e— -p— f~ •,— Z Z >- QJ OJ _J 4-> <4 •r 1_ -O . J_> *J o -Z > QJ JC CJ QJ t_ re OJ jx; o cn •• >-• s: L_ -r— ^~ jc ra ra ajojcuzo r co jr; c e 4-> I h- O QJ or z re —i 3 QJ j^ -vc > ra in -p o cr <: or or > c _j cnj c rera-r-raa1 j •O -*-> QJ -»-J i— rd or u. -U •1— ^ OJ •• a> T ra i lorE * cz ra •— •— 14- 3 i— ;> or o -u zt j^ u ra _i 3 -r ra T3 H— *n cr z m 1—1 •4-J 4-J i— o -r -a rareraN_i o jrr LU CO or J£ 4 O) O o T3 cr ra cjcncjQ J •^ >» oj d in ^ CJ p—* o QJ 3 it •— ra ra -^ QJ 3 QJ 4-> "o I •z. u. re •*-* O" L_ E cr cr 3 c>>c:c:.— o x> o <4 ra o p^ re re o QJ ra ra QJ ojraojo4 ' " cr QJ ••— cj E CJ co s orz re cj cj st COCJCOZ3 cn u T3 (-m ra cr QJ o o CO CL QJ j_i u_ i»_ CJ •. » co m 4— o o_ re m re in XJ o 1- in I I_ T3 L. -01 E ra o E QJ \r> O) OJ -i— QJ T QJ o c ••— ra in \^* i- u > OJ > QJ > u in o i i— 3 CT) o> cr •t c: -r C •• >4— QJ ••— O_ Cn t—i t > cu or o or o or -M -4-J o in LU •r— L. T3 ra ro T3 4-> 3 31 or ^ oj • o « o QJ o s_ QJ ra o I CO oj ra cn re CD •i— -i— ra d EE -i— or -*: _J QJ in cj QJ o QJ <4— -D Q. cr o > LU CJ > OJ OJ 5 O ^ •i— CT QJ •'— t— QJ t— re • ~f~ cr cm c m -0 i— s jc -cr L o; OJ OJ O OJ O O (X CO CJ Q. I 3 co to O CO CO s: i ra -Q u -a I 6-20B PGE00047172 I I Although not directly comparable with previously cited studies that report PCB concentrations in whole body samples, concentra- I tions reported by Spagnoli and Skinner (1977) appear high. This is particularly true in light of the lower lipid content of edible I fish portions as compared to other tissues of the body. Whole body would therefore be expected to contain higher PCB concentrations • than levels in edible portions (Schmitt et al. 1981). Neverthe- • less, Spagnoli and Skinner (1977) reported that averages for total _ PCBs in edible flesh were approximately 4 to 6 ug/g in areas of | local industry and approximately 1 to 2 ug/g in some nonindustrial areas with no known PCB sources. They note that detectable amounts • may result from atmospheric fallout, accident spills, or landfill leachate. The values reported in this study were generally similar I to the range of those reported in fish from the Housatonic (Section 6.2.3.1) and again demonstrate the tendency of fish to accumulate • substantial PCB concentrations (>1.0 ug/g) in areas that are for H the most part undeveloped or rural and for which no known or detec table sources of PCBs exist. I

With the exception of the study of Spagnoli and Skinner (1977), the I previously cited studies measure PCB levels in whole fish samples and are not readily comparable to fish measurements made in the • Housatonic, which are reported for edible portions (fillets) only. Even if edible portions were measured, comparisons from one site to • another based on fish data are problematic, and comparisons of fish I concentration levels from different sites are difficult to inter pret. This is due to differences in eating habits, lipid content, | age, sex, size, and movement, all of which affect PCB uptake. As such, it is difficult to determine the exact percentage of bocy I burden in Housatonic River fish attributable to background (non-GE) sources. Nevertheless, background sources in drainage areas with • geographic and land use characteristics similar to those of the ™ I

~~6-21 hauler Matusky Skollv Kno-inccrs I PGE00047173 I~~

I study area account for a non-negligible PCB concentration and I assist in the determination of endpoint fish concentrations. While not directly comparable to the Housatonic fish data, the I above-mentioned studies show that PCBs in fish tissue are ubiqui tous in this country, particularly in the east. Concentrations I approaching 2 ug/g in lake trout, which are relatively efficient bioaccumulators, have been cited even from pristine areas (Swain 1978). Higher levels are cited in fish sampled from rivers and I waterways receiving industrial contributions (Hesse 1975; Veith 1975; Veith et al. 1979a; Winger et al. 1984). According to I Schmitt (pers. commun.), fish taken from any major river in the United States with at least one urban source is expected to contain I from 0.2 to 0.5 ug/g PCBs. Concentrations in fish from the heavy industrialized areas of the northeast and the Great Lakes are ex I pected to be much higher. 6.2.3.1 PCB Trends in Housatonic River Fishes. As part of an on I going effort to quantify trends in PCBs in Housatonic River fishes over time and space, studies were conducted during 1977, 1979, I 1983, 1984, and 1986. The details of each study are addressed in some detail below, since differences in collection locations, meth I ods, and analytical procedures must be considered in the interpre- I tation of observed temporal and spatial trends. 1977 sampling program. The 1977 study was conducted by the State of Connecticut. Sixteen trout were analyzed from the Cornwall sec I tion of the Housatonic River while 25 "warmwater fishes" were analyzed from Lakes Lillinonah and Zoar. Only the edible portions of I the fish, with the skin left on, were analyzed for total PCBs. Because of the small sample size and limited spacial coverage, I these values will be discussed only in a very qualitative manner. I I 6-22 I Liiwlor Matu.sky " Skolly Kinjinoers PGE00047174 I I 1979 sampling program. The 1979 study was also conducted by the State of Connecticut. While fish were collected from June through • August, most were taken during June. Approximately equal numbers of "warm water" fish specimens were analyzed from Bulls Bridge • (91), Lake Housatonic (104), Lake Lillinonah (100), and Lake Zoar • (103). Thirteen species were analyzed: American eel, black • crappie, brown bullhead, carp, chain pickerel, largemouth bass, • Lepomis spp., smallmouth bass, white catfish, white crappie, white perch, white sucker, and yellow perch. Thirty-nine trout (20 rain | bow, 19 brown) were collected from the Cornwall section of the Housatonic. •

Specimens were collected using a variety of techniques: angling, • gill net, seine, dip net, and electroshocking. Specimens were • either scaled and filleted or skinned and filleted before being shipped to the Connecticut Department of Health for analysis. I Results were reported as total PCB (based on Aroclor 1260).

1983 sampling program. During August 1983 fishes were collected from Bulls Bridge (40), Lake Lillinonah (44), and Lake Zoar (41). • A total of 17 species - American eel, black crappie, bluegill, brown bullhead, brown trout, chain pickerel, largemouth bass, • pumpkinseed, rainbow trout, redbreast sunfish, rock bass, small- • mouth bass, white catfish, white crappie, white perch, yellow bull- — head, and yellow perch - were analyzed by IT Analytical Services | (ITAS), a Stewart Laboratories Division, for PCBs. The method of PCB analyses was different than that used subsequently, and this is I considered in later analyses. Results were reported as Aroclor 1242 and/or 1016, 1254, 1260, and total Aroclors. •

~~1984 sampling program. In 1984 the Academy of Natural Sciences of Philadelphia (ANSP) began a program that was to be conducted in I 1984, 1986, and 1988. This program focused on (1) specimens and _~~

~~6-23 I Ljuvlfjr Matusk !>k~~

PGE00047175 I I sizes likely to be caught by sport fishermen, (2) species likely to have elevated tissue concentrations (based on other studies in the I Housatonic River and elsewhere), (3) species with diverse feeding habits, physiologies, lipid levels, and potential pathways of up- I take, and (4) sampling sites with significant sports fishing and/or I previous indications of unusual PCB concentrations. Sampling during 1984 was conducted at Bulls Bridge (76 specimens), I Cornwall (52), Lake Lillinonah (79), and Lake Zoar (71). Brown trout and smallmouth bass were collected from the Cornwall section by electroshocking during July and October. Bulls Bridge and Lake I Zoar were sampled during August and October while Lake Lillinonah was sampled during May through August. A wide variety of collec I tion methods, including electroshocking, gill nets, hoop nets, traps, and trot lines, were used at these locations. Species used I for PCB analysis from these three locations were brown bullhead, I smallmouth bass, white catfish, white perch, and yellow perch. Only the edible portions of specimens were analyzed for PCBs. Bass and perchllke fishes were scaled and filleted; trout were filleted I but not scaled. Catfish and bullheads were skinned and filleted. I ANSP used a gas chromatograph method to obtain total Aroclors. Aroclors 1254 and 1260 were separated by capillary column chroma I tography followed by electron capture detection. I 1986 sampling program. ANSP used generally the same sampling and analytical techniques in 1986 as in 1984. A total of 152 specimens were examined from Bulls Bridge (43), Cornwall (37), Lake Lilli- I nonah (56), and Lake Zoar (16). Bulls Bridge was sampled only in August; all but two of the 43 specimens from Cornwall were also I collected in August. Lake Zoar was sampled in August and October; I Lake Lillinonah was sampled in June, August, aid October. I 6-24 I La\vlor Matusky okellx Kno moors PGE00047176 I I Results and Discussion. The inconsistencies in sampling locations, sampling dates, species composition, sampling methods, and analyti- • cal techniques make unbiased estimates of temporal and spatial trends difficult to obtain. To understand how these factors may • influence PCB levels, one must examine the underlying factors regu- * lating PCB levels in fish. •

It is thought that PCB levels in fishes are influenced by a variety of factors; however, most are related to a fish's metabolic pro- | cesses and the duration and level of exposure. Norstrom et al. (1976) and Jensen et al. (1982) provide a more detailed explanation • of these processes. In general, the change in body burden of PCBs is equal to the rate of uptake from food plus the rate of uptake • from water minus the rate of clearance from the body. The rate of intake from food is dependent on the efficiency of assimilation, • which varies with season (temperature) and size of ration (Warren I and Davis 1967); the rate of food consumption, which increases with body size; and the concentration of PCBs in the food, which varies • with such factors as distance from source.

The rate of intake from the water is dependent on the efficiency of assimilation, the concentration of PCBs in the water, and the rate • of water flowing past the gills. This ventilation rate is related ™ to gill surface area and activity. Additionally, Price (1931), us- « ing smal imouth bass, demonstrated that the gill area increases with | body weight and that young fish have proportionally greater gill areas than larger fish. I

The rate of clearance from the body has been demonstrated empiri- I cally to be proportional to the body burden and the weight of the fish, with the clearance rate decreasing with increasing body weight. Niimi and Oliver (1983), among others, have demonstrated I

~~that elimination rates are related to lipid levels and to the type —~~

6-25 I La\vli;r Maiusky SkHly Kni>iru:(;r.s I PGE00047177 I I of PCB congener. Depuration rates are greater for congeners with lower chlorine content, with no chlorine substitutions in the ortho I positions, and those with two unsubstituted carbons that are adja cent on the biphenyl. Since PCBs are highly lipophilic, individ I uals and species with higher lipid content will tend to have higher PCB levels. Lipid levels may be influenced by age (length and/or I weight), sex, and reproductive condition (time of year). PCB levels may also be reduced through spawning since lipids are accu I mulated in eggs. As evidenced from the above discussion, many of the metabolic fac- I tors are closely related to size (age, length, or weight) of the individual. The interrelationship among these factors can be I demonstrated from the withln-group (data segregated by species) correlations among age, length, weight, lipid, and total PCB con I centration given in Table 6.2-15. All comparisons, except age and Hpid, Indicate highly significant positive correlations. The age and lipid comparison was also positive, but not significantly dif I ferent from zero correlation. In addition, the matrix of correlation coefficients among groups (data not segregated by species) I shows similar correlations except for lipid, which has a significant relationship to PCB concentration (Table 6.2-15). The PCB I lipid partitioning that has been found in many cases cited in the literature is also apparent for the Housatonic River when all fish I are analyzed together. The consequence of these interrelations is that any difference in I mean age, length, weight, or lipid content, whether arising from changes in sampling season, sampling location, collection gear, or I even by chance, will bias a comparison of PCBs from other times or locations. To minimize this bias, analysis of covariance (ANCOVA) I is typically used for statistical tests. This procedure adjusts I the mean PCB levels to a common standard before comparisons are I 6-26 I Lawlcr Matusky Skellv K PGE00047178 I I I

rvi n- •o • UJ in .-H d) ^* 0 0 l/l (3 * * « o ce >—t •<** • to* •—* i in-K ,.^ ^1 Z LU LU rx Co cr> c> v. ,__ ^^ O a. 3 to c\j to •—i 10 aj 1 1 0 c C3 CD CD CD " -* CO Z ^_ CJ < Q. I_

• O • ^~ (/} i * * * * *•>! -k l^ CL •• C • 'I—< *t in oo ro n c\j co to '— U. t— h- i^ to f^ r^^ in co r— OJ UJ O C3 cr> to •—i en to o in Q£ C3 O h- Z *—t o "=c uj LU o a o CD CD CD i U. 1— C _l CM -o • Z 1— LU o *• 10 o T: _i j-~ •—' C3 LU *• l/> ^M UJ H— •—* (/O j_> .^ ^1 1 • c -K * * 4~> LJ 4—> DC *f. 1 * * +C -K ***•(< rcf O 1— O ^C r^^ CM Co c\l r^ r^^ •—i o~t d CU "•• o o >-« — inr^oco ^rcvjom t— 0- O CO ^^ ^^ if) f\i ^^ ^j ^j ^j- O ••- 1 _ ^_J rs c _i 0. OCDOO OOOCD J2 5 >_ o •— C >r o; r— r. »• fy o TD uo • i 31 o (— CU ro '1 Z I- 03 tO -S" d CQ c CQ cr • • -Q E Oft t—< c ;D cj oj a. o QJ in *-« oo CB a: •— O OL. •— ^J O. i— t— o; •*-> o -»-> CD O " ^ >—< C3 *— '^ ^T Oi •— f~~ -C. OO CD O '— .C OT ^r \ (O <^3 CT) ^1 CD *O *O Ol ^3 ro s^ | ^| i— oo i\ Z 4-^ -*-» -r— T— | 4.J 4^ *r— •—i O O (L> Q.CDOOOJ Q. a> i*- -M 3: -W 4J » -r-Z-4->4-> 5 .1- ii n n => e • 1— -«c:c:c:cji Scctzc n CZ * -K CO oo • * ra ™I I I 6-26A I PGE00047179 I I made. This common standard could be any one or more of the variables Influencing PCB levels, such as length, weight, age and/or I lipid content. I Temporal-spatial trends. ANSP (1985), in their analysis of the 1984 sampling program data, used ANCOVA to investigate the effects I of station, lipid content, age, sex, and season on PCB levels in Housatonic River fishes. They found no single factor or set of I factors to be significant for all, or even most, species; a differ ent set of factors was seen for each. For brown trout only the period of time in the river was found to be significant. Since I these are generally hatchery fish, annulus age minus the age at stocking is more appropriate than is annulus age as a measure of I exposure duration. For smallmouth bass concentrations at Bulls Bridge and Cornwall were significantly greater than those at Lakes I Lillinonah and Zoar. No other effects were found to be significant for pooled smallmouth bass data. White perch displayed a highly I significant station difference; no other factors were found to be significant. White catfish was found to have a highly significant age effect, with highest levels in the oldest individuals. Station I and lipid effects were not significant. Seasonal effects could not be tested owing to the limited time period during which individuals I were captured. I Since largemouth bass were captured only at Bulls Bridge, station effects could not be tested. No seasonal effect was noted; how ever, weakly significant age and lipid effects were noted. Yellow I perch indicated significant station, age, and station-age interaction effects. Using only Bulls Bridge specimens, significant I age, sex, and lipid effects were also observed. Brown bullhead in dicated a significantly higher PCB level from Lake Lillinonah than I from Bulls Bridge and Lake Zoar. I I 6-27 I Lawlor Malusky Jjkelly Engineers PGE00047180 I I AN5P (1985) noted that the most prominent pattern was the differ ence among stations with a decrease in PCB concentrations In a • downstream direction. Additionally, they noted that concentrations increased with age (or time in the river) for several species. • Lipid and sex effects, observed in some cases, were generally not ™ strong. They could not reach firm conclusions with respect to the • variation among species, i.e., ecological, physiological, and/or I trophic differences. Since only 1984 data were analyzed, no state ments regarding changes in PCB levels over time could be made. I

By pooling the 1979-1986 data sets and using ANCOVA to compute I least-squares means, IMS is able to obtain estimates of temporal and spatial effects for a number of species. Total specimen length (TL) was chosen as the sole covariate because (1) it is consis- I tently available for all collection years and for every specimen, _ and (2) would be easily and cheaply obtained during any future | monitoring program. Lipid values were not available for the 1979 samples. While age demonstrates a slightly higher correlation to I PCB concentrations, it requires considerably more skill and effort to obtain. Season and sex effects were ignored since ANSP and our I graphical displays (Appendix A) indicated that any effect caused by these factors was weak. Age and weight effects were not included • because of their strong correlation with TL. Least-squares means • were used for comparisons since they represent the PCB concentra- _ tion after adjustment to a common size, i.e., the mean TL. Both TL | and PCB concentrations were natural log (In) transformed to better meet the assumptions of normality, additivity, and homogeneity of • variances. Although for hatchery raised brown trout the time in the river (RIVER AGE), i.e., annulus age minus stocking age, would I be a more appropriate covariate due to its higher correlation to PCB concentration, substitution of total length (TL) made only a • minor difference. RIVER AGE data are limited to brown trout in • 1984 and 1986 at Cornwall. These data adjusted to an average RIVER I 6-28 I Lnwlor Mntusky Sknllv Kntjincc-rs I PGE00047181 I

I AGE yielded PCB concentration estimates of 2.70 and 4.40 ppm for 1984 and 1986, respectively, while adjustment to an average TL I yielded 2.56 and 4.75 ppm, respectively. Hence the temporal trend based on length adjustment is similar to that based on RIVER AGE I adjustment. I In general, ANCOVA results indicate highly significant year and location effects (Table 6.2-16). Of 12 tests for year effects, 10 were found to be significant, and of 13 tests for location effects, I 10 were found to be significant. Four of nine tests indicated significant year-location interaction effects, i.e., that the change I in year-to-year PCB levels is different at one or more of the river I locations. Graphical presentation (Figures 6.2-2 through 6.2-16) of the least- I squares geometric mean PCB levels over time and river location demonstrate a fairly consistent temporal-spatial trend of decreas ing concentrations downriver and a generally decreasing concentra I tion over time. This latter trend appears to be brought about primarily by a large decrease in PCB levels from 1979 through 1983; I 1983-1986 levels appear to have remained relatively constant or have increased slightly. It should also be noted that the 1977 I levels, while not directly comparable, tend to be very similar to the 1979 values. For trout from Cornwall during 1977, values I ranged from 4.6 to 43.0 ppm wet weight, and averaged 17.9 (5=10.8, n=16), while wanmwater fish from Lakes Lillinonah and Zoar ranged I from 0.3 to 26.0 ppm and averaged 5.3 (S=5.3, n=l6). The least squares geometric mean (GM) PCB concentration for brown I trout and rainbow trout (with skin) for the Cornwall section during 1979 averaged 14.2 and 15.8 ppm, respectively. (Note: the GM I values given in Table 6.2-16 represent average values of all obser- I vations. Because 20 of the 21 rainbow trout were taken in 1979, I 6-29 I Lawler, Matusky okelly Kngin<:«!rs PGE00047182 I I p JJ O r*sO»inrnioo*o^cooo*j »-to^'cn CO c a. cniotomcMOOcninmi-iincncnoto rv oo-^mcncMootooi-icnivcMi-i o — cn I — —i Z X • inoooocntooocnrvtocncn^cnmin CM U JJO o>«rcMCOfvcnrvcncnocNjcororvco to lls inoi-tcnr-iooo^Oi-icMinr-io O j ""en I n X coiv^rini-icoioinoO'ncocncnioiv rv 10 oin«inmiOoo-»o.-''»-O'*rco -i CM to § < coo^*^-cM*-*oo*rOi-'cniOi-<^ o o r-t I to co cn o i * cn cn* *o o o cMto «cMin o o *CM tv •i cn »• coco o» «a- •! cn o CM CM cn cn i—i in cn O I oo •* CMIO o cn o O 1 i i i I p i 1 X oo o oo oo o o t- EC o «c LLJ JJ U- >- U- I LU * * «« *«« «« « « «« «« « «« * tooi-tino i-tii cni-i^* CM~* * CMOOOOIV O O COO O OC^ o *roo*r*r oo ooo oo o i-iooiorv oo ooo oo 0 o i . . i ... i . 1 «J —J I oooo o o o ooo o o o o «« « «« «« « «*«« ** ** « a toin^H.—!•—»m»-ir- i «-i(t f—i inoooivivo o oo o o to ce < cnOOOOOCOOO OO O o 0.JJ oooocntMo o oo o o I II 1 1 I ^ or oooooooo oo o CD o « « « « « * « i O>cn*-ii-<<-4toocMoomin<--icnoocncnio^rocOLnrn oo f L: ivcntooooiOtOfntOi-tomino CO CM I z •-icsjooc^^rocoocnootoivo ooooooooooooooo o I

~~^fvmivcnoi-iiocncMcn^-cn^t-i> n in to cn^HivivcM^ CMCOi^cncDrn»-iCMCn O "»• 0- mcMCM.csiincncn»-imi-«CMcncncMcn CM CM I j 5 -CM^^'cn^j'i— i^^cMcninmcM*n*»' " :j I j~~

01 0 LU fe 1 cc CCMCM^r^cMCMmfncMi-i^^rcM^ •-* ^ .r LJ I o O E i. c: E -ttvi-t^-cMOom^^t-^cocMini-iinio 11 O ro • — j^H^-rvoo^i-'COivcMi-icooO'-iom «< o ~ — U ^H •-> »—4 E — > I i. C71 CJ n U — U t/i 1/1 T3 T3 •— vt K> CD cute cota +* (oj=a> GJ ,— -* tu i-ja.^ jaw^— v- CUQ.JZ-*-* CU Q.O ^-Q.-CC U •— U in ^ 0)O.i— 3 .X.C:(JLI- jr«*-D_u^£ *— L_ O CD o • c I 3 a> mot ) C L. *3 L. — 3 «3ral-l» 3 f") Q, o o IO U ^ +J CX. O ""I 3C (O O U U Q. f) CLd.' — -i: C C CO*E.O f— CD CD CD « g I II II E <— i_i-io£=40Qj1i{)OE ft) •< co ca cn o o i i| cc ce ixi .z 3c iz !I * *« I I G-29A I PGE00047183 I I I I

~~I QO en I m CO o» I W > I-H -00 I 05 en CM I e\j 0) I<0 0) UJ c 00 L_ O en D § O (U g 'l_ I 0> E "• E CN 03~~

~~I O -oo I ffi O) o ffi O I en Q.~~

~~en C I -h- (0 en x: I 1C I I s.QOd I 6-29B PGE00047184 CO 0) O~~

~~x » c c o -C to .2 m aO.~~

~~o H >, a c a~~

~~s,aod~~

~~PGE00047185 I I I I I I CO I 03 en~~

~~•CD I PS cn~~

~~D CJ U 0) I . . JZ ID D CO o tc. JO CD~~

~~a c I 1-00 c h-i cn MH o £ a I a> m o a.~~

~~I •D o~~

~~a c I a~~

~~4) I £1 3 I C~~

~~O I z I 6-29D PGE00047186 V I en -g 'k_ _c CD D u C c: o o I "5 c m o o N I~~

~~- 00 I I~~

~~00 en I uJ oo I CT1 I CD = 1 00 (Jl o Hi V (C o I I~~

~~(N 00 ui en I D o I 03 CTi I m O o CO Q. cn I~~

~~0 N _>. a en c I o I o I s.GOd |D; I I 6-29L I PGE00047187 I I I I~~

~~CO l-oo I cn I oo 01 I w~~

~~HH 00 I cn PS~~

~~CO i ICM ro CD •00 cn III o Icr o O H CM •oo I cn CO a £1~~

I o oo a cn 10 I £ O I Oo. I I I I I s.GOd I 6-29F PGE00047188 to r-OO K O) W ^>^ HH r-CD CD i^ K w r i H 2 ^ » UJ t—J_H^ -u^ -00 oc Z "5- cn o D ^—< V g 0 | s CM \-00 5^ • cn CO a t> £> o a J

~~6-29G~~

~~PGE00047189 I I I I~~

~~to I l-oo I m K l-oo I o> I -~~

~~00 IO4 <£> •oo 01 O) ^ O I o E en E-i CN D •oo I O) CD CD U> .O I O x c ffi c o c. I « co O I o. •o I o~~

~~I CD CM a a I t> H I z I 6-29H PGE00047190 CO ro en~~

~~K en W~~

~~00 en 05 I C\J CO CO 01 oo cr. en D D (D (5~~

~~-oo CO en D O -oo tn en~~

~~CO o 0.~~

~~s o ,aod z~~

~~6-291~~

~~PGE00047191 1 1 1 1~~

~~CD 1 hoo 1 m hoo K en 1 W~~

~~HH oo 1 PH en o ^^^^ I o 7) .-» 1 » (n 1~~

~~a en c 1 ^ a en~~

~~1 CD 1~~

~~O 1 z 1 6-29J PGE00047192 W~~

~~CO to o -O CJ CO o 0) § O J o E CO D o a~~

~~£01~~

~~s.BOd~~

~~6-29K~~

~~PGE00047193 I I I I I I in oo I 01~~

~~oo I 01 CVJ I ". o « • (»D HH 33 UJ o •oo o O) O gor <^H " 0) I i O £~~

~~CM 00 I CO o I Ko u o w I JO Co D I a.~~

~~c I a~~

~~o I o f I c o I z I C-29L PGE00047194 -oo K en W~~

~~-co en CO 7 0) CM 'a m -oo o en v § O~~

~~-00 cn eg CD o~~

~~o c -oo c en B o~~

~~ffi o 0.~~

~~•O C9 N 1^5v c a~~

~~G-P9M~~

~~PGE00047195 I I I I I I m oo I w~~

~~h-H I K -oo I u o (U -DO D. cn D I o 0) H -oo cn I CO a o~~

~~X O I O •00 I M en I m e N _> I cCO~~

~~o I CD o « I I I o I 6-29U PGE00047196 I I I I~~

~~m o CL~~

~~T3 0 _ <0 c a I~~

~~CD I oe I~~

~~s,BOd o I 6-290 I~~

~~PGE00047197 , 10 h°o K 01~~

~~oo O) (O~~

~~Y— e\j CJ e CO i> Q. to LU oo DC en 3 g O u. H CM co en~~

~~O C~~

O a .c*

~~ffl O a.~~

~~•o c~~

~~a c~~

~~o « Zl E 3 C O £ I-~~

~~a z~~

~~6-29P~~

~~PGE00047198 I I the GM average is heavily weighted toward the higher PCB levels that were present in previous years.) tt~~

Although least-squares means could be obtained for 15 species, only the following seven species were collected over a broad enough tem- I poral-spatial region to derive quantitative estimates of trends: brown bullhead, largemouth bass, Lepomis spp., smallmouth bass, J white catfish, white perch, and yellow perch. Even for these spe cies care must be exercised in forming generalizations regarding • temporal-spatial trends since not all species were collected from all locations from each year. It should be emphasized that the • following concentrations are based on least-squares means and * therefore refer only to the average-size individual. Smaller in- « dividuals will have lower concentrations; larger individuals, m higher concentrations. Estimates of the rate of change in PCB con centration over time were estimated using ordinary least-squares | multiple regression. The dependent variable was the least-squares In PCB concentration. The independent variables, year and river • kilometer, were taken as years from 1979 and distance (km) down river from Cornwall Bridge, respectively. For the latter, loca- • tions for Cornwall, Bulls Bridge, Lake Lillinonah, Lake Zoar, and ™ Lake Housatonic were taken as 0, 18.6, 45.3, 57.5, and 68.9 krr, « respectively. Regression results are summarized in Table 6.2-17. • Brown bullhead. A total of 74 brown bullhead were analyzed from • the years 1979, 1983, 1984, and 1986 with a geometric mean size of 273 mm TL. The highest geometric mean PCB concentration, 3.99 ppm, I was observed at Lake Lillinonah during 1979; the second highest, 2.44 ppm, at Bulls Bridge during the same year. The lowest concen • tration was 0.42 ppm from Lake Zoar during 1984. Concentrations • generally decrease in a downriver direction and in more recent _ years. The greatest decrease was from 1979 to 1983. During 1933 | the highest concentration, 1.22 ppm, was at Bulls Bridge. From

~~6-30 I La\vlor Matusk\ (>ki>ll\ Lnoimicrs I PGE00047199 I OJ z t~^ to to CO **- CM CO M~~

U DC •—I CM «—i CM •«!• CO CM 1 1 1 1 1 1 1 I LLJ QJ t— C? o I "0. CL JJ < X IO QJ i: c_j a: > i ^-f ro ^j- pv CM in cn ^3- to iJ V^. Q_ LU ^f LO i—' cn to in co ro •~H | 1 I 1 i i i i o -a >i O Ct: ro TT rv. TC cn —i iv. cr cr < o OJ O I 1- LU in ro rv. 0 LO CO «-> 10 CM r* .-1 CM to to i i i 1 1 1 1 ./— L_ Of. ro ct to ce c: cu LU CO o >- cn oc •»• in LO o o ^j- tv. I cc CD -^ ro CD CM CM -H .a QJ LU 1—4 i—4 1-4 T-i CM CM i-l O c: LU cn O O O 0 O O O o o cn 0 0 0 O 0000 c cu o t— < +1 +1 +1 + 1 +1 +1 -H o 1-) L- I O • 0 h- d <» •-« o; LU LO in O h- UJ U. m o in U_ QJ LU f— CO cn c_> c X LU o co in «S- CM tO LO OJ O- O I L U o V/ 1—1 to *r o CD rv. in ro rv CJ r-i rv c_3 1-1 CM CM ro ro ro ro "O 4-J *-« -Z.-Z. o o o o o o o QJ O) rt3 I O O L^ CM CJ U_ o o o CD O O CD Q. 1 1 1 1 1 1 1 I to co C cu c_> LJ LU Q- O _i a: m z o: to t— o ce in ro cn 1—4 |v, LO LO QJ CO I o: tO O 1-4 "3" «—i CM ID cno LU CO fv* fv, CO o ^ O O T-4 CD CD CD CD cr 0 0 O CD O CD CD CD CD in OJ -p— O h— Z J^ o «t •H +| +| + 1 +1 +| +| 4J QJ I LU r— to cn O C cu O IO LU •—< jz: t— t— C U < ct: z o CC Of cC LU ro CM ^t rv. rv, CD ro •f- OJ I LU LU H-. tO ^" ^J" rv, CM o «r rtJ 10 cn CD >• >- 0 Ifl ^^ £f) CM |v, CD TJ- C 10 IO 1-4 1—4 1—4 CM i—< CM CM QJ 4-> ii -5 » L- c: u_ o o o O O O CD o> cn QJ LU 1 1 1 1 1 1 1 O cn QJu O I o •.- QJ O L. i_ Q. m o t— tl- QJ (X to LO ro LO CM Tl- CM LU co ro to T CO in CD T3 I O CO CD IO co CM CM cn 0> C T3 oc. ro o rj »—• rv, 1—4 PO QJ -O QJ

t— .-4^0 CM ro ro -4 •z. O ••- r— • C-) "+— 4-J QJ cn *+— c "o I 10 rv E cu cu ,/> in O cn o o ••- ra i—I 1- O U 4-J (a JT QJ 4J 14- -D ••- J=. 0 o to o n ba : jr f- o i_ C t- QJ QJ t^O I Ihea c 4-J 4-> 1— OJ It) •.— QJ it- O CO =3 IO O> Q.

~~bu l >/> E >> i—~~

IO •*— •!— t 4-J QJ -r— QJ - i^ >- T3 3 Largemout h Lepomi s I Brow n IO J3 O T3 I 6-30A PGE00047200 I I 1983 through 1986 there appears to be a slight increase in PCB levels for Bulls Bridge and Lake Llllinonah. Overall, PCB concen trations decreased at an average rate of 15.3% per year, and de creased 1.6% per kilometer downriver. •

Largemouth bass. A total 83 largemouth bass were analyzed from — 1979, 1983, and 1984 with a geometric mean size of 321 mm TL. The | highest observed geometric mean PCB concentration was 2.30 ppm at Bulls Bridge during 1979; the lowest, 0.16 ppm at Lake Zoar during • 1983. The general trend is for decreasing concentrations in a downriver direction. There was a major decrease in PCB levels from • 1979 to 1983, but an increase from 1983 through 1984. The overall rate of change was an average decrease of 13.4% per year and 2.6% • per kilometer downriver. •

Lepomis spp. A total of 71 Lepomis spp., Including bluegill, red- | breast sunfish, and pumpkinseed, were examined from 1979, 1983, and 1984. The geometric mean size was 186 mm TL. The highest observed • mean concentration was 1.14 ppm from Bulls Bridge during 1984; the lowest, 0.06 ppm from Lake Zoar during 1983. A trend toward fl decreasing concentrations in a downriver direction is apparent. There was a major decrease in PCB levels from 1979 to 1983. Levels m during 1979 and 1984 are similar to one another. The average • annual decrease was 17. 7'/., while downriver concentrations decreased — at a rate of 2.0% per kilometer. |

Smallmouth bass. Smallmouth bass provide the most complete tem- • poral and spatial coverage of any of the species examined. A total of 182 specimens, with an average size of 289 mm TL, were examined • from 1979, 1983, 1984, and 1986. The highest observed geometric ™ mean PCB concentration was 8.79 ppm at Bulls Bridge during 1979. « The lowest concentration, 0.17 ppm, was observed at Lake Zoar dur- • ing 1983. In general, highest concentrations were observed during

6-31 Lawlor Maluskv i»kell\ I mnnocrs I PGE00047201 I I 1979, lowest levels during 1983. PCB levels appear to have remained relatively constant or to have increased slightly between I 1983 and 1986. A general overall trend toward decreasing concen trations in a downriver direction is apparent. This decrease ave I raged 20.4% annually. Downriver concentrations decreased at an I average rate of 3.0% per kilometer. White catfish. A total of 85 white catfish were examined during the years 1979, 1983, 1984, and 1986. The average size was 334 mm I TL. The highest geometric mean concentration, 12.34 ppm, was observed during 1979 at Lake Lillinonah; the lowest, 1.62 ppm, during I 1983 at Lake Zoar. As with the previous species, there appears to be a trend toward lowest concentrations downriver and during 1983. I The average rate of annual decline was 15.9%, while the downriver I concentrations declined at a rate of 3.7% per kilometer. White perch. During 1979, 1983, 1984, and 1986 a total of 105 white perch were analyzed for PCBs. The geometric mean size was I 224 mm TL. The highest geometric mean concentrations were observed during 1979: 4.85, 3.80, and 3.03 ppm at Lakes Lillinonah, Zoar, I and Housatonic, respectively. The lowest concentration was 0.72 ppm at Lake Zoar during 1983. Concentrations decrease in a down I river direction. PCB levels for 1983 through 1986 have remained relatively constant or have increased slightly. Over the entire I period PCB concentrations decreased at an average rate of 18.1% per year and 3.5% per kilometer in a downriver direction.

I Yellow perch. A total of 110 yellow perch, with an average total length of 246 mm, were examined in 1979, 1983, 1984, and 1986. The I highest observed geometric mean PCB concentration was 1.16 ppm at Bulls Bridge during 1979; the lowest, 0.08 ppm, at Lake Zoar during I 1984. PCB levels generally decrease in a downriver direction. I Overall levels were highest during 1979, then dropped to their low- 6-32 I LawJen Matuskv okelly Kn«fin(!«rs I PGE00047202 I I est level in 1983. Concentrations In 1984 and 1986 tended to be higher than those observed during 1983, but not as high as those • observed during 1979. From 1979 through 1986 the average rate of decreasdecreasee wass 21.7%%pe rr year., Downriver concentrations decreased at • a rate of 3.3% per kilometer.

Summary. The consistency in the annual decrease in PCB concentra- I tions among species, ranging from 13.4 through 21.7%, suggests that despite year-to-year changes in sampling methods, catch size com • position, and analytical techniques, there is a degree of accuracy in the observed trends. However, some caution must be exercised B since some of the apparent trend may be attributable to the low concentrations observed in 1983. Despite precautions to minimize • the influence of such factors as length, weight, age, and lipids • through the use of ANCOVA, some effect may still remain. As demon _ strated in Table 6.2-18, average length, weight, andlipid content | were consistently low for each species during 1983. This may be a reflection of the change 1n sampling emphasis when ANSP began • sampling in 1984. More importantly, however, a 1985comparison of split samples analyzed by ITAS and ANSP suggests that the lower • concentrations in 1983 may be a result of differences in the ™ methods used by the different laboratories. The effect of removing m the 1983 data, which are shown in Table 6.2-17, is a general • decrease in the rate of change, particularly for Lepomis spp.

A rapid decrease in PCB concentrations, such as is apparent between 1979 and 1983, has been noted in other populations. PCB levels in • Hudson River striped bass have displayed an initial rapid drop from 1978 to 1980 followed by relatively slight declines from 1980 to • 1985. Post 1980 declines have averaged 1.6% annually (Sloan and * Horn 1986). The rapid decrease in the Hudson River striped bass H was accompanied by a substantial shift in theAroclor composition. • In 1978 Aroclor 1254 made up 43% of the total PCB mix, but by1985

6-33 Lawlur. Matusky TJkelly Kuqinccrs I PGE00047203 I I TABLE 6.2-18 AVERAGE LENGTH (TL mm), PERCENT LIPID, AND WEIGHT (g) OF I SEVEN SPECIES OF HOUSATONIC RIVER FISHES EXAMINED FOR PCBs I DURING 1979-1986 ARITHMETIC MEAN I 1979 1983 1984 1985 TOTAL LENGTH (imt) I Brown bullhead 278 268 266 302 Largemouth bass 336 299 323 - Lepomis spp. 186 184 187 - I Smallmouth bass 304 246 287 301 White catfish 329 255 345 384 White perch 250 221 211 229 I Yellow perch 242 234 245 267 PERCENT LIPID _ I Brown bullhead 0.77 0.81 1.08 Largemouth bass - 0.26 0.82 - Lepomis spp. - 0.34 0.87 - Smallmouth bass - 0.54 0.94 1.02 I White catfish - 2.43 2.75 2.96 White perch - 2.45 3.39 5.16 I Yellow perch - 0.34 0.78 1.02 WEIGHT (g) Brown bullhead 281 304 271 427 I - Largemouth bass 581 565 592 Lepomis spp. 186 147 138 - Smallmouth bass 338 204 316 408 I White catfish 434 228 684 880 White perch 279 188 145 194 I Yel low perch 164 165 195 307 I I I I 6-33A I PGE00047204 I I it made up 81% of the total. This change was brought about by a decrease in the Aroclor 1016 component and suggests that the light- • ly chlorinated biphenyls are depurated or metabolized more rapid ly. Such a conclusion is consistent with the findings of Niimi and • Oliver (1983). Based on the 1984 and 1986 data, the Aroclor break- • down for fish is approximately 70% of 1260 and 30% of 1254, similar _ to the breakdown for the sediments. The 1983 ITAS data on Aroclor | composition was similar, with the addition of a small amount (approximately 3%) of Aroclor 1242. Unfortunately, the complete I Aroclor composition of the 1979 Housatonic River samples was not identified; only Aroclor 1260 was quantified. This does indicate, • however, that the decrease observed from 1979 to 1983 in the Housa tonic River could not have resulted solely from the loss of the m lighter PCB congeners. •

Trophic trends. First, the trophic position within fish are re | viewed for trends; second, limited data on PCB levels in inverte brates are presented. Owing to the presence of temporal-spatial • trends in PCB concentrations as well as the differences in collec tion locations and methods, generalizations regarding trophic A differences among species are difficult to confirm. No readily apparent trends were noted. Highest PCB concentrations tended to a occur in bottom feeding omnivores such as white catfish and carp; • however, these species also contained the largest and oldest individuals and tended to have high lipid concentrations. It | appears that PCB concentrations are more closely related to fish size and/or age than to diet. •

Sampling by CDEP at Cornwall from 1978 through 1981 and from 1984 • through 1987 provides data on PCB concentration in predatory in- ™ sects and hydropsycid caddisfly (Figure 6.2-17). These insects were generally sampled in late May or early June. The average PCE I concentrations of trout also sampled at Cornwall generally later in

~~6-34 | Lawler Matuskv Skollv Ki I PGE00047205 I I PCS (mg/kg - wet weight) I I I~~

TJ I 70 m o H O o I *70 5 GO I m TJ 3 O 00 TI I i—i Q 2 c m •—i ~, I *-> f •^ fo I CO 1 -< tcj ^j I O O 70 O H -o I 00 o o CD CO o I > a o g oo I ~n I 5 I I I I PGE00047206 I I the year (i.e., summer or early fall) are also shown in this fig ure. As the trout data originated from various sources and the sample size varied from three composites in 1982 to 39 individuals in 1979, any trophic effect cannot be evaluated quantitatively. m Nevertheless, the general trend from 1977 to 1985 in the inverte- m brates is similar to that observed in the trout. Interestingly, the higher PCB concentrations in the invertebrates found in 1986 • and 1987 are also found in the trout. This departure from the pre vious trend may be explained by natural variability such as river • flow or PCB concentration of the water and sediment to which the insects were exposed. The similar trends in the invertebrate and • trout data suggest qualitatively that the trophic pathway for bio- • accumulation of PCB is evident at Cornwall. Forthcoming data on _ PCB concentrations in river water from the expanded monitoring j§ recommended in Section 6.5.3 will allow for a more rigorous evalua tion of the trend. • I I I I I I I I 6-35 L«wl«r \laluskv 'ikclh Knyinerrs

~~PGE00047207 I I 6.3 PCB FATE AND TRANSPORT MODEL~~

~~I 6.3.1 Description of WASTOX Model~~

I The model applied to the Housatonic River, WASTOX (an acronym for water quality analysis simulation of toxics), was developed by Man I hattan College for the U.S. Environmental Protection Agency (EPA) (Connolly and Winfield 1984). IMS uses the first part of the I WASTOX model, which analyzes the exposure concentration or the levels of toxicant to which the biota are exposed. Instead of using the second part of the model, which analyzes the bioaccumula- I tion of a toxicant and requires data on the food chain of the Housatonic River that are not available, we have used a simpler I approach to evaluate PCB levels in the fish. I The model was developed to assist in answering the question, "How long will it take a contaminated natural water system to recover to some specified level?" The time-dependent behavior of the chemical I in the system is the key perspective. To address this question vis-a-vis the no action plan, the physical, chemical, and biologi- I cal characteristics of the Housatonic River and its PCBs must be specified as Input to the model. These characteristics determine I the fate of the chemical by defining the rates of transport, transfer, and reaction as defined below: I • Transport - physical movement of PCB caused by net advective movement of water (river flow), mixing, and the settling and resuspension of solids to I which PCBs are adsorbed • Transfer - movement of PCBs through the solid, water, and air phases of the system, e.g., adsorp I tion/desorption and volatilization • Reaction - transformation or degradation of PCBs, I e.g., biodegradation I 6-36 I Lawler Matusky okelly Kno-iiuMir.s I PGE00047208 I The general expression for the mass balance equation about a speci fied volume, V, is: • (6-1) I where: C = concentration of PCBs m t = time m J = transport through the volume T = transfer from one phase to another • R reactions within the volume W = wasteload inputs fj

The model is assumed to be one-dimensional along the longitudinal • axis of the Housatonic River; however, it represents each water segment overlying a segment of active bed sediment. River flow • advects water through the overlying water segments and 1s the driv ™ ing force for resuspension from and deposition of solids to the bed m segment. We assume that there is no direct advection from one bed • sediment segment to another, that is, bedload is negligible.

~~The total concentration of PCBs consists of the dissolved and par ticulate components: • Cfc = C(j + m Cn (6-2) I where: Ct = total concentration of PCBs (m 1~3)* • 9- porosity (I3 1~3)~~

*Units are denoted as m = mass, 1 = length, t = time. I 6-37 I Laulur Matuskx ok<;ll\ K I PGE00047209 I

I C(j = water-specific concentration of dissolved PCBs (m I'3) I m = solids concentration (m 1~3) Cp = solid-specific concentration of dissolved particu I late PCBs (m nrl) I For PCBs in natural water systems, the sorption isotherm is as sumed to be linear and the relationship between the particulate and dissolved components is governed by the partition coefficient I (I3 m-1). I *=§£ (6~ 3) I The bulk fraction of PCBs in the dissolved phase, fj, is expressed as the water-specific concentration of dissolved PCBs and the po I rosity based on the two previous equations: I The bulk fraction of PCBs in the particulate phase, fp, is similarly expressed as: I fp = *+ m?r

~~I Since the wasteload inputs are assumed to be zero at present, the mass balance equation for PCBs in a water column segment is then I written as:~~

dCt ,<# _ _ Q Ct,w + E dCt-,w - vs fp»w A C^,^ (6-5) I dt ~ dx u D * b t * b "" t * w t * w I where: first subscripts t, p, or d denote total, particulate, I or dissolved phase; second subscripts w or b denote I water or bed segment. 6-38 I Lawler Matusky okclly hnaino<:rs I PGE00047210 I I 0 = river flow (I3 t"1) x = longitudinal distance (1) p E = dispersion coefficient (I2 t"1) ^ 1 vs = solids settling rate (1f ) " A = surface area between water and sediment segments • (I2) I

~~vu = solids resuspension rate (1 t~l) M 1 kt,w = transfer/reaction rate coefficient (f )~~

~~The five terms of this equation represent advection, dispersion, I settling, resuspension, and transfer/reaction of PCBs, respec tively.~~

The bed sediment is modeled explicitly as an active layer of sedi- m ment of some prescribed depth (h) below which is the model boundary with the inactive bed. The active layer is assumed to be at equi- • librium so that a solids mass balance states that the net settling and resuspension flux at the sediment-water interface is equal to m the burial flux of sediment from the active to the inactive layer: •

~~0 = r (vs "V ~ v u mb ~ vb mb) (6-6) • where: • Vfc = solids burial rate (1 t"1)~~

Assuming the depth of the active sediment layer is constant and the • long term trend is an accretion of bed sediments (at least in tie ^ lakes and impoundments), the solids burial rate may be viewed jg as the rate at which the sediment water interface moves upward away from a fixed datum such as bedrock. • I

6-39 • Lsuvle.r Matusky Skully Kr I PGE00047211 I I The fluxes of PCB associated with the active bed segment are shown I schematically in Figure 6.3-1 and expressed by the equation: 7~t = h* fP'w Ct' w ~ h f I DA (6' 7) I - TT (fd»bct*b - fd.w ct»w) where: I D = diffusion rate coefficient (1^ t"1) I 6.3.2 Parameter Evaluation The model parameters defining the rates of physical transport, I chemical transfers, and biochemical reactions are evaluated spe cifically for the Housatonic River.

I 6.3.2.1 Segmentation and Hydrology. The model study area of the Housatonic River extends from the Massachusetts-Connecticut border I (mile point* 83.1) to the Stevenson Dam (MP 19.5) at the downstream end of Lake Zoar (Figure 6.1-1). The model segmentation separates I the river into segments having similar physical and hydrological characteristics (Table 6.3-1). Seven water column segments are I used to represent the Falls Village and Bulls Bridge impoundments, Lakes Lillinonah and Zoar segments, and the interconnecting riverine reaches segments (Figure 6.3-2). Additional model segments are I designated for the active bottom sediment layer below each water column segment. These bed sediment segments are numbered 8 through I 14 or an increment of seven more than the associated water column segment number. A flow balance of the river is based on USGS aver I age flow data for gaging stations on the mainstream Housatonic and I *Mile point system adopted from Frink et al. (1982). I I 6-40 I Lawler. Matusky 'Skolly Knjjinoors PGE00047212 FK3URE 6.3-1 I FLUXES OF PCB ASSOCIATED WITH THE BED I I I WATER COLUMN I I f ACTIVE SEDIMENT I INACTIVE SEDIMENT I Particulate PCB flux Dissolved PCB flux I PROCESS MASS FLUX TERM © DEPOSITION OF SOLIDS I ©SCOURING OF SOLDS I (3) NET SEDIMENTATION OR BURIAL ©DIFFUSION OF DISSOLVED CHEMICAL I I I I I I 6-40A I I PGE00047213 I I I TABLE 6.3-1 I Drainage Areas and Long-term Average Flows I Drainage Avg. Mile Area Flow Station Pt. Sq. miles cfs I Great Harrington , MA * 100.4 280 530 MA - CT State line 83.1 545 910 Blackberry R. 82.0 48 74 Falls Village Res. 81.2 580 990 I Salmon Creek * 79.0 29.4 43.5 Hollenbeck R. 77.3 24 40 Falls Village Dam * 75.9 635 1087 I Bulls Bridge Res. 57.7 745 1278 Bulls Bridge Dam 53.1 791 ]353 Tenmile River * 52.1 203 303 Gaylordsville * 50.6 993 1701 I Still River 40.6 70 116 Lake Lillinonah 40.1 1224 1930 Shepaug River 31.8 133 236 I Shepaug Dam 29.6 1392 2380 Pootatuck R. 27.2 24 .2 48 Pomperaug R. * 26.1 75. 1 128 I Stevenson Dam 19.5 1544 2620 I * USGS daily flow data available I I I I I 6-40B

~~I Lawlcr, Matusky ? fikelly E I PGE00047214 1 FIGURE 6.3-2~~

~~SCHEMATIC OF MODEL SEGMENTS I AND RIVER FLOWS OF HOUSATONIC RIVER 1~~

~~1 ADDITIONAL FLOW (cfs) AVG. AVG. • TO SEGMENT NOT TRIBUTARY MAINSTREAM • ASSOCIATED WITH FLOW FLOW NAMED TRIBUTARY TRIBUTARY (cfa) SEGMENT BOUNDARY (cts) ^~~

~~01 • a IV MASS STATE LNE a ' U 6 ( BLACKBERRY R. 74 1 * _ G;5 FALLS VLLAGE RES. 990 • 8.5 SALMON CK 4.5 ! 5 1 s HOLLENBECK R. 4 y FALLS VILLAGE DAM 1 087 -J 2g /~~

~~k 1 101 £J, ^ *u 1~~

~~75 BULLS BRIDGE RES 1278 fl ** C m ©' M 1 TEN MILE R. 303 _ «• $•> BULLS BRiOGE DAM 1353~~

a 1 ® 208 IT M ^ B ^ LAKE LILLINONAH 1980 • STILL R. 116 _ 3 / 164 g 2 - ® it 1 8HEPAUQ R. 236 . t SHEPAUG DAM 2380 1 POOTATVJCKR. 48 3 " ,. K> p i POMERAUG R. 128 £ k o 9 , u.^.u^,^.,un ^ ^^^.. » 1 c Fn •

~~? "^ 1 6-40C • 1 PGE00047215 I~~

I its tributaries. The model processes tributary flows as entering I the upstream boundary of the segment. The physical characteristics of the model segments were evaluated I based on available data (Aylor and Frink 1980; Frink et al. 1982), as summarized in Table 6.3-2.

I In the interest of specifically modeling the fate and transport of PCBs, Segment 6 (Lake Lillinonah) excludes the Shepaug Arm because I PCBs do not significantly penetrate this area (Frink et al. 1982). The volume and surface area of Lake Lillinonah (including the I Shepaug Arm) are reduced by approximately 15% to account for this I annexation. 6.3.2.2 Bed Sediment Characteristics. Model segments 8 through 14 represent the active bed sediment volumes beneath the seven water I segments. The total organic carbon, sand content, bulk density, and PCB concentration data are summarized in Table 6.3-3. The dif I ference in sediment properties between the lake and riverine sedi ments is exemplified by higher organic carbon, lower sand, and I lower bulk density in the lakes than in the riverine segments. The sediments in the Bulls Bridge impoundment are similar to the river I ine segments, whereas the Falls Village impoundment appears to have relatively more fine material, but not as much as the lakes.

I In addition to data from Frink et al. (1982) on PCB concentration, six core samples of sediments (one in each impoundment and two in I each lake) were collected by divers in October 1986. Each core was sliced into 1-in. increments and sent to York Laboratories of I Monroe, Connecticut, for analysis of PCBs and total organic carbon. The laboratory reports are included as Appendix B. PCB I determinations for Aroclors 1016, 1221, 1232, and 1242 were less I than the detection limit (0.1 mg/kg) for all 100 sediment samples; 6-41 I Lawlcr Maluskv Skolly Knginoc-rs I PGE00047216 I

~~ra PO J •*—' I C t~ •—I «* CO «=t —• HJ Hfj in en in oo o oo CM (/, o o •«*• in r~ in o co • r— ^ «—« r>^ CT) CM ^.4-^J I O) 4->~~

~~_ — O) I c „ a> LU Q. Q -— to CM r^ ^* m o> oo *>o >1 4J *+— CM o CM in oo oo rr , I C *"* T-I ^-i 00 CM rO J c •t rO 1/5 I~~

~~u> — *J o in CM oo in «a en 3 ** oo 10 «* o CM c-o ir> o CM I-H CM CM CM O OO cr I c 10~~

0 T3 I CM CM t! I • 4-J o >— «*- Q. 00 CJ »-• IO CJ Jj to in i—< ^r •—i CM r^ f~ 00 «t co to co CM r^ *r r^ rO t, I 1 LU LO l^~ LO CM CO CM CO —I 00 ^^ r"** •—* oo o •—< 4_> 03 •-" ^c •^ CM CO o D_ O I «f a> ce ^ 1 LL ro LU ce u CJ O a> I J_ 2 f*1 CD ^^ tO r*»*. f~i LO li in o oo •-! en CD r-~- E i ^^ -^ in —i CM in en oo 00 LU T—1 <^ > - Of. en 10 CD QJ I LU z: '-I in OO «3" CD CM C3 T3 —i — 0 'r— I = o 3E 0 ^ CM oo 'a in 10 r^ C3 Z I/) LU ro o I OO 0) -C o cn . r— CO i— •— L. I OL •- CZ (O ••— i— fO ^(11 LU »- ro t/> ^ to ^ -r— O E t—

cr _J 1 ^ LU o CD ro 10 to r-x »-i ra m 4-> I O CO CO tr CM CM CD CT) r3 t— i— •— cr -—? O • • • • • • • CT o .— ra E >— t- CO O r U— QJ -*-* "O Z O *~* "U QJ o E cr T3 m I I-H u ^ 03 d — ^£ in ro ro in TI i cr ro ja •f or JD 10 ^" ro CM at. O ro i • i • i • • ra OJ CD tn c_ — >- _J CD CD •—' •—• J3 -tr rO CD C71 z 4-> i— CJ UJ cr -— o. i- I CJ O in r o o , "O ~O "X3 f~ o CJ CM O 10 O> I-H CO QJ 1— i/> T3 Z^^ rr—O to r^ •-* o ir> LT> 4J i— in CD CQ • • * • i • * CJ a: 4J o o o o o o O • ra ro c_ I Q I1 QJ E in u_ E o O OJ CL CJ J^ L — QJ rO O ra 1-

^T o> r*** CTI co *—* CNJ o> • o in I E ^ O ^ «-H GO »—t OO OJ C OJ 4-1 c r— 0 ^ n c^ c^ ^j" n 10 co • E » QJ o Z > cr T3 r3 !_<*-•— UJ t— —• QJ .— . jr: o *J •s: CD O C 4-> ra cs CO ro > f in L. UJ z L ^ ^ | t I CO UJ —. a> .—. T» r— m cr oro > in ra rjj * +j ro in to r^ r\ «-t ro TJ O- tn fi o UJ ^* •—* n I-H cr> o •i— o ja cr • t—i i •>. o ID JO CO ^t OO CO »-» f**> < o —1 O CJ ra 21 CO •— to co co co ro ^f rO l/> T3 O I ^-^ cr "^ ^~^ CQ r*"*j u_ V .—,^ cr> ^»: ro to cr> — 4-> .—i .— ra i O t— *-< cr i_ —-. — _i ro >—i o •^ •— 3 QJ rjj ra I r— z LU .— ^r C3 C3 ^J •—i C\J CO c/1 H~ ^^ i • • • • • » • - -a cr .— 4 o i—i tn Z —• ^^ ro co ID t—» ^* f— in cr rs ro o t- UJ <"v 0 CD CO OO CO CVJ CO ro r3 _J LU o C O TD 4-> QJ CT) CO r— 4-» r— CX QJ QJ CT)"^. r ^ a«^: z i— i— ra i- ra • o i- o *3 < C. r >, ^ T3 0 CO >t cT) ^t* O"> ro oo o-j UJ CtC —• 1 • * • • • • JO O ra CT QJ rO •Si —J •—•t •—* CD CNJ cr> r^ in ro 1- ro T3 E »—t o T3 4-> OJ jr ^r ^^ r o — oj >•> o. cr ~ 4-* 4-* ra UJ D 4-> L. o rjj ra r3 .— cr I 1/5 t-T) 4->ocrcxi-o o cr o E "=C cr Q D->t t O c O •— UJ U_ QJ O O CJ r U •— CO f"i rO l_ • .— • •— CO in .—- in CQ.— cr m t— CC LU »n in ~o QJ CJ *— o QJ i I LU _l QJ rO C CD O. _J i— ^. CO Q. CM CD ^T LO 1 CO CSI •— E ro ra 4-> ro QJ ^y ^r ro csj Q_ in 1— QJ m ra .— j»£ ID < E 4-» o o QJ z: o E>t • •— >+ cj E CT) CD Z co en CD ^H cxi ro ^i r •— ^— I to OJ ro + rocrjnojojz: m > .O4" ^-QJ^H m o f^ t_ 4«) _i^ ^J r— ra "O QJ -Q lO CJl "O C CT) r— cr »—* 4-> .cr cr u m I — TJ t 0 in •— ra 4-J ra i •*— »— i— -i o cr i_ c: t ii i_ 4-> o Z *r— ^-~ t 4^_ r— oaju- 4->in-c r *>^ 0 > ^ co ^ r 1_ it- T3 •— Q cr QJ rjj rjj QJ a) cr ro •- — ra CD QJ ,— CT) E E •— a >— ra »/> ^ «/> Z »" CD ra ^ cr cj n ~o '— r3 o_ LU i/> I —i O U- O CO Z _l _J ra J2 CJ T3 I 6-41B PGE00047218 I I detectable concentrations for Aroclors 1248, 1254, and 1260 were reported, with Aroclor 1260 averaging about 75% of the total. Due I to the potential weathering (desorption/diffusion) of PCB isomer groups, Aroclor concentrations are somewhat questionable and the sum of the concentrations for the three Aroclors is used as the I total PCB concentration. _

Both sources of data show greater PCB concentration of surficial sediments in the lakes than in the Impoundments, which are farther • upstream and nearer to the source (Table 6.3-3). PCB concentra tions of the surficial 3 in. of sediment reported by York are gen- M orally greater than the values from Prink et al. (1982), although the elapsed years between the two samplings would suggest that the • more recent data should be slightly lower due to the burial of • higher PCB concentrations by cleaner sediment (Figure 6.3-3). Nevertheless, an average of the Frink et al. and York PCB concen- | trations is used in the model for segments 9 and 11. The PCB con centration for model segment 12, which was not sampled for PCB • analysis, 1s assumed equal to that of segment 11. There were two sediment analyses by Frink et al. in the short model segment 8 so • that the PCB concentration for this model segment was set in agree- " ment with these data and the model data for the next downstream • segment. As the average of four PCB analyses by Frink et al. in m segment 10 was relatively low, the average concentration of seg ments 9 and 11 was used in the model to have a consistent trend in J PCB concentration. The PCB concentration of the sediments in Lakes Lillinonah and Zoar is listed in Table 6. 3. -3 as a sediment- • weighted concentration computed by Frink et al. and footnoted as the average of the PCB concentration reported for all grab sam pies. The PCB concentration for the model was calculated as the average of the sediment-weighted concentration and the average of » the eight surficial sediment samples analyzed by York. As an • alternative calculation, the average of the Frink et al. grab sam

~~6-42 I Ljwler Malusk Skelh K I PGE00047219 FIGURE 6.3-3~~

~~SEDIMENT PCB vs. MILE POINT J T 2.8 i 3.16. 2.6 2.4 ~~

~~2.2 T -f I 2 !~~

~~PCB 1' (mg/kg)1.6~~

~~1.4~~

~~1.2~~

~~0.8~~

~~0.6~~

~~0.4~~

~~0.2~~

~~0 80 60 40 20 0 MILE POIN" nSEGMENT~~

~~f T irRINK 1982~~

~~O # OF SAMPLES~~

~~1 YORK 1 986 + SURRCIAL LAYER OF CORE~~

~~6-42A~~

~~PGE00047220 I pies, pooled with York's data, yields PCB concentrations within 0.08 mg/kg of the model input data for these lakes. ft~~

TOC and PCB concentrations and the ratio of PCB concentrations to g TOC concentrations are plotted vs sediment depth in Figures 6.3-4 • to 6.3-6. Because PCBs are hydrophobic, and organic matter is their primary sorbing constituent (Karickhoff 1984), PCB concentra- £ tion normalized to organic carbon concentration filters out the variation in PCBs over sediment depth that is due solely to varia- tt tions in TOC concentration. The profile of normalized PCB concen tration over sediment depth reveals the trend of decreasing PCB M concentration over time, reflecting the change in PCB discharge ™

from GE's Pittsfield plant from the period of PCB use (1930s — through 1977) to the cessation of discharge over the past 10 | years. The maximum normalized PCB concentration in the deeper sediments of the Falls Village and Bulls Bridge impoundments is ap- • proximately three times the surficial PCB concentration. The maxi mum normalized PCB concentration of the two cores in Lake Li Hi- fl nonah is at a depth of approximately 15 in. and is four to five times the surficial sediment concentration. The peak normalized M PCBs in the deeper sediments of the Lake Zoar cores are approxi- »

mately twice the surface concentrations. These PCB data show a — dilution of PCBs in the active layer of the bed sediment and are f later used to estimate sedimentation rates. In addition, labora tory chromatograms were analyzed for any evidence of biodegradation • (Section 6.3.2.7).

6.3.2.3 Settling, Resuspension. and Burial. The fluxes of solids across the sediment-water interface and the active-inactive sedi- mt ment interface are important in determining the transport of PCBs. • As stated previously, the net solids settling minus the resuspen sion flux at the sediment-water interface is assumed to be equal to • the burial flux at the active-inactive sediment interface. Because I 6-43

~~Lawlur Mcilu.skv ok<:l I I PGE00047221 FIGURE 6.3-4 1 SEDIMENT PCB and TOC vs. DEPTH Falls Village Bulls Bridge 1 Impoundment (MP 76.1) Impoundment (MP 53.2)~~

~~1 TOTAL PCB CONC. TOTAL PCB CONC (mg/kg) (mg/kg) ) 0.5 1 0 1 2 0 ' ' ': 1 1 •] •~~

~~• 4 1 » DEPTH DEPTH (in.) I f(in.) • 4 ' 6 •~~

•

~~• 12 . • 1 • • 1 TOC CONC. TOC CONC. (a/kg) (gAg) ) 5 10 3 15 30 45 0 H 1 .j • • •~~

~~• • 1 4 - i • DEPTH — DEPTH (in.) • (in.) • • 4 • 6 ~~

•

~~• • , 1 12 • , • i • i 1~~

~~TOTAL PCB/TOC TOTAL PCB/TOC (mg/g) (mg/g) 1 D 0.2 0 4 0 0.1 0.2 0.3 0.4 0.5 0 H 0 ~~

~~1 » "•' ' ! 2 4 i • I DEPTH DEPTH • * (in.) • (in.) • i 4 a • m • 1 » 6 12 • • •~~

~~6-43A~~

~~1 PGE00047222 FIGURE 6.3-b SEDIMENT PCB and TOC vs. DEP~H I~~

~~Lak e Lillinonah — Rt. 133 La Ke Lillinonah — Snepcug • Bridge (MP 32.3) Dem (\jp 29.8)~~

~~TOTAL PCB CONC | TOTAL PCB CONC. (mg/~~

~~6 6- - DEPTH 1 DEPTH (in.) • (in.) ; 12 • 1 12 * i~~

~~" 16 1 18 • • on * ' ^~~

~~TOC CONC 1 TOC CONC. (9Ag) 1 41 51 40 65 90 I O j -i ."j ; j~~

~~1 J 6 i 1 6 • DEPTH ", DEPTH • (m.) , (in.) -I ,2j H 1 12 J J • 1 - ^ " ' . 1~~

~~18 m £, £ 1~~

~~TOTAL PCBAOC TOTAL PCB/TOC _ (mg/g) (rng/ o) ^1 0 0.1 0.2 0 01 02 ™• 0 . • I 1 6 • 6 DEPTH DEPTH On.) On.) 1 12 * 12 * 1 18 • • 1~~

~~• t 18 • ^^K~~

~~6-43B 1 PGE00047223 | SEDIMENT PCB and TOC vs. DEPTH~~

~~Lake Zoar - Rt. 84 Lal~~

~~TOTAL PCB CONC. TOTAL PCB CONC. 1 (mg/kg) (mg/kg)~~

*t „ ° °'5 1-° ° . . 4 , 8 ^B 0 * \J m 9

~~m m m • B DEPTH 1• DEPTH (in.) (in.) * » 16 • . • * 1 - . * 24 ~~

~~1 . : •«~~

~~| TOC CONC. TOC CONC. (gAg) (g/kg) 5 0.8 1,1 20 60 100 0 1 «°i • B~~

~~• 1 B • DEPTH * DEPTH * _ (in.) (in.) * . * 16 " " 1 -1 • - . • 24 • 1 -. '~~

~~t O I TOTAL PCB/TOC TOTAL PCB/TOC (mg/g) (mg/g) 3 0.5 1.0 3 0.1 0.2 1 0 - ' '•i ' ' ' ' ' ' '~~

~~• . • 1 8 2 . DEPTH . * DEPTH (in.) • ! (in.) . • 16 4~~

«* . 1 24 6 _. • *. 1 32 6-43C 1 PGE00047224 I I the mass of PCBs in the river is stored predominantly in the bed sediments of the lakes and impoundments, these model segments were • the focus of the sediment transport analyses. Because these seg ments are depositional areas, radioactive tracers from atmospheric • fallout were used to analyze the sedimentation history and date key • episodes of cesium fallout. A sediment core from each impoundment — and lake was sampled using 3-in. diameter cylindrical cartridges | deployed by divers in October 1986 at locations approximately 1000 ft upstream of each of the four dams. The decision to sample by • divers was made to minimize disturbance of the surficial sedi ments. The depths of penetration and sediment recovery in the ft cores analyzed for radioactive tracers were:

~~SAMPLING MILE PENETRATION RECOVERY STATION POINT DEPTH (in.) DEPTH (in.) I~~

Falls Village 76.1 17 12 Bulls Bridge 53.3 17 10 I Lake Lil linonah 29.8 44 18 Lake Zoar 19.7 48 25 I I Why the recovery depth is approximately half the penetration depth in the lake sediments but approximately two-thirds the penetration • depth in the impoundments appears to be related to the difference in sediment composition between these areas. The finer and higher m porosity sediments of the lakes probably oozed from the bottom of • the core cartridges when the cartridges were retracted. Another explanation, that sediments were compacted, seems unlikely because | the wall friction from manual deployment would probably not cause compaction to extend over a 3-in. diameter. Therefore, the re- • covered sediments are assumed to be representative of the top por tion of the bed sediments and, as shown by the PCB/TOC profiles, • are sufficiently long to provide information on sedimentation over * most of the period since PCB usage. I 6-44 I LauU;r Molu.sk> Skulk Kiicrimjors I PGE00047225 I I Selected increments of these four cores were analyzed by Teledyne Isotopes, Westwood, New Jersey, for Pb-210 and Cs-137, the radio I active tracers commonly used to date deposition. The Pb-210 technique necessitates having approximately 10 subsamples of the core I from the Inactive layer and is generally useful for long-term sedi mentation rates over the past 50 to 100 years. Pb-210 measurements I were performed only on the lake sediments; however, the interpreta- tive analysis of these data was not conclusive on the deposition I rate for the last 30 to 50 years. Cesium, which has an affinity for fine organic sediment similar to the sorption properties of PCB, was the better radioactive tracer for the lakes. However, all I cesium measurements in Falls Village and Bulls Bridge cores were below the detection limit because of the low content of fine I organic sediment (Table 6.3-4). 1 The atmospheric fallout rate of cesium from the atomic bomb-testing period 1954-1972 exhibits a relatively sharp peak in 1963 (Figure I 6.3-7). Sediment-dating techniques generally focus on the deepest appearance of cesium as a marker of the first year of fallout, and the sedimentation or burial rate is computed on this basis alone. I However, cesium measurements of the study lake sediments covered an order of magnitude range that allowed use of the peak concentra I tion and the pattern of the cesium profile for evaluation of sedi mentation. Furthermore, the depth of the active layer of sedi I ments, where mixing through bioturbation and bottom currents occurs, was also quantifiable from these data.

I The evaluation procedure used a two-tiered approach. The conven- tion established for both methods was to match the computed maximum I cesium concentration to the observed maximum in each core. The first method assumes no mixing of sediments, that is, the fallout I and settling of cesium-adsorbed solids occur in the same year and I sediment is advected away from the sediment-water interface. This I 6-45 I Lawlor Matusky Skolly Kntrimuirs PGE00047226 1

00 LO CO «* 'T -—I tO oo m n CM^T to ^rcMtxjcM en 1 LU JJ r< O r-i^-l CMCn —I C3OO O CtL •^x N^ ~^S V O c ^ O -I2 oc LU 1 > •—H C 0 >—» 3 z 1 o "• OO CT> ro CD co to ^ r^ ^j' h- —I • t^. CD LO O CM CO f^x CD < _i en • • • • • • • • £/> O O •—1 CM CM i-< C3 •—< => _j o 0- z LU Z 1 i^ •— Lu •z. •—i CO cvjCMnncMmcocM—< 1 y* , i n ro ooaoooooo n h— to c_> CO UD U. _J 0-

~~UJ LT ^5 1 _l OO 00~~

~~< OS t— o U, 1 > UJ l_ to \S "^ \/ ^^ \f \/ \s \s a. _j a. _i z: < «— u_ 1~~

~~^— o •- 1 6-45A 1~~

1 PGE00047227 FIGURE 6.3-7 I PROCEDURE FOR EVALUATING SEDIMENT BURIAL RATE I AND DEPTH OF ACTIVE SEDIMENT BASED ON CESIUM Atmospheric Fallout of Cs I 0.5 1.0 1.5 2.0 -H— —I— —\~ I 1972 1970 I 1968 I 1966 1964 YEAR Cs spike I 1962 I 1960 1958 I 1956 I 1954 137 Cs half life * 30.11 yrs. decay rate coefficient. X = .023/yr.

I TWO-TERED EVALUATION PROCEDURE: I 1st Method: Estimate burial rate from peak Cs in core corresponding to deposition of Cs spike. Compute Cs profile based on decay but no mixing. I Compare computed and observed results. 2nd Method: Assume sediment mixing such that earliest fallout (32 yrs ago) was mixed I through the depth of the active layer (h) in 1954. Deepest Cs is measured at a depth corresponding to 32 + N yrs, where

h = vbN. I Compute Cs profile based on mixing and decay so that variance between I computed and observed is minimized for selected v^ and h. I Note: Computed maximum Cs is matched to observed maximum in both methods. I I 6-45B PGE00047228 I I method resulted in the deepest observed appearance of cesium and peak concentration agreeing with the computed counterparts in Lake • Zoar but a discrepancy on the deepest appearance in Lake Lillinonah * (Figure 6.3-8). Since the construction of the Shepaug Dam (1955) f almost coincides with the earliest cesium fallout, and because the • deepest sediment was not recovered in sampling, this discrepancy is not surprising. The comparison of both computed profiles with all • observed results shows that the cesium concentration probably spread through the sediment by mixing rather than purely advective • transport.

The second method assumes that the earliest fallout of cesium, in • 1954, was mixed through the depth of active sediments rather than ^ confined to the sediment-water Interface, as assumed inthe first | method. Although Lake Lillinonah does not fit this assumption, the method is generally applicable for subsequent years. The analysis I proceeds by taking the annual fallout for each year from 1955 through 1986, computing the cesium concentration in (1) an active I layer of assumed depth based on mixing and radioactive decay and (2) the inactive layer by advection and decay (Robbins and Edging £ ton 1975). The values for the burial rate and depth of the active " layer are selected by minimizing the square of the residuals be — tween observed and computed data. The computed results from the | second method show improved agreement with the observed data in both lakes (Figure 6.3-8). The reason that observed concentrations • within the top 5 in. of sediment are greater than the computed cesium is that the input of cesium from 1973 through 1986 is 8 assumed to be zero, but probably is not because of the scour of up stream bed material and its deposition in these lakes. Neverthe- m less, the active layer is deeper in Lake Lillinonah than in Lake • Zoar because of the finer, lower bulk density sediment of the former. Lake Zoar had a slightly higher burial rate than Lake Li1 - | 1inonah. I 6-46 Laulcr Mnluskv Skcllv I I PGE00047229 MUUKt 6.3-»

1 Sedimen.•me -r —i i i i i w < > t* Cesiuw -r m vsi ••- .I Dept™ ™- p. h 1 Without Mixing LAKE LILLINONAH LAKE ZOAR 1 Cs (pCi/gm) Cs (pCi/gm) (3 1 2 .3 C) 1 2 3 0 • *

~~4 ' * 1 8 • -> 8 I « Depth ; Depth 12 - 1 ' § (in.) ; (in.) 12 • 16 ' ^=== ==^ ^-^^ » ^>—-~^^ * 1 20 • 16 • * » VL= 0.46 in/yr v, = 0.63 in/yr b 24- * b~~

~~1 With Mixing 1 LAKE LILLINONAH LAKE ZOAR~~

~~Cs (pCi/gm) Cs (pCi/gm) 0 1 2 3 0 1 2 3 o 1~~

~~4 1 N. * V 8 Depth 12 ' (in.) 12 ,2 16 • 1 ^^^—•~""-^ •* 20 16 ./ * v, = 0.46 in/yr / » D '' v. =0.51 in/yr 1 ^>h = 5.5 in 24 * h = 4.0 in. 20~~

~~1 + observed — computed 1 l - 6-46A~~

PGE00047230 I I According to data from Frink et al. (1982) on the decreasing cross- sectional area of sediment with distance upstream of the dam and • the lake width from USGS maps, the average burial rates for model * segments 6 and 7 were estimated as 70 and 63%, respectively, of the ^ burial rate near the dam. This yielded a burial rate of 0.32 in./ • year for both lake segments.

The burial rates for the Falls Village and Bulls Bridge impound ments were extrapolated from the depth profiles of PCBs presented j| in Figure 6.3-4. The depth of maximum cesium concentration in the lakes was found to be approximately 64% of the peak PCB concentra- • tion. This percentage was applied to the depth of the maximum PCBs • in Bulls Bridge and Falls Village to estimate the 1963 sediment ~ horizon and compute a burial rate. As the pools from the impound- | ments are less extensive and shallower than the lakes, the average burial rate for those segments was estimated as 907. of the rate I near the dam. The depth of the active layer was estimated assuming a mixing period, N, of 8 years, similar to that for Lake Zoar I (Table 6.3-5).

As resuspension of sediment in the lakes is assumed to be zero, •• the sediment flux from the active to the inactive bed sediments is ^ equal to the settling flux at the water-sediment interface. Set | tling rates for the lakes were computed accordingly and checked for agreement with the settling velocity computed by Stokes law. A • long-term equilibrium between solids settling and sediment erosion is assumed for the riverine segments, such that the average resus I pension solids flux equals the settling solids flux. As the physi cal characteristics of the Falls Village and Bulls Bridge impound • ments are intermediate between those of the lakes and the riverine • segments, the average resuspension flux for these impoundments was _ assumed to be approximately 40% of their settling fluxes. USGS | data on total suspended solids and river flow at Great Barringtor, I 6-47 Laulr,r Miituskv i>kt;llv I I PGE00047231 I I

~~Q LL_ LLJ I O CQ >>~~

~~3: LU • O ^H 0 ^H O CO ID c. CU 1-4 -r- •—I ^-4 —I CM ^H •** rO ill I *»M>. I Q 0 c: T3 o cu~~

CO 10 .— «j I _j t_ £> l/l *f LU >, O *}• O CO O CM CM o •—"I O *-t O CM o ro n en LU i -a CO ^ cc c O O 0 O n o ca •• CO O O to LI I cu a> Q. i_ •z. 3 c: 0 o> o I o CO »—. CO LU LU> «»• «-H CO CO Cf> O O Q- >— ^ PO t-H IO •—* CO CD C3 =3 a: c O 0 O O 0 O 0 CO •• I UL1 1J 1 ^*—• o <-> ce .c vi .c tn c (O O CD I 1 a _l LU TJ ro Tf Tj* in r^ ^H m o CO •«- CO to i_ en Z in ta CU O LU o i CL. LU L- LU U 4-> CU I CO O 0 CQ < CO u_ •—-o ; o: CO «J Tf CO r-^ n CM LU CfL LU LU in CD ro —i CO ^" »—* Z5 < I— 1- ^-« in ^H CM ro r^ CO LU C3 o reJ I u -a ro to to in ~a t _,_ a> i- Z • O '-t CNJ n «r I t_ o 3E Z n} >/> CO C3 •-H CM ro •«*• in to r^~ > o T3 to a> a> o I CO a> on cu O 3 TO C7> T3 C •i- in •— T3 L- O Q CO O C. I LiJ ^E~ CO O =«• — CQ C CO •1- r— f(J in o o: i— Z -i- O 3 -.-> LU «*- < c 0) • • o: o I 0 Ll^ O CQ Z _l _l I I I 6-47A PGE00047232 I I Canaan, Falls Village, Kent, Gaylordsvi 1 le, and New Milford were analyzed to obtain a relationship between the resuspension rate and • river flow for the five model segments (Figure 6.3-9). The resus pension rates input to the model were determined by reading these fl plots for the given monthly flow. The settling, resuspension, and burial rates estimated as described were adjusted slightly in the • calibration process (Section 6.3.3), and the final set of values is 9 presented in Table 6.3-5. _

The significance of algae in the overall solids balance was ana lyzed by reviewing USGS data on chlorophyll a concentrations in • Lakes Lillinonah and Zoar. These concentrations are generally be low 15 ug/1 and, based on the ratio of chlorophyll a to dry weight fl of 143 (Canale 1974), the T5S concentrations associated with the ™ algae is at most 2 mg/1. Since this is approximately 10% of aver- * age TSS concentrations, the model does not have to account specifi- I cally for algal dynamics in simulating solids fluxes within the Housatonic River. I

6.3.2.4 Sediment-Water Partitioning. The transfer of PCBs between I the particulate (sorbed) and dissolved phases is analyzed by a par tition (a phase distribution) coefficient as defined in Equation 6-3. The model assumes that adsorption-desorption are reversible I and instantaneous, although PCBs may have a resistant component _ that is not readily desorbed (DiToro 1985). For hydrophobic or- | ganic pollutants such as PCBs, the partitioning can be a priori estimated based on the chemical's affinity for water or the octa- •

~~nol/water partition coefficient (Kow). In natural water bodies organic matter is the dominant sorbent, and partition coefficients I~~

indexed to carbon content (Koc) are relatively independent of other sediment characteristics (e.g., geographical origin of soil) • (Karickhoff 1984). Experimental data for a set of 22 hydrocarbons • I

6-48 I Lawlor Malusky Skollv l.noincers I PGE00047233 1 FIGURE 6.3-9 1 Flow Dependent Resuspension Relationship I Segment 1 Segment 3 1.5 ...... — ___ •^ *

~~1 "u1- u ~- 1.1 - ^% 1.2 . I ! "3 £ ' • e jj *• D.B ~~

* 1 j li •B jj o.» - £ 0.1 - ,

~~| • :: , 0.4 0.3 • OJ • 1 0., " • 0.2 • 0.1 • 1 "•:; 0 • 0 1 3 3 4 5 ( 1 2 J 4~~

~~(Thousand*) fThousondi) 1 UWpoM of now Clan (cfa) Midpoint of no* Clo.» (cf«)~~

~~1 v Segments 2 and 4 Segment 5 13.~~

~~1.4 1 :- U 1.2 - t.J 1.1 • 1 „ ':: 1 ^ •?; 0.9 - | 0., H~~

~~=• 0.8 j "•• f 1 0.7 - 1 0.7 > 0.8 - > ti. c c -5 0.5 - 1 OJ I • I -.. 1 0.4 1 1 I ; 0.3 - * OJ - X *• g 0.2 - 0.2 • 0.1 .- • • ' O.I " i * i * t~~

~~(Thou*and3) (ThouiandB) | Midpoint of How Clan (cf:) M<)poln( of Flo. Cta» (c(.;~~

~~. 6-48A~~

~~PGE00047234 I~~

~~and chlorinated hydrocarbons yield this relationship with r^=0.986 (EPA 1983): •~~

~~Log Koc = 0.942 log Kow 0.144 (6-8) m~~

~~Based on log 1COW for Aroclors 1254 and 1260 of 6.47 and 6.91,re _~~

~~spectively (EPA1983), a log Kow of 6.79 was estimated for total £~~

~~PCBs in the Housatonic River and a log Koc of 6.25 is computed.~~

The sediment-water partition coefficient (T) is equal to the Koc • times the fraction organic carbon content (foc)« (r°r Housatonic River bed sediments with an organic content between 1 and 10%, the • partition coefficient ranges from approximately 104 to 105 I/kg.

The partition coefficient has been found to be inversely related to B the solids concentration such that the partition coefficient at a ^ typical water column suspended solids concentration of 20 mg/1 is | greater than that of typical bed solids concentrations of 600,000 mg/1 (O'Connor and Connolly 1980). As the total carbon concentra • tion of suspended sediment in the Housatonic River is generally higher than that of the bed sediment, the lower solids and higher • organic carbon concentrations of the water column consistently invoke a higher partition coefficient in the water column than in • the bed. Thus, a solids-dependent partition coefficient was I adopted for the model based on * equal to 10^ I/kg and 104 I/kg at solids concentrations of 20 mg/1 and 600,000 mg/1,respectively. || The equation for the partition coefficient as a function of solids in the model concentration is: •

= 2 x 105 nr 0 -223 (6-9) 1 The implication of the tenfold difference in partition coefficient between the bed and the water column segments is that if the par ticulate PCB concentration in the bed and the water column is the I

~~6-49 • Lawlor Matusky okrllv Knq-iiuM;rs I PGE00047235 I~~

I same, the gradient in the dissolved PCB concentration is from the bed to the water column. This gradient affects the diffusion of I PCBs from the bed to the water column. I 6.3.2.5 Bed Sediment-Water Column Diffusion. The diffusion of dissolved PCBs between the bed sediment and water column depends on 1 the molecular diffusion coefficient of PCBs in water and the porosity of the bed sediment. According to Thibodeaux (1979) and Chapra and Reckhow (1983), the diffusion coefficient (the parameter 0 in I Equation 6-7) for PCBs in the Housatonic River is estimated at 6 x 1Q-6 cm^/sec. The significance of diffusion as a flux of PCB is I described in Section 6.3.3. I 6.3.2.6 Volatilization. The two-phase resistance theory is used by WASTOX to simulate the volatilization of a toxicant. The value I of Henry's constant for PCB is 7.1 x 10~3 atm m3/mole which for a temperature of 20*C (293'K) converts to a unitless Henry's constant of 0.295 (Mills et al. 1982). The transfer of PCBs is similar to I dissolved oxygen gas transfer in that the liquid phase resistance controls. The model evaluates the PCB volatilization rate coeffi- I cient analogously to the reaeration rate coefficient. In the riv erine segments (1, 3, and 5) the volatilization rate coefficient I averages between 4 and 5.5 ft/day; wind effects in the lakes and impoundments produce a value between 1.0 and 1.4 ft/day in these I four segments. 6.3.2.7 Biodegradation. The packed column gas chromatograms for I the 100 1-in. layers of sediments from six cores were carefully examined by Dr. John F. Brown, Jr. of GE's Research and Development I Center, and six representative samples (No's. 57538, 57552, 57561, 57565, 57575, and 57583) selected for detailed characterization by I DB-1 capillary gas chromatorgraphy. All of these samples appeared I to have originally been composed predominantly of roughly 2:1 Aro- I 6-50 I Lawler Matusky ] JJkelly Kngincers PGE00047236 I I dor 1260:1254 mixtures, along with smaller amounts of Aroclor 1242 and DDT metabolites (e.g., DDE). There were clear evidences of • dechlorinatlon in a pattern similar to Pattern H, with 10-50% con version of the susceptible congeners, as was also seen at Woods m Pond (Brown 1987). m

Pattern H is a dechlorlnation system that has recently been seen in £ the sediments of New Bedford Harbor, Escambia Bay, Woods Pond, and the mid-Hudson (Brown, unpubl.). It presumably is an anaerobic • bacterial degradation that converts PCB congeners carrying 3,4-, 2,3,4-, 2,3,4,5-, or 2,3,4,6-chlorophenyl groups to the correspond H ing congeners carrying 3-, 2,4-, 2,3,5-, or 2,4,6-chlorophenyl groups, respectively, and thus can accomplish quite effective • detoxication, but only limited dechlorination. Accordingly, zero m biodegradative loss of total PCB was assumed in all segments of the mathematical model. I

6.3.2.8 Upstream and Tributary Suspended Solids and PCBs. Sedi I ment and PCB transport in the Massachusetts portion of the Housa tonic River were measured by ITAS (Stewart Laboratories of Knox fl ville, Tennessee) in 1982 and 1983. Data on TSS and flow at Great ™ Barrington and Andrus Road (near the Connecticut border) provide a • basis for adjusting the USGS data from the 18-month period of mea I surement at Great Barrington (Frink et al. 1982) to produce the model's upstream boundary for the model calibration (Stewart Labo | ratories 1982, 1983). For river flow greater than approximately 500 cfs TSS concentration at Andrus Road averaged about four times • the TSS concentration at Great Barrington. This is attributable to the sediment yield from the predominantly agricultural land that drains to the river over this reach. For flows less than 500 cfs I the TSS concentration at Andrus Road was similar to that at Great M Barrington. 1 I

~~6-51 • Liiulcr Matusky Jikcillv Knirinot:rs I PGE00047237 I~~

I A mass balance of suspended solids and PCBs between Great Barring- ton and Andrus Road based on ITAS data showed that there is negli- I gible scour from the riverbed. That is, the increase in sediment transport is due predominantly to "off-river" rather than "in I river" sediments. The monthly TSS concentration of the model's upstream boundary (mu) was set as follows according to the monthly I average flow at the upstream boundary (Qu) and the monthly average TSS at Great Barrington

I Qu < 500 cfs mu = (6-10) I Qu :> 500 cfs mu = 4 x TSS concentration data for tributaries were available from USGS I annual reports during 1979 to 1986 for the Still River and the Shepaug River. For the limited data available, TSS concentration did not appear to be directly related to river flow; consequently, I a constant TSS concentration (18 mg/1) equal to the mean concentra tion of the Shepaug and Still rivers was used for all tributaries I except the Still River, whose mean TSS concentration of 25 mg/1 was I used in the model . The PCB concentration of the upstream and tributary inflows was es I timated from limited available data. Based on sampling and analy sis of PCBs and gaging of flow at Andrus Road, Stewart Labs estimated the average annual PCB concentration to be 0.024 ug/1. Ac- I cording to Stewart Labs, the relationship between suspended solids and flow and assuming the solids-dependent partition coefficient I and a constant particulate PCB concentration, the total PCB concen tration was virtually constant over the range of flow. The total I PCB concentration for the upstream boundary used in the model cali- I bration was 0.02 ug/1 . I 6-52 I Lawler. Malusky SkcIIy Kn^ineers I PGE00047238 I I No data were available on the PCB concentration of the water column for tributaries to the Housatonic. The USGS data presented in Sec I tion 6.2.1 for Connecticut, Massachusetts, New York, and New Jersey showed the percentage of detectable PCB measurements for the diver sity of waterways to be less than half that of detectable PCB mea I surements at Falls Village and Gaylordsville, Connecticut, from 1979 through 1980. The total PCB concentration of tributaries in I the model was set at 0.008 ug/1 except that the Still River, which showed relatively high PCB concentrations in bed sediments, was set I at 0.010 ug/1. In summary, the upstream and tributary suspended solids and PCB concentrations used In the model are: I UPSTREAM STILL RIVER OTHER TRIBUTARIES I Suspended solids Flow-dependent Constant Constant Average TSS (mg/1) 22 25 18 I Total PCBs (ug/1) 0.020 0.010 0.008 Dissolved PCBs (ug/1) 0.004 0.003 0.003 I Particulate PCBs 0.38 0.28 0.28 (mg/kg) I 6.3.3 Model Calibration I The process of testing the model for accuracy and adjusting param 1 eters within reasonable bounds to attain agreement with field data is known as model calibration. Available data on sediment and PCB transport in the study area were compiled as the basis for calibra I tion. The most intensive sampling of suspended solids and PCB con I I I 6-53 I La\vlor Matuskv J> I PGE00047239 I I centrations was conducted during an 18-month period from April 1979 through September 1980 (Frink et al. 1982). This 18-month period I was therefore simulated by the model and its results were compared I with the available field data. Data such as upstream boundary and tributary flows for these 18 1 months were input to the model and a linear interpolation was used in the model computations at a time step of 0.05 days. Average monthly flow data for six USGS gaging stations (listed in Table I 6.3-1) were used to construct a flow balance for the system by estimating ungaged tributary flows and interpolating/extrapolating i mainstem Housatonic River flows on the basis of drainage area i (Table 6.3-6). The model results, presented graphically for all segments in an up i stream to downstream sequence, are spatially continuous; that is, only the concentration at the upstream boundary of the first seg ment was set as input. Concentrations of all other segments were i computed. The displayed model results are the outcome of certain adjustments to the initial parameter evaluations described in Sec i tion 6.3-2. Sensitivity analyses were performed to determine the degree of influence that certain parameters have on resulting su i spended solids and PCB concentrations. i 6.3.3.1 Solids. Suspended solids concentrations measured during the 18-month period at the following locations were available for i comparison with model results: i i i i 6-54 i Lavvler Matusky vikelly K PGE00047240 I I I I »- r- »- f%* (M I I I I

~~^^feO^«*k£i/so O-o-oiNKtoo-r^-. v- r- *- r- f\t f\J I I ^O 3~~

~~o o -T in o o O O vi N-i — O I to eo BO o •- ^t I 1 coaoohOirtmocQ ^- •-•-»- r>j I~~

~~--*-»— f\» fM OJ Kl~~

*O C ~ I sc s

~~Noo- N-00 Km> 4St °co o r>j K\ ec KI •- r- r- •- rg I~~

s i — o c I £ m M * * ~ !ll C ^ £ ^ > it O fl) fl O 3 »J H 00 LJ U. LJ CO X I I PGE00047241 I I SAMPLING MILE MODEL SAMPLING I STATIONS POINT SEGMENT INTERVAL Canaan 80.0 2 Monthly Falls Village Dam 75.9 2 Daily I Gaylordsville 50.6 5 Daily New Mil ford 46.0 5 Monthly Brookfield Center 37.0 6 Monthly I Riverside 25.1 7 Monthly I A depth-integrated sample was collected daily and analyzed for sus pended solids concentration at Falls Village and Gaylordsville. In i general, only one depth-integrated sample per month is available for the other four USGS stations. Figure 6.3-10, the graph of sus pended solids concentrations In segment 1, shows the inputted i upstream boundary labeled as "observed at Andrus" and the model result for segment 1. The model results for segment 2 show good i agreement with the measured data. The comparisons of the model results and measured data for segment 5 are generally close. The i model shows an attenuation in suspended solids concentration for segment 6 that is also demonstrated by the measurements in Lake Lillinonah at Brookfield Center (Figure 6.3-11). The model- i computed suspended solids concentrations for segment 7 are gener ally lower than those measured in Lake Zoar at Riverside. Because t the sampling station is located closer to the lake's headwater than to Stevenson Dam, the field data reflect a detention time at the i sampling point that is less than half the full segment detention time used by the model to compute the solids settling flux. This i explains why the model results are slightly lower than the observed data in the last segment.

i The computed solids concentrations of the bed segments remained relatively stable over the 18-month simulation as expected in the i prototype system. The settling, resuspension, and burial rates as i well as the depths of the active bed sediment used in the model i 6-55 i Lawlor Malusky Gkelly PGE00047242 FIGURE 6.3-10 Suspended solids colibration

~~S*gm«nt 1~~

~~Jtoi tt «ut 1O~~

~~Dot. Dot.~~

Otu«v*d • Anoni*

~~- Uoi)*l~~

~~100~~

~~£ 60 ~~

~~•o T>~~

~~Dole I~~

~~Uodd + ObMivtd • Canaan FVD~~

~~Model I~~

~~I 6-55A I PGE00047243 FIGURE 6.3-11 I Suspended solids calibration I S*gm*nt 5~~

~~I 100 100 BO M I BO eo 70 70 ~~

~~10 M ~~

~~I SO so i I X 20 I 10 0~~

I Date Date I + Ob«nr«d • Ulllford • Ob»«rv«d • Coylord. • Obwvtd • Brookfld I Uadtl I I I I I I o Rvrald* I Modal I 6-55B I PGE00047244 I I yielded stable solids concentrations in bed segments. The overall comparison between the model and field data is deemed adequate for I an acceptable calibration of suspended solids.

6.3.3.2 PCBs. Sampling and analyses of PCB concentrations in • Housatonic River water during the same 18-month period were per- « formed primarily at high flows at the Falls Village and Gaylords- K ville locations. The sum of the Aroclor concentrations reported by Frink et al. (1982) for the whole water and filtered samples (re- I ferred to as "total" and "dissolved," respectively) is listed in Table 6.3-7 along with the more recent data collected at the sam- I pling stations recently installed at Kent. Six of the 14 observa tions at Falls Village and one of 13 at Gaylordsvil le in 1979-1980 • showed a dissolved PCB concentration greater than or equal to 0.1 • ug/1 (detection limit) and a total PCB of 0.0 ug/1 - an obvious _ contradiction. These contradictions indicated that the particu- | late/dissolved PCB concentration breakdown is questionable and led to the consideration of the total PCB data alone. I

The total PCB concentration was found to be at or above the detec- ft tion limit in 43% of the samples taken at Falls Village and 157= at Gaylordsville in 1979-1980. These data are of limited usefulness • in defining precisely the total PCB concentration for model cali- P bration. Nevertheless, the water column PCB concentration appears to have been unusually lower than the detection limit and probably | to have decreased with distance downstream. I The more recent PCB data collected at the Kent sampling station, which was established cooperatively by USGS , CDEP, and GE as an • outcome of the HRA, have all been below the detection limit. The only other sampling during the past three years (besides that at • Kent) was at Falls Village, and a PCB concentration of 0.1 ug/1 was I found. As a means of analyzing for temporal trends in PCB concen

~~6-56 l.a\vl<".r Matusky Skcllv Knijinoors I PGE00047245 TABLE 6.3-7 I PCB CONCENTRATIONS MEASURED IN WATER COLUMN OF HOUSATONIC RIVER IN CONNECTICUT~~

I PCB CONCENTRATION tug/1) FLOW TSS I STATION DATE TIME TOTAL DISSOLVED3 (cfs) (mg/1) Falls 100479 1230 0.1 0 3240 Village 112779 940 O.I 0 1850 24 I (MP 75.9) 31880 1115 0 0.1 2660 100 31880 1215 0 0.1 2820 31880 1315 0 0.1 2985 31880 1415 0.1 0 3150 I 31880 1515 0 0.1 3310 210 32280 820 0.1 0 7550 242 32280 930 0 0 6920 I 40480 1230 0 0 2180 19 41080 930 0 0.1 4740 41080 1230 0 0.4 4740 I 41080 1550 0.2 0.1 5500 128 63080 1315 0.2 0 1110 12786 1430 0.1 0 4740 130 I Gaylords- 100479 1500 0.1 0 4000 ville 112779 1250 0 0 3780 43 (MP 50.6) 31880 1100 0 0.1 3390 I 31880 1200 0 0 3560 31880 1300 0 0 3900 31880 1400 0 0 4250 31880 1500 0 0 4590 I 32280 1040 0.1 0 4930 208 40480 1315 0 0 14800 870 41080 1045 0 0 3790 I 41080 1300 0 0 9190 41080 1500 0 0 9190 I 63080 1245 0 0 9190 Kent 60284 1530 0 0 15800 116 (MP 58.1) 60284 1100 0 7520 142 60284 1130 0 7520 188 I 80185 1200 0 0 2550 114 82785 1115 0 0 264 48 92885 1100 0 0 3640 141 I 92885 1330 0 0 3930 164 12786 1115 0 5635 150 60686 0 3865 130 I Source: USGS, 1987 Hartford, CT. I Note: PCB concentration lower than detection limit (0.1 ug/1) is listed as 0. I aSample water passing through a 45 urn pore diameter filter. I 6-56A I PGE00047246 I I tration, the pooled 1979-1980 data, showing detectable PCBs in approximately 30% of the samples, are compared with the pooled I 1984-1986 data (collected primarily at a sampling station between the 1979-1980 sampling stations), showing only 11% of the samples • having detectable PCBs. This indication that PCB concentration is • probably decreasing with time is consistent with the temporal trend _ in fish data and will be relevant 1n the modeling projections dis- | cussed 1n Section 6.3.4.3. I The model results for total PCB concentration in water segments and the participate PCBs in the water and underlying bed segments are I displayed in Figures 6.3-12 through 6.3-15. The extent to which the measurements for PCBs can be used to calibrate the model is: I • Total PCB concentrations computed by the model fall into the below-detection-limit range that • resulted from approximately 70% of the measure- | ments during the period. • Th decrease e in PCB concentration from segment 2 I to segment 5 agrees with the decrease in observed • PCB concentration from Falls Village to Gaylords ville. •

~~Unfortunately, the PCB measurements in Connecticut were not per formed at a sufficiently low detection limit to allow for more I rigorous comparisons.~~

According to the model, the particulate PCB concentration in the I bed is in the same range as that of the overlying water column for • segments 1/8* through 5/12. In contrast, the particulate PCB con- • centrations in the water columns of Lakes Lillinonah and Zoar are I

*Notation refers to pair of water column and bed sediment segments, I e.g., segments 1/8 refers to water segment 1 and underlying bed segment 8.

~~6-57 I Liiivlcr Matusk oki:ll\ K I PGE00047247 FIGURE 6.3-12 I PCB Calibration Results I In Water Column and Sediment~~

~~Segment 1 Segment 2 Total Totol I O.M o.ow •.021 I 0.014 O.OH • •^ t.n • •x. I 0> 0410 ^ "" 0.011 ID ID O 0. a. 0.014 0.01] I CJ1 •~~

~~I 0.004 o.on o Imb-tO «pr-W tof-00 I Date Date I I~~

~~S«gm«nts 1 & 8 (WC Bed) Segments 2 & 9 (WC& Bed) I Partlculate Portlculate~~

~~1.4 I U ~~

~~m I CJ a. 0.0 0.7 O.I I ro OJ a. 0.4 OJ 0.3 I o.t ~~

~~0 Aw-7* Oe*-7» 0«e-7« *a»-»0~~

~~Date I Wotw Column + B.d I I I 6-57A PGE00047248 FIGURE 6.3-13 PCB Cal bration Results In Water Column and Sediment • Segment 3 Segment 4 H Total Total H U n M ^—^_^ —^_____^_ — .~~

~~N OO2I ~~

~~M • OO24 ~~

~~14 0034 • •n • ami n •~~

11 0> 0011 • "A /A 1 14 • °°" " ^ /\ \ / \ rj 14 j£ OJ14 • ni - \ /V / \ 'o 0.011 \• A V\ / \\ / \\ "o \ A. / \ «i - t OJ)I • \ / ^~~\ / \ at • V-w/ \yV V^~^__ OJXM M OAH v v_ 104 OJ04 • 1 101 0.002 •

0 |iii|>ip|iii(iit[ii»|iii|iii;«ii|il ..T.,,T,.,1pl ,^,,.,, • i | T r * | r •» i •• | i i-r- | TT T j F -r » | • • • 7 • ' ^^ Apr- 71 Jim— 71 Auo-7t 0*t-7* D**~7* r«o— 10 A*r-tO Jw-M Auf-00 V- Date Dote |

~~Segments 3 A 10 (WC 4 Bed) Segments 4 Ac 11 (WC & Bed) M Participate Particular H 1.4 i• .. ^™ u 14 ~~

~~1.4 1.4 IJ tJ 1 IJ r~, IJ " ' 1.1 \^ 1.1 ~~

1 m u OJB 5 » 1 a. OJ ^ OJ V 0.7 •5 =-7 a 0 »•• ?o OJ 1 " 1 a. 0.4 0- .0 4 . /A N ' OJ \/~^—"~ '~^vw ^ ~~ 0.1 0.1 1 01 01 n tf-n Jon-7l *u^7» Od-71 0~-71 F*-«0 If-tO Jun-H «uf-iO Vh-7» Jw»—7 > Auf-7ft Orf-Tt Dae— ?• Fafe-SO Apr— €0 JLH— «O iug-M flB Date Date • Water Column 4 Bod *olor Column - B.i

~~1 6-57B 1 PGE00047249 FIGURE 6.3-14 I PCB Calibration Results I n Water Column and Sediment~~

~~Segment 5 Segment 6 I Totol Total IJHI~~

~~0.031 *.aru I 0.014 *.«4 O.OM ojn tun I 0.011 •C noil 0.011 • ? Uli m o 0.014 • ID a. o 3 ojni a. £ OJJI I 1 0.000~~

~~I OJ04 0.005~~

~~Apr-7« Jui-71 log-n CM-71 I »-7» r«fc-w VC-M ju>-«o lut-ia tfi-n laf-tO I Date Dote I I~~

~~Segments 5 & 12 (WC fc Bad) Segments 6 Si 13 (WC & Bad) Portlculote Portlculat* I l.i IJ ~~

~~1.4 I IJ IJ ~~

~~1.1 1 I OJ m o u o. 0. 0.7 0.7 ~~

~~O.I 0.1 I OJ OJ ? o 0.4 0.4 OL OJ ~~

~~0.2 I 0.1 ~~

~~tfr~tt Ju>-n luf-71 Ort-Tf D»-7I r*- Jun-U iuf-iO -71 Jio-n tuf-n Ort-7» Du-71 I -10 v—'f~~

~~Dote Date I - Watv Column •> B«d Water Column -I B.d I I I 6-57C PGE00047250 FIGURE 6.3-15 PCB Calibration Results I In Water Column and Sediment I~~

~~Segment 7 Total I I I I I I I I~~

~~Segments 7 it 14 (WC tc Bed) Parttculot* I I \ t.1 • m Sj§ I O. M « 5 "-7 3 « t ,. o I »• 0.4 OJ~~

~~D.2 D.1 I~~

Date - WaUr Cokicnn •» B.d I I I 6-57D I PGE00047251 I I substantially lower than their respective beds so that the PCB concentrations of the lakebeds decrease monotonically over the I 18-month period. This examination of participate PCB concentration provides a glimpse of the rates of change in sediment PCB concen- I trations for the upstream and downstream portions of the river, to I be projected later. The model sensitivity to three parameters that affect transfers of I PCBs was investigated individually to further test the model's validity. If the partition coefficient were an order of magnitude lower than that used in the model, PCBs in the water column would I be approximately 70 to 80% in the dissolved phase as opposed to the 20 to 30% range that is more typical of surface waters (Schroeder I and Barnes 1983). In addition, the particulate concentration of the water column would be substantially below the bed's particulate I concentration in all segments. Since these differences were considered unlikely, the solid-dependent partition coefficient based I on the literature as described in Section 6.3.2.4 was deemed appro priate for the model.

I If volatilization were zero throughout the system, total PCB con centrations in the water column would be approximately 10 to 15% I greater. Although the precision of PCB measurements does not per mit testing this hypothesis, the inclusion of volatilization, which I is documented in the scientific literature (Chapra and Reckhow 1983), casts minimal effect on the model results.

I Finally, the elimination of diffusion of dissolved PCB from the bed to the water column has a negligible effect on the computed re I sults. I I I 6-58 I Liiwlor Malusky Skelly Kncrineurs PGE00047252 I I 6.3.4 Model Projections of PCBs I The usefulness of the model as a tool for projecting PCS levels in Housatonic River sediment and water hinges on setting reasonable conditions for the driving forces of the system. Two such impor I tant factors are the river flow and the upstream PCB concentra tion. Evaluation of the expected river flow and total PCB concen I tration of water influent to the study area is discussed in Sec tions 6.3.4.1 and 6.3.4.2. The model projections of PCBs in sedi I ment and water are presented in Section 6.3.4.3, and the extrapola tion of PCB levels in fish is discussed in Section 6.3.4.4. I 6.3.4.1 Long-Term Hydrological Period. Because river flow affects the PCB entering the Connecticut portion of the Housatonic River I from Massachusetts and also controls the resuspension of PCBs, we looked for any cyclical trend in hydrology that might regulate I long-term PCB levels. Housatonic River flow data at Falls Village, recorded by USGS from 1912 through 1986, were analyzed by a multi I ple periodic regression of monthly mean flows (Figure 6.3-16). No significant long-term period was found. For example, periods of 20 and 30 years account for approximately 1% of the variance as com I pared to the well-recognized annual (one-year) period that accounts for over 22% of the variance. Hence, the mean monthly flows at I Falls Village computed from the 74 years of data and used for the projections are: I FLOW FLOW I MONTH (cfs) MONTH (Cfs) Jan 1105 Jul 571 Feb 1101 Aug 505 I Mar 1974 Sep 513 Apr 2463 Oct 569 May 1385 Nov 906 I Jun 892 Dec 1078 I 6-59 I L.wler Maluskv Sknllv hninru I PGE00047253 I

~~I o I en~~

~~o I oo 01~~

~~I w I o »-J t—i > I o I K en~~

~~CO Oin IUJ CT) oc o Iu. o I I rt- CJ~~

I O H I CO D O I a I I q O I to" in cv (spuosnoqi) I (sp) I 6-59A PGE00047254 I I Like the model calibration, the upstream boundary and tributary flows for each month were estimated on the basis of drainage area • (Section 6.3.2.8). A 50-year period was selected for the modeling projections as a reasonable time span to study the river's natural • recovery under the no action plan. Although the field data that * established initial conditions for the model were collected from _ 1979 through 1986, for simplicity, time zero is assumed to be 1986. I

6.3.4.2 Upstream and Tributary Inflows of PCBs. Because the total I PCB concentration of the Housatonic River near the Connecticut- Massachusetts border is generally below the detection limit, there I is considerable uncertainty in estimating influent concentrations for the future. Nevertheless, the approach to do so entailed (1) determining whether the upstream source will be constant or will I diminish with time, (2) if the latter, estimating the rate of de- « crease, and (3) estimating a minimum boundary PCB concentration | that upstream and tributary waters will not be below.

The data-supported decreases in PCB concentration of the Housatonic River's water, bed sediments, insects, and fish indicate that the I upstream boundary PCB concentration will continue to decrease. The temporal trend in PCB concentration was assumed to decrease expo- • nentially with a time rate coefficient of 0.05/year. According to • the literature on background PCB levels (Section 6.2), the presence _ of PCB in remote, nonindustnal ized areas suggests that there is a | minimum concentration for influent surface water. As PCBs in the environment become more and more dispersed through fluvial and I atmospheric transport, the tributaries as well as the upstream boundary are assumed to reach a minimum boundary PCB concentra- I tion. This minimum total PCB concentration (Cf w) ^or the up \ L t W / i~ stream and tributary inputs to the model was set at 0.002 ug/1 and • is reached in 28 years for the tributaries and in 46 years for the • upstream boundary (Figure 6.3-17). To provide more conservative or I 6-60 I Liiwlor Maluskx Skelh Fruimores I PGE00047255 FIGURE 6.3-17 I Decaying PCB Boundary qn_d I Tributa ry Con centrati ons I Upstream PCB Boundary Conditions I I

~~m o a. I Minimum BC(C, J-.OOSug/l | I I I I~~

Tributary PCB Boundary Conditions I 0.02 0.019 0.018 I 0.017 0.016 0.015 I 0.014 0.013 0.012 0.011 ID I O 0.01 0. 0.009 0.008 0.007 I Ulnlmum BC(Ct>w)-.005ug/l 0.006 O.OOS 0.004 I 0.003 0.002 0.001 I 0 I I 6-60A PGE00047256 I I cautious projections, a second model projection was performed with the minimum boundary concentration set at 0.005 ug/1 and a lower I rate of decrease.

The only other change in model input from the calibration was to use the average suspended solids concentration of 22 mg/1 for the upstream boundary rather than the monthly variation in order to I avoid lengthy input data files for the 50-year simulation. Com parison of the constant and time-varying solids boundary condition showed a negligible difference in the results over several years. I 6.3.4.3 Projections of PCBs in Sediment and Water. The modeling projections can be viewed as an extension of the model calibration, • which was a simulation of 18 months, to a 50-year simulation. The results are presented in terms of the commonly measured forms of • PCBs, namely, total PCBs in the water column and particulate PCBs I in the bed sediment. The trends in PCB concentration projected for the Housatonic River from the Falls Village impoundment (segments 2 J and 9) through Lake Zoar (segments 7 and 14) are presented graphi cally in Figures 6.3-18 through 6.3-20. Results for the very short I segments 1 and 8, which are not presented, are similar to segments 2 and 9. Concentrations are projected to decrease monotomcal ly • with time for all segments except the bed sediment in the Bulls ™ Bridge impoundment and the next downstream segment (segments 11 and • 12), where the particulate PCB concentration of the suspended I solids is initially greater than that of the bed sediment. In view of the scarce PCB data available for these segments and the short I time (less than seven years) projected until these bed sediments reach their maximum, overall declining PCB concentrations are ex- • pected throughout the study area. The percentage reductions in water and bed PCB concentrations were arithmetically averaged for • segments 2 through 7 and 9 through 14 to provide a riverwide out- * look. The time to reach 50'/. of the initial PCB concentration is I 6-61 I Lawlor Matuskx Skcllv I [mincers I PGE00047257 FIGURE 6.3-18 I PCB Projection Under No Action Plan

~~I Total PCB in Water Column~~

~~I Segment 2 Segment 3~~

~~I m 01 o o o. a. I~~

~~I Year* from 1986 Years from 1986~~

~~I Particulate PCB in Bed~~

~~I Segment 9 Segment 1 0~~

~~I I n.i oj 0.7 CD ID I u U a. 0.1 OL flj I 0.4 o a. OJ~~

~~0.1~~

~~I 0.1 Minimum BC(Ct|»)-.002u9/l~~

~~I Years from 1986 Years from 19E6~~

~~6-D IM I PGE00047258 FIGURE 6.3-19 I J?rojection Under_ No_Ac_t_ion Plan I Total PCB in Water Column Segment 4 Segment 5 I I I m ao. Minimum BC(Cj w)-.005 ug/l : I I I Yean from 1986 Years from 1986 I~~

~~Particulate PCB in Bed I~~

~~Segment 10 Segment 1 2 1.1 I 1.1 I I o.t I OJ ~~

~~a> m 07 o o I a. a. O.t ~~

~~OJ '•e 0.4 O I Ulnimum BC(Ct w)-.005 ug/l I CL OJ - Ulnlmum BC(Cj w)-005 jg/l I Ulnlmum BC(G,4W)-.002u9/l~~

~~Years from 1986 Years from 1986 I I I 6-GIB I PGE00047259 FIGURE 6.3-20 I PCB Projection Under No Action Plan~~

~~I Total PCB in Water Column I Segment 6 Segment 7 I I ID o a. I 1 Ulnknum BC(C|.J - .OOSug/l i Minimum BC(C,,J - .005ug/l I~~

~~I Year* from 1986 Yeore from 1 985 I I Particulate PC3 in Bed I Segment 13 Segment 14 I~~

~~m CD I o U a. 0. Minimum BC(C(.J - .OOSug/l !~~

~~Minimum BC(Ct.«) - .OO5ug/l I o a.~~

~~Minimum BC(C.,J - .002ug/l I Minimum BC(C,.J - .002ug/l~~

I Years from 1986 Years from 1986 I I 6-61C I PGE00047260 I based on the average of the projections using the higher and lower I boundary conditions. A reduction of 50% of the total PCB concen- I tration of the water is projected for 17 years from 1986 and the same reduction in the particulate PCB concentration of the bed is • projected for 22 years. The exponential PCB decay rate that re- • suited from the model computations is approximately 0.04/year, _ which is equal to the average of the decay rates used for the up | stream boundary PCB concentration. An 80% reduction in the PCB concentration of the water and bed is projected in the year 2036. | The reduction in PCB concentrations of the water and bed are close to equivalent over the 50 years, and a combination of both shows I the decline in PCB levels to which the fish are directly and indi rectly (through the food chain) exposed. I The upstream loading of PCBs is a greater source than the feedback _ of PCBs from bed sediments via resuspension and diffusion, particu- | larly in the early stages of the projection period. The difference in the total PCB concentration of the upstream boundary increasing- • ly reaches 0.003 ug/1 (at 50 years, difference equals 0.005 minus 0.002 ug/1), or 60% of 0.005 ug/1 (the greater of the minimum boun- • dary concentrations). This difference between the upper and lower estimates of influent PCB concentration produces a difference of • approximately 15% in the projected percentage reduction in water • and bed concentrations at the end of the projection period. I The relative importance of the external (upstream and tributary) loadings vs the sedimentation and burial of PCBs in the river's re- I covery was investigated by keeping the external PCB loadings con stant over the 50 years. In this case a minimal change in PCB con- • centration of the water column is projected, and only the bed sedi ment concentrations of Lakes Lillinonah and Zoar would decrease • substantially due to sediment burial. Hence, the projected re- • covery of the Housatonic River in total relies primarily on the I 6-62 I Laulor Matuskv Skellv I'ncrmrcrs I PGE00047261 I I diminishing source of PCBs from upstream. The recovery of the two lakes also depends on sediment burial and declining tributary load I ings. I 6.3.4.4 Extrapolation of PCB Concentrations in Fish. The ap proach to projecting PCB concentrations in fish is to extrapolate I from the model projections and also to use the regression analyses that evaluated the temporal change in Housatonic River fish from I 1979 through 1986. The direct relationship between the PCB concen tration in fish to the PCB concentration in the water and/or sediment to which the fish 1s exposed is generally well accepted (EPA I 1983). Substantial effort has focused on quantifying bioconcentra tion factors (BCF) or ratios of fish PCB concentrations to the I exposure concentration, generally through experimental work. Three such bioconcentration relationships were applied to a range of PCB I concentrations for Housatonic River water and sediment representative of the past several years (1979-1986) to examine their appli I cability (Table 6.3-8). Veith et al. (1979b) found the BCF to be related to the octanol- I water partition coefficient through laboratory experiments. For an order of magnitude range in the PCB concentration of water, the PCB I concentrations of fish are estimated to be from 0.6 to 5.9 ug/g, which underestimates the actual data (Section 6.2.3.1). Schnoor I (1982) also related the BCF to the octanol-water partition coeffi cient, but used field data to empirically derive a 1ipid-normalized BCF. The resulting range in fish PCB concentrations for a range in I lipid fractions representative of Housatonic River fish is similar to the actual data. This is probably attributable to Schnoor1s use I of field data that include effects of biomagnification as well as bioconcentration; hence, the term "ecological magnification factor" I may be more appropriate. McFarland and Clark (1986) combine two I equations relating the BCF and the sediment-water partition coeffi- I 6-63 I Lawlor. Malusky okelly Kntrinec;r.s PGE00047262 1

~~o O it)~~

t— QJ 1 CD • 4-* QJ •- u. it) f 3C S z CO cn r~\ f\j CD LO CD ^" OJ ro c; LU •-4 CT IT CD to o ^-4 . n r^ 4-> a) a) Z CT 0 LO f^ p^ <^ f^ £3 ^-< o co cr -u • 0 Z 3 .— ID in O • c_> u, C\J QJ LJ QJ • a> .-4 L L ^ CO 0 4-> a. c: a5 o IT5 u, QJ O Cu 3: o L. _o _c • T* to z . c o a) — CT> O ^1 -a u. ZD 0 i—4 t—* u c c: 4-J • •—1 • * CZ QJ (O C • CQ I— t— •— 0 0 o r-~ o aj • O ^c Z CT II II 0 TT C CQ O • Cu ry • O CJ C h- LU — •3: < < < <: < o o co to a cu o Z ^4—4 CT3 •2 :z zzzz <• <* o ° »» LU a «— Q_ II 4-1 QJ • LU LU -C o o C > CO • to o 0 O O f QJ -r O • O 4-> O E 4-J Q. 0 o f- 4 i—4 i^ • -i— (O

1. CO ^_ to CT) OJ QJd C •)• o o p— •»— O > i/l QJ rd ^| Cu o; —. U• > LO LO LO LO It- r— -r 10 • to LU CT C> LO O O O CD cr: 73 QJ o H LU LU 1— 3 c> o oooo ^ O L_ i—( O t/> rn) L^l 1— Cu 1_ •r O -*-J ^C *s* f— ^B to X > T3 4-J 4-J U. rO 4-J 4-J VJjV LU LU L* (O o o in c TJ 4-J L (O •— 3 • QJ O ^^ C 4J <4 O O .—.£ in 0 OJ C O 1C «/l CT> • c QJ c: L. in QJ "O • h- OJ O O «f u, QJ in QJ • T3 C T O i— 4J !• ce 0 -D i_ .i 0) O 4-> u, cr •- r— LU 0 CT •i- (O to cn co *J i\ rj i_ • O •r— cj c: • . ~—• QJ QJ • z i a> aj a CJ Q. aj 4-J UD > in • o 3: 4-> CT CT> -v. O u c cr • o ^ o 1^ *^« ^^ CT H- i»- c: QJ ii o u. -C o Z ^O 4-> S CT CT> 3 0 O -r ^ 4-J ao VX QJ CT1 U 4_l ^4-> L ^ cc •—4 •» cn o o ^" o o) o *4— ^H cn ~~ II •r -r 14- ^ 1_ U_ O • < Cu 0 CTc n 4-J fl (4— 4-J ^1 LL a: «-4 r^— ^-™. 3 O .C ra QJ cn c >s ty" i— i CT o CT> u. "O O O QJ • CD ^c Z CO ID 3 O 4J •^* QJ 4_) . U 1- .—1 CO LU Z CO • -C OJ M C r3 • i o o • 1 CT u, -t— c QJ O O >) • 0 CD UD I U) ••— O •— O E <-> 1^1 r— • z O 1— C3 T-< jd O 0) *j ra •• 3 o H o f— c/l o C fll • o 4—4 LU -II C M 4J **— o 4-J ^C f~^ 03 E QJ »—• CD CC U. f ~ 5 f^ QJ CT) h- O '- • O • J »—4 CO -* ^ •—4 H- •*—• II CT ^ or II — II -. a. 4J D. V_ CD <-) O CL ^ cn Ll- LL. CT ra-i-OJ • ^ C CX T3 o_ 0 CJ O M ^ 3 •— •— cO CO (j| Lu 4J (d n3 '— 4-J • CT) C II J IICT 1r~ ~ QJ to ^1 CU 1 U. r L ,- • u) .— -— S O —I rtf •— ^ C Li_ O O • *—^ O CJ C ^^ •— m i/) QJ (O J^ —. 4J T3 L -^ O 4-> O>4-» -i^ O O H LU COOO — >— — J4-J4J ^ CU O> ^ 1— CM r— O tO rd L rv O O CO ~O u, 13 T3 II II II CC J= CO o CD t— (^ ~o cr* LU [J t~ »—4 QJ Ii 1— r— jQ Lu 7 O ^ Lu ^0 *~ •*—" 'r— tj 4O ~~~* f& O 0 0 • Lu CJ —1 U Lu ID —1 CD ^ L>_ • Ct >• *— CO ^^ Z • rO -Q 0 • I 6-G1A I PGE00047263 I I dent to the octanol-water partition coefficient and derives a relationship between the lipld-normalized fish concentration and the I organic-carbon-normalized sediment concentration. Since McFarland and Clark's equation is essentially based on laboratory studies of I bioconcentration and do not include biomagnification, the estimated I PCB concentrations for fish underestimates the actual data. The foregoing leads to the conclusion that direct bioconcentration and indirect food chain biomagnification both influence PCB concen I trations in Housatonic River fish. The relative contributions from the two pathways cannot be quantified accurately using these simple I relationships, but would necessitate modeling PCBs through the food chain. However, Thomann and Connolly's (1984) food chain model of I PCBs showed that the concentration of PCBs in lake trout is linear ly related to the dissolved PCB concentration 1n the water. That I is, the reduction in fish body burden is directly proportional to the reduction in dissolved PCB concentration. If indirect PCB exposure through food were predominant over the water route, then I the only difference from direct water predominance would be a lag in the time it would take for the fish concentrations to decrease I due to the purging of PCBs from organisms in lower trophic levels. This lag of a few years would be relatively minor 1n relation to I the 50-year projection period. I Because the fish of interest include bottom feeders, foragers, and predators, the average percentage reduction in PCB concentrations of the water and bed is used to extrapolate the PCB concentrations I of fish. The geometric mean PCB concentration of each fish species, based on 1979-1986 data, is projected to decrease according I to sediment and water reductions (Figure 6.3-21); any lag time is implicitly accounted for by using 1986 as year zero. For example, I the brown trout and white catfish, which have geometric means of I 3.16 and 3.33 ug/g, respectively, will reach the FDA action level I 6-64 I Lawler Matusky Skolly K PGE00047264 FIGURE 6.3-21

~~PCB PROJECTIONS UNDER NO ACTION PLAN I Average of Model Segments I Z 2 rs I _i MIN BOUNDARY Ct w = .002 ug/l O O £E UJ I~~

WIN BOUNDARY C, w= .005 ug/l I CD O Q_ I I O Z o p I o o UJ DC I I Z UJ S O I UJ MIN BOUNDARY C, w = .002 ug/l w o UJ CO I 2 00 o .OO5 ug/l 0. I UJ I o a. I u. O Z Q I i— o D O UJ I DC 20 30 I TIME, < years) 6-64A I PGE00047265 I

I of 2 ug/g in approximately 12 years. Limited fish data on the white crappie and American eel (geometric means of 6.49 and 8.08 I ug/g, respectively) indicate that approximately half these species will have lower concentrations of PCBs than the FDA action level in I 35 to 40 years. In the year 2036 the geometric means of all fish are projected to decrease to approximately 20% of their present I concentrations. Decreases in fish concentrations based on WASTOX projections are I lower than those found in the statistical analyses of Housatonic River fish data presented in Section 6.2.3.1. For example, the I multiple regressions of 1979 through 1986 data showed an average half-life, or time to reach a 50% reduction in PCB concentration, I of approximately four years (or somewhat longer for largemouth bass and Lepomis spp. when the suspect 1983 data are deleted). An I explanation for this is that a large portion of the fish analyzed in 1979 were exposed to higher PCB levels from GE's direct PCB dis charge before the mid-1970s as compared with a smaller portion in I 1983-1986. The abrupt decrease in PCB levels around 1980 is naturally greater than the gradual recovery projected over the next 50 I years. Signs of this may already be apparent: when the 1986 and 1984 fish data are compared, they show a leveling of PCB concen I trations. Therefore, the model projections are considered appro- I priate and reasonably conservative for the long term. An alternate view focuses on the percentage of fish in species of concern that will exceed the FDA action level in 50 years. Fre I quency histograms of selected species of fish with relatively high PCB concentrations show considerable variation in the percentage of I the sampled species that exceeds the FDA action level among the four years of data collection (Table 6.3-9). This variation is I largely attributable to the random catching of very old and large I fish (see Appendix A for scatter plots of dat?). The natural mor- 6-65 I Liiwler Malusky Sknlly E I PGE00047266 1 1 TABLE 6.3-9 1 PERCENTAGE OF FISH ABOVE FDA ACTION LEVEL OF 2 uq/q 1 1979 1983 1984 1986 ALL American eel 100a 50a 88a 1 Brown bullhead 45 9a 16a 50b 33 Brown trout 33b 52 96 67 1 Carp 54 80b 57 1 Largemouth bass 15 Oa 26 18 Lepprnis spp. 7 Oa Oa 4 1 Smallmouth bass 80 Ob 22 39 38 White catfish 85a 20b 56 74 66 1 White crappie 100b Ob 91 1 White perch 97 17a 36 47a 51 Yellow perch 15 Oa 15 3 10 1

~~aBased on 25 or fewer specimens, bBased on 10 or fewer specimens. 1 1 1 1 1 1~~

~~G-65A 1 1 PGE00047267 I~~

I tality of these old fish, such as carp and white catfish from 10 to 17 years old, is an additional mechanism for decreasing fish con I centrations, which should be superimposed on the model projec tions. The natural mortality effect of these old fish was esti I mated along with the 80% reduction in exposure concentrations pro I jected by the model. The percentages of certain species expected to exceed the FDA ac tion level in 2036 were estimated by a review of the data for indi I vidual fish with PCB concentrations greater than 10 ug/g. The sta tistical interpretation of this concept 1s that the variance of PCB I concentration about the mean, i.e., the distribution of PCB concen trations, will decrease proportionally to the decrease in the mean I concentration. Because this natural mortality effect, which is probably reflected in the 1979 through 1986 data, has a key impact I on the upper tail of the PCB distribution, an extension of the sta tistical approach was also used. Based on the regressions of the In of PCB concentration vs time and the coefficient of variation, I the percentage of fish exceeding 2 ug/g was also estimated statistically and found to be similar to the previous model extra I polation. I The resulting average of both predictive methods are summarized:

•/. >10 ug/g % >2 ug/g I SPECIES 1979-1986 PROJECTED 2036 White catfish 15 5 I Brown trout 10 4 Carp 10 4 Smallmouth bass 4 2 I Brown bullhead 1 White perch 1 Lepomis spp. 0 0 Largemouth bass 0 0 I Yel low perch 0 0 I 6-66 I 7 I Lawler. Malusky Sk<:lly Kiitrincjors PGE00047268 I I Two species, American eel and white crappie, sampled primarily In 1979 but infrequently in 1983, did not have sufficient data for • these projections. Nevertheless, the high PCB concentrations found during those years indicate that a large percentage of these spe- • cies may continue to exceed the FDA action level 50 years from now. *

It should also be noted that a projection of no fish of certain I species exceeding the FDA action level in the future is optimisti cally based on a decrease in background PCB levels. In light of • the data presented in Section 6.2.3, citing fish PCB concentrations close to 2 ug/g in areas remote from industrial point sources, • e.g., Montezuma's Wildlife Refuge, a decrease in background levels is as important as the diminishing upstream source and sediment • burial for this ultimate outcome. • I I I I I I I I I

~~6-67 • Lawler \l;itusk\ ok<;llv Kntrin~~

I The two major uses of the Housatonic River described in the interim report are recreation - fishing, boating, and swirmiing - and hydro I electric power generation. The potential direct impact of the PCB advisory on recreation is discussed in Section 6.4.1. The discus I sion focuses on socioeconomic aspects as CDEP has already addressed the health concerns in its advisory that "warns against the con I sumption of any fish from the study area of the river" (CDEP 1984). Although the low level of PCBs in the river does not directly I affect the use of it for hydroelectric generation, there is a po tentially indirect effect that became apparent with Northeast Util I ities' proposed Bulls Bridge Hydroelectric Redevelopment Project. The PCB contamination of river sediments, and in this case sedi I ments in the power canal, caused concern about proposed plans to excavate portions of the canal. Similar situations may exist for the dredging of river sediments in conjunction with cable or pipe I line crossing. At CDEP's request in February 1986 this study was extended to encompass future developments involving sediment and I PCB movement, discussed in Section 6.4.2. I 6.4.1 Socioeconomic Impacts I 6.4.1.1 Recreational Impacts. This section describes (1) the fishing, boating, hiking, camping, and other recreational activities on the Housatonic River in Connecticut and (2) the impacts on I these activities caused by the presence of PCBs in the water, sedi- I ment, and biota. Fishing. The river supports an active sport fishery consisting I primarily of trout and bass angling in the upper riverine section I I 6-68 I Lawler Maluskv Skelly Knoince PGE00047270 I I and panfish (white perch and yellow perch) and bass angling in the lower impoundments: Lake Lillinonah and Lake Zoar. In 1986 there • were approximately 32,000 angler visits to the Housatonic River.

The 10-mile section of the river between the Route 112 bridge at • Lime Rock and the Route 4 bridge at Cornwall (Cornwall Bridge) is a trout management area (TMA) with a zero creel policy (return of I captured fish to the river) for trout. The TMA receives relatively high angling pressure (30% of total trips to the river). In the • section of the TMA along Housatonic Meadows State Park angling is restricted to fly fishing. As detailed below, high catch rates for 8 sizable trout in the TMA attract people from far away (average one- way travel is about 40 miles), Including a number from out of state • (18%, vs 77. for the remainder of the river), indicative of the • value of this section of the river. I There is heavy angling pressure in Lake Lillinonah (41% of all vis its to the river), somewhat less in Lake Zoar (17% of all visits). • Unlike the riverine section, angling at these two lakes is year- round, with an active ice fishery (6% of all visits). •

Boating. Open canoes and kayaks are popular in both the free-flow ing and the impounded sections. During the extended dry weather I periods typically seen in the summer, the average flow in the river _ is often not high enough to allow satisfactory canoeing in many of | the riverine sections. At Falls Village, however, the hydroelec tric station is operated so as to maximize weekend generation dur- • ing hours that are also popular for canoeing. By concentrating the river flow during these hours, the facility extends the white-water • season, making this river unique in the region. I I

~~6-69 • Lau'ler Mnluskv 'okt;llv Kn~~

I The section below the hydropower dam at Falls Village and south to the headwaters of the Bulls Bridge impoundment at Kent supports I most of the use by canoeists, although there are no known use sta tistics. The rapids below the covered bridge at West Cornwall are I the site of white-water canoe competition. South of Kent the river run canoeing is hampered by the Bulls Bridge, Shepaug, and I Stevenson dams; existing and planned (Northeast Utilities Service Co. 1982) portages support river run canoeing downriver to Lake I Zoar. There are no known statistics on this activity. The only known motorboat activity is on Lakes Lillinonah and Zoar. I There are several state, local, muncipal, and private boat launches on these two water bodies (Tables 6.4-1 and 6.4-2), but no compre I hensive statistics exist on the use of these waters by motor- I boaters. Camping. Public camping is provided in two state parks - Housa tonic Meadows State Park on the west shore of the river between I Cornwall Bridge and West Cornwall and Kettletown State Park on Lake Zoar. Many of the campers are believed to use the river for both I fishing and boating. I Other activities. The river supports swimming, hiking, and sight- seeing at the impoundments (Tables 6.4-1 and 6.4-2) and in the I riverine section. The covered bridges at West Cornwall and Bulls Bridge are noted tourist attractions.

I Impacts associated with PCBs. In 1976 PCB concentrations in fish collected from the Housatonic River were found to exceed the U.S. I FDA standard at that time of 5 ppm in food (the present standard is I 2 ppm). Approximately one year later, the Connecticut commis- I I 6-70 I Lau'ler, Matusky ' Skelly Engineers PGE00047272 I I TABLE 6.4-1 EXISTING RECREATIONAL FACILITIES ADJACENT TO LAKE ZOAR I

RECREATIONAL FACILITY MAJOR I BY TOWNSHIP ACTIVITY OWNERSHIP MONROE I Town Park Swimming Town (residents) Stamford Boys' Club Field sports Private use Zoar Beach Boating, Commercial I swimming OXFORD I Town Park Swimming Town (residents) NEWTOWN I Eichler's Cove Marina Boating Commercial Paugusett State Forest Undeveloped State I SOUTHBURY I Southbury Town Beach Swimming Town (residents) Kettletown State Park Swimming, State camping State Boat Launch (Lakeside) Boating State I

~~Source: Northeast Utilities Service Co. 1982. I I I I I I 6-70A I I PGE00047273 1 1 TABLE 6.4-2 1 EXISTING RECREATIONAL FACILITIES ADJACENT TO LAKE LILLINONAH~~

1 RECREATIONAL FACILITY MAJOR (RESOURCE) ACTIVITY OWNERSHIP 1 SOUTHBURY George C. Waldo State Park Undeveloped State Southbury Town Park Hiking Town (residents) Waterbury YMCA and Pear Hiking Private club Street Neighborhood Camps 1 ROXBURY Roxbury Town Park Boating Town (residents) BRIDGEWATER Town Park Picnicking Town (residents) 1 State Boat Launch (Route Boating State 133) Driftwood Marina - Cooper Road Boating Commercial

11BV NEWTOWN Town Park Boating Town (residents) 1 State Boat Launch - Pond Brook Boating State 1 BROOKFIELD Town Park Hiking Town (residents) 1 NEW MILFORD Town Park Undeveloped Town (residents) 1 Source: Northeast Utilities Service Co. 1982. 1 1 1 6-70B

1 PGE00047274 I I sioners of Health Services and Environmental Protection issued an advisory against eating fish caught upstream of the Stevenson Dam. f| Compliance with the advisory is voluntary, however, as it is not a regulatory prohibition. I As detailed in the following section, there are no known data that quantify the angling pressure that existed prior to the advisory. I Thus, the magnitude of any drop in angling pressure after the advi sory was issued cannot be estimated directly. Actually, angler • pressure has probably increased as a result of the highly success ful trout management strategies implemented by CDEP approximately • four years after the advisory was issued. Three responses to the advisory are postulated: I 1. Some individuals ceased, either completely or partially, to fish the Housatonic River in favor M of a less desirable fishing location, a more re- | mote location of comparable quality, alternative recreational pursuits, or a mix of the above. _ 2. Some individuals have chosen to ignore the advis- • ory and continue to consume fish caught from the Housatonic River. • 3. Some individuals continue to fish on the Housa tonic River, but return their catch to the m river. For those anglers most commonly observed • in the TMA, eating the daily catch does not appear to be a key aspect of the fishing experience.

Any one angler can exhibit a combination of the above three re sponses. For example, individuals who continue to consume their catch from the Housatonic River (Response 2) may do so less fre- I quently (as compared to pre-advisory behavior) and may have accord ingly increased their visits to more remote locations (Response i). I I I

~~6-71 • Lawler Malusky Tikelly K I PGE00047275 I~~

I Response 1, above, would result in reduced angling pressure on the Housatonic river. Although a biological model for this river has I not been established, it is expected that the reduced pressure has resulted in an increase both in numbers and sizes of fish in the I river. For post-advisory anglers, the catch rate in terms of num bers of fish caught per hour of angling should therefore be high I er. Higher catch rates typically attract greater numbers of anglers. If it is assumed that Response 1 is in fact operable and has reduced some visits to the river, then it should likewise be I assumed that a compensating mechanism of higher post-advisory catch rates is manifested in a fourth response - increased visits by I anglers attracted by a favorable catch rate. I Other recreational activities. There does not appear to be a mechanism by which the low PCS levels in the waterway could di I rectly impact upon nonfishing recreational activities - boating, swimming, camping, or hiking. However, to the extent that the advisory has reduced angling by individuals who enjoy multipurpose I visits, e.g., fishing and boating, there should be a concomitant drop in the number of, e.g., boater-days. Since PCB impacts on I these activities are directly associated with the fishery, they will not be discussed further, and the following sections will con I centrate on sport fishing. I Fisheries management on the Housatonic River. For the purposes of fisheries investigation, the Housatonic River has been divided into I six sections (Barry, n.d.): I I I I 6-72 I Lawlcr Matusky okelly PGE00047276 I I RIVER No./ LENGTH SECTION LOCATION (mi) SPORTFISHERY I 1 Falls CT/MA border to 10 Trout, smallmouth bass Village Rtes 7/112 Bridge Pool I 2 TMA Rtes 7/112 Bridge 10 Trout, smallmouth bass to Rte 4 Cornwall I Bridge 3 Stanley Rte 4 Bridge to 8 Trout, smallmouth bass Works Rte 341 Bridge I Property at

I Fingerlings were found to have low (9-36%, depending on strain) first-year survival, hypothesized to be a result of predation by I smallmouth bass. For the Bitterroot strain, however, once the I I 6-74 I Lawler. Matusk Skcll E PGE00047278 1 1 TABLE 6.4-3 CDEP TROUT STOCKING ON THE HOUSATONIC RIVER •

YEARLING/ 1 YEAR No. ADULT F1NGERLING TOTAL 1976 NA NA 21,830 1 1977 NA NA 22,257 1978 NA NA 6,044 1 1979 NA NA 11,045 1 1980 0 0 1981 9,000 3,000 1 1982 9,000 3,000 1983 2,500 6,000 1 1984 2,250 5,000 1 1985 2,600 10,000

Sources: Moulton 1980; Orciari and Phillips 1986; and CDEP, n.d. 1 (a). Note: Numbers do not include breeder stocking by Housatonic River 1 Fly Fishermen's Assn. NA - Breakdown between adults and fingerlings not available. 1 1 1 1 1 6-74A 1 1 PGE00047279 I

I stocked fingerlings reached adult size they exhibited higher survival rates than stocked adults and achieved comparable growth I rates. I Based on these findings, the river is no longer a put-and-take trout fishery. Sport catch rates (detailed in Section 6.4.1.2) are I generally higher now than they were before the zero creel policy, even though fewer catchable trout are stocked. The chances for an angler to catch trophy-sized (16-in.) trout are substantially I greater, which appears to be an important element in attracting I anglers to the TMA (Barry, pers. commun; Hyatt, pers. commun.). It should be noted that the fly-fishing-only policy for the Housa I tonic Meadows portion of the river has been in place since the 1940s. This restriction was extended to the entire TMA in 1981, I but in 1982 bait and single-hook lures were again allowed upstream of Housatonic Meadows. The latest angler's guide (CDEP, n.d. [b]) does not restrict the method of angling in the upstream section I (except for statewide-imposed restrictions against, e.g., snagging I and spearing). In 1984 the TMA was opened to year-round fishing. Only portions I within 100 ft of the mouths of tributary streams are closed to all fishing during July and August. This restriction is required be I cause fish will congregate near the cool feeder streams when the river temperatures are elevated.

I Trout is the target species for the large majority (about 95%) of anglers in Sections 1 and 2 (Barry, pers. commun.). Even so, the I river has a sizable smallmouth bass population, and substantial I numbers are caught by trout anglers. For instance, in the TMA, I I 6-75 I L awl or Matusky SkuIIy Kntrinee PGE00047280 I I where over 92% of the summer anglers fly-fish, the catch rate for bass is 2.7 fish/hr; the trout rate, on the other hand, is only 0.32 fish/hr. Throughout the riverine sections, the smallmouth I bass is a significant element of the catch, as indicated below (Barry, n.d.): I

PERCENTAGE OF TOTAL CATCH I FALLS VIL. TMA STANLEY GAYLORDS SEASON SPECIES POOL (%) (%) WORKS ('/.) VILLE (%) Early Trout 0 100 0 0 I spring Spring Trout 100 78 8 0 I Bass 0 22 92 73 Pan/game 0 0 0 27 I Summer Trout 7 29 0 0 Bass 93 71 100 73 Pan/game 0 0 0 27 I Fall Trout 0 79 25 0 Bass 0 21 75 28 Pan/game 0 0 0 72 I

These rates are high enough for CDEP staff (Barry, pers. commun.) I to suggest consideration of smallmouth bass management to further enhance the resource. CDEP staff (Barry, pers. commun.; Orciari I and Phillips 1986) do not foresee abandonment of the zero creel policy should the advisory be rescinded; however, lifting the advi I sory against consumption (of smallmouth bass) may increase angling pressure by those seeking to fish for large, throwback trout and at the same time take advantage of the attractive bass catch rate for I the purpose of keeping fish for personal consumption. I Freshwater fisheries in the western Connecticut region. The three- county region (Fairfield, Litchfield, and New Haven) that encom I I 6-76 I Lawl«r Malusky Skelly Knjrinr.ers I PGE00047281 I I passes the Housatonic River has many freshwater fishing locations comparable, though not completely identical, to the Housatonic I (Figure 6.4-1). I In Litchfield County, which roughly encompasses the riverine por tion of the river, 16 ponds and 37 streams are stocked with approx I imately 19,000 brook trout, 116,000 brown trout, and 63,000 rainbow trout of adult, catchable size. These figures include the 2600 I adult brown trout stocked in the Housatonic TMA. Half of these ponds and two-thirds of the streams are located within 15 miles of the Housatonic and they receive 52, 65, and 78% of the adult stock I for the three types of trout (130,000 total stocked fish). I In Fairfield County, which borders the west shore of the Housatonic River impoundments, somewhat less stocking takes place: 10,000 I brook, 59,000 brown, and 23,000 rainbow trout of adult size in six ponds and 19 streams, one of which is a trout management area (Mianus River). In New Haven County the total stock is 52,000 I adult fish, lower than Fairfield's figures. With the exception of Atlantic salmon in the Farmington River, a tributary of the Con I necticut River, only trout are stocked in the waters of these coun- I ties. Fishing conditions and angler use of the nearby ponds and streams I apparently have not been studied intensively. The Farmington River was studied for the 1982-1984 period (Hyatt, n.d.), and a creel survey was reportedly performed at Bantam Lake, but the report has I not yet been issued. Fishing conditions and angler use of these water bodies were therefore reviewed in a meeting with CDEP Western I District staff (Barry, pers. commun.). Unlike the Housatonic I River, most of the trout streams support only put-and-take fisher 1 I 6-77 I Lawler Matusky 'fJkelly Knaineurs PGE00047282 r~r—i I I

~~I I I~~

~~I I~~

FIGURE 6.4-1 I RECREATIONAL AREAS ON THE HOUSATONIC RIVER | AND ITS ADJACENT WATER R 10 MILES I I PGE00047283 I I ies with minimal year-to-year holdover. Accordingly, there are few large fish In these streams. However, there is some holdover in I the Saugatuck River upstream and downstream of the Saugatuck Reser voir, Pootatuck River (which also supports natural reproduction), I and the Pomperaug River. I Most of the ponds support warmwater fisheries similar to those in Lakes Lillinonah and Zoar. The largest is Candlewood Lake, in I which brown trout, white and yellow perch, and smallmouth and largemouth bass are actively fished. A number of bass tournaments are held on this lake, which is reported to be among the top 10 in I the state in terms of bass size and catch rates. As with Lakes Lillinonah and Zoar, there is an active ice fishery on Candlewood I Lake. I The three impoundments are within 12 miles of each other. As a result, there appears to be movement of anglers between them during a I single fishing day, based upon perceived fishing quality and catch rates (Barry, pers. commun.). The bulk of angling on Candlewood Lake is from boats, however, apparently a result of Its large sur- I face area and the residential development along the shore. At Lakes Lillanonah and Zoar angling from the banks predominates. The I shorelines of these lakes 1s not nearly so developed, a result of I their steep topography. Other lakes supporting comparable fishing are Saugatuck Reservoir (access only by permit from a water company), Bantam Lake, East I Twin Lake, and Tyler Pond. All but Saugatuck Reservoir are more I remote from population centers than Lakes Lillinonah and Zoar. 6.4.1.2 Angler Surveys. The various angler surveys conducted I either on the Housatonic River or in Connecticut statewide are sum- I I 6-78 I Lauler Matusky Skellv Kninnurs PGE00047284 I I marized below, together with rates for visits, catch rates, and expenditures. Where possible, comparisons between the various sur- • veys are made.

Statewide. Statewide surveys of Connecticut resident anglers are ™ conducted every five years by the U.S. Department of Interior's Fish and Wildlife Service (FWS) as part of the agency's national I fishing and hunting survey. Results for the most recent (1985) survey are not available. The published report on the 1980 survey I (FWS 1982) provides limited statewide data on total freshwater anglers (311,000 aged 16 or over) and resident freshwater angling • pressure (6,123,000 angler-days). Information on expenditures is grouped for salt- and freshwater anglers. •

A contractor's report for the 1975 FWS survey provides a greater _ level of detail on statewide freshwater fishing (National Analysts, | n.d.). Unlike the subsequent survey, the reporting base is anglers aged nine and over: •

TYPE OF FISHING COLD WATER WARM WATER I No. PRESSURE No. PRESSURE LOCATION PARTICIPANTS (ANGLER-DAYS) PARTICIPANTS (ANGLER-DAYS) Lakes 191,000 1,972,000 203,000 1,557,000 I Reservoirs 131,000 813,000 125,000 1,112,000 • Farm ponds - - 32,000 389,000 Tailwaters 17,000 125,000 2,000 13,000 | Streams 214,000 1,914,000 107,000 1,077,000 _ All 329,000a 4,824,000 335,000 4,148,000 • a Al1 participants do not add to given total because individuals I fish at more than one location. • I

6-79 • Ltm'ler Matuskv fJkulK Knqrinoor.s I PGE00047285 1 1 For warmwater fishermen, fishing for bass and panfish accounted for I the majority (35% bass; 24% panfish) of angling for specific spe • cies. Ice fishing was pursued by 42,000 anglers in 1,398,000 trips.

Since the population bases for the 1975 and 1980 surveys are dif 1 ferent, they cannot be compared directly; however, an increase in the numbers of anglers and angling pressure from 1975 to 1980 is 1 indicated. 1 As detailed below, the Housatonic River supported 12,000, 8800, and • 12,000 angler-days for trout, bass, and pan/game fish, respec tively. The river thus appears to support relatively more panfish angling (about 1% of the statewide total) than bass angling (less 1 than 1%). Total expenditures for coldwater and warmwater fishing were simi 1 lar, and averaged around $6.50 per angler-day (1975 dollars). Moulton trout management study. Catch rates reported by mail re turns of trout diary holders utilizing the Housatonic River were analyzed for the 1976-1979 period (Moulton 1980):

ANGLER CATCH (No.) TOTAL PRESSURE NEW CATCH RATE 1 YEAR (hrs) STOCK HOLDOVERS TOTAL (No./hr) 1976 1804 1713 226 1939 1.07 1977 2692 2896 165 3061 1.14 1 1978 1094 896 366 1262 1.15 1 1979 2074 2453 174 2627 1.25 1 6-80 1 Lawler Matusky Skelly PGE00047286 I The total catch represents only a fraction of the stock for these I years (Table 6.4-3). Since diary holders typically have higher catch rates than nondiary holders, it is difficult to extrapolate I Moulton's figures to the entire body of anglers using the river. The 1976 figures are the earliest known data on angling pressure on the Housatonic River, and it was in that year that the presence of I PCBs in the river first gained public awareness. As the above tab ulation indicates, in 1977, when the advisory was issued, fishing I pressure increased rather than decreased; it declined drastically in the following year and then recovered past the average of the I prior years. Because of the nature of the survey, whether these figures are representative of overall angling pressure for that I period or whether they reflect only random cooperation by diary holders cannot be ascertained. If it is assumed that the advisory was partly responsible for the 1977-1978 decline, then it appears I that its impact was temporary. I Evaluation of the TMA. The catch and release policy in the TMA was evaluated based on data collected during 1981-1984 (Oman and I Phillips 1986). Part of that effort entailed trout angler surveys for each of these years along with analyses of angler diaries I returned by mail. The 1981 data represent the earliest comprehen sive figures on use of the Housatonic River by trout anglers. Data were gathered for the trout season (third Saturday in April to I October 15) throughout the TMA: I ANGLER HRS ANGLING/ TOTAL TOTAL CATCH/ YEAR TRIPS TRIP HRS CATCH HR I 1981 3,200 2.48 7,900 8,000 0.75 1982 6,100 2.60 16 ,000 19,500 0.87 1983 5,700 2.25 13 ,000 13,600 0.73 I 1984 3,500 2.41 8,400 8,400 0.73 Average 4,600 2.44 11 ,000 12,400 0.77 I I 6-81 I Lauler Matuskv Jikclh l,uoin<;rrs I PGE00047287 I

I The cause of the drop in angling pressure during 1983 and 1984 is not known, but may be a result of unusually high spring and summer I river flows. I Because the above figures are representative of all anglers, they cannot be compared with those reported for the 1976-1979 period by I Moulton (1980). As expected, catch rates for the average 1981-1984 angler are lower than those for the more expert 1976-1979 diary holder. Catch rates for 1981-1984 diary holders averaged 1.47 I trout/hr, however, about 30% higher than the diary holder 1976-1979 rates, a result of the replacement of put-and-take management with I the zero creel policy. I The average one-way travel distance by anglers to the TMA was about 45 miles. Use of the TMA by nonresident (out-of-state) anglers I increased from 6.4% in 1981 to 11.7% in 1984, indicative of the enhanced drawing power of this portion of the stream.

I 1985 angler survey. This survey by Barry (n.d.) covered the period 1 December 1984 to 31 November 1985. Unlike Orciari and Phillips I (1986), Barry collected angler use data for the full nontidal river I length, not just the TMA, and for trout and nontrout angling. I Data were aggregated by season, defined below: SEASON PERIOD Early spring 1 Mar - 19 Apr I Spring 20 Apr - 15 Jun Summer 16 Jun - 3 Sep Fall 4 Sep - 31 Oct I Late fall 1 Nov - 31 Dec I I 6-82 I Lawler, Matusky SkelJy Enq-mners I PGE00047288 I I In addition, a winter season was defined to include the period dur ing which conditions were suitable for ice fishing. Data were also I aggregated geographically within the six river sections defined previously. •

Barry reported statistics on angling pressure (total visits and hours of angling), type of fish sought (trout, bass, and pan/game | fish), type and numbers actually caught, method of fishing (bait, lure, and fly), and expenditures (variable and fixed). Some sta- I tistics were reported by place of residence (in state vs out of state). •

~~Angling pressure. Barry's estimated year 1985 angling pressure by river section is given below: I~~

RIVER No./ PRESSURE SECTION (ANGLER-DAYS) I 1 Falls Village Pool 961 I 2 TMA 11,536 • 3 Stanley Works 1,214 Property • 4 Gaylordsville 2,274 • 5 Lake Lillinonah 10,635 6 Lake Zoar 6,824 • Total 33,444

~~As indicated, one-third of the angling pressure took place in the TMA. Over 82% of angling took place during the spring and summer: I I I I~~

6-83 • La\vler Matusky Skelh• I]rn»mm:rs I PGE00047289 I I PRESSURE I SEASON (ANGLER-DAYS) Winter 938 Early spring 1,066 I Spring 15,887 Summer 11,482 Fall 3,507 I Late fall 204 I The winter angling represents ice fishing on Lakes Zoar and Lilli nonah. The early spring and late fall angling takes place only in the TMA. Barry's report does not provide breakouts of angler pres I sure by both river section and season, together. However, unpub lished work tables with these breakouts were made available by I Barry to LMS. In the TMA the seasonal breakouts are 1066 (early spring), 4019 (spring), 4606 (summer), 1640 (fall), and 204 (late I fall). Aggregation of the spring-to-fall visits allows comparison of Barry's survey with that of Orciari and Phillips (1986). As I indicated below, Barry's estimate is more than double the Orciari and Phillips estimate for the 1981-1984 period:

I PRESSURE (ANGLER-DAYS) ORCIARI AND I PERIOD PHILLIPS BARRY 1981 3200 1982 6100 I 1983 5700 1984 3500 I 1985 10,266 An increase in angler activity as a result of increased popularity I and public awareness of the TMA can be expected to account for some I of the change from 1984 to 1985. The difference in pressure be I 6-84 I Lawlcr Nhituskv okellv Kncrincers I PGE00047290 I I tween the two surveys is too great to be accounted for by popular ity alone, however. I

One possible cause for the difference is unusually high river flows during the summer of 1984; however, the unpublished work sheets by I Barry indicate that the summer pressure is 45% of the total spring- _ summer-fall pressure on the TMA. Therefore, even a total loss of | the 1984 summer angling effort because of the high flows cannot account for the difference between the two surveys. I

As noted previously, the TMA exhibits attractive catch rates for 8 smallmouth bass, particularly in the summer. Therefore, another possible cause for the difference is that Orciari and Phillips restricted their work to trout angling. Barry, on the other hand, I covered all types of fish. Barry (pers. commun.) has indicated, _ however, that essentially all anglers in the TMA fish for trout, | even though a lot of bass are caught.

~~In summary, no reasons have been found to explain adequately the difference between the Orciari and Phillips estimates of angling I pressure and the Barry estimates.~~

Catch. The 1985 catch was estimated by Barry to be about 256,000 • fish, of which trout, bass (smallmouth and largemouth), and mis- _ cellaneous pan/game fish constituted 17, 36, and 47%, respectively. |

Winter fishing on Lakes Zoar and Lillinonah resulted in a catch of • 12,000 miscellaneous panfish and gamefish. All early spring and late fall angling took place in the TMA and resulted in catches of • about 4000 and 800 trout, respectively, for these two seasons. * (Only 150 bass were caught in late fall.) During spring, summer, I I

~~6-85 •~~

Lauler Mfituskv SkclK I I PGE00047291 I I and fall, all species of fish were caught and there was angling activity in all six river sections (Table 6.4-4). The reported I angling pressure and catch were not broken out by target species, I season, and river section, however. Expenditures. As indicated below, Barry found distinct differences I in 1985 expenditures between the river sections: EXPENDITURES I RIVER No./ ($/ANGLER-DAY) SECTION VARIABLE FIXED TOTAL I 1 Falls Village Pool 14.07 5.29 19.36 2 TMA 24.40 10.50 34.90 3 Stanley Works Property 19.37 4.15 23.52 4 Gaylordsville 16.28 1.45 17.73 I 5 Lake Lillinonah 12.55 8.34 20.89 6 Lake Zoar 9.97 9.29 19.26 I Weighted average 16.66 8.57 25.23 I Both per-day variable (travel, bait, etc.) and fixed (equipment, boats, etc.) costs were highest in the TMA. Total variable and I fixed expenditures were about $560,000 and $290,000, respectively. I Travel costs ranged from 37 to 57% of total variable expenditures for Connecticut anglers and 53 to 94% for nonresident anglers:

I RESIDENTS NONRESIDENTS % OF VARIABLE EXP. % OF VARIABLE EXP. RIVER No./ TOTAL f$/day) TOTAL ($/day) I SECTION EFFORT TOTAL TRAVEL EFFORT TOTAL TRAVEL 1 Falls 2.4 13.38 4.95 0.5 17.40 10.44 Vi 1 lage Pool I 2 TMA 29.3 20.66 11.78 5.1 47.22 26.44 I (Continued) I 6-86 I Lawl«r Matuskv Skollv ti I PGE00047292 1

~~o; o 1~~

~~(O LU caoca CDIO»*- onr^ CM CD CNJ f-. CD CO —1 1 1 3:~~

o LU Z LO v: •—> ooo or^n CDUDLO xf, i CM co »r tn ^ in 1 o _i i CO *—*

~~UU CO Q 1 z LU~~

~~0 > CO 1 1— o s o "*• CC o n r** o r*~t I "* o co CM LU r** CM p^. CM CM ^^ CO~~

~~OH LU 1 ? C3 >•~~

~~O£ I cr 1 *r LU O LU 1— to cj »: ro z LU LU COCMO OOC3 LOtnO LU D_ CD O CM r~^ i U -i CO CC T—* CQ i: Q. 1 h- CQ I O I—~~

~~r-tCD O> •—< O L— CM s: t*^ CM CM r^ r^ CM o i— L— 1 u_ o LU o 4 cz 1 t— 2? Ck. LU CO O _J LU a: •r-t _l C. ooo r-.no ooo LU o a> u <—i 1 a.~~

~~CD d) CTLO -M CP*n 4-J o>iO (_3 ^H ZJlO-^-r- rjlO^.r- UtO^^-t— Ctl C. OfC"*— OulC"*- OW1C1 * «C u. ^— fO rO L. fo (O C— 'O fO I— a. >— CQCL. 1— CQ Q- »— CQd.~~

01 1 C C 0) " 1 u~ e • u Q. U fO o^ (S CO CO U_ 1 6-C6A 1 PGE00047293 I I (Continued) RESIDENTS NONRESIDENTS % OF VARIABLE EXP. % OF VARIABLE EXP. I RIVER No./ ($/day) TOTAL ($/dav) TOTAL SECTION EFFORT TOTAL TRAVEL EFFORT TOTAL TRAVEL I 3 Stanley Works 3.1 16.08 7.40 0.6 37.11 19.67 Property 4 Gaylordsvi 1 le 6.4 16.68 7.34 0.2 25.60 24.06 I 5 L. Lillinonah 29.5 11.80 6.73 2.4 21.89 17.95 6 L. Zoar 20.5 9.97 5.48 0.0 I 91.2 14.72 8.09 8.8 37.44 22.69 I The 8.8% participation rate by nonresidents indicated above is at the midpoint of the range reported by Orciari and Phillips (1986) for the 1981-1984 period. Travel costs for Connecticut residents I were significantly higher for angling at the TMA than for other river sections, a comment on the distance anglers are willing to I travel to this area (and dollars they are willing to spend). The lowest travel costs (about half the TMA costs) were for Falls Vil I lage Pool and Lakes Lillinonah and Zoar, indicating that these sec I tions are attractions to local, rather than regional, populations. Nonresident travel costs and the percentage of nonresident effort were also highest for the TMA. Nonresidents accounted for approxi I mately 15% of angling on the TMA, further indication of the dollars anglers are willing to spend to fish at this location. There was I only nominal nonresident use of the remaining sections. I In his survey Barry asked three questions designed to obtain data on fishery values (Section 6.4.1.3): I 1. How much greater do you think your total ex penses for today's trip would have to become before you would probably have decided not to I have gone fishing today? I I 6-87 I Lawlor Malusky Skclly Kn

~~Results for these questions were not reported.~~

1986 angler survey. The report on this work by Barry had not been I issued as of this writing. Handwritten work tables on angling pressure were reviewed to provide information on the no action alternative. I Angling Pressure. The same survey methodology was used by Barry for both the 1985 and the 1986 surveys. Overall, there was little • change in angling effort from 1985 to 1986, although for individual * sections there appear to be more substantive trends: I FISHING PRESSURE RIVER No./ (ANGLER-DAYS) • SECTION 1985 1986 % CHANGE • 1 Falls Village Pool 961 983 +2 2 TMA 11,536 9,721 -16 3 Stanley Works 1,214 1,285 +6 I Property 4 Gaylordsville 2,274 1,549 -32 5 Lake Li 1 1 inonah 10,635 13,067 +23 I 6 Lake Zoar 6,824 5,337 -22 Total 33,444 31,942 -4 I

As indicated above, the decline in angler activity registered for the TMA, Gaylordsville, and Lake Zoar river segments was counter balanced by the increase at Lake Lillinonah. The statistical con- • I

6-88 • La\vl<;r Malusky SkclK I PGE00047295 I I fidence interval for 1986 is not yet available, however, so the I significance of these changes cannot be determined. Unlike the 1985 report, the 1986 report tables break out angling I pressure by season and river section. Table 6.4-5 indicates that winter fishing (Lakes Zoar and Lillinonah) increased by 58%, from I 938 angler-days (1985) to 1478 angler-days (1986). The changes in the TMA from 1985 to 1986 can be compared since the 1985 work I tables have been made available to IMS:

TMA FISHING PRESSURE i (ANGLER-DAYS) SEASON 1985 1986 % CHANGE I Winter 0 0 0 Early spring 1,066 642 -40 Spring 4,019 3,980 - 1 Summer 4,606 2,710 -41 I Fall 1,640 2,236 36 Late fall 204 152 -25 I Total 11,535 9,720 -16

I As indicated, the decline in summer angling activity in the TMA I accounts for most of the loss in this section. Consumption. For the 1986 survey anglers were asked about con I sumption of fish caught from the Housatonic. % OF ANGLERS EATING CATCH I TYPE OF LAKE LAKE ANGLER RIVERINE LILLINONAH ZOAR I Bait 54 67 65 I Lure 22 15 I I 6-89 I Lawler Matusky Skelly hncr PGE00047296 I I —i n O <• cr> CD CD r-s co CM CO «*• us m ro i— CD r^ CM in o ro 01 o 1 i— CD •—t i—l CO LO •—< •-• ro

~~1 1 i 1 — i CD CM O O O CO CD ^C *i LO .-i co ro «-< CM Tf 1~~

n _J co co ro *± LO n CD _J TJ n n LO en ir> •—i C71 t^~ •-< O> ro CD ••I ^~ ^f r- LO ^r co CM CD cu LU ""^ f. f. r. *. to a: in CM *T CM ^H r-H n s- I LU n C£ LU ~^ a; >^ in Q. C3 in CD CD T—» CM r^ i—t i—( LU CD to co co co o CD rv. \n at Z o; ^" CD rO CD ^ CO O 10 • i a. a. •> • « « > • > ^ 1 *—t *3 ~T~ co ro ID «—* ro o C3 O ,__! a. CD Z

~~LU JL £ —i in CQ "-H t i < U. C3 O~~

CD C3 CM O O O CD CM CO en IS UD CD LU in ! I CD X o • to 4-> • a: LU " TD • i— O O O O CD CD CO _Q "O z LO ^-i r^ fO fO "* •—i ^5 , I tf— ro ^B ra S j-j L_ .— CD i— Q. 0 0 . 0 • 0 I- -a c: B Q_ Q. jr !2 fti c o 0 J^ .— O 1— tO 1- •— C CJ i— O -i— ••— LU L. CU | in •— to ,— 10 re ^D "~ >• >> T3 -r o m e o to i— O i— z ^^ *^C f^ ^1 \^ \^ 4_J s S o: fO ^^ *J (O ro rt O 0 •• 1• UJ u i— in co —i _J i— L a> >. 13 *J H-< o o o; ^n CM ro ^r m CD in ^ • I I PGE00047297 I I Bait anglers show a much higher tendency to eat their catch than lure anglers. This may be correlated to family income levels, I which were investigated in 1986: I AVERAGE 1986 TYPE OF PERSONAL INCOME ($) I ANGLER FOR RIVERINE ANGLERS Bait 24,000 Lure 31,000 I Fly 33,000 1 Only 3% of 1986 bait anglers reported incomes in excess of $50,000; the percentages for lure and fly anglers were 12 and 18%, respectively. About 50% of the anglers who eat their catch were aware of I the health advisory.

I The upcoming report on the 1986 survey will tabulate several eco nomic statistics. One will depict, by municipality, the residence I of the anglers (the 1985 report only distinguishes between in-state and out-of-state anglers). Such information can be used to deter I mine economic values (consumers' surplus) by the travel cost method.

I During the survey, anglers were asked (1) how much they would have to be compensated if they were denied fishing on the Housatonic, I and (2) how much they would be willing to pay, e.g., as a gate fee, to fish on this river. Answers to such questions have also been I used to establish economic values, as discussed below. I 6.4.1.3 Resource Evaluation Angling pressure. The data presented in the previous section indi I cate that the Housatonic River is actively fished throughout the I I 6-90 I Lawlcr Malusky Skclly Knoinocrs PGE00047298 I I year by anglers in pursuit of trout, other gamefish (notably bass), and panfish. The river supports about 1% of the angling statewide I for each of these three classes of fish. In recent years Housa tonic River anglers have been intensively surveyed, and a fairly • good depiction of the present use of the river now exists. *

Unfortunately, the data collected prior to the PCB advisory is Q limited and not fully comparable to the more recent, comprehensive compilations. Only Moulton's (1980) report on diary holders pro J vides comparable pre- and post-advisory data on trout angling pres sure. As noted previously, in year 1977, when the PCB advisory was • issued, there was an increase in angler pressure (by 49%) over the previous year's level rather than the expected decrease. A decline • in angling pressure during year 1978 was registered, but by 1979 ™ angler use more than recovered relative to the 1976-1978 average. • One car conclude that the discovery of PCBs in the Housatonic River • had little, if any, discernible impact on use of the river by trout anglers. I

Expectations by CDEP staff are that the zero creel policy in the I TMA will continue even after the PCB advisory is lifted. As the TMA supports virtually all of the trout angling on the river, the • presence of PCBs will continue to cause little (if any) diminish • ment of this activity. I For the remaining species, primarily smallmouth bass and panfish, there are no available data allowing direct quantification of a • PCB-induced reduction in angling pressure. As described below, there is an analytical framework that might allow indirect calcula B tion of the reduction in angling demand for the Housatonic River. However, the data necessary to implement this framework have not been collected. I I

~~6-91 I Lawlor Mutuskv JikHK l!ncrini:*'rs I PGE00047299 I~~

I Valuation. With little basis at present for providing any estimate of the reduced use of the river by anglers, there is even less I basis for quantifying the dollar value of this reduction. The additional problems in this quantification are associated with a I lack of agreement between individuals on the definition of value and the difficulty in compiling the necessary data even if a defi I nition is agreed upon. I The common definitions of fisheries values are: o Option value and opportunity demand I o Existence value I o Expenditures and regional economic value o Compensation value I a Consumers' surplus Values for non-anglers. Option value, opportunity demand, and I existence value refer to the values ascribed by individuals who do not directly use the fishery. In this regard individuals may want I an intact fishery in case they decide to utilize it in the future (option value) or in case they believe others may have such a desire (opportunity demand). Other non-anglers receive a benefit I in knowing that there is an intact or unpolluted fishery (existence demand). These values are typically not quantified in terms of I dollars in fisheries management studies. I Expenditures. Expenditures by anglers have value to local mer chants and producers of the goods marketed by these merchants. As I indicated previously, 1985 expenditures by anglers visiting the Housatonic River amounted to $850,000. About two-thirds of this I was for gasoline, meals, and lodging (variable expenditures) in- I I 6-92 I Lawler Matusky Skelly K PGE00047300 I I curred for individual trips. The remaining dollars were a propor tion of fixed expenditures for boats, fishing gear, etc., that are • used for multiple trips. '

These expenditures benefit local and state economies by more than | just the $850,000 per year noted above. These monies are spent by local merchants for labor, services, goods, and imports. The de I gree to which the anglers' expenditures generate additional busi ness in the state is quantified by the "regional multiplier," which I has an estimated value of 1.5 (Hyatt, n.d.). Therefore, the re gional (statewide) economic value of Housatonic River angling is • estimated to be $l,300,000/year. •

Money not spent in pursuit of angling on this river because of the I PCBs presumably will be spent elsewhere in the state economy. Therefore, assuming a PCB-induced change in angling pressure could I ever be quantified, the diversion of expenditures would not be an estimate of the actual lost value. I

Compensation value. This is the value in dollars that anglers would accept in compensation if not allowed to fish on the Housa I tomc River. Arguments have been made for utilizing compensation value in cost/benefit analyses (Stabler 1980). However, the argu | ments are largely rhetorical and appear predicated on the fact that this form of value is typically far greater than the commonly used • consumers' surplus (discussed below).

There are also problems in measuring compensation value. Must ™ studies typically involve angler questions such as, "How much would you have to be compensated if you were not allowed to fish on this I river9" Answers to such hypothetical questions are subject to bias. Only one known study (Bishop and Heberlein 1979) quantifies I 6-93 Laulor \latusk\ Skolh Inquirers I PGE00047301 I I this bias. Actual cash offers were made to purchase scarce hunting permits. The average willingness to sell by hunters was $63 per I season based on licences actually returned. A companion mail ques tionnaire with a hypothetical question phrased similarly to the one I above resulted in a $101 per season estimate for compensation I value. Consumers' surplus. Consumers' surplus is defined as the addi tional amount that anglers are willing to expend for fishing beyond I what they actually must pay for the experience. This is the form of value used in standard cost/benefit analyses where fisheries I management is concerned. Note that expenditures are excluded from the calculations, as only the value beyond what is spent is rele I vant. I Two methods are used to measure consumers' surplus. One is based on hypothetical questions posed to anglers, the other on observa tions of differences in actual angler behavior in response to vary I ing costs for this form of recreation.

I The hypothetical method asks one of two questions: 1. At what price level would fishing be so expensive I that the angler would cease to fish? 2. If not allowed to fish, what would be the cost of I the next preferred form of recreation? The first question leads to contingent valuation; the second, to I alternative recreational value.

I Both types of questions have been criticized as being subject to several forms of bias: hypothetical, strategic, starting point, I payment-vehicle, and information. A few studies have attempted to I I 6-94 I LcJ\vlor Mntusk Gkelly K PGE00047302 I I quantify the magnitude of these several biases, but results are mixed. I

Barry's 1985 survey asked the following: "How much greater do you • think your total expenses for today's trip would have to become be- ™ fore you probably would have decided not to have gone fishing to day?" Once tabulated, results will give an indication on the over- I all value of fishing. If the response by some prior Housatonic River anglers to the PCB advisory is to cease fishing altogether, • then this overall value may approximate their loss. As stated pre viously, there are numerous substitute fishing locations in the • vicinity of the Housatonic River (Candlewood Lake, in particular). If the response by anglers to the PCB advisory 1s to fish at alter- • native locations, then the contingent value may overstate the • actual loss. •

The empirical method can be implemented by several means, but the most commonly used is the travel cost technique, which assumes that I travel cost can be used as a proxy for willingness by anglers to pay for fishing at a specific site. Variations in rates for visits I and travel costs from different population zones are used as the basis for calculating the consumers' surplus. •

The technique has also been criticized for several reasons. One is that the value of travel time is usually not measured, and is I therefore not always integrated into the calculations for deter mining overall values. The result of this omission will be an I underestimate of the consumers' surplus. This shortcoming is typi cally addressed by assuming a time value based on some assumed • fraction of the angler's wages or by reference to nonwork time value estimates given in studies of proposed highway improvements. I I

~~6-95 •~~

Liwler Miilusk Skellv K I PGE00047303 I I Another criticism of the travel cost technique is that there is no explicit consideration of alternative fishing locations or the I quality, e.g., catch rates, of the fishing site investigated. Refinements to the technique have been formulated to address these I limitations, but they require more comprehensive data and more I sophisticated analyses than commonly used. The travel cost technique could be used to estimate the impact of PCBs on angler use of the Housatonic River. Creel consensus or a I regional mailing to anglers would be required to evaluate the significance of alternative fishing locations in the region. These I data may indicate that anglers are traveling farther from home to fish at more remote, alternative locations otherwise similar to the I Housatonic River. Assuming this excess travel would not occur if there were no PCB advisory, then the loss in angling pressure and I consumers' surplus could be directly computed. Given that there is little pre-advisory data (none for gamefish and I panfish, which support most of the angling), no estimate of the loss in angling value can be made until such regionwide information I is collected. I 6.4.2 Future Development I Subsequent to the submittal and review of the interim report and the specification of the work scope for the no action plan, CDEP requested that this study be expanded to incorporate the consider- I ation of future developments in the Housatonic River. At that time Northeast Utilities' sediment management plan for the Bulls I Bridge hydroelectric redevelopment project was the issue. Discus I sions among CDEP, GE, Northeast Utilities, and IMS helped to define I 6-96 I Liiivlnr Malusky Skelly Engineers I PGE00047304 I I the environmental concerns, monitoring, and mitigation that need to be addressed and to clarify the integration of such developments 8 with the no action plan. This section begins by reviewing the regulatory process for approving developments, including actual • case studies that demonstrate how the regulatory process operates, ™ as well as some of the monitoring and mitigation alternatives that _ are incorporated into the process. Potential PCB releases from | construction activity are then evaluated technically with a discus sion of a hypothetical project example. Finally, the alternatives • for dredging and disposal are reviewed and a recommendation consis tent with the no action plan is made. •

6.4.2.1 Regulatory Process. Development of waterways in Connecti cut is dependent upon approval from the New England Division (NED) I of the Army Corps of Engineers (COE) and one or more divisions of _ CDEP. After receiving a complete application for development, the | foregoing agencies will begin the review process to determine the suitability of the proposed project to the area in which the work • is to take place as well as compliance of the work with state and Federal statutes. Local agencies or municipalities may also be • involved in the review process.

The review and permitting of projects by state and federal agencies I is done on a case-by-case basis, as indicated in Table 6.4-6. This table shows several recently approved and pending permit | applications that involve development by third parties of river- ways or harbor areas containing PCB-contaminated sediments. It is I evident from this table that even similar project activities (such as marina development) differ greatly from one site to another. • The issuance of permits is specific to each project, depending on ™ the type of activity proposed, site-specific conditions, and agency I I 6-97 I Laulor \IalusKv Skcllv I nqinc'crs I PGE00047305 1 g a> (/)••— « C 1 Q) 4-* C o " _Q «LJ) 1/t1 ^O- •**—J 1 U_o La.. «T—> trpt ^«3 o •*J *J .— (U

~~1 1— . «J « •""* -t-> CNJ L. O t- (J i— 1 C *o +J Q> a) o O LO m i/> t/> u. o_ O~~

~~1 0) L C a>r- a> o •R1 -Q • -P ~~ u~~

1 tl" q " " 8 « ^^ F ^ 1 ^ P) +J ^- (O C *3 ^ fl> o) "Q o --:P x«J"Oc zITIGATIO N o £ jjjj||i| ff|| • ! g a i j i-"!!"* *!^ O *" ^- O O v> L O ? 8 ^~ L rt •»— c **~~ LP. O *^~ C •-i te a. 1 COQc^ UJ I fl 5fi: to t/i in «— ^. Ij __i ^-> : : (O C «^> (O C 1 L_OO CO L- O »— 4-* — H- Q> 4-J •— (TJ O c C7> Pi -4-> C rn 1/1 Qj 4) «*— **~- (D Q) O 3 S O i- "D *o O 1- Q. h- p- 1 Lu S. a> o g g 6 "5. •-< E 5 U"> fc" V o S m *a o o • TJ to CM «-4 0 E f-i Q £jf 1 *j a> p. v I 1 CJ S L o cs< *-> ^ ^*- e R ^ IIi o ^ i"s" I o. § 1 >• S | 8 ^^

~~G 5 _ 1 g M 3° «a ~o X) to~~

~~XI O UT O r- . •— <«— o **- to L. • l_~~

~~8 o 'o >> 6 CT i- l^ 4-> tt> C 24-» U1 •*- *O 0 « U (U U C C C7> 1 •r- p > ra «- <« -p 1 «L- -CP m Qfa^. l_ITJ C. 1 t 3t « r-« U ^ -P T3 Io B o !L> c j£ TJ 031 1 • i I § g 8 ^ u. T-i~~

• 6-97A1 PGE00047306 I x?85 ? •— "Q at *n +-> 3 ^~ _ (O J TJ flJ rO ecc.ec ± c -M O) XJ •S L^ ^ "^ g s2£%~ •*—>•— I/) .» O) O 8i o QJ > TI &. --a e « s *0 S = — c f> ro o i^^gjlo*^c^ c-^ e n) ^-r- "- p "5 *D- C O .*-» C 0) •* vt •»— rti +j ro a 2 r7,e'«2"8 -M C Q .3 V> ^j E •— c «^ : f» i~ P C O b ir C N+J >>O ft. iu> a «-_i_ -- «J>*— O IO (/) t/» •*-* O •»— E: a>

nj T cv — O) — Q J. O O w ^ - Q.D *-">> ou E t^ •-o— 5 o a> fo 1< hrr +-~ »v> •^^g -g ^«OFs&^ - 4-> a) C !*3 *" Z3 •- tT 5 O *r {D f> ~ •• — 0 C— i r— Q) i— (O C i_ »~ O e pt B ty o £ o LO O E O O s to « = i r?1 •i— U k_ C> 3 (O C QQ. 3 U S icil o Q, C5 u*J v» CC< QQ. O O 10 ffo Q_ it- ^r- b O H~ fO •*- U O

~~-• o CM II : > B8LL CL~~

~~I 8 j> oo-TjLpoo P rkoj m •«•-• ^— s Io q in «- i~~

~~Q) **— CT)*J c C — T3 3 ZO ft) "Q x» c "p > c ~ 0)13" S OQ13 ^"O Coce^ ••— QJe z*^ ~~

~~m u £u> ~o£ "i_ -er™ >>^-> -»-> o'o 'a.~~

~~c ra ID un E OJ~~

6-97A2 PGE00047307 I I concerns with regard to the specific projects (Table 6.4-7). Depending on the stability of each site and the size of the pro- I posed project (volume of sediment to be moved), more or less effort must be put into the development of a management plan, including I short- and long-term mitigation plans and on-site monitoring. I Both of the marina projects approved for dredging call for disposal in Long Island Sound, which is adjacent to the location of these projects. Sediments from both sites are either similar to or I slightly greater in PCB concentration than the Housatonic River sediments and consequently capping at the disposal site is re- I quired. Since there is no navigational route from the Housatonic River study area to Long Island Sound, other disposal alternatives I are proposed for the Bulls Bridge Hydro and Route 341 bridge re I placement projects. The review process consists of (1) the submittal of an application for a Permit to Develop; (2) determination of the suitability and I compliance of the proposed project by review agencies, including suggested changes or mitigations to the original as proposed by I these agencies and agreed upon by the applicant; and (3) in cases in which all potential impacts have been satisfactorily considered, I the granting of permits needed for commencement of the proposed activity. This process may also be preceded by an optional prepro I posal meeting between the applicant and representatives of the various regulatory agencies involved. Such a meeting is helpful in defining the overall feasibility of the activity to be proposed as I well as areas of special concern and specific points that need to be addressed, and aids the applicant in preparing the permit appli- I cation. In general, applications must fully describe the existing habitat and the desired work activity as well as the mitigation of I potential impacts to the environment. I I 6-98 I La\vler Matu.sky Skelly Kn-inner.s PGE00047308 uj _ji 2M:J OZ .Q 1 L_uj *LL* 3CJ £^r^ i— h— i—

E X X X X X X QJ 1 o > CO t/> CO U X 1 5 3 •£ T? QJ i) c_) t J2 ^ h — 1 s ^ 5 § S a X X S § 3 *~ 1^1 •< UJ Q) _d L. fl> 5 S 4-1 *-* H— 2? o z ^ ^ 1 ^ a <£ Q GJ in O ^r QJ iii (i^ !j X X QJ •* _) 3C P LU pE ± o 5 QJ ""T en T3 c: 1 3 O 1/T; Q rf Q >— «—• 0 CL ^5 rx o •—• UJ en U I o: tt Q X X X 1 *o O 10 < 3 8 LU LU ate° o a —t O 1 1/1 o c/> >-, ra c CL O a ULJ < ^j •—< _J O O •—• " -" 1 S B U- X X X X X L -t-i £ & U ae: L_ U o ^ JJ 3 >-, QJ LU XI -O cr 1 g _> ^^ CC O LU *— Q) - X X X X X

to 31 Q_ O Z 51 iii o rf o O E Q t^ O 1 -«-» o O LU ^ I UJ to t— r-g t^j X X X X X X "S "^ 1 S K 2 H S a. o u_ LU z _J LU •-« U- O l/> C X X X X X X 5 ° H VI b^ O »— O CJ LU LU HI ^-« Q£ < O Q h to CL 1 5 1 rn i PC: LU oo z cC" -^ X X X X X X >» 0 1 ""J S o —• P O u_ < t*— — oo O o

~~1 U, 1 Z UJ 1— LU O 3*7 LT> i—* O O *~ X X X X X X 1 r: ,_ S^^^^ggfe~~

~~o~ ^p~~

,— VI l/> l/> .— »/» •*— l/> O Lj U Ul -*-> »+ L_ L. TJ OJ -^^ --7 o. c: a> *~* 4-Jp3 p o-j,— pC r 'r-io r, r •»~> r^ *J T) rfl ^— L. 1_ ro TO C. -•— i-— -*-" O flJCrfi. QJT ) *J.r c (Do— CT) cr> QJ »/i cp c: ••— o r— 85 o i— •«— C v "v p •" "O w '— i— r- U *^ "^ **~ r— W S C) "C W ^— '— Q_ S ^ 1 u7 CL Q Cl O => " -C° 6-98A 1 PGE00047309 I

I A list of permits applicable to the sediment management plan and the agencies responsible for issuance of these permits is shown in I Table 6.4-8. Different agencies often work simultaneously in re view of submitted material and the granting of permits. For exam- I ple, a 401 Water Quality Certification (as issued by CDEP) must I always accompany a 404 Clean Water Act (as issued by COE). Currently, as defined by the Interim Plan for the Disposal of Dredged Material from Long Island Sound (NERBC 1980), PCB I concentrations of 1 ppm or more in sediment is considered high and will result in a Class III classification. Furthermore, the I highest level of PCBs detected and not an average concentration dictate the PCB concentrations .at which sediments will be I classified and form the basis for disposal decisions (Deshefy 1985). Class III sediments may be "toxic" to biota, and are I therefore subject to greater scrutiny than Class I or II sediments with regard to disturbance and disposal. This may include I elutriate and bioassay testing. Conditions under which projects involving Class III sediments are I approved may involve both temporal and seasonal restrictions to the proposed activity as well as defined limitations regarding disposal I or stabilization of sediments. Even if impacts from a proposed disposal activity are deemed "acceptable," it is necessary to exam I ine and consider practical alternatives with even fewer damaging environmental impacts, giving consideration to economic, technical, and logistical feasibility as well (NERBC 1980). Possible alterna- I tives include land disposal, containment in certain harbors or I waterways, and under certain circumstances no dumping at all. Overall, the current regulatory process appears to be compatible I with the no action plan. As previously stated, projects involving I I 6-99 I Laivl«r Mfilusky okelly Knjrinecrs PGE00047310 I I TABLE 6.4-8 PERMITS APPLICABLE TO SEDIMENT MANAGEMENT PLAN

PERMITS APPLICABLE DURING: AGENCY I Rivers and Harbors Construction, excavation, COE Act of 1899 deposition, obstruction, or I (Section 10) alteration Clean Water Act Discharge of dredged or COE (Section 404) fill material Rivers and Harbors Construction of bridges or COE and Dept. Act of 1899 causeways of Transporta I (Section 9) tion Inland Wetlands Filling, dredging, Local municipal and Watercourses obstructing, or discharging wetland agency I (CGS Section pollutants or CDEP 22a-36 through 22a-45) I Water Quality Requisite to application CDEP Certification for COE permit Section 404 I (Section 401 Clean Water Act) I I I I I I I 6-99A I I PGE00047311 I I greater concentration of PCBs or volumes of material will undergo greater scrutiny and require greater measures of mitigation on a I case-by-case basis. While the regulatory process is broad in its scope and review, the Sediment Management Plan for the Housatonic I River defines a narrower class of activity with fewer viable op tions. The specificity of the river, its characteristics, and I PCB, which is a primary concern, call for a more customized evalua- I tion procedure that fits within the existing regulatory process. 6.4.2.2 Evaluation of Potential PCB Losses From Dredging and Fill- ing Projects. Construction activities that physically affect river I sediments generally fall into two broad categories: dredging/dis posal and filling operations. A cable or pipeline crossing of a I river usually involves side casting, which can be viewed as a com bination of the two, that is, removing sediments to make a trench I and then replacing the sediments to bury the line. Filling operations pose minimal potential adverse impact on the release of PCBs, assuming that clean fill material is placed so that the stability I of the existing substrate is maintained.

I Studies of short- and long-term releases of sediment-bound con taminants from dredging and disposal operations provide an assess I ment of dredging and disposal methods (Phillips and Malek 1987). Under most circumstances conventional hydraulic dredges produce I less resuspension and have a higher removal efficiency for liquids and solids than mechanical dredges, e.g., clamshell bucket. In light of the hydrophobic characteristics of PCBs and the limited I access to the Housatonic River, hydraulic dredging (for which por- table units are available) would typically be the preferred I method. However, in areas that can be dewatered, such as a power canal, excavation "in the dry" (using mechanical equipment) would I practically eliminate the release of PCBs to the river. I I 6-100 I Lawler. Malusk Skelly L PGE00047312 I I The disposal options for a dredging project are either subaquatic (relocation of PCB-laden sediments within the river) or upland • (introduction of PCBs and containment on land). The upland alter native has a higher environmental risk in that it increases the • potential pathways of PCB transfer through plant and animals to • humans. Furthermore, an upland alternative consisting of a clay- lined containment area with nonporous cover has some, though mini- | mal, potential for PCB release because:

• Dewatenng of sediments produces large volumes of • water having soluble and collodial PCBs that would not be fully removed through treatment. • • Potential failure of the clay lining, nonporous cover, or any leachate collection and treatment « controls would result in some PCB discharge to the • envi ronment.

The potential release of PCBs from subaquatic disposal of dredged J sediments is similarly minimal. The short- and long-term releases of PCB were technically evaluated for a hypothetical dredging proj I ect and the detailed computations are included as Appendix C. The evaluation is based on the COE method for analyzing dredged maten • al dispersion around openwater disposal operations (Barnard 1978) ' in conjunction with the PCB analysis of sediment-water partitioning • and dissolution/diffusion from the redeposited sediments. I

~~The conditions assumed for the hypothetical dredging project are • as follows:~~

A marina in Lake Lillinonah has to dredge an average ™ of 3 ft from a surface area of 4.34 acres for a total volume of 21,000 yd^. The profile of PCB concentra- • tion over the dredging depth, adapted from the core | data near the Shepaug Dam, shows a PCB concentration of 1 mg/kg in the active layer (top 5 in.) and an _ average concentration of 3 mg/kg over the full 3-ft • I 6-101 I Laulor \Ialuskv Skcllv l.nq-iruM'is I PGE00047313 I I depth. In the process of hydraullcally dredging, a slurry of this material is pumped through a pipeline to an area of the lake having an average depth of 38 I ft. Dredging/disposal is performed for approximately 9 hrs per day for five consecutive days. A silt cur tain (impervious, floating barrier extending verti- I cally from the water surface to a specified depth) deployed to assist in confining the turbidity plume fosters flocculation and settling of sediment into a mud mound with an area of 31.1 acres and a maximum I height of 3.3 ft.

I According to COE (Barnard 1978), at most 5% of the dredged sediment remains suspended in the water column; 95% is deposited as a fluid I mud. The application of the solids-dependent partitioning coefficient (Section 6.3.2.4) to compute the particulate and soluble PCB I components in the plume and mud assumes that absorption-desorption is reversible. This tends to be a conservative assumption in that it probably overestimates the dissolved PCB component. Less than I 0.5% of the total PCB mass dredged would be solubilized during the dredging and disposal operation (Table 6.4-9). The plume has an I average TSS concentration of 139 mg/1 above background and occupies approximately 1.4% of the lake volume. As the plume dissipates I within one to seven days after the dredging operation (Bohlen et al. 1979; Onuschak 1982; Tavalaro and Mansky 1984), practically all of the PCBs are reabsorbed onto solids. Hence, the short-term bio I logical uptake of PCBs is minimally affected by the operation because the spatial and temporal extents of the plume are neglig I ible relative to the lakewide annual uptake by biota. I Evaluation of the incremental impact of the hypothetical dredging project compares the PCB release rate after dredging and disposal I with that from no dredging. The dredging and disposal areas have particulate concentrations of 1 mg/kg without dredging and 0.1 and I 2.6 mg/kg, respectively, after dredging (Table 6.4-10). According I I 6-102 I Lawlor Matusk Skell PGE00047314 CM 0 rj O T3 1— _l r O> O CM ra < •"*^. ro «r r~ cr IT) CO l_ C3 3 •N ra 1— ^—• .—1 CQ D c _ o 0 < LU *—* TJ =» CO •<*• «-t QJ — _l •^. C\J O CNJ ^y C/) o cn • • • fl3 LU to 3 o o o _Q CJ to **_> IO O Q 1— •^J CJ O) LU cn CQ T3 s: CJ LU D >—• L. 1— t_ o < -a LU _J a (/} !^ Jkl to O o 10 r-v */> a f—i O LO LO TJ LU 1— I C3 a: n• CN*J CM* a < CQ LU 0 Q- cc z O o •—« Q_ CD ro o n lt- Z Q .. LU 0 if tO 0 oe «n Q 0< ^r «—i m LO 1 CO 2 CM CM «*• O C3 • a. z II C^ 1—H >. CO Z m in to "0. < O 0 I/I i—i 1— i— Z t/1 9_i CQ C t— <; CJ to ID Q_ f-._j ^ CQ LU 0 •a •-i CC u. CD o: o o T3 i— * ' c in LU CO LU OJ 1-1 CQ to t— o «t <: Q. O LO l-x 3 1 0 CM LO to r^. •—' ^t" o to i— •cr —i ro CO (— o; CM CM CJ LU «q o_ a: Q. |«— o <• a. o > to ^c to 1 < ra V ro ,— rO C3 Q. O 4_) z o z i—< •*-» i—i O e> ca 4-> 0 L. Q OJ LU O LU > O£ Z 3O 13 r^ O <•-«•— 3 E cr> u_ ~- o: o_ s ^5 ,—i LU 3 C_J *•—• * CQ a

~~6-102A~~

~~PGE00047315 1 1 TABLE 6.4-10 LONG-TERM PCB RELEASE WITH AND WITHOUT HYPOTHETICAL DREDGING PROJECT IN LAKE LILLINONAH~~

YEAR 1 AFTER DREDGING DISPOSAL REMAINDER DREDGING AREA AREA OF LAKE TOTAL 1 Surface area (104m2) 1.758 12.57 639.3 653.6 Pre-dredging/No-dredging 1 Sol ids cone. (g/cm^) 0.513 0.513 0.513 Partic. PCB cone. 1.0 1.0 1.0 (rug/kg) PCB release rate 1 0.019 0.140 6.970 7.129 (g/day) 1 [ug/m^ day] [1.10] [1.10] [1.10] 1 10 0.006 0.044 2.238 2.288 [0.35] [0.35] [0.35] 1 70 0.002 0.016 0.831 0.849 [0.13] [0.13] [0.13] 1 Post-dredqinq Solids cone, (g/cm^) 0.513 0.280 0.513

1 •V Partic. PCB cone, (mg/kg) 0.1 2.57 1.0 PCB release rate 1 .002 0.248 6.970 7.220 1 (g/dav) [ug/m^day] [0-11] [1.97] [1.10] 1 10 .001 0.076 2.238 2.315 [0.03] [0.63] [0.35] 70 .000 0.030 0.831 0.861 1 [0.01] [0.24] [0.13] 1 1 C-102B

PGE00047316 I I to Thibodeaux (1979), the dissolution and diffusion of PCBs from bottom sediments is modeled as a leaching process dependent on the I participate and dissolved PCB concentrations, the solids concentra tions, and the diffusion rate coefficient (Appendix C). The com • bined PCB release from the dredging and disposal areas would ™ increase by 57% in the first year after dredging; however, the M lakewide release of PCBs would increase by only 1% as a result of I the dredging/disposal operation (Table 6.4-10). Projections show a negligible difference in the long-term total PCB release from Lake g Lillinonah with and without the hypothetical marina dredging and subaquatic disposal (Figure 6.4-2). It should be noted that this I analysis overestimates the PCB release in all areas because sedi mentation and burial are not factored into this comparative evalua • tion. The burial effect would tend to progressively decrease the * difference between the dredging and no dredging conditions as the • deposition of cleaner sediments forms a more homogeneous active • sediment layer over the entire lake.

6.4.2.3 Future Development and the No Action Plan. The preced ing evaluation of a hypothetical development having a relatively I large dredging volume suggests that subaquatic disposal is a vi able option. The analytical framework presented in detail is • applicable to other projects having different sediment characteris- ' tics, dredged volumes, and disposal operations. This evaluation _ procedure is intended to supplement CDEP's technical review for | specifically addressing PCB impact. The potential impact on benthic habitat and other considerations must also be addressed for • each project. Although the short-term impact on benthic organisms within the disposal area may be relatively significant, total re- • covery has been observed to take three to 18 months (Barnard 1978). Substitute habitats are available in the interim. Because • the potential PCB release is accountably low and limited to the • I

6-103 • LiW'lor. Matuskv Skollv Kncrinoor.s I PGE00047317 FIGURE 6-4.2 I TOTAL PCB RELEASE FROM LAKE LILLINONAH WITH AND WITHOUT HYPOTHETICAL MARINA DREDGING I AND SUBAQUATIC DISPOSAL I I I I I I PCB RELEASE I ( 0 /day) I I I I

~~I DREDGING/DISPOSAL I I 1 -i NO DREDGING/DISPOSAL I~~

I I 20 40 60 80 100 I TIME AFTER DREDGING/DISPOSAL (years) PGE00047318 I I aquatic environment (whereas the upland disposal alternative has greater environmental risk), subaquatic disposal is recommended. • The potential for cumulative impacts from many dredging projects over a relatively short period does not appear likely due to the • nondeveloped nature of the river banks. Nevertheless, regulatory agencies can take appropriate steps, such as limiting disposal to certain areas. I

Subaquatic disposal is not necessarily confined to deep water. For | narrower, riverine reaches such as the area of the Route 341 bridge in Kent, disposal along the bank, but below the water surface, may • be feasible if a silt curtain or other barrier is used to confine the mud mound. The use of fill as a cap over the dredged material • may be a mitigation, provided that the cross-sectional area for " conveyance is large enough so that the cap will not scour. Of • course, the dredging volume in a riverine section would have to be I relatively small for disposal along an adjacent bank; otherwise, pumping to deeper areas may be required. The main point is that • subaquatic disposal has a lower environmental risk than upland dis posal and is therefore recommended as part of the no action plan. I

Monitoring during and after dredging should be planned together with remediations that would be necessitated under certain pre- I scribed conditions. For example, turbidity and suspended solids « may be measured in the disposal area during dredging. If levels I are not sufficiently low downstream of the disposal area, an alter native pipeline orientation, pumping rate, or submerged diffuser • would be used. For the initial years following disposal, the mud mound may be mapped using staff gages, divers, or a side-scan sonar I survey. If substantial movement is found, a cap of clean material may be added. I I

~~6-104 I Lawler Maluskv 'Jkellv Liujiiiccrs I PGE00047319 I I 6.5 PROPOSED MONITORING PROGRAM~~

~~I 6.5.1 Objectives and General Principles~~

I The development of a monitoring program was specified by CDEP to be part of the no action plan. The objectives of the long-term moni I toring program are to: • Assess the trends in PCB concentrations in the sediment, water, and fish and thereby track the I river's natural recovery. • Detect any potential reversal of projected de I creasing PCB trends so mitigation, if necessary, can be implemented promptly. I o Gain a better understanding of the PCB fate and transport processes to provide a sound technical I basis for the river's management. The modeling projections of PCB concentrations in sediment, water, and fish provide guidance on the design of this monitoring pro I gram. For example, the rate of change in PCB concentration projected in the sediment provides an estimate of the time to allow I until a measurable difference actually takes place. I While PCB levels in the fish are not expected to increase, it is important to have a system in place that safeguards against this I going undetected. The model showed that the reduction in PCB levels in the water and sediment in Connecticut is largely dependent on a diminishing upstream source of PCBs. An unexpected in I crease in upstream PCB levels could be traced directly to an otherwise unnoticeable breach of a dam in Massachusetts. Alternatively, I implementation of the recommendation from the 135 Day Interim Report (Blasland and Bouck, 1985) on remedial alternatives (i.e., I Flow and Sedimentation Control at the Woods Pond Dam) would pro- I I 6-106 I Lawler Matuskx \ Skclly Engineers PGE00047320 I I bably decrease PCB transport from upstream and would also be impor tant to measure. As some aspects of the modeling analysis are less • certain than others, additional data are needed to develop a more complete understanding of the fate of PCBs in the Housatonic River. •

~~In recommending what monitoring should be done, several principles and considerations have been applied: I~~

• Baseline survey data onPCB concentrations inbed • sediments andd fisfish h havhave ebee been n acquiredacnuired. . • The most important compartment of the environment al system to monitor is the fish because of the I interfaces with humans. Sediments, which do not have to be monitored as • frequently as fish and water, should exhibit a — gradual but steady decline in PCBs, and thereby indicate cumulative changes. • PCB transport measurements should continue at the Kent sampling station; although the existing PCB • detection limit will probably rarely be exceeded, | these measurements will safeguard against unex pected increases in PCB. A complete definition of the PCB concentration of • background, or control, fish populations (includ ing their sediment and water exposure levels) • should be developed to set the ultimate recovery p endpoints. This report proposes a timetable for sampling the sediment, water, I and fish, and sets a schedule for a comprehensive data review and reassessment. Subsequent stages of monitoring would be planned in • accordance with the findings of the reassessment. The next three sections describe the plans for sampling the sediment, water, and I fish; the last section briefly describes when and how these data should be assessed. I I

~~6-107 • Lawlor Mnluskv 'Jikellv Knyiminr.s I PGE00047321 I I 6.5.2 Sediments~~

I The analytical variability between USGS and the Connecticut Agri cultural Experiment Station laboratories in the PCB measurements of I sediments, as well as the spatial variability in PCB concentrations within the two impoundments and two lakes, has been reported in I Frink et al. (1982). A shift in the mean PCB concentration of these four segments that would be statistically significant (95% confidence limit) translates into a projected number of years I according to the model results. As the baseline sediment data were based primarily on the 1979-1980 survey by Frink et al. and secon- I darily on the core sampling in 1986, the year 1982 is ascribed to the baseline. Assuming that the number of grab sediment samples I would be the same as that of the 1979-1980 survey (see Figure 6.3-3), a statistically significant change in the PCB concentration I of the lakes would occur in 10 years. Because the impoundments and riverine segments are expected to show I a more gradual change, more than doubling the number of samples, which amounted to only 20 samples in the five upstream segments, is I recommended to attain a desirably lower standard error of the mean. Sampling stations from the 1979-80 survey (see Appendix A of I Frink et al. 1982) should be reoccupied and additional new stations I should be interspersed among them in the upstream segments. The recommended number of sediment grab sampling stations by seg- I ment are: LOCATION SEGMENT NO. NUMBER OF STATIONS I Canaan 1 4 Falls Village 2 18 Cornwall 3 8 I Bulls Bridge 4 10 I (Continued) I 6-108 I Lawler. Malusky Skollv Kngineors PGE00047322 I I LOCATION SEGMENT NO. NUMBER OF STATIONS (Continued) I New Mil ford 5 8 L. Lillinonah 6 36 L. Zoar 7 22 I I The next comprehensive sediment survey of the Housatonic River in Connecticut is proposed for 1992 when discernibly lower PCB concen- • trations in surficial bed sediments are expected. In addition to the sampling of surficial sediment, a core sediment sample should • be taken in each impoundment and two in each lake, similar to the • 1986 sampling. The grab sediment samples and 1-in. increments of _ the four cores should be analyzed for total PCB concentration on a | dry-weight basis. I 6.5.3 Water

The model results indicate that the primary source of PCBs in the water in the study area is from upstream. The flow gaging and sam pling station for suspended solids and PCB concentrations, in- I stalled by USGS at Kent and operational since 1985, has shown that:

o PCB concentration at this location is lower than that measured upstream at Great Barrington, MA. _ The overwhelming majority of samples from Kent are below the PCB detection limit of 0.1 ug/1. I Periodic monitoring of the PCB transport at Kent 1s essential to keep a "pulse" on the situation. Although the data currently • available were of limited use in quantifying the upstream boundary condition in the model, additional data would allow more sophisti- • cated analyses. For example, statistical techniques are available m to extrapolate from the distribution of data above the detection

~~6-109 •~~

Lauler Mntusky Skolh I1 iii>im:crs I PGE00047323 I I limit to estimate the PCB concentration of samples in the subdetection limit range. These techniques have to be exercised I cautiously but may assist in evaluating the PCB concentration and I its trend. As PCB concentrations are consistently below the detection limit I particularly in the downstream segments, caged insect larvae provide a measure of the spatial and temporal trends in waterborne PCB concentration. Chironomus tentans, which have been used success I fully in the Hudson River, bioaccumulate PCB very rapidly such that a seven day exposure is sufficient to reach an equilibrium with the I water. This allows for a measure of short-term fluctuations in PCB concentration as well as long-term monitoring. Although experi- I mental analyses of this dipteran larvae have shown that its biocon centration factor varies for different PCB congeners, the biocon I centration factors for the isomer groups that are the principal components of Aroclor 1260, which is most prevalent in the Housatonic River, fall into a relatively narrow range (Wood et al. I 1987). Concurrent measurements of PCB concentration in the water and caged insect larvae would provide additional data on site- I specific bioconcentration factors. Daily water samples composited over the exposure duration would account for any fluctuations in I the exposure concentration. The PCB analysis of water warrants low level measurements to accurately determine bioconcentration fac I tors. As part of the recent Remedial Investigation/Feasibility Study for the New Bedford Harbor Superfund site, analytical methods for measuring PCB concentrations in water achieved detection limits I of approximately 0.01 ug/1. Similar methods should be employed to analyze water samples in conjunction with the caged insect larvae I monitoring. I I I 6-110 I Lawlen Matusky Skelly Enffine«r.s PGE00047324 I I consistency of bioconcentration factors among stations and times of sampling will not only provide for an assessment of rela- • tive trends but also an estimate of the river's PCB concentration in the subdetection limit range that would be germane to the model • verification. In addition, these data provide a point of departure ™ to analyze the ecological transfer of PCB through the food chain. _

A culture of Chironomus tentans is available from the New York State Department of Environmental Conservation (NYSDEC) and can be • maintained in a rearing system similar to that described in Simpson et al. 1986. The cages for placing the larvae in the river are I made from a monofi lament screening fabric folded and stapled to form an envelope packet 6.5 x 12 on. Several packets are placed in • a steel mesh basket that 1s suspended from a float at a depth of • approximately 3 ft and anchored to a cinder block on the river bot- _ torn with wire cable. Control specimens placed in packets can be | brought to the site, then removed and frozen without being put into the river. •

~~The annual sampling of predatory insect larvae and hydropsycid I caddisfly at Cornwall by CDEP is a distinct monitoring effort that should continue independent of the caged insect larvae sampling. •~~

~~Quarterly caged insect larvae (Chironomus tentans) measurements are recommended at four sampling stations: I Falls Village • Kent • Lake Lillinonah (Route 133 Bridge) Lake Zoar (Route 84 Bridge) m~~

~~In addition, one sampling station on a tributary should be measured in each survey to quantify background levels of PCB. Caged insect larvae sampling for each of the next four years should have four I~~

~~6-111 • Liiwlor Matuskv Skclh I I PGE00047325 I~~

I quarterly surveys on each of the following tributaries: Tenmile, I Still, Shepaug, and Pomperaug rivers. The PCB data on insect larvae and water should be reviewed after I each survey to evaluate spatial trends and bioconcentration fac tors. The insect larvae sampling should commence in the spring of I 1989 and continue for four years until the next data assessment. I 6.5.4 Fish As part of the HRA,a 1988 fish survey will be conducted by ANSP. I We recommend that sampling and analysis of fish exposed to back ground PCB concentrations be added to the 1988 survey. Control I stations should be located on tributaries to the Housatonic up stream of a migration barrier, i.e., dam. To select control sam I pling stations that have a representative selection of those sam pled in the main stem, riverine and lacustrine sites should be ex I plored. Three tributaries that have dams in proximity to the I Housatonic (D. Majors pers. comnun.) are recommended: DISTANCE FROM DAM DAM TO HOUSATONIC HEIGHT I TRIBUTARY DAM (mi) (ft) TOWN Salmon Creek Lime rock 1.6 11 Sal isbury I Shepaug River Botsford Road 8.3 10 Roxbury I Pomperaug River Pomperaug River 3.2 13 Southbury I In addition, control stations in the Upper Shepaug Reservoir, Lake I Waramoug, and Woodride Lake would be suitable for lacustrine sam I I 6-112 I Lawler Matuskv 'Ukellv Unwinders PGE00047326 I pi ing. Sediment and water samples from these stations should also be analyzed for PCBs to evaluate the exposure levels. These data I will assist in quantifying the background levels of PCBs in fish at present and also in the future. I Sampling and PCB analysis of a single species and a certain size or age class was considered as a potential way of filtering some of I the variability from the PCB trend analysis. Smallmouth bass, the species analyzed most often for PCB concentration, are generally I present at the four sampling stations i.e., Cornwall, Bulls Bridge, Lakes Lillinonah and Zoar. The practicality of analyzing inten I sively a single age of smallmouth bass hinges on the number of specimens needed to detect a certain change in PCB concentration. •

The size-age distribution of smallmouth bass sampled by ANSP in _ 1984 and 1986 showed the highest percent of fish in the 200-250 mm | range of length; age three and four year olds account for approxi mately 40 percent each of the fish in this size class. The coeffi- • cient of variation of PCB concentration for this size class is ap proximately 50 percent. Furthermore, the three and four year olds • have similar coefficients. Approximately 100 specimens of either age would have to be analyzed to detect a 25 percent change in PCB • concentration (at " = 0.05 and 0 = 0.95). This is equivalent to a • sevenfold increase in the average number of specimens analyzed per station by ANSP. As fish collection would be geared to size, the g numbers caught would have to be two to three times the required number of a certain age. These numbers of three and four year old • smallmouth bass probably could not be caught at Bulls Bridge and Cornwall and may affect the bass populations of Lakes Lillinonah • and Zoars. I I

6-113 • Law I or Nhitusky Skully Mnoimu;r.s I PGE00047327 I I The ANSP fish monitoring program has evolved into a minimum number of species of interest to cover the diverse modes of PCB uptake and I sport fishery importance. In addition, PCB data on the full range of fish age and size are important to address concerns such as the I FDA action level. While narrowing the program further seems im practical, it may be possible to allocate the number of individuals I within certain species and age strata to attain greater statistical power for trend analysis.

I Certain species, i.e., American eel and white crappie, had relatively high PCB concentrations, according to the 1979 and 1983 I data, but were not sampled thereafter. Sampling for these species I is recommended to achieve a broad-based monitoring program. In light of the lack of a substantial change in PCB concentration I between 1984 and 1986, the next survey following the scheduled 1988 survey should be in 1992. This would coincide with the planned sediment survey and provide synoptic data on these two environmen I tal compartments. The sampling protocol for the 1992 survey should remain the same as that currently followed (except for the addition I of background stations and possible reallocation of individuals in species and size strata) to provide program stability for a long I term assessment. I 6.5.5. Data Review and Assessment The recommended sampling described in the three previous sections I is summarized in Table 6.5-1. For each type of survey (sediment, water, caged insect larvae, and fish), the data should be reviewed I shortly after they are available. This is basically a routine quality control-type procedure to be sure that the survey and lab I oratory analyses are reliable. I I 6-114 I Lnwlcr Matusky Skelly E PGE00047328 I

TD •• •>— I oe LU in x: Q. »/— TJ 4-> -r— LU ^ r— • 1— cnr— ^r o o cj a> ex c «^ 0 0 O > OJ • ce 1— 1— H- O» •— •— *-> I Q. -5 T f • en CQ CO CQ O CQ ^ CO CO DO ••— a> O 0 O — O 0 O o o QJ en D_ Q. O-

-O m ^ C 0> O rd E -a^ 1- ra (O -M tn a> •— c I CO c L. ^ tt> JT o jC ry O *J to en «3 o QJ fT3 C •r tn 4-> rd c "~-^ en c: ^* 4-> Q. Cl r— O T> TJ T3 O cj t— id 3 QJ r— c c: C" .p— c™ Ctl r^ 3 .— V •• CD K l/>C-i->' O >• •— 1— O ,— O i— CO i— 1— I O CD CO •r CZ C •— 10 l_ L_ rO r— n3 o; c it o 3 m •. o en r5^ cn ^ in •<— o Q. • . ca> £ o +-> •— •*-» ii>j-i^ c ^: c i— i rvi •i-L-cncL C •— c o !_ U 1_ i— CD a I— O CJ E CD (O QJ • * *C O ra o 3 • • jz E U-£l/> i—i ±i U.^_I_JCO CJ CQ O CO —I —1 I cc co o i—. O) 1 z -o o c I 2: o> t3

CQ H I—I O 0 Q. O O. a> 0 OJ 3 -M J-> < 1— cc o; • r— I/) I LU to *j a> m in 4-) i > U— in -a o reJ O L. 0 (U i —• o LL. O Q> ^ Q> C) > O. flJ Q. r— 1 or O •a L r^. cnini_T36-*-> e 1— >~i L. o i lacngcioi a o •r— CJ t LU tu o O {_>"-•— io o 5 o LL. —. (T3 Q_ I O E I c a:

in >) t\J > CM C\J J (M r— CM ^-^ CO I uj -^ CO C7> O CT> L. CO C CO >• Q Cf> O> »— CTi (L) CO 3 w—1 iy i »—1 *—4 V •—4 -M »—* *~3 —^ ; 1 1 L. 1 1 *• LO C. CO J= CO ra CO >, co t/1 co en co 3 CO (O CO I CO --~ CT^ O" co s: CO I—H ^-^* T-H

~~r~i ~ L I 4-J h- c u C a> a> f~ 6 i- rO in l/> 1 .^ a> 4J C •r— T3 -I-1 O •-• U.~~

~~6-114A I I PGE00047329 I~~

I A comprehensive evaluation of the river system, consistent with the analytical framework developed by this study, should be completed I following the 1992 sediment and fish surveys. The goal of this assessment would be a confirmation of the PCB projections; however, I the results may also aid in adjusting PCB projections based on up dated data. No matter what the outcome, the periodic revaluation I is essential for reliable planning and environmental management. The model would be applied to simulate the Housatonic River from I the baseline (1982) to the reassessment milestone (1992). Actual data, such as river flows, suspended solids, and PCB concentrations I entering the system, will be available for model input. The model output would be compared with the monitoring data, and any differ I ences will be critically evaluated. Modeling assumptions and pa rameter evaluations would be scrutinized to identify the source of I any deviations in the model. Parameters may be reevaluated and assumptions may be adjusted within reasonable bounds to rectify the model. This process of model verification or recalibration would I be followed by modeling projections and lead to recommendations on I the next stage of monitoring and reassessment. I I I I I I I 6-115 I Luwler Matusk okull K PGE00047330 I I I REFERENCES CITED Academy of Natural Sciences of Philadelphia (ANSP). 1985. Interim report on PCB concentrations in fish from the Housatonic River, I Connecticut. Prepared for General Electric. Report No. 85-10F. 35 p. I Aylor, D.E. and C.R. Frink. 1980. RVRFLO: A hydraulic simulator of water quality in the Housatonic River in Connecticut. Connecticut Agricultural Experiment Station Bulletin. I Barnard, William D. 1978. Prediction and control of dredged mate rial dispersion around dredging and open-water pipeline dis posal operations. U.S. Army Engineer Waterways Experiment Sta I tion, Tech. Rep. DS-78-13. Barry, T. n.d. An angler survey and economic study of the Housa I tonic River Fishery Resource, 1 Dec 1984 to 31 Nov 1985. Pre pared for Department of Environmental Protection, Bureau of Fisheries, Hartford, Connecticut. I Bishop, R.C., and T.A. Heberlein. 1979. Measuring values of extramarket goods: are indirect measures biased? Amer. J. I Agr. Econ. 61(5). Blasland and Bouch. 1985. Housatonic River Study 135 Day Interim Report Assessment of Remedial Alternatives. Prepared for I General Electric Co., Pittsfield, MA. Bohlen, W.F., D.F. Cundy, and J.M. Tramontane. 1979. Suspended material distributions in the wake of estuarine channel dredg I ing operations. Estuarine and Coastal Marine Science 9,699-711. I Brown, J.F. 1987. Letter re: PCB dechlorination in Wood's Pond sediments dated 18 March 1987 to J.H. Thayer (GE, Pittsfield, MA). I Bruner, R.J., and D.W. Hill. 1977. Ambient concentrations of PCBs in the southeast from STORET data and selected EPA studies. I U.S. Environmental Protection Agency, 21 p. Canale, Raymond P., and A.H. Vogel. 1974. Effects of temperature on phytoplankton growth. 100 (EE). I Chapra, S.C. and K.H. Reckhow. 1983. Engineering approaches for lake management Volume 2: mechanistic modeling. Biutterworth I Publishers. I R-l I Lawler, Matusky &f Skelly Engineers PGE00047331 I I

~~REFERENCES CITED I (Continued)~~

Connecticut Department of Environmental Protection (CDEP). No date (a).1984-1985 Connecticut fish distribution report. Bureau of m Fisheries, Hartford, CT. • Connecticut Department of Environmental Protection (CDEP) No date. (b). 1987 Connecticut angler's guide. Bureau of Fisheries, • Hartford, CT. | Connolly, J.P., and R. P. Winfield. 1984. WASTOX, a framework for « modeling the fate of toxic chemicals in aquatic environments • Part 1: exposure concentration. Performed for U.S. Environ — mental Protection Agency. EPA-6001 3-84-007. Cordle, F., P. Corneliussen, C. Jelinek, B. Hackley, R. Lehman, J. • Mclaughlin, R. Rhoden, and R. Shapiro. 1978. Human exposure to polychlorinated biphenyls and polybrominated biphenyls. Environ. Health Perspec. 24:157-172. I

Crump-Wiesner, H.J., H.R. Feltz, and M.L. Yates. 1973. A study of — the distribution of polychlorinated biphenyls in the aquatic • environment. Jour. Research U.S. Geol. Survey 1(5):603-607. • Dennis, D.S. 1974. Polychlorinated biphenyls in the surface • waters and bottom sediments of the major drainage basins of the | United States. Prepared for the U.S. Environmental Protection Agency, Office of Pesticide Programs, NTIS No. PB-276313/4.12p. g Deshefy, S. 1985. Proposed addition of Unit 7 at Bulls Bridge. • Letter Department of Environmental Protection, Office of Planning and Coordination. B Dexter, R.N., W.G. Nines, E. Quinlan, and S.P. Pavlou. 1978. Dynamics of polychlorinated biphenyls in the upper Mississippi • River. Task 2: evaluation of compiled information. Fish and I Wildlife Service, U.S. Department of the Interior. DiToro, D.M. 1985.A particle interaction model of reversible • organic chemical sorption. Chemosphere 14(1). • Finlay, D.J., F.H. Siff, and V.J. Decarlo. 1976. Review of PCB • levels in the environment. Environmental Protection Agency. | 143 p. Frink, C.R. 1978. Distribution of PCBs in sediments. The Conn. I Agricultural Experiment Station, New Haven, CT. 5 p. — I

~~Lawler, Matusky Sf Skelly Engineers |I PGE00047332 I I I REFERENCES CITED (Continued)~~

I Frink, C.R., B.L. Sawhney, K.P. Kulp, and C.G. Fredette.1982. Polychlorinated biphenyls in Housatonic River sediments in Massachusetts and Connecticut: determination, distribution, I and transport. A cooperative study by the Conn. Agricultural Experiment Station, the Conn. Department of Environmental Pro- tection, and the U.S. Geological Survey. I Raines, T.A. 1983. Organochlorine residues in brook trout from remote lakes in the northeastern United States. Water, Air, I and Soil Pollution 20:47-54. Henderson, C., A. Inglis, and W.L. Johnson. 1971. Residues in fish,wildlife, and estuaries. Pesticide Monitoring Journal I 5(1): 1-11. Hesse, J.L. 1975. Contaminants in Great Lakes fish. Michigan Water Resources Commission, Staff Report 15 p. (eds.). New I York Academy of Sciences, New York. Horn, E.G., L.S. Hetling, and T.J. Tofflemire. 1979. the problem I of PCBs in the Hudson River system. p 591-609. In Health effects of halogenated aromatic hyudrocarbons. W.J. Nicholson and J.A. Moore I Hyatt, W. No date. An angler survey and economic study of the Farmington River fishery resource - final report. Prepared for I CDEP, Hartford, CT. Jensen, A.L., S.A. Spigarelli, and M.M. Thommes. 1983. PCB uptake I by five species of fish 1n Lake Michigan, Green Bay of Lake. Karickhoff, S.W. 1984. Organic polllutant sorption in aquatic systems. ASCE Journal of Hydraulic Engineering. 110(6). I Lawler, Matusky & Skelly Engineers (LMS). 1985a. Housatonic River PCB sediment management study. Prepared for General Electric Co., Fairfield, CT. I b Lawler, Matusky & Skelly Engineers (LMS). 1985. Water quality modeling and wasteload allocation workshop manual. Prepared I for American Petroleum Institute. McFarland, V.A., and J.U. Clark. 1986. Bioavailability and sedi ment quality criteria. Seventh annual Meeting of the Society I of Environmental Toxicology and Chemistry, Alexandria, VA. I R-3 I Lawler, Vtatunky Sf Skelly Engineers PGE00047333 I I REFERENCES CITED I (Continued)

Michigan, and Cayuga Lake, New York. Can. J. Fish. Aquat. Sc: I 39:700-709. Mills, W.B., J.D. Dean, O.B. Porcella, S.A. Gherini, R.J.M. Hud I son, W.E. Friek, G.L. Rupp, and G.F. Bowie. 1982. Water qua lity assessment: a screening procedure for toxic and conven- • tional pollutants. Prepared for U.S. EPA. p EPA—600/6-82-004a. Moulton, J.C. 1980. Management of the Housatonic River as a I "catch and release" area. Interdepartment Message to R.A. " Jones, Chief,CDEP - Fisheries. National Analysts, n.d. 1975. National survey of hunting, fish- I ing, and wildlife - associated recreation. Prepared for U.S. Department of Interior, Fish and Wildlife Service. • National Academy of Science (NAS). 1979. Polychlorinated bi phenyls. Report by the Comm. on Assessment of PCBs in the environment. Nat. Acad. Sci., Washington, D.C. • New England River Basins Commission (NERBC). 1980. Interim plan for the disposal of dredged material from Long Island Sound. • Niimi, A.J., and B.C. Oliver. 1983. Biological half lives of polychlorinated biphenyl (PCB) congenessrin whole fish and _ muscle of rainbow trout (Salmo gairdneri). Can. J. Fish. • Aquat. Sci. 40:1388-1394. • Nisbet, I.C. 1976. Criteria document for PCBS. Massachusetts • Audubon Society. Prepared for U.S. Environmental Protection • Agency. Nisbet, I.C. and A.F. Sarofim. 1972. Rates and routes of tran- I sport of PCBs in the environment. Environ. Health Perspec. 1:21-38. Norstrom, R.J., A.E. McKinnon, and A.S.W. DeFreitas. 12976 A bio- • energetics based model for pollutant accumulation by fish. Simultation of PCB and methylmercury residue levels in Ottawa • River yellow perch (Pevca flavescens). J. Fish, Res. Board I Com. 33:248-267. Norvell, W.A., C.R. Frink, and D.E. Hill. 1979. Phosphorus in B Connecticut lakes predicted by land use. Proc. Natl. Acad. • Sci. 76:5426-5429. I

~~Lawler, Matusky Sf Gkelly Engineers I jg~~

~~PGE00047334 I I~~

~~REFERENCES CITED I (Continued)~~

I Northeast Utilities Service Company. 1982. Report on recreational resources Housatonic project. Submitted to the Federal Energy I Regulatory Commission Project No. 2576. O'Connor, D.J. and J.P. Connolly. 1980. The effect of concen tration of adsorbing solids on the partition coefficients, I Water Research, 14:1517-1523. Orciari R.t and C. Phillips. 1986. Establishment and evalluation of two trout management areas on the Housatonic and Williman I ticx Rivers. Connecticut Department of Environmental Protec- tion, Hartford, CT. Onuschak, Emil Jr. 1982. Distribution of suspended sediment in I the Patuxtent River, Maryland during dredging operations for construction of a pipeline. Bulletin of the Association of I Engineering Geologists Vol XIX, Nol. 1, 1982 pp 25-34. Pavlou, S.P, and R.N. Dezter. 1979. Distribution of polychlori nated biphenyls (PCB) in estuarine ecosystems. Testing the I concept of equilibrium paritioning in the marine environment. Environ. Sci. Technol. 13:65-71. Phillips, K.E., and J.F. Malek. 1987. Evaluation of dredging as a I remedial technology for the commencement bay Superfund site. Proceedings of llth U.S./Japan Experts Meeting, 4-6 November I 1985. Seattle, WA. T.R. Patin, (ed.) Price, J.W. 1931. Gross and gill development in the small-mouthed black bass, micropterus dolomieu, Lac'epj'de contributions of the Ohio State University, Franz Theodore Stone Laboratory. I Ohio State Univ. Press, Collumbus, OH. Robbins, J.A., and D.N. Edgington. 1975. Determination of recent I sedimentation rates in Lake Michigan using Pb-210 and Cs-137 (Geochimlcaet Cosmochimica Acta). Jour, of the Geochemical I Society and the Meteoritical Society. 39(3). Sawhney, B.L., C.R. Frink, and W. Glowa. 1981. PCBs in the Housatonic River: determination and distribution. J. I Environ. Qual. 10(4):444-448. Schmitt, C.J., J.L. Ludke, and D.F. Walsh. 1981. Organochlonne, residues in fish: National Pesticide Monitoring Program, I 1970-74. Pesticides Monitoring Journal. 14(4):136-206. I R-5 I Lawler, Matusky Sf Skelly Engineers PGE00047335 I

REFERENCES CITED I (Continued) I Schmitt, C.J, M.A. Ribick, J.L. Ludke, and T.W. May. 1983. Orga nochlorine residues in freshwater fish, 1976-1979: National Pesticide Monitoring Program. U.S. Fish and Wildlife Service, • Washington, DC Resour. Publ. 152.62 p. • Schmitt, C.J., J.L. Zajicek, and M.A. Ribick. 1985. National • Pesticide Monitoring Program: residues of organochlorine | chemicals in freshwater fish, 1980-81. Arch. Environ. Contam. Toxicol. 14:225-260. _ Schnoor, J.L. Field validation of water quality criteria for ~ hydrophobic pollutants. Aquatic toxicology and hazard assess ment: Fifth Conference, ASTM, STP 766, J.G. Pearson, R.B. • Foster, and W.E. Bishop, eds. American Society for Testing and • Materials, pp 302-315. Simpson, K.W., M.A. Novak, and A.A. Reilly. 1986. Biomonitoring J of PCBs in the Hudson River. I. Results of long-term moni toring using caddisfly (Insecta: Trichoptera: Hydropsychidae) larvae and multiplate residues. II. Development of field pro- • tocol for monitoring PCS uptake by caged live Chironomus ten- • tans (Insecta: Diptera; Chironomidae) larvae during dredging operations performed as part of the Hudson River Reclamation • Demonstration Project. | Steward Laboratories, Inc. 1982. Housatonic River Study 1980 and _ 1982 Investigations. Final Report. Prepared for General Elec- • trie Company. * Sloan, R.J., and E.G. Horn. 1986. Contaminants in Hudson River • striped bass: 1978-1985. New York State Department of • Environmental Conservation. Division of fish and wildlife. Technical Report 86-2 (BEP). • Spagnoli, J.J., and L.C. Skinner. 1977. PCBs in fish from select ed waters of New York State. Pesticides Monitoring Journal. ll(2):69-87. • Stabler, M.J. 1980. Estimation of the economic benefits of fish ing: a review not. In allocation of fishery resources. Pro- • ceedings of the Technical Consultation on Allocation of Fishery | Resources Held in Vichy, France, 20-23 Apr. 1980. J.H. Grover (ed.) Department of Fisheries and Allied Aquacul tures, Auburn Univ., AL. I I R-6 g Law lor, Matusky Skelly Kntpnours p

~~PGE00047336 I I I REFERENCES CITED (Continued)~~

I Swackhamer, D.L., and D.E. Armstrong. 1986. Estimation of the atmospheric and non atmospheric contributions and losses of polychlorinated biphenyls for Lake Michigan on the basis of I sediment records of remote lakes. Environ. Sci. Technol. 20(9):879-883. Swain, W.R. 1983. An overview of the scientific basis for concern I with polychlorinated biphenyls in the Great Lakes, pp 11-48. In PCBs: hjumas and environmental hazards. P.M. D'Intri, and I M.A. MKamrin (eds.) Butterworths, Woburn, MA. Swain, W.R. 1978. Chlorinated organic residues in fish, water, and precipitation from the vicinity of Isle Royale, Laker I Superior. J. Great Lakes Res. Internat. Assoc., Great Lakes Res. 4(3-4):398-407. Thomann R.V. and J.P. Connolly. 1984. Model of PCB in the Lake I Michigan lake trout food chain. Environmental Science & Tech- nology 18:65-71. I U.S. Environmental Protection Agency (EPA). 1978. Polychlorinated biphenyls ambient water quality criteria. U.S. Environmental Protection Agency (EPA). 1983. Environmental I transport and transformation of polychlorinated biphenyls. EPA 560/5-83-025. I U.S. Fish and Wildlife Service (FWS). 1982. 1980 national survey of fishing, hunting, and wildlife - associated recreation. I U.S. Govt. Printing Office, Washington, DC. Veith G.D. 1975. Baseline concentrations of polychlorinated bi phenyls and DDT in Lake Michigan fish, 1971. Pesticide. Moni I toring Journal 9(l):21-29. Veith, G.D., D.L. DeFoe., and B.V. Bergstedt. 1979b. Measuring and estimating the bioconcentration factor of chemicals in I fish. J. Fish. Res. Board Can. 36:1040-1048. Veith, G.D., D.W. Kuehl, E.N. Leonard, F.A. Puglisi, and A.E. Lemke. 1979. Fish, wildlife, and estuaries polychlorinated I biphenyls and other organic chemical residues in fish from major watersheds of the United States, 1976. Pesticides Moni I toring Journal . 13(l):l-8. I R-7 I Luwlor, Mntusky 'Skelly 1CnDinners PGE00047337 I I

~~REFERENCES CITED I (Continued)~~

Warren, C.E., and G.E. Davis. 1967. Laboratory studies on the feeding, bioenergetlcs, and growth of fish. The Biological * Basis of freshwater fish production, p. 175. S.D. Gerkin (ed.) • Blackwell, Oxford. Winger, P.U., C. Sieckman, T.W. May, and W.W. Johnson. 1984. • Residues of organochlorine insecticides, polychlorinated • biphenyls, and heavy metals in biota from Apalachicola River, FL. 0. Assoc. Off. Anal. Chem. 67(2):325-333. • Wood, L.W., G.Y. Rhee, B. Bush, and E. Barnard. 1987. Sediment desorption .of PCB congeners and their bio-uptake by dipteran ^ larvae. Water Resources. Vol. 21, No. 8, pp. 875-884. • I I I I I I I I I I R-8 • La\vh«r Matusky "'JJkuIly Kn^inoors • PGE00047338 I I I I I I I I APPENDIX A ™ SCATTER PLOTS OF FISH PCB CONCENTRATION (ug/g) VERSUS TOTAL LENGTH (mm)

~~I I I I I I I I I PGE00047339 1~~

~~1 »" a~~

.-> O-. • Tro 1 O c X > C, -^ n ^ c *• i •*• a - ° u I C * 0 o-*. 0° " > i O X » (< ° ^ •* 0 c.X.,*o » "4 ^ 0 •> 1 « o* o y •. •* j ° 1 •i C"* r »v* ° "CD 1 _'' 'gd rt x * °°" . a •• o «i * „ 0 a Jl 0 a 0 0 o °- D x p^, 1 I ^ °o ' " 0\ 0••"• o° . Z3 X o a OJ -CD pj •f o

~~O o "s i Q 1 I X 0 > C_ S X ~i a C a « OJ Q. a CM I CO I CD [ — -• ' r-^ , , r—, , I " "" OJ '•• -~~-- = -.--,. ~ =33~~

~~1 trj — c: •—cL.f-ica —' CT »- C- tLJ JD~~

~~' [1~~

~ •*» tri OJM r «% « i o ^, *** o 1 * * X - X * o ° * * * « « J* 1 * ^ o 0 * 0 « " ° ", C. * ** 0 *« « a '= O x ° >t . -s>"- \f *-* * 1 x *-* "CD A * )f o o M ro i 0° ^ ' f* o > * ^*° o * 0 %"** " -•« rn o a 1 I -! M . »r< o° o -* , < 0 M *< • s (=; •o _O M r = K o X 0 « O 1 1 CQ 1 a » ITi Q. ^ (\j fM 1=1 1 »

~~[ , , , , —, T -, .— . r~~

~~' 1 - i *t f .~~

~~PGE00047340 I I 7= 1~~

~~o m ••«• Y~~

* »« a c_ V *B " : a"? * "' CD -b"-K ° . ca "*• T ° ' I I I a I PGE00047341 1 | • i 1 - ! +• 1 ,_ ' » i 1 . - , 1 0 ™ 0 1 •> a _5 ° -A n ^ 1 ° * ^ o D 1 0 o 13 a u ' .=-,/'"' " ^ 1 <• " —a3 ' 0 * ! ? 1 S v I O ' f ^ D

~~wi —' %A> * " ° o 1 1 * * "~3 i < » - ^ C2. X A *•> * t .X c X " »-—J ! o iJ | _ 1 S c ^ - • f^ i i t o> x 1 1 6 » -: a a "s "^ ™ C\J I~~

~~T- i— r~~

~~" CJ w «r^*.^-.— " — • — - ~~

~~•CT* " ~ ~ 4s w> i 1 , . C. < CL, i. t -Jl 1 o 1~~

~~1 0 ! O -^ x I • 0 " 1 0 ^ ° » x x . ^ ° 1 0 „ o a 1 Q x 1 o u o x o * ~,;» - a «vox "! * |~~

~~•» I >f 1 O rfl O >f «» ^ ** • o l> « ° S •= , x ^ - - *« r 1 (0e- (_J D o~~

~~O 1 *^ 1° tri e i/1. E m Q— . a , J 1 CM CM ' » T • •• t f 1 1— , 1 ' 1 1 a ' T T~ T r~ —i T i r T i r- r~~

~~o 1 .-<: i— a_(_-icn —> c »— a_t_>cn PGE00047342 1~~

~~I 1 1~~

~~1 I ? o u ' <> X » ° ' = • .;. i LI ^f o -<>< -t X X ^ *" ,.~~

*• u i 0 » ": s :'. ",-.v ' 1 uH 0 « " " D * o o = " , ..-• I cm o 0« ° ° ™ o J ~ J — — »• E 1 * 5 I N CM ~ti a

~~1 ' ' ' l ' ' «"------*U-.L— .-.-. -, ^|~~

~~• -. c i n. L-j^r: *—. cr i— n Ljtn ? : 1 1 1~~

~~9 x~~

~~0 o o " ' " .-•'" " ' 1 K "V « 0 •<* ' ^^ H a 0 - . ° ° " o o O 0~~

X O A 0 LO o ^> o* Ao rat 1 CTD X *fc xx^?o^» ^..> ^ cn ^o %« x x -/ C* ( | *>x ^ j «] O .cr *M * « ° •5 * .- -0 " ' • /" - K » " " 0 • ra x o F ~ °^L • * CD \ x *** -s ° ° t= cu o o a U 0 . cn o c_ DO ra O | 1 r^i | 1 " CM 1 ^ H .. • • cn i •

~~<> c ""* ^ H •-» C *— Q. *_J CO - ~^ m PGE00047343 1 1~~

~~O " -4. 1 1 1 t ""• 1~~

•> 1 D * 1 "" i O 0 a o : o 1 ^0 h-. <; o_. \ l o CD. •- O- * i x CO n 0 ° ^ U1 o *

0 t-n 0 0 h F=! 1 s 1 0 9t ~ ^ £ rj 1 D n >f *^ x CD ., * x ** CD n o CD * * + oi y x F f«Ci o « F— ° * ^S. D i 9 '"" CD ' * x d_ ?, i a° *> * 1 * 1 a o 0 DJ< f- -' I E a r™ i a. 1 0 a CM i i i

~~1 0 1 x | .S t-: •» 1 ••~~

~~O i >, 0 x 1 " w» r m~~

~~o o X, o X o o ->o 1 c, ^^ c^_ o X 1 = 0 X LH 0 O X X~~

~~o i P! x «K> -= o *•% X D 1 C"D A ^x CD ' 0 0 *D X 0 *^0 F= o°o % o »"« %, c " hS " - * -• f V ' 0°0 °* " ' * 1 ^ 0 * 0 E s Q. ^ = &a N' 1 t "r1 1 " • 1~~

~~CT> 1 cn~~

~~o —• d i— o_tjinn — c. t— a.c_>cn 1 PGE00047344 I I I I _, ^~» < ~^"' «^l o* ° " A" *° _ rJ" s .K-.,-Tf* .t->*•*.•.. I~~

~~fl^-^^'> J^? tj t^ •>'* " - ^ ^^B °o V\* h"" * flx Q~~

~~/R <^ o°* » o o 0 X ~ 1 A r-j X •" tr s E a. CL a CM C\J ; ! i~~

~~T r o ^M r~ JT ^ r ' ~^ ' - • ~j~i u ... 1~~

*a OJ r? S3 L ^K

~~^- c~ — -v (_im 1 1~~

~~— 1- j. MB " 1 + ™ a » U3 ** +• MI X ^ O1~~

~~c o -* 0 o <, 0 » O 0> X H< en o o o*-u ^H en *** wxx v o ^" 1W » * ° M ro o ° -"rto O j^. f«?r-* *>j^* ., H A ^ o •* a 1 1 •«* „ •" ^H~~

~~O -f O "• (=! (•^^/•^> <**1 «h || n° 8tt^ o i^ o o -, * * \ 0 1*0 x fw °a "~~

*» ™ : " 1 ' '

~~rn 1 U1~~

~~^c ^^^ra — « cz •— O. Ljm ^B~~

~~PGE00047345 1 1 ( _~~

~~a. M M M 5 • in a o a ~^- ~ 1 OOB»t^ CM oo«ou CJ i :' M _ U 1 K~~

0 * • 0* 1" r n r ° X "« A v U IU O •• 1 4 U _ _ _ ° "•"ITi» 0 Q 4* b *J « ti ^ 0 U t. a L, ^ ' • O X '^ * u u J x= co OUD •r-> Q | u° oo i4«* 0 o o iiJdty u ; X (. */** '^ J3 o J (i ^^ a o a u v ° ^^ o o o o^ X J 0*" J J 8 *» * H* ° •: 1=-' * »" . ' '^ ' S o0 0 f v* B° ' ,r° '

~~0 0 " ^~ M - * ' o-. - I JH lj • /—I o •£ § * ." M°° <> ~* M 1 M S o .J Q u , —~~

~~I * r~ ,,,,,. i • i i ^~~

~~• ~^~~

~~j5)~~

~~I •^ 1 •—» c >— n,<_,m ^C ^-Q-t-LO 1 1 E E T Q. 1 -,, § 00 O d u W OO«BOV CM ^^ K X >«*"»'+. +• +~~

1 fu " >. ^ -*o , •»• * * .- ° « 'r" +• -r* + • V^ * j + ^tX T . x 1 a o * u **»/ •• _cz *u o + , r **" Q .»( *lt * K X -*K W if 1 s v ro « * -»>+-*^* ^ - -f* + a u a o » < ^M* 8 -Z f^ QJ

~~1 T , ——i—-.: , , , , , , ,—, r—, ,—r •~~

~~m U 1 -I PGE00047346 I~~

~~I— I I~~

~~o o o>) Q * O '""T, I~~

~~C- i ir B « » U_) I X ** X a. ' I *«« v cu ! X O~~

~~• r- -t ' -CZ ! I I I~~

~~—• c. •—a- t_>to I I I ** (- 1 f~~

~~0 00 c. « XX i-' O X I ^ 0 a x x 0 ° -* * ! , o +• -~ 0 xj •O O 1 ' 1 X x o v + ° « 0 I . ° OOP u . X X «L"~~

X »* o 0 fyV. K X — * » ^^. V"*^ ^ O O(^ ** = I Q ; x iL ^ " 1 X c < 0 0 QJ N^ «« » 0 x M * X ^• ^° d. u w X K 0 " xx x i=! ° "o o 0 .» <^o ^ I — ** * X I o X . I ;H § ' E o. o C3 a • CL a CJ OJ i \ I 1 « p> nj rJ — — 0 0 I I PGE00047347 1

~~_ - _^ ______— 1 ^ f [ 1 c~~

a a a o 1 OJ —i c) o " c 0 "u O a \ . Q 0 o +• - £ x 1 o ** » ir C o O K x O a x x X * + -3 o c_ M S ^ o X D a ^ [? 1 ^ a ^(_) ^ LT) - i l I i r •pr ! "s. 1 I E ~ "^ a i Q. a a CM ••" a CM

~~1 1 r 1 1 , „ ~ ^ • * ~ OJ •77?:"* ca 1 5 0 •—> c: >— o- 1—J EIJ ~- c: *— a-cjea 1 1~~

** f

~~o o I I * 0 X "* X •o ^*o a~~

~~0 O 4 a o o " a * 1 x~~

~~c_ O .^ A o X A V * o 0~~

~~1 1 "^ rJ •* * LO >~~

~~QJ ' •> 1 I~~

~~.- O 1 CM~~

•*

' * ' *

~~en (.11~~

~~_ -0.^-cr, 1 <—i CI t— CL.CJOD c PGE00047348 I I I I ttJ u c _ fLJ I~~

~~CD I~~

~~- I a. i I I I I I I I L ~ I *** tH *^ C I~~

~~L J N OM „ x o C_ I K " \ ni ' Q_ I I I I~~

~~•• €Z t— n_ CJCD I PGE00047349 I I I I I I I I APPENDIX B PCB ANALYSIS OF ONE INCH INCREMENTS OF SEDIMENT FROM _ CORE SAMPLES OF HOUSATONIC RIVER IN 1986~~

I I I I I I I I I PGE00047350 I I I a • O en I to •-< I to a —< . . DC CO O CO I o (O M w-i O O «* '-• CM • CO (£> LO LT) LO LO LU p-v i^ r^ r-^ r^ rv. r^ ' LU • LO LO LO LO LO LO LO o; I— I a: o

~~I < co tO 01 CM cooor^oooocoooi— I o to LU •-I CO to I > I CO —I en 01 toincvir— CO CO • •-•OCMCM CD I" CM O HJ Q_ to n I •-t Q. Of 10_ 1 O o I o~~

~~ce. CD r»1CM;CM'-«O.-«'-~~

~~I HI to co t-^ ir> to roCT >n O> O> O> <7> O) <7> CD C_Ol CL Lr-O~ Lr-O^ LrO^ LrO^ Lr-Os lfr>^ IrO^ LO LO kO LO LO If) LO I <~~

I Q. LU C i ii fv* O Q O I I I I I I I I I I I I I I I I I I I I I I LU Z I CO •-< I PGE00047351 I I LABORATORIES DIVISION I I I CERTIFIED REPORT TRANSMITTAL I 30870-0267 REPORT NUMBER I February 12 1987 DATE I Lawler, Matusky & Skelly Enpineers CLIENT One Blue Hill Plaza I Pearl River, NY 10965 I ATTENTION Mr. Guy Apicella The above referenced report is enclosed Copies or this report and supporting data I will be retained in our files in the event they are required for future reference If there are any questions concerning this report, please do not hesitate to contact us Any samples submitted to our Laboratory will be retained for a maximum of sixty (60) I days from receipt of this report, unless other arrangements are desired Naturally, as in the past, our staff will be pleased to quote on any future requirements you may have. In addition to the service provided, we also offer the following I Hazardous Waste Analyses Product Evaluation/R&D Water and Wastewater Analyses 1 Air and Process Gas Analyses Industrial Hygiene Surveys Metallurgical Analyses I Microbiological Analyses Mass Specrrometry Services I Very Truly Yours.

~~fcokt 0 Aj lV)/\ I RobertQ Bradley Vice President I I PGE00047352 I I I February 12, 1987~~

I 30870-0267 LAWLER MATUSXY & SKELLY ENGINEERS One Blue Hill Plaza I Pearl River, New York 10965 Attention- Mr. Guy Apicella I Client Project I.D.: (GE) 337-017 I PURPOSE One hundred (100) soil samples were submitted to York Laborato ries Division of Y'VC, Inc. by Lawler, '.fatusky % Skelly Engineers. I The client requester! the samples be analyzed for polychlorinated biphenyls (PCB's). I METHODOLOGY The samples were extracted and analyzed according to protocols supplied by the client. Basically samples were Soxhlet extracted I with 1:1 acetone/hexane. Sample extracts were cleaned up using Florisil and/or mercury if needed. It was noted that most sam ples required exhaustive mercury cleanuo due to extremely hipjh I sulfur concentrations. Soil blank matrix spikes and matrix spike duplicates were spiked at regular intervals to indicate method recovery and precision. I The instrumentation used was a Perkin-Elmer Model Sigma 110 ga-5 chromatograph equipped with an electron capture detector. I RESULTS The results are presented in the following Tables. Attached as I Appendix A are appropriate QA/QC and raw data. I Prepared by: ' < _, ( /s.*.',,^. Jeffrey C. "CurraTi Laboratory Manager I v ,' I JCC/md The liability of YWC, Inc. is limited to the actual dollar value I of this project. I PGE00047353 1 1

~~TABLE 1.0 I 30870-0267 LAWLER, UATDSKY 2: SKELLY ENGINEERS PCB RESULTS 1 AH values are mR/T~~

~~Lab I.D. 267-008 267-009 267-010 267-011 267-012 267-013 267-014 1~~

Type 57503 57518 57502 57501 57523 57528 57531 1 PCB 1016 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 PCB 1221 <0.10 <0.10 <0. 10 <0. 10 <0.10 <0.10 <0.10 1 PCB 1232 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 1 PCB 1242 <0.10 <0.10 <0. 10 <0.10 <0.10 <0.10 <0.10 PCB 1248 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 1 PCB 1254 <0.10 <0.10 <0. 10 <0. 10 <0.10 <0.10

1 TABLE 1.1 30870-0267 LAffLER, T4ATUSKY & SKELLY ENGINEERS 1 PCB RESULTS 1 All values are mg/Kp; dry basis. 1 Lab I .D. 267-015 267-016 267-017 267-018 267-019 267-020 267-021 1 Type 57516 57505 57522 57527 57515 57530 57517 PCB 1016 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10

PCB 1242 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 1C) <0. 10 1 PCB 1248 0.35 0.28 0.29 0.50 <0. 10 <0. 10 <0.10 PCB 1254 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1260 1.1 0.89 1.0 1.4 1. 1 1. 0 1.3 1 1 Lab I.D. 267-022 267-023 267-024 267-025 267-026 267-027 267-028 Type 57521 57525 57526 57510 57509 57506 57507 1 PCB 1016 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 PCB 1221 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 1 PCB 1232 <0.10 <0.10 <0.10 <0. 10

~~TABLE 1.2 1 30870-0267 LAWLER, MATUSKY & SKELLY ENGINEERS PCB RESULTS 1~~

All values are mg/K{» dry basis . 1 Lab I .D. 267-029 267-030 267-031 267-032 267-033 267-034 267-035 1 Typo 57508 57511 57512 57533 57534 57535 57536 1 PCB 1016 <0. 10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0.10 PCB 1221 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 1 PCB 1232 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10

~~PCB 1242 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 1~~

PCB 1248 1.1 1.1 1. 4 1.0 0. 25 0. 68 O.RO 1 PCB 1254 <0.10 <0. 10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 PCB 1260 3.2 3.7 3. 4 2.0 0. 71 1 .2 1.4 1 1 Lab I.D. 267-036 267-037 267-038 267-039 267-040 267-041 267-042 1 Type 57537 57538 57539 57540 57541 57542 57545 PCB 1016 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 1 PCB 1221 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0.10 PCB 1232 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0. 10 <0.10 1 PCB 1242 <0. 10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0.10

~~PCB 1248 0.84 0.87 <0. 10 0.30 0. 62 0. 67 0.59 PCB 1254 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0.10 1 PCB 1260 1.3 1.1 0. 12 0.80 1. 1 1. 1 0.90 1 1 PGE00047356 1 1~~

1 TABLE 1.3 30870-0267 LAWLER, MATUSKY & SKELLY ENGINEERS 1 PCB RESULTS 1 All values are mg/Kg dry basis . 1 Lab I.D. 267-043 267-044 267-045 267-046 267-047 267-048 267-049 1 Type 57544 57543 57546 57547 57550 57549 57552 PCB 1016 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1221 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 PCB 1232 <0.10 <0. 10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1242 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 <0, 10 1 PCB 1248 <0.10 0.36 0.30 <0.10 0. 42 0. 76 0.58 PCB 1254 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0.10 1 PCB 1260 0.95 1.9 2.3 0.52 1. 8 2. 6 2.3 1 1 Lab I.D. 267-050 267-051 267-052 267-053 267-054 267-055 267-056 Type 57553 57554 57555 57556 57557 57558 57559 1 PCB 1016 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 PCB 1221 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1232 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 PCB 1242 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10

~~PCB 1248 0.51 <0.10 0.47 0.31 <0. 10 <0. 10 0.34 1 PCB 1254 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1260 3.2 2.4 1.52 2.8 1. 6 0. 85 1 .2 1 PGE00047357 1 1~~

~~TABLE 1.4 1 30870-0267 LAWLER, MATUSKY & SKELLY ENGINEERS PCB RESULTS 1~~

~~All values are mg/Kg dry basi s . 1~~

Lab I.D. 267-057 267-058 267-059 267-060 267-061 267-062 267-06 3 1 Type 57560 57561 57562 57563 57564 57565 57566 1 PCB 1016 <0.10 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0.10 PCB 1221 <0.10 <0.10 <0.10 <0.10 <0.10

PCB 1242 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 1 PCB 1248 0.23 1.2 0.84 0.32 <0.10 <0. 10 <0. 10 1 PCB 1254 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0. 10 <0. 10 PCB 1260 0.89 4.7 3.3 2.7 0.23 <0. 10 <0. 10 1 1 Lab I .D. 267-064 267-065 267-066 267-067 267-068 267-069 267-07 0 1 Type 57567 57589 57585 57576 57586 57581 57582 PCB 1016 <0.10 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 1 PCB 1221 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10

~~PCB 1232 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 1~~

PCB 1242 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1248 <0.10 <0. 10 <0.10 <0.10 <0. 10 <0.10 <0. 10 PCB 1254 <0. 10 <0.10 <0. 10 <0. 10 <0.10 <0. 10 <0. 10 1 PCB 1260 <0. 10 1.2 1.0 1.1 1.2 1.2 1.5 1 1 PGE00047358 1 1 1 TABLE 1.5 30870-0287 LAWLER, :iATDSKY & SKELLY ENGINEERS 1 PCB RESULTS

1 All values are mg/Kg dry basis. 1 Lab I.D. 267-071 267-072 267-073 267-074 267-075 267-07 6 267-07 7 1 Type 57587 57578 57588 57579 57583 57580 57584 PCB 1016 <0.10 <0.10 <0. 10 <0.10 <0.10 <0. 10 <0. 10 1 PCB 1221 <0.10 <0.10 <0. 10 <0.10 <0.10 <0.10 <0. 10 PCB 1232 <0.10 <0.10 <0. 10 <0. 10 <0.10 <0.10 <0. 10

PCB 1242 <0.10 <0.10 <0. 10 <0.10 <0.10 <0.10 <0. 10 1 PCB 1248 0.48 2.2 0. 63 0.58 0.79 0.68 1. 2 PCB 1254 <0.10 <0.10 <0. 10 <0.10 <0.10 <0.10 <0. 10 1 PCB 1260 1.6 2.1 2. 4 2.9 3.5 2.2 3. 0 1 1 Lab I.D. 267-078 267-079 267-080 267-081 267-082 267-08 3 267-08 4 Type 57577 57569 57568 57574 57575 57573 57572 1 PCB 1016 <0.10 <0.10 <0. 10 <0.10 <0.10 <0. 10 <0. 10 PCB 1221 <0.10 <0.10 <0. 10 <0.10 <0.10 <0.10 <0. 10

~~1 PCB 1232~~

~~All values are mg/Xg dry basis . 1~~

Lab I.D. 267-085 267-086 267-087 267-088 267-089 267-090 267-09 1 1 Type 57571 575750 57602 57601 57600 57590 57594 1 PCB 1016 <0.10 <0.10 <0. 10 <0. 10 <0. 10 <0. 10 <0. 10 PCB 1221 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 1 PCB 1232 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 1 PCB 1242 <0. 10 <0. 10 <0.10 <0.10 <0.10 <0. 10 <0. 10 PCB 1248 <0.10 <0.10 <0.10 <0.10 <0.10 <0.10 < 0 . 1 0 1 PCB 1254 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 PCB 1260 0.63 0.50 <0.10 <0.10 <0.10 <0. 10 <0. 10 1 1 Lab I.D. 267-092 267-093 267-094 267-095 267-096 267-097 267-09 8 1 Type 57591 57592 57598 57597 57595 57596 57593 PCB 1016 <0.10 <0.10 <0.10 <0. 10 <0.10 <0. 10 <0. 10 1 PCB 1221 <0.10 <0.10 <0. 10 <0. 10 <0.10 <0. 10 <0.10 PCB 1232 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 1 PCB 1242 <0.10 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 1 PCB 1248 <0.10 <0.10 <0.10 <0.10 <0. 10 <0. 10 <0. 10 PCB 1254 <0.10 <0. 10 <0. 10 <0.10 <0. 10 <0. 10 <0. 10 1 PCB 1260 <0.10 <0.10 0.29 0.32 0.30 0.19 0.22 1 1

~~PGE00047360 1 1 1 TABLE 1.7 30870-0267 LAWLER, MATOSKY & SKELLY ENGINEERS 1 PCB RESULTS~~

~~1 All values are ing/Kg dry basis~~

Lab I.D. 267-099 267-100 1 ^v 1 Type 57599 57603 PCB 1016 <0.10 <0. 10 1 PCB 1221 <0. 10 <0. 10 PCB 1232 <0.10 <0. 10 1 PCB 1242 <0.10 <0.10 1 PCB 1248 <0.10 <0. 10 PCB 1254 <0.10 <0. 10 1 PCB 1260 0.16 0. 19 1 1 1 1 1 1 1 1 PGE00047361 I I I I I I I I APPENDIX C • ANALYTICAL FRAMEWORK FOR THE EVALUATION OF PCB RELEASE AND ITS APPLICATION TO A HYPOTHETICAL DREDGING PROJECT WITH • SUBAQUATIC DISPOSAL IN LAKE LILLINONAH

I I I I I I I I I PGE00047362 I / / j -v- •/--* LMWLtn, IVIA i UOK.Y 01 ^JktLLY tMUINttMi / i ^ r,.,, /V%tY^/ ^Mcrr n-rt ' o^ ~ I If ' ' / ' ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS ^ _ _ ^ POST OFFI"--: oox 1503 PEARL Riven ME\^ VO*=K 1O9G5

~~; ; -r /-'ir.-/i • ?/.j ^..'^-^"^ . "!;;<-^--",';ftvr+'i,f-ir~}l'fti,:; J ;,^^^~~

~~/ /it *"• ' j rV)^^ *1. ^ I / / / 1 OOO yj " I "'^'^ D^p-i-^ d ^ &~~

~~VS. PCB I evy -.223 I "PCS~~

~~a.na r 0 0 ' f litter^ I i *21£L<22.3J. * I IO> I f (Ji isroU'£)~~

~~I v Mass -for vo/iw<2- 'to be — 79 •? If = 2.X I rr ic?~~

~~10,100 e~~

~~- C?.Z? 3 X C?- 7^f A \(y •~~

~~M,, ..1 I LAWLER, MATUSKY & SKELLY ENGINEERS SMtCT NO £NV»RONMENfAL SCIENCE & ENGINEERING CONSULTANTS~~

~~ONE BLUE I--LL PLAZA ^>G5T O^P'CE BOX ' SOa 1 ,RL «'VEW New YOOK 13~~

~~.w — I I I I I I I I I I I I I I PGE00047364 I LAWLER, MATUSKY&SKELLY ENGINEERS ~, j SMUT NO Or ENVIRONMENTAL SCIENCE * ENGINEERING CONSULTANTS~~

~~I CHKD BV OATC -JNfc BLUE MLL PLA^A I ——-'>nsr OFFCF BOX IDOO .~~

~~SUBJECT~~

~~I^^ / 7 / ^T~^sj\ i / vi^ | - 2-ZSW.744) + Zsg. 5-^000^*4.~~

~~C,b - 0.2. f 153^.0 - i I JT. /, ° SSU>W MdfAJi'c ofn?~~

~~flpe ///ie ^tameief = IZ I Ex^f- V(?lc?flw - iS-f^o Dr^qc £<\4rairt3 /ai^ vua-fer in raf/o~~

~~prosify c^ JnfcUg,~~

~~. '7 95/4 (/ ^ump^J? 5-|ctffU Se ffjes to "tKe tc-ffan as a rvuiuU (S.H-~~

~~5y^> cC -jK*.~~

~~^?M.t«xie "Dre-d^iha ^5/i~~

•

t ' V_jni2_ = I6x/o"/ 5*13,0002i ,-4- , '-s "o , ? l(7jinr ^ I ^^ ^^/^ y /r^fl77T J "* •te* - 4-O. 3 hrs ((approx/ffwfrr o^ 1 a. rli^Kvie. PGE00047365 I LAWLER, MATUSKY & SKELLY ENGINEERS SMCCT NQ ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS

~~CMMD •¥ - DATE __ ,-...- orvjE fSLOC *l_L PLAZA TQST OFFTd BOK -5OS C-FAP»L FWER Nt W YO^K 'C365 J~~

~~- i * ->3.r c 5ol/ciS T'C~~

~~- i-'f^,. I~~

~~a^,.t=?^r . t-' 7 ^CsS I 6' ^ ^ •~~

~~s^e , z'^ C c/ 7 (>u =~~

~~IP 76~~

~~rft ^ ts'l" ^ ,^ |40 I "p S?f^- 3Q5"iia, • — r" 'n J- A0q H^cior 7. 7)Sh-/c.M = UT = / c/- r^ •~~

~~— -o~~

~~-for _uj_- O.Z^^ f y- (.0 u CSF - O.\ ~i~~

~~PGE00047366 I LAWLER, MATUSKY & SKELLY ENGINEERS SHE CT NO ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS~~

CHKO. »V DATE . t irsJE BLUE >—LL F>_AZA I •^racT orrcc BOX 1509 I DgARL PtVE« ^tW YOP»K me I -for = o. 1 -y I r 10 O.S7 I.c9 370 I 33Z 3-7 7.5 I 500 /^. c9 ,20 looo I 10,000 ^70, .0/3 I I I C-) I I I ft 0,3 6 I U I ,= AX O - \O ;o 3 10,000 3o 3 CO. I /o -4-0 3o IS / ooo 4O-7S" 35- 3S 7S^? 7S 6^ 30O 3,100 \, 530, 000. I /so /20 1 SO /5O )^^ 5-5 &,2SD \,(,Cff.OC>P. £ '.fO^ I ^/U ml 7-1 m' "P - ^ /Lt 1 -J ir^^t- Z. ^ I

~~LAWLER. MATUSKY & SKELLY ENGINEERS ,j_ iMttT r*O ENVIRONMENTAL SCIENCE «. ENGINEERING CONSULTANTS I CMXO.BY OATI ~^f Hi _.t -i* - Pi A7A -•US'" (JPFICE 0~~

~~V V-J % D c?.^".* D~~

~~iss ej- ;/i~~

~~dr^ci^ //no -titfig P^^ f\rs / s foP ty'S t*S W -/ // I solids 10 a (re>%-tr\aJre>J s; = ^ vp I~~

~~cf •su^ "fc o S 1 '~~

~~.s is -"" £X 2~~

~~IO -~~

~~TTp =~~

~~if>k zumua.vela.hv . J -, M*+>~~

~~PGE00047368 I LAWLER. MATUSKY & SKELLY ENGINEERS SHCCT NO ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS I CMKO. BY DATE , nrsiF B lE HILL PLAZA ^.Jb* OFFICE BOX 1SO9 =>eARL t^VCH NE-A/ ^ORK 1O3G5~~

~~r J— - D 093 _ Td,p /o.-fe ^-^)-^~~

~~f ^ p;r~~

~~• ^P^~~

~~H?>?~~

~~x rf ^~~

~~I "T?~~

•

~~B f~~

~~A (_ \ L- ro/™ r^ // 4i co —»^-c i i* i*\rA l V— ft un*" 'fy\ * **b^ ' @ Ail ' " ' I —C "™l^^ i VAJ' I ^—L/T^- i i <-v i tvx I i ' ' ••»'-' -**- i v^"'t- /" • " 3 "u c/ ^ t; Ctjp~~

~~I PGE00047369 I LAWLER, MATUSKY & SKELLY ENGINEERS tH CCT NO ENVIRONMENTAL SCIENCE & ENGINEERING CONSUL ANTS~~

~~CHKO my - OATS rjF e^g _^ L PLA^ ' 3ST OFf^ C- E3OX Ifa ] I CWFB Nf^ A vn*=K~~

~~SUBJECT J~~

~~b. is COvp<,-.£ I~~

~~K \o "a~~

~~C' I~~

~~I I PGE00047370 I LAWLER. MATUSKY & SKELLY ENGINEERS SMCCT NO '4 ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS~~

CHKD BY DATE. ON.E BL-JC -ilLl. t^L I POS- O^P'CE BOX I ,PL OiVE» NEW VQi- I I , rr, t.rt/1 I I I Par-fi -or 3 T£ O. xj I "ft. 'A^ J>(?ctr/« I,/25 MS 2 I 6 2,57 (?. 2/ I -tcfa\ lr PCJB r\ -^% I I DeeJ f U of PCS f-^ "tng I -fluid /Mud n

~~/O ^^ and radius of ri^ r- I Fr <-UAC. 3G) I For £W c I O r>\ I z n - -rr r -- Tr~~

~~_ v rvr ^ I H ~ "A" ~ PGE00047371 I~~

~~LAWLER. MAT;. 0,v i SKELLV ^NGINEERS SM KCT NO C\VIH(>NMEfJTAL SCI E N( t i ENOIMEESI'J , ' uNSb^TANTS~~

~~J N0 I CMKO Of OATC r h hi i *.*V °" I~~

~~1 '.L > " " --'-' • . f •~~

~~r ( -, L I~~

*-?>•

~~/- I~~

~~[fa t'i ^C-ti ^crr.-h/ i.n -f:r-;uOSify i --//C7. I~~

~~fir *ja^ r ( x I . f ^ v Cl J T = abroUte -b^mricre. " (^io»/»-c "^5 = JO . J > (/~~

~~PGE00047372 I LAWLER, MATUSKY & SKELLY ENGINEERS~~

~~ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS I niM^ Rl UP Ml l £>UAZA I -*ut i uf-^tfc BOX sua~~

~~I "(L - ~^~ ~~

~~I/ I /Z. I r I T I x 1~~

~~I •~*-s-!^ I 25,700^ -0.24? I >•£»-'<~~

~~I t - 7cy< I l- CkT-yl d i-h o/o.~~

~~I , ' C I \O, looz_ I 0 /-?~~

~~I -/ r I ror r~^Ce v o I f PGE00047373 I LAWLER, MATUSKY & SKELLY ENGINEERS ' ' ENVIRONMENTAL SCIENCE & ENGINEERING CONSULTANTS CHKO »Y DATP C»NE BLUE MC^- PLA^-A J~~

~~--".-} "••/' T"~~

~~I I r-^ p T- T\ ^. <-\ Sl :<_: r ? t a S\_ . S-^^-cC'' i-'h'r~~

~~16 Ui ^'jj iif) /i>iL_ u cy I~~

~~of IA'H£ J i-7Sg 12.57~~

~~0. SI3 J?-5/3 ' \.o~~

~~(9/W,) i 0.019 O.io ^.97o -7. / lj |0 a ^~~

~~70 0.202. O.D/6^ 3J~~

* o.

•

~~PGE00047374 I I I I I I1,0 •• I I~~

~~I I . "D r(?~~

~~\ : I : \ I \ I I I . rT" < 'ii v/ J ,00\ I /o PGE00047375 ;~~

~~LAKE L~~

~~7 • 1 i i U • i 20 (£. gO (00 • PGE00047376 " " ••-"- trt- I PART III: PREDICTION OF DREDGED MATERIAL DISPERSION:~~

~~I OPEN-WATER PIPELINE DISPOSAL OPERATIONS~~

I 67. During the maintenance dredging of channels located in rivers and estuaries, fine-grained dredged material is typically disposed within I designated open-water or side-channel disposal areas located 300 to 1,000 m from the channel in water depths of 1 to 6 m. On most large I maintenance operations, a cutterhead dredge may be used to excavate the sediment, which is subsequently pumped as a slurry through a pontoon- I supported pipeline at velocities of A to 6 m/sec to a disposal area I adjacent to the channel (Figure 14). Due to the variability in I I I I I Figure 14. Typical channel rraintcnance dredging operation I with open-water pipeline disposal depth of cut, rate of bwing, and stepping technique used on a particular I operation, the dredged material slurry will usually have a highly variable solids content ranging from 0 to 40 percent solids by weight; I 15 percent solids by weight is a typical average value. Dissolved oxygen levels in the fine-grained slurry are essentially ?ero. The end I of the pipeline may be either above water or submerged at an angle of 0 I to 90 deg relative to the water surface and may be equipped with a 52

I PGE00047377 I deflector plate. As the dredge advances down the channel the discharge point is usually noved periodically to other disposal areas adjacent to _ the channel. The dredging operation is normally continuous, but rrav be interrupted by riechanical breakdown, ship traffic, or bad weather. ^ Modes of Dredged Material Dispersal •

68. The discharged dredged material slurry is generally dispersed I in three modes. Any coarse material, such as gravel, clay balls, or coarse sand, will immediately settle to the bottom of the disposal area • and usually accumulate directly beneath the discharge point. The vast * -uajority of the fine-grained material in the slurry also descends rapidly to the bottom where it forms a low gradient circular or ellipti- I cal fluid mud mound. A small percentage (1 to 3 percent) of the discharged material is stripped away from the outside of the slurry jet I as it hits the water surface and descends through the water column and remains suspended in the water column as a turbidity plune. These • latter two forms of dredged material dispersal will be evaluated in more detail in the following discussion. •

Turbidity Plumes I .13,53 Plume characteristics 69. The levels of suspended solids in the water column above the I fluid mud layer generally range fron a few tens of milligrams per litre to a few hundred milligrams per litre. Concentrations rapidly decrease B with increasing distance downstream from the discharge point (Figure 15) and laterally away from the plume center line due to settling and hori • zontal dispersion of the suspended solids. Since solids concentrations ™ in the plune often increase with increasing depth, the plume boundaries • may be more distinct in the near-bottom portions of the water column. W

Under tidal conditions, the plume will extend inland during the incoming M (flood) tide and seaward during the outgoing (ebb) tide. I he plume | length will usually be only slightly longer than the maximum c'ittance of I 53 I PGE00047378 I

~~I DOE3CC SIZE s-iPt'orn o o 71 £0 30- '00 i e o C APALACHICOLA SAY 4t 18 2S-50 3 3 ft- — -— -~~

~~200 40O 60O 800 1000 '^00 "00 630 '8OC 2COD I OISTAMCE FRO»J OISCMARGC, rn~~

Figure 15. Relationship between suspended solids concentration along I the plume center line and distance do'.mcurrent frrm several open-water pipeline disposal operations r.easured at the I indicated water depths I I PGE00047379 I one tidal excursion (i.e., the distance that the suspended sediment is transported during an ebb or flood tide). In other words, as the tide I changes direction a new pluae will form downcurrent from the discharge point and will be superimposed on the older plume formed during the last • tidal cycle; the new plume will continue to grow until the tide again changes direction. In rivers where the flow is unidirectional, the ft plume length is controlled by the strength of the current and the set tling properties of the suspended material. In both estuarine and riverine environments the natural levels of turbulence and the fluctua- I tions in the rate of slurry discharge will usually cause the idealized _ teardrop-shaped plume to be distorted by gyres or eddylike patterns f (Figure 16). Factors controlling plurre characteristics • 70. The large degree of variability characteristic of most tur bidity plunes can be traced to several major factors: the discharge • rate, character of the dredged material slurry, water depth, hydro dynamic regime, and discharge configuration. The first four factors M will be discussed below and used as input for a relatively simple plume • model; the discharge configuration and its affect on slurry dispersal will be discussed in Part IV. I 71. Particle settling rates. The nature and persistence of tur bidity plumes are controlled largely by the settling rates of the • material suspended in the water column. Low concentrations of silt and clay (with diameters of less than 0.03 mm) settle very slowly causing • large, persistent turbidity plumes. Under certain conditions clay particles may collide to form aggregates or floes with diameters of 0.1 • to 2 mm. If the suspended particles are coarse-grained or conposed of ™ -arge floes, they will settle relatively rapidly; the suspended solids _ concentrations in the resulting plume will be relatively lew and de • crease very rapidly with distance from the discharge point. 72. Turbidity plumes will be relatively persistent ^n fresh water, • because fine-grained particles at low concentrations do not readily form 54 ' • rlocs. However, the degree of flocculation increases very rapidly as • salt concentrations increase from 0 to 10 g/JL and regains essentially

~~55 I I PGE00047380 I~~

~~I1 I~~

~~f I~~

PGE00047381 I constant between concentration of 10 g/£. and seawater concentrations of 35 g/i. The clay mineralogy of the dredged sediment may exert a I subtle influence on settling behavior but its importance is relatively small due to the presence of naturally occurring organic n.iterial which • apparently coats the clay particles. In fact, in salt water the set tling rate of the suspended material generally increases as the organic 54 I content increases. Regardless of the sediment/water composition, settling rates generally increase with increasing solids concentrations B up to approximately 10 to 20 g/2; ' ' at higher solids cor.centra tions the particle settling rates are "hindered" or reduced due to contact with adjacent sediment particles or floes. I 73. Effect of various factors on plume characteristics. The particle settling rates, slurry discharge rate, water depth, current | velocities, and the diffusion velocity (describing horizontal dispersion) all interact to control the characteristics of the turbidity plume • 53 during the disposal operation. As the current velocity increases, the plume (as defined by a specified level of suspended solids in excess I of background) will grow longer. With increasing depth of water in the disposal area, the average concentration of suspended solids in the plume will tend to decrease. As the dredge size increases or particle I settling rates decrease, the plume size and suspended solids concentre- _ tions will tend to increase. In addition, as the diffusion velocity | increases for a given current velocity, the plume becorces longer and wider, while the solids concentrations in the plume decrease. (However, • it there is no resuspension of bottom sediment, the total amount of solids in the plume will remain the same.) Finally, with a decrease in • diffusion velocity or particle settling velocity, or an inciease in water depth, the length of time required for the plume to dissipate • after the disposal operation has ceased will increase. ™ Turbidity plume model 74. A simple method for predicting plume characteristic^ h.is been I developed based on a theoretical hydraulic model. This model has subse quently been verified and empirically refined using field data collected • around three typical open-water pipeline disposal operaticns in estuiirine I I PGE00047382 environments. Only six input parameters are necessary: the size of the I -iredge, water depth and average current velocity in the disposal area, -_ean diameter or settling velocity of the sediment being dredged, an I estimate of the diffusion velocity, and the "age" of the plune. The todel will provide an approximate "worst case" prediction of the shape I z.nd dimensions of the plume, the corresponding r-^erage excess roncen -ration of suspended solids (above background) along the plume center I line as a function of distance from the discharge point, and the per sistence of the plume after the disposal operation has ceased. Factors cuch as discharge configuration, waves, and wind, although important, I ^re. not considered in the model due to their complex and quantitatively unpredictable effect on the plume characteristics. I 75. Procedure. The following stepwise procedure indicates how the six input parameters are used to calculate several nonriinensional I variables that are plotted on the accompanying nomographs. The non- dimensional numbers are then converted to physical units of metres and I milligrams per litre using distance and concentration scaling factors, respectively.

Lakes and rivers 0.3 Medium estuaries 1.0 I (e.g., Corpus Christ! Bay) Large estuaries l.i (e.g., Chesapeake Bay) I Determine the mean particle settling velocity (v ) A size analysis of a typical sediment sarr.ple from the I dredging site should be performed using disposal site water; no dispersing agent should be used. From the <-ize data determine the mean diameter of the sediment. The I mean particle settling velocity v (cm/sec) can thet. be approximated using Figure 17. I I I I I I I I Figure 17. Settling velocity versus mean particle diameter I Determine the age of the plume (t): The age of the plaine is the tiitie (t) that is required for the plume to r^^ch its maximum length. In an estuary, where t'le currtnt I reverses direction, this will occur at the end of ,in e^b or flood tide. In those estuaries where there are two ha/n and two low tides/day the age of the plutr.e is: I

~~59 I I PGE00047384 -, -i r, T >JL —• ~- 10 ^e\.~~

~~'."hero there is one hi.-i: ;:r.c . Lc\; tide j;er i! .": r(sec.) = 12 IT ~ -i I')" :--c~~

>n rivers the flow Is unidirectional: thererore, ?iu e i_:c is 'Jcrined as t'-;e length or t L~e that is reauircc icr a particle v.ith .1 settling volocitv (v ) to setti< a specified vertical distance rel; tivo' to the ce;i'_'r. i f water (C). The surrace plu:^e '.-'ill probablv not be visible after the suspended particles uavt- settled approximately 10 cm below the surr.ice; iriere on , i JT the surr.'.cc r.are:

~~For the near-bottom pluire: D t = —~~

T)c-t errr.ine the average current ''elocitv t or estuaries the overa^o current velocitv over the tirae interval t can be measured in tne : ielc (over several tidal c'.'clesj or c-st in.Ued using a value of raxirrram current velocif/ (u ) obtained from tide tr.bles published bv the N'aticnal rax Oceanic and Atmospheric Administration. 1 he ..verr.ue current velocity Is Ihen estimated using the lollowinc; f omul a:

~~u(cni/sec) =~~

~~For rivers u is sir.iply the averai;e current velocity over the length ot the disposal operation.~~

76. Knowing these six parameters, nondinensicna 1 ritios and scal ing factors can be developed and then used to calculate vorst case estimates of vertically averaged suspended solids concent ra 11 cms (,ihovc background) alon^ the plurr.e center line a= a function of di.-,lance L ro-ii the discharge point. a. Deterinine the ^'.ll'ae of -• — u b. Determine the valae of -, where: v t •<--T and round otf to 0.1, 1, j.2, a \0.

~~f)0~~

~~PGE00047385 I~~

_cerr:in_ t..u I I Lnter i L^ure 13 it rhe c i l<_u latea vilue the appropriate curve, ,.na :hen ver t I ar.u (vertical) scale tii obtain '.he val..e or I Loneen trat -en ("":/ ) it c intent i I ".\:~. cwin^ c.iis vdlu<_ an cat- L oncen t r.i t : ^ n Lit c^n thon be caliul I lo dettT-ine the discanre "• (en) jcwnstream .-> charge point '.-/here tne 'liarae i L ntfr iip.e coru'-.. './"ill bo a specified Y (mg/.), alu v-1 ^acl.^r.nip.i. I (1) Calculate: I Concentration at distance ut (Step <^)

~~(2) Using this ratio enter Figure 19, 20, .' '. . depending on value or - , aloaj, the lei: I~~

(3) Move horizontally to the w'u curve clost. value calculateu in Stop a_ and cown to ; •i - louer I . ... . I'Lo^iJUUlA scale. Obtain a v.ilue ror —-,, OS r where tae average concentration wLLl he I Distance " "his value of multiplied bv t:; equal to Distance X.

~~To determine the concentration Y (:ni;/O aijuve 'i at ,i specified distance X (cm) downstream fro'i I fh.ir ge point , Hist, in i e X (1) Calculate: I I I I PGE00047386 !00~~

~~0 01~~

~~0 OOI u o z o 0 OOOI r-~~

~~OOOOOI~~

~~0 00000! L 001 100 OJ/U~~

~~. . , . , Figure Li> • AL lat lU'isrup hucween ^/u and Solids cone cntracisr it (list nrc t r'sr for =0.1, 1, j.2, .aid ID~~

~~PGE00047387 = Oi,!0,IO~~

~~DISTANCE X~~

~~Relt.tionshin i _ • u• o betweeu n Distanc- e X and~~

~~^olivJs conri'ntr at icn at distance X •ioJ ids concentration at distance ut tor r;ual to '1.1, 1 and 10 and -, =0.1~~

~~PGE00047388 I I I 10,000~~

~~7 = I 1000 I uj/u = 01, 10. 10 I 100~~

~~UJ I (J~~

~~I o K I £ I~~

~~I o oooi 001 <00~~

~~I DBF~~

~~.,,, , . , . , Distance X I rigur c 10. rRelationshi n p between —— and Solids concentration at distance .< tor ^/u equal "o 0.1, ! I Solids conccntrat ior. at distance UL I and 10 md y - 1~~

~~I PGE00047389 r = 3 2 J/u = 0 I , , 0~~

~~0 01 0 I 3ISTANCF X DSP~~

~~, ") i > t inci. lationsnip -—~~

~~cinr ilLLjJ_^ <^lrjjMj:^K_jj^ L ltu olids coriLntrjLion 7t~oT^r~i7i, , — c '^ /u . R~~

~~PGE00047390 I I~~

~~I 10,00 0 F~~

~~I y = 10 1000 L I = 0 I , I 0, 1C I 10 0 X I bJ I I cr o I cr I o i I 00 1 I I 0 00 1 •=~~

~~0 0001 I 001 0 I 00 DISTANCE X I DSF „ _ Distance X Figure 22. Relationship between .md~~

I Sulidj concer.tration at distance X for _/u equal to O.I Solids concentration at distance ut I and 10 .inc -t = 10 I PGE00047391 I I I (2) I'sirr-j; iais r^tio enter ! i~ure 19. 21. _ 1 . r 2_ , c ^n the "a.jj of . .. .,"!.: *:.<_• ..-r:.: T.ti I ( j ] 'Aovf. vi-rtiC2ll'- upward tc t.ie .; n- rorr i.i.L-3 M , i: t •/ > . then Lor i mentally across to trie V"rtie'.i ax: ;. '-ri.r. Concentration at .:istatue '•'. a value lor •— I uoncentrac-jn ac dista-.cc >.L

~~(4) This ratio multiplied bv the concern rat - "n .it c.t B (Step £) gives the average center ''ae cer.jcnt-.ti n at distance X.~~

To determine the anproxir.ate naxi::iu^i •.•;idt;; ,T '. :t. pi1 I '•..'inen occurs at an ipproximate distance ut ) .- -~-iec V' 3 solids concentration above background, ;'_ti ^1" center line length of the plu~e las cerinec i • t :.it cantration) by ttie a >prmriate factor liat^o % • >•.. : respective _/u values.

~~,./'i 'Viet:. Iact.;r I~~

~~0.1 C.25 0.3 0.6 I 1.0 1.3~~

idealized plume shades generated by the mocel ^re -, i.'vn • in Figure 23 for various hydroJynamic conditions. .\ r values of u/u less than 0.1 the plume width will ie k^i, _ than 0.25 times the length. As -.j/u increases bevor.d 1.0 • the plu^ie will approach a circular "patch" ~.hap? ci-ntcrt-c! on the discharge point. Plume shapes are similar tor different values of -, . I Txample. To illustrate the use of this Tiodol the toLlowinj, 3 given. I a_. Operation : (1) 61-cm (24-in.) pipeline w_th a radiu.s = 50 cm

~~(2) Velocitv = 5-9 cm/sec (13 f t''.-.o. 1 •' I~~

(3) Solids content - 15 percent ;olids 1-y U WO-^ilt" • -(30)" • 5«'.9 v 0.15 = 2827 cm" • V*9 > 0.15 = 2J2,SO} '/-.LC 232,303 < 0.05 = 11,640 ,;/sec = 11.t> - K)'' mi;/-,IT I ipproximate values may he used if better totl^iati'S are

I PGE00047392 I I I I I I I I I I I Figure 23. Idealized plume shapes generated, by the model I _b. D: A m = 400 cm (Fora water dcnth of 8 m) c. Medium size estuary: 10 = 1 cm/^ec I ^1. Mean grain diameter: 4 * 10 cm; v - 0.001 en/sec _e. Tidal half-cycle = 6 hr x 3600 -p^ =2.1 6 < 10 sec

~~I ~2 x 104 bee I I I I I 68 PGE00047393 I I "a:-:i:-.nm t i:rrjnt • e j. n cr i s ec I 10 J 'if 2 I |,a A~~

~~LSr = UG i-in/ 10 ' sec) = 32 10 ' i , r ^.2 r.j J~~

~~i_sr = " (1 cm/sec ) "~ (-CO (.."')(? • 10 sec; I "ollc'.-: (Jottea line in figure 2^a lor . / u ' > ( ' ' Co'x en Lr it irr. ' —'_• / • ) ,'t j_i_r T.a , u . «; : ..~~

~~Concentration at 3.2 kT = 0.05 "• CkST = i0.05 :i' . »r i i).023 rv/ci = 23 r,-, |~~

~~To determine the distance !•'. C^m) downs t rt an .- i:u Jis charge point where trie plu-e crater li:u .-on.^.-.i: it i. • i 1 1 • !' .-> ^33 .1 ncmg» / .., , •~~

~~(1) ^y-^yf = ^.i' - ^~~

(2) Enter Figure 2^b (foi , =0.1) alcr.ti jeit-h.inc scale t at 2.2 ("C") t ' (3) Move horizontally to (.he ^./u = 0.1 cui - the lower scale ("D") I Distance X _ „ -^ I (A) Distance X = 0.9 • 3.2 = 2.9 km

~~This can be repeated for ,inv spec it led t or.ceiH r it ion.~~

~~_ The above steps in 1_ can tie levc-rsed to o'lta.i: i i _• ,oli.U i on- I centration at a specilied Jist.ince X, s I~~

~~t (?) Pnter r-'iuure J ,b (for . 0.1) en lover c..lc :t "I" I I I PGE00047394 I I I~~

~~I 0 o I 1 II I I I I I~~

~~I X 33NV1.S Q 17 T/0-" •NOi I I I I I I I I I I PGE00047395 I I~~

3) |jvc v.-rtiL ill" -•> to I L'.-I t-^'irm tale t 71 I .3 .-/• 1 -.) Concentration at 1 km = 7 ' 2j = '61 ..; I The concentration of excess •> jsoendeu < ol id , •..as 23 ny/ at a distance on 3.2 km rrun ilii' point. .'ith .*/i: less tha.i 0.1 the p.axinjm '.< i d L K I plume as defined by the 23 n?/ contcur '.-il! • •_• 0.25 ' 1. 2 km = 0.8 kn. I -he shace of an idealized < _arr.e under ctcsu "vi: ••ill aporjximate the biiace ^no\vm in r'i'urc _ • '.

'.he average center line Concentration for thio i' L er;ti n * I - .r. ci distance front the disnar^e ijint 13; Distance , icn Co n r or, t r r. t..on . .j I 0.5 368 1.0 Ini 2.0 71 I 3 . 0 37 1.2 2 3 I .ifter the disposal operation has ceased, the susneniied r.at l.i .. ie plume will settle (v ) ,.ncl diffase laterally (* ). 1 ie I _ '..jar-surface plume will usually disperse within i period 01 I r 'Equation 3); ' however, the subsurface plume may tacoiet I tor a few days Jeocnaing on the water depth, settKu. ^tf/ of the suspended particles, and the diffusion velocity. .'v I .oj .or estimating the rate of decrease in the plume t • uicrnt ra 11 ons r. ) ..ettling and/or diffusion is given by Schubel .>t al. •

Fluid Mud PisDerjion I .liereas a small percentage of the f me-'rainei Jroci.u'i' '.LI. ^ } • • '-i-TV discharped during open-water pipeline disposi! 'pt>r.H- :i., > or . c I \\ the water column as a tarbiditv ;>lurr.e, t ie vat t r . i - i >:>.j./ descends to tlie bottom or the disposal iroa .vnoie i t I

71 I PGE00047396 1 I I accumulates under the discharge point in the form of a low gradient fluid mud mound overlying the existing bottom sediment. If the dis charge is moved as the dredge advances, a series of mounds will develop. I The majority of the mounded material is usually high-density (nonflow ing) fluid mud that is covered by a surface layer of low-density (flow I ing or nonflowing) fluid mud. The short- and long-term dispersion characteristics of the discharged slurry depend on many factors, in 1 cluding the nature and rate of slurry discharge, the discharge configura tion, and the hydrodynamic regime and bottom topography in the disposal I area. These factors will be evaluated in more detail in the following discussion of fluid mud accumulation and dispersion. I Dispersion of low-density fluid mud 81. At a typical open-water pipeline disposal operation, an esti mated 97 to 99 percent of the fine-grained dredged material slurry I descends rapidly through the water column and impacts on the bottom. During the descent and impact phases, the slurry usually entrains water I and may initially flow radially away from the discharge point over the bottom or surface of the existing mound as a fragmented flow oflow- I density fluid mud. The fluid mud front propagates in the form of a I near-bottom headwave ' (Figure 25). Under quiescent conditions more

~~I UUO FLOW ' ^_ ^ I r~~

' COHCfH T>* TIOH I -VELOCITY I INCREASED VELOCITY AND SEDIMENT CONCENTRATION — I Figur e 25. Velocity/sediment concentration distribution within the headwave of a low-density fluid mudflow. (Redrawn from Reference 56 and used through courtesy of the Pacific Section SEPM, P. 0. Box 70344, Ambassador Station, Los Angeles, Calif. 90070.) 72

~~PGE00047397 I I~~

•nan 98 percent of the sediment in the mudflow remains in tne rluic nua 'jyer at concentrations greater than 10 g/i, while the remaining 2 -2i:cent is resuspended in a turbid layer behind the heaawave by turbj •nce and upward mixing of the sediment suspended at the upper surface ^: tae mud layer. High sediment concentrations and flow velocities :-aracterize the lower levels within the low density fluid mud, whereas relatively low solids concentrations and velocities are present in the I overlying turbid layer. ' Areal and temporal fluctuations in this • rlow of low-density fluid mud are probably caused in part by pulses of • ' 'ligher density dredged material slurry discharged from the pipe. 82. Material characteristics. The size distribution ot the cis .iargea material will strongly influence its flowing/mounding behavior, I ery coarse sand, gravel, and clayballs from "new work" dredging proj- ^ cts will rapidly settle out at the discharge point independently of anv B j.ne-gramed material. However, in slurries of fine-grained material Containing less than 30 percent medium to fine sand, the sand apparentlv B ces not settle out, but tends to flow with the slurry. As the sand Content increases above 30 percent, the fluid mud will tend to flow less B 12 V ind mounding will increase. The relative percentages of bilt, clay, ind organics may have a marginal effect on the behavior of the tluid "iud; however, no such relationship has been documented.' ™ 83. Neither laboratory nor field data indicate any obvious differ ences between the flow characteristics of freshwater and estuarine fluid I 13 57 nud. ' Under similar flume test conditions freshwater and saltwater sediments appeared to generate mudflows of similar thickness and flow • velocity; however, the overlying turbid layer generated above saltwater fluid mudflows maybe somewhat thinner due to flocculation of fine- • Drained material. 84. Bottom slope.57 The slope of the bottom probably has the greatest influence on the flow characteristics of low-density fluimud.mud. ' Mudflows propagating uphill decelerate very rapidly due to ir.creat.ed settling of the suspended sediment. However, if the bottom has a sufficient downslope, the velocity of the flowing mud will increase until it reaches a constant terminal velocity. For slopes less Chan •

~~I PGE00047398 I I~~

I 2 deg (1:30), the terminal velocity generally increases as both the solids concentration of the fluid mud and the slope increase. Under I quiescent flume conditions, the minimum critical downslope angle at which a channelized (i.e., nonexpanding) headwave will maintain a I constant velocity is approximately 0.75 deg (1:76). Under field conditions, this critical angle may be less than 0.75 deg depending on the pipeline configuration at the discharge, the slurry discharge rate, and I the hydrodynamic regime in the disposal area. In other words, if fine- grained dredged material slurry is discharged in open water where the I bottom slopes are greater then 0.75 deg, the fluid mud material will flow downslope at velocities of approximately 0.1 to 0.3 m/sec as long I as that slope is maintained. At slopes of less than 2 deg the thickness of the fluid med layer remains approximately the same regardless of the I slope; however, the thickness of the turbid layer overlying the fluid mud layer tends to increase as the downslope angle increases. 85. Discharge rate. The flow characteristics of fluid mud also I depend in part on the discharge rate of the dredged material slurry. High discharge velocities produce maximum levels of dispersion, both I areally and throughout the water column. As the size of the dredge and discharge rates increase, the thickness of both the fluid mud and turbid I layers increase; however, the thickness does not appear to be significantly dependent on the solids concentration of the discharged slurry. I 86. Currents. Under laboratory flume conditions the flow characteristics of low-density fluid mud are not significantly affected by currents up to velocities as high as 3 cm/sec; however, the thickness of I the overlying turbid layer will tend to increase when current velocities exceed 1.8 cm/sec. Under field conditions dominated by low current I velocities, this turbid layer is relatively thin with concentrations increasing from 1 to 10 g/i within a vertical interval of 5 to 10 cm; at I current velocities of 50 cm/sec the same transition zone may extend over 13 an interval of approximately 50 cm. This indicates that resuspension I and subsequent downcurrent movement of some material at the surface of I the low-density fluid mud layer may occur during periods dominated by I 74 PGE00047399 I I

'ilgh currents. However, significant resuspension and upward -.ixin^ OL I Jluid mud throughout the upper water column apparently dees not occur. 57 ^- ^7aves. Waves generated by weak to moderate winas interfere I vary little with the overall motion or velocity of low-density t'luic ?.udflows. In fact, the near-bottom orbital notions induced by waves B are damped significantly by the presence of a fluid mud layer. Hcwever, .:hen orbital velocities exceed 1.8 cm/sec, the thickness of both the B fluid mud and turbid layers will tend to increase. The amount of ^ resuspension and the height to which fine-grained sediment is suspended iapends primarily on the depth of water, wave size, and length of time I j | chat the fluid mud is exposed to wave activity. As these factors in '% irease in magnitude, so does the degree of sediment resuspension. In B ".ost cases, typical tidal current velocities (e.g., 5 to 10 en/sec) ara much greater than wave-induced velocities except during periods of B ~.igh winds and/or wave activity when dredging activities usually cease, "luid mud mound characteristics 38. If the bottom slopes are not steep enough to maintain low- I density fluid mudflows, the sediment suspended in the fluid mud laver '.'ill tend to settle and the flow velocity of the headwave will I Decrease. When suspended sediment concentrations exceed 200 g/*., -he fluid mud is no longer capable of flowing freely, but instead will accumulate under the discharge point in the form of a low gradient (e.g., 1:500) circular or elliptical fluid mud mound. I 39. Solids concentrations. At the water column/fluid mud interface, the solids concentrations increase very rapidly from approxi- B mate levels of a few hundred milligrams per litre to 200 g/i. Below the 200 g/i concentration level the solids concentration within the high- density fluid mud increases at a slower rate with increasing depth. I Tluid mud layers may be stratified in sublayers of 20 to 30 cm thick with each layer becoming increasingly more dense with depth in the fluid mud | Mound. Concentrations at the base of the mound may be as high as 500 ;;/1 depending on the thickness of the mound and its state of B Consolidation.

75 I PGE00047400 II— 11 I I and predominant direction of the current and the configuration of the •iischarge. For a vertical discharge in an environment without I significant currents, the fluid mud mound will have a conical shape "ith the apex centered on the discharge point (Figure 26a). Vhere current velocities are greater than a few centimetres per second the I nound will be skewed in the direction of the predominant current. Mound slopes on the downcurrent side will also be less than those I facing the predominant current direction (Figure 26b) . Under low cur rent conditions a similarly skewed mound will result if the discharge I is oriented at a low angle relative to the water surface (Figure 26c). The degree of mound elongation is controlled by the discharge ccnfieura I ;ion and the strength of the currents, both of which affect the dispersion of the fluid mud (Figure 26d). 93. Mound consolidation. When the solids concentration of the I :luid mud exceeds 200 g/£, the material will begin to undergo self- '/aight consolidation. During this process, the bulk density of the I iedircent increases, the height and slopes of the mound decrease, and the rate of consolidation decreases. In high energy environments the thick I ness of the mound may also be reduced significantly by resuspension and erosion of low-density fluid mud at the surface of the mound by waves I ^nd currents. 94. The time that is required for the mound to reach its ultimate state of consolidation depends primarily on the characteristics of the I faterial and the thickness of the deposit. As the thickness of the ;.-.ound increases, the amount of time that will be required for the a material to reach its final state of consolidation will increase; doubling the thickness of the layer will decrease the rate of consolida i tion by as much as four times. However, the thicker the layer, the higher will be the final bulk density of the material after it has i undergone consolidation. Depending on the sedimentation/consolidation characteristics of the dredged sediment, complete consolidation of a i -luid mud mound may continue from one to several years. ' ' i i 77 PGE00047401 I I

90. Flow characteristics. As the disposal operation contini.es, • the thickness and radius of the mound will increase due to the addition of dredged material slurry and the settling of the dredged material • suspended in the water column; however, the rate of expansion of the mound decreases exponentially with time. Flow velocities at the water • column/fluid mud interface generally range from zero to a few centi- H metres per second indicating that recently discharged slurry may flow ^ away from the discharge point along the surface of the existing ^iound as £ i a fragmented sheet of low-density fluid mud. Velocities decrease rapid- ! ly with depth as the bulk density of the fluid mud increases. This • pattern indicates that high-density fluid mud within the mound prooablv moves away from the discharge point by a very slow creeping process or • occasional sudden failure (that was not detected by the field measure ments) when the slope of the mound exceeds a critical angle and/or the fluid mud loses its strength due to rapid deposition of sediment and I consequent generation of excess pore pressures. r 91. Mound slopes. The slopes on the fluid mud mound are con trolled primarily by the dispersive characteristics of the discharged dredged material slurry. During the disposal operation typical mouna • 13 55 slopes may average about 1:500; ' however, the surface of the mound close to the discharge point may be pocked with conical hills and scour pits with maximum slopes of 1:50 and a relief of approximately 0.5 m. If the slurry is widely dispersed, mound slopes will probably range from 1:500 to 1:2,000. With a low degree of dispersion, the fluid mud nound i 13 will have slopes ranging from 1:100 to 1:500. Unfortunately, the dynamic nature of open-water environments, the high degree of varia i bility associated with the disposal operation itself, and the lack of extensive fluid mud field data make it very difficult to accurately predict the mounding characteristics of the discharged dredged material. f However, the amount of slurry dispersion can be controlled by using various pipeline configurations at the discharge point; this is dis- r cussed in Part IV. 92. Mound shapes. The areal and cross-sectional shape of a fluid mud mound on a nonsloping bottom depends primarily on the strength I 76 i PGE00047402 I I I I 1 I I

~~I NO CURRENT C. I I I miM&i$y$, I 1~~

~~PREDOMINANT I CURRENT I VERTICAL DISCHARGE HORIZONTAL DISCHARGE I Figure 26. Effect of discharge angle and predominant current direction I on Che shape of a fluid mud mound 1 I 78 PGE00047403 I I~~

:\-\RT IV: METHODS FOR CONTROLLING DREDGED MATERIAL DISPERSION: I OPEN-WATER PIPELINE DISPOSAL OPERATIONS I :J5. Probably the most promising method for controlling the disper •ion of dredged material slurry at open-water pipeline disposal opera I .ions involves modifying the pipeline configuration at the discharge ?oiat. In addition, under certain circumstances silt curtains may be • •sed to control the dispersion of water-column turbidity by modifying ^ ihe current flow patterns in the vicinity of certain types of dredging irid disposal operations. Flocculants may be injected into the pipeline I .0 increase the settling rate of the fine-grained material, but this -2chnique is not recommended. I

~~13,53,55,57,61 Pipeline Discharge Configurations I~~

Dredged material dispersal II 96. Of all the environmental and operational factors affecting * :he dispersion of dredged material slurry during open-water pipeline . Lsposal operations, the configuration of the pipeline at the discharge I -Oint appears to be the only parameter that, from a practical point of lew, can be varied to effectively control the characteristics of | Dispersion. The pattern of dredged material dispersal is apparently -routrolled by the configuration of the pipeline at the discharge point • ?•=; well as the angle and height of the discharge relative to the water ."surface (for above water discharge) or bottom (for submerged discharge). • 97. Generally speaking, pipeline configurations that minimize ^ '"ater-rolumn turbidity tend to produce fluid mud mounds with steep side « •slopes, maximum thickness, and minimal areal coverage. Conversely, those • "onf igurations that generate maximum levels of water-column turbidity .-roduce relatively thin fluid mud mounds of maximum areal extent. As • -!-own i.n Table 3, as the height of the mound decreases by a factor nf '-'•••'

~~79 I~~

~~PGE00047404 I I I Table 3 Fluid Mud Mound Characteristics (Mound Volume: 250,000 cu m)~~

I Height, m Areal coverage, sq m 4 187,500 2 375,000 I 1 750,000 I 0.5 1,500,000 height and areal extent/resuspension potential should be considered when I evaluating the potential short- and long-term impact of a particular disposal operation. Unfortunately, there is no "best" pipeline configu ration; the design chosen should be based on the desired dispersal of I dredged material in the water column and on the bottom. Typical configurations I 98. The dispersal characteristics of several typical discharge I configurations (Table 4) are described below the table. Table 4 I Effect of Pipeline Configuration on Dredged Material Dispersion WATER COLUMN FLUID MUD MOUND ! TURBIDITY AREA RATE OF | I TYPICAL PIPELINE CONFIGURATIONS' SURFACE MID-DEPTH HEIGHT SLOPE COVERED CONSOLIDATION DIFFUSER-SJBMERGEO LOW LOW HIGH HIGH LOW LO* 90 ELBOW WITH CONICAL EX LOW LOW/MEDIUM PANSION SECTION-SUBMERGED I 90 ELBOW-SUBMERGED LOW MEDIUM 90 ELBOW WITH SPLASHPLATE LOW HIGH 0 WITH SPLASHPLATE-SUBMERGED MEDIUM HIGH I 0 WITH SPLASHPLATE-ABOVE MEDIUM HIGH 20" OPEN END-SUBMERGED MEDIUM HIGH 0 OPEN ENO-SUBMERGED HIGH HIGH r f I 0 OPEN END-ABOVE HIGH HIGH L DW L(3 W HI ;H HIGH PIPELINE ANGLE RELATIVE TO WATER SURFACE.

I With a simple open-ended pipeline discharging slurry at 4 to 6 ra/sec parallel to the water surface, the high momentum levels cause a great deal of entrainment of I disposal site water as the slurry jet descends through 1 the water column and impacts on the bottom. Turbidity I 80 PGE00047405 I I

~~levels are generally high and the fluid mud layer Is relatively thin and widely dispersed. I~~

b_. Submerging the discharge just below the water surface may reduce the degree of slurry dispersion; however, based on I the field data, it is difficult to determine how signifi cant this reduction may be. If the discharge pipe is submerged to a sufficient depth below the surface, a I visible plume may not be apparent.

c_. At low discharge angles a significant reduction in the slurry momentum can be achieved with a deflector or I splashplate mounted at the end of the pipe perpendicular n to the slurry flow. Although this tends to disperse the slurry as it is discharged, the momentum loss is I apparently significant enough to cause the dispersed slurry to settle to the bottom relatively quickly, there by generating less water-column turbidity. I cl. The amount of water-column turbidity generated by a sim ple submerged discharge decreases as the angle of the pipeline increases from 0 to 90 deg. With a simple 1 90-deg elbow on the end of the pipeline, the slurry is discharged vertically towards the bottom with less entrainment of disposal site water. Upon impact on the I bottom, the vertical motion of the slurry is translated into a horizontal flow, which spreads radially frora the impact point. In areas where current velocities are less than 10 cm/sec, this configuration produces near- I surface turbidity plumes that are very diffuse, witn occasional "puddles" of higher solids concentrations at varying distances from the discharge point, I

e_. By adding a splashplate to the simple 90-deg elbow, the amount of slurry dispersion can be increased. With the I end of a 69-cm (27-in.) pipeline discharging at a depth of 1 m against a splashplate positioned at a depth of 2m, the slurry is dispersed at the depth of the splashplate with traces of surface turbidity visible only within I 100 m of the discharge point.55 t_. By adding a 15-deg conical section at the end of the I simple 90-deg elbow, the effective velocity of the dis charged slurry can be reduced by a factor of 2 or 3 (with out affecting the dredge's production rate). This will I tend to decrease the levels of water-column turbidity and increase the mounding tendency of the fluid mud. I Communication, 3 January 1978, William B. Cronin, Research cientist, Cheasapeake Bay Institute, Baltimore, Md.

PGE00047406 I I Submerged diffuser system I 99. The amount of water-column turbidity generated by an open- water pipeline disposal operation can probably be minimized most effec I tively by using a submerged diffuser system (Figure 27) that has been developed through extensive laboratory flume tests. (Unfortunately, the diffuser system has not been field tests.) This system has been I designed to eliminate all interaction between the slurry and upper water column by radially discharging the slurry parallel to and just above the 1 bottom at a low velocity. The entire discharge system is composed of a submerged diffuser and an anchored support barge attached to the end of I the discharge pipeline that positions the diffuser relative to the bottom. As it is presently designed, the diffuser/barge system can be I used in water depths up to 9 m. 100. The primary purpose of the diffuser (Figure 28) is to reduce the velocity and turbulence associated with the discharged slurry. This I is accomplished by routing the flow through a vertically oriented, 15-deg axial diffuser with a cross-sectional area ratio of 4:1 followed by a I combined turning and radial diffuser section that increases the overall area ratio to 16:1. Therefore, the flow velocity of the slurry prior to I discharge is reduced by a factor of 16, yet the dredge's discharge rate (i.e., slurry flow velocity x the pipeline cross-sectional area) is not I affected in any way by the diffuser. The conical and turning/radial diffuser sections are joined to form the diffuser assembly, which is flange mounted to the discharge pipeline. An abrasion-resistant im- t pingement plate is supported from the diffuser assembly by 4 to 6 struts. The parallel conical surfaces of the radial diffuser and im I pingement plate slope downward at an angle of 10 deg from the horizontal so that stones and debris can roll down the sloped surface and auto- 1 matically clear the diffuser. The radial discharge area of the diffuser can be adjusted by changing the length of the struts supporting the I impingement plate. In this manner both the thickness and velocity of the discharged slurry can be controlled. The strut length, which deter I mines not only the slurry discharge velocity, but also the maximum diameter of an object that will pass through the diffuser, should be

~~I 82~~

~~PGE00047407 > 1-4 -r-1 O U Jl <_/~~

~~_- II T— O :— C j; P^ C v; • j o > l-i M -^~~

~~ii-i - -a ra o -a x -rt a o > tc in o M -H a a -o E J .£> -a E 3 C T1~~

~~D CO~~

~~PGE00047408 1 I I COUNTING FLANGE SLURRY FLOW~~

~~CONICAL DIFFUSER 1 SECTION 1 GAS VENT~~

~~Vl. TURNING 4 RADIAL /1\ DIFFUSER SECTION 1 / r\ GAS SHROUD~~

~~1 -SUPPORT STRUT I -RADIAL DISCHARGE • IMPINGMENT PLATE~~

~~i /flT ff/f //// BOTTOM SEDIMENT~~

~~i Figure 28. Submerged diffuser~~

approximately five-sixths of the pipe diameter. Since the gas concent t of bottom sediment is often high (e.g., 5 to 30 percent of the in situ volume), the diffuser is also equipped with a gas collection shroud i around the circumference of the radial diffuser section to trap any sediment-covered gas bubbles before the slurry is discharged. The gas I is vented to the atmosphere through a hose extending from the shroud to the top of the derrick. The diffuser for a 46-cra (18-in.) pipeline is I approximately 2.4 m tall from impingement plate to mounting flange and 2.4 m in diameter at its base. 101. A discharge barge must be used in conjunction with the dif i fuser to provide both support and the capability for lowering the dif fuser to within 1 m of the bottom at the beginning of the disposal i operation and raising it as the fluid mud accumulates under the diffu ser. The barge also provides a platform for the diffuser while it is i being adjusted, serviced, or moved to a new site. i i 84 PGE00047409 I

102. Because the dredged material slurry discharged from ;he I Jiffuser will form a low-gradient fluid mud mound similar to chat torned / any typical open-water disposal operation, a schedule for Lifting the • .iffuser as the mound accumulates has been developed by equating the •olume of slurry discharged from the diffuser to the resulting volume of • ~he mound. I ; Volume of solids in the Volume of solids ini ththe discharged slurry fluid mud mound m

f d2 V T B = 7 R2 H B (11) A s 3 m Mere: I d = inside diameter of the pipeline V = flow velocity of theslurry in thepipeline • T = total length of time of the disposal operation (at the same site) m 3^ = solids ratio (by volume) of the slurry s R = radius of the fluid mud mound H = height of the fluid mud mound I B solids ratio (by volume) of the fluid mud mound ..ssuming that the slope (S) of the fluid mud mound (S = —) is equal to "R I i/200, the average slurry velocity in the pipeline is 5.5 m/sec, and , -he solids contents of the slurry and fluid mud mound are equal to I 15 and 25 percent (by weight), respectively, Equation 11 can be as: •

B 3 H3 / -f-2- = 2008 \ <12> Bs S2d2V d2 I '••'here: • I' is expressed in days P il is expressed in metres ^ d is expressed in centimetres || formula can be used to develop a schedule for raising the ditfuser noving it to a new site to prevent interference with themounded •

85 I PGE00047410 I I material. Table 5 shows a typical schedule for adjusting the heiuht o: I diffuser above the bottom for several different discharne pipeline sizes. The mound configurations at the ends of the time periods are

~~1 . 13^8 5 S-jbcnerneJ Diff'-iser Miypmpnt ;'ch*"1u,R~~

I ,-ecosmer.a.ed Height of Tine Tiffuser Above Bctton J3 ;n -J -n jl ca ol C3 . :n Perl^g A^ regir.nir.g cf ?ir:e Per: A, ~ 1 r? i".. .6 i^ . c3 in. J- in. 23 lr.. 1 1.0 0. 3.2 3.1" 2 1.5 3.3 2.1 1.3

~~3 5.1 13.3 "".3 -.5 1 » i.T 30.3 17.3 10.6~~

..ate: 7c 5e conservative all r.unters fceyona the first ^ecisai place are i^aured. I * ircr pipeline sizes exceeding 51 ", tnc iK'f'iser sr.ou^d be ir.i'-ially positior.e^l 1.; a abcve I the bottom. shown in Figure 27. In most cases the movement of the diffuser to a new site would probably be determined by the advancement of the dredge I rather than excessive mounding of dredged material at the disposal site. 103. Mathematical scaling techniques based on flume test data : I can be used to compare the initial dispersive characteristics of fine- grained dredged material slurry discharged from the diffuser with that } I of a simple submerged pipeline. For a 61-cm (24-in.) dredge discharging j slurry at 5.5 m/sec in 3.7 m of water through a pipe section submerged jj 1.2 m below the water surface at an angle of 20 deg, the flud mud ^ I .5 layer a short distance (30 to 60 m) from the discharge point would be .{ I approximately 1.5 m thick and would move away from the discharge point j at a velocity of about 0.3 m/sec. A 0.7-m-thick turbid layer mi^ht be jj '?t present above the surface of the fluid mud layer. Using a diffuser | I system positioned 1 m above the bottom, the slurry would be discharged | ; at a velocity slightly greater than 0.3 m/sec. A short distance from the J I diffuser the headwave velocity of the flowing mud would probably be less { than 0.1 m/sec. The fluid mud layer would be approximately 1 m thick f or a reduction of about 33 percent over the 20-deg submerged discharge } I 7 I configuration. The thickness of the turbid layer overlying the rluid mud I 1 86 PGE00047411' I I

layer would probably be about 0.6 m thick so that the total thickness ^: • :he fluid mud/turbid water layer would be approximately 1.6m. [n both cases the thickness of the fluid mud/turbid water layer will do- V .rease with increasing distance from the discharge point as the flow .-xpands radially and the bulk density of the fluid mud increases due to Battling of the suspended sediment. I i i 104. Although the diffuser has not been field tested, it has a ^ i ^reat deal of potential for eliminating turbidity in the water colunn M j -.nd maximizing the mounding tendency of the discharged dredged material.

; aw velocity near the bottom, thus eliminating all interaction of the I .lurry with thewater column above thediffuser. This effectively • _.i::iinates water-column turbidity as well as any depression of the S3 62 Dissolved oxygen levels in the water column. ' Unfortunately, using :>.e diffuser does not eeliminate the impact of the fluid mud on the 9 ~c>nthic organisms, ' nor does it eliminate the possible resuspension r low-density material at the surface of the fluid mud mound by waves ;.id ambient currents. I Silt Curtains

105. One method for physically controlling the dispersion of near -iriace turbid water in the vicinity of open-water pipeline disposal M operations, effluent discharges from upland containment areas, and ™ possibly clamshell dredging operations in quiescent environments in- £ volves placing a silt curtain or turbidity barrier either downcurrent || ~rom or around the operation. Silt curtains are not recommended for Derations in theopen ocean, in currents exceeding 50 cm/sec (1 knot), • -n areas frequently exposed to high winds and large breaking waves, or •.round hopper or cutterhead dredges where frequent curtain movement M be necessary. ••I'.eral Description • •06. Silt curtains (Figure 29) are impervious, floating barriers ™ •'!i-it extend vertically from the water surface to a specified water I 87

~~PGE00047412 I I~~

~~01 I 111 £XTRA FLOTATION TO COMPENSATE FOB HEIGHT -HANDHOLD DESIGN OF END CONNECTGFK WATERLINE BUOYANCY FLOAT 1 J_! | 1 0 9~~

~~FLDTAT'°£EGMENT^ M 9 1—1 i *—* 0 I -TENSION CABLE — 10 1 al 0- uj 1 < O SKIRT- t 1 SKIRT o| 1 |0 o: 1 01 10 el >- END CONNECTOR ' 1 CROMMET^ SALLAST 0 0 O 0 O T CHAIN~~

3ALLAST CHAIN 1 VIEW A-A I I Figure 29. Construction of a typical center tension silt curtain section depth. The flexible, nylon-reinforced polyvinyl chloride (PVC) fabric forming the barrier is maintained in a vertical position by flotation 1 segments at the top and a ballast chain along the bottom. A tension cable is often built into the curtain immediately above or below the I flotation segments (top tension) or approximately 0.5 m below the flotation (center tension) to absorb stress imposed by currents and I other hydrodynamic forces. The curtains are usually manufactured in 30-m sections that can be joined together at a particular site to pro I vide a curtain of specified length. Anchored lines hold the curtain in a deployed configuration that is usually U-shaped or circular. Processes affecting silt curtain performance I 107. In many cases the concentration of fine-grained suspended solids inside the silt curtain enclosure may be relatively high (i.e., i in excess of 1 g/J.) or the suspended material may be composed of rela tively large, rapidly settling floes. In the case of a typical pipeline i disposal operation surrounded by a silt curtain (Figure 30), the vast majority (95. to 99 percent) of the fine-grained material descends

~~i 88 i PGE00047413 I I~~

~~^PIPELINE I I I I I~~

~~BOTTOM SEDIMENT I~~

~~~ure 30. Processes affecting the performance of silt curtains in controlling dredged material dispersion I~~

to the bottom where it forms a low gradient fluid mud mound. I nile the curtain provides an enclosure where some of the remaining ~ Lne-'^ rained suspended material may flocculate and/or settle, most of I ^na turbid water and fluid mud flow under the curtain. The silt curtain Joes not indefinitely contain turbid water, but instead diverts its flov I Oder the curtain, thereby minimizing the turbidity in the upper water •olunn outside the silc curtain. 1 108. Whereas properly deployed and maintained silt curtains can Affectively control the flow of turbid water, they are not designed to contain or control fluid mud. In fact, when the accumulation of fluid I n.ud reaches the depth of the ballast chain, the curtain must be moved away from the discharge; otherwise, sediment accumulation on the lower I -"Jce of the skirt will pull the curtain underwater and eventually bury it. Consequently, the rate of fluid mud accumulation relative to I -li.inges in water depth due to tides must be considered during a silt Curtain operation. When bottom slopes exceed 0.75 deg, the fluid I ~.u

~~PGE00047415 I~~

•asoec.c -o currant velocity, water ccrth (relative r:: ti.ial ran^c > , I ••.jctCTn slope and sediment types. t.r«.i "^ssibl" ^ackyrcnirm j^veis or :.rr:: ditv. Vaxinum surface currents --"»r r. iidai i_^cie ;.:.<-•.. 1J r I -. houid ."irst be established from c'.irrent T.easureTndntj or -: .:>•

.:blcs. In addition, information on current direction and •.-.•<-t,tr ~^.r:i\ .jnce nay also indicate potential deployment nroDlens and/or the uest I ,onfiruration(s) to use. • 112. If the hydrodynamic regime appears to be concucivc to silt ™ -•-irtain deployment (i.e., current velocities are less than 50 c-v'sec), M survey of the water depths over the entire site and surrouncir.c | nas is required so that a curtain with a proper skirt deoth can he ji^czad and its initial and future placement geometries determined. • .- .".•Ires where the tidal range (i.e., the difference in depth be eve en 11 * ''.r .; tidal cvcle, the survey data must be adjusted to account for ..anges in water depth that will occur during the operation. F'.ie I -r:^T.um depths (at the lowest low tide) are then used to determine the "-•cessary skirt depth allowing about 0.5 m clearance between the lower B -i ie of the skirt and the existing botton. Unfortunately, this 0.5-m ;p between the skirt and bottom may be difficult to maintain in very • ."allow water. The effect of fluid mud accumulation on water depth ^ ''ell as the proposed schedule for moving the silt curtain to prevent B '-iri.il should also be considered in selecting the curtain skirt depth. * 113. The character of the bottom sediment/vegetation at the pro- » •'osed deployment site should also be established us ins a j;rnb .sampler or • •'''•'i'.i'A device to determine the type of anchors to use and convenient ^ • 'ehor points on the outer limits of the deployment site. The potential •

~~' : - ct 01 boat traffic and boat-generated waves on the. proposed depLov~~

1 n» jimi^uration and mooring system should also he considered. • • "• e ! I'.inchini; and retrieving tlie silt curtain will require (.he use ; .>;rrle truck and a boat(s), a launching r.imp r.nd possibly ,' crnno •

91 I PGE00047416 I I should be located as near the site as possible. If an evaluation of I silt curtain effectiveness relative to preoperation background conditions is desired, background turbidity levels at various depths must be I determined, preferably under a variety of current and wave conditions. 114. Deployment configurations. After the deployment site has I been surveyed, the length and geometry of the proposed curtain should be determined based on the type of silt curtain application, the hydro 1 dynamic regime at the deployment site, any environmental policies regulating allowable turbidity levels as a function of distance from the operation, and such factors as boat traffic. Curtain lengths for typi I cal operations might be 150 to 450 m for the U-shaped or semicircular configurations, 300 to 900 m for the circular/elliptical case. 1 115. In most cases a silt curtain can be deployed in an open-water environment in the form of a maze, semicircle or U-shapc, or a circle I or ellipse: a_. The maze configuration ("A," Figure 31) has been used on rivers where boat traffic is present, but appears to be I relatively ineffective due to direct flow through the gap between the two separate curtains.

b^. On a river where the current does not reverse, a U-shaped I configuration ("B," Figure 31) is acceptable, but the distance between the anchored ends of the curtain should be large enough to prevent leakage of turbid water around I the ends of the U. c_. Where turbidity is being generated by effluent from a I containment area or a pipeline disposal operation close to the shoreline, the curtain can be anchored in a semicircular or U-shaped configuration ("C," Figure 31) with the ends of the curtain anchored onshore approxi- I mately equidistant from the discharge point. The re quired radius of the configuration is determined by the type and volume of material being disposed inside the I curtain as well as the water depth. d_. In a tidal situation with reversing currents, a circular or elliptical configuration ("D," Figure 31) is neces I sary. Unfortunately, this latter case requires an extensive mooring system. I 116. Silt curtain specifications. The silt curtain can now I be selected based on the appropriate deployment geometry and the 1 92

~~i PGE00047417 I I -^ I I (NOT RECOMMENDED)!/ N H> I LEGEND I I~~

~~U-SHAPED I IN STREAM I 1~~

~~CURTAIN MOVEMENT DUE TO REVERSING CURREN I~~

•C" I U-SHAPED ANCHORED ON-SHORE I r r i I ESTUARY I I i CIRCULAR OR ELLIPTICAL i "ic:ure 31. Typical silt curtain deployment configurations i 93 i PGE00047418 I I characteristics of the deployment site. The following recommendations

V resulted from an evaluation of silt curtain performance under various field conditions. A 30-m section of silt curtain with these specifica I tions and a skirt depth of 1.5 m could be purchased in 1977 at an approximate cost of $700 (no tension member) to ?1300 (center tension;. I a_. The silt curtain should have a skirt depth such that the lower edge is about 0.5 m off the bottom at low water; however, the skirt depth should not exceed 3 m unless the current velocities at the site are negligible. The I I fabric should be a nylon-reinforced PVC material (or ; equivalent) with a tensile strength of 52,538 N'/m (300 j lb/in.); a fabric weight of 610 g/sq in (18 oz/sq yd) for I low current conditions, 746 g/sq m (22 oz/sq yd) for high current conditions; a tear strength of 445 N (100 Ib) or 890 N (200 Ib) for 610- or 746-g/sq m I fabric, respectively; and a tensile strength after abrasion of greater than 35,025 N/m (200 lb/in.). The fabric surface should be easily cleaned and resistant to marine growth, ultraviolet light, and mildew. All fabric seams I should be heat sealed.

b_. Segments of solid, closed-cell, plastic foam flotation I material should be sealed into a fabric pocket and pro vide a bouyancy ratio (bouyant force/curtain weight) of greater than 5. Each flotation segment should be less " I than 3 m long so the curtain may be easily folded for storage or transport. Handholds should be located along , the top of the curtain between the flotation segments for d ease in handling. .; I i c_. In low current situations (i.e., less than 5 cm/sec) most •; available connectors for joining 30-m sections probably -j I maintain adequate physical contact along the entire skirt « joint. If current velocities exceed 5 cm/sec, aluminum S extrusion (or equivalent) load transfer connectors are g I recommended (Figure 32). CONNECTOR-x. j

~~I j~~

~~r- l CURTAIN -^~~

~~I 94 I PGE00047419 I I~~

'Tien current velocities are negligible, HP tension mcnber I (other than the tabric itself) is necessary. ror current velocities between 5 and 50 cm/sec, a galvanized or ^tainless steel wire rope snould be used as a COP or I center tension member; the center tension curtain pro vides a somewhat greater effective skirt depcn, but requires stronger tension members as well as ^orc erfcc- tive anchor systems. The noncorrosive, ballast chain I should have a weight ranging from approxinatelv 1.5 r?/r (1 Ib/lin ft) for a 1.5-m skirt depth up to 3.0 k'-;/n (2 Ib/lin ft) ror a 3-m skirt depth. I

117. Transportation. When transporting silt curtains fron a I :orage facility to an unloading site, thev should be furled (Ficure 33), _ca ';ith lightweight straps or rope every 1 to 1.5 n, conpactlv foldec I FLOTATION -STRAP FLOTATION I SKIRT 1 A -* BALLAST CHAIN BALLAST CHAIN I ''IE\V A-A Figure 33. Furling of the curtain skirt for deployment I and/or recovery of silt curtains

.ecordion style, packaged into large bundles, carefully lifted into the I :nnsport vehicle, and transported to the unloading dock. At the -nloaaing dock, the truck is backed down the ramp so that the tailgate I if .-'s close to the water as possible; the curtain is then carefully 'ailed out of the truck (like a string of sausages). After all the 30-m I Actions have been payed out and joined, the curtain can be towed bv '•at to the site at speeds of approximately 1 m/sec. The curtain should r-n.iin furled except near the end connectors until it has been deployed I

L '-!ie operation bite. '13. An alternative method involves maneuvering the1 curtain onto I ." < -un-tiecked workboat or barge, transporting the curtain to rh • I I PGE00047420 I site, and, finally, uCf-loading the curtain in sections. The sections I are then joined and the curtain deployed. 119. Mooring. Improper and/or inadequate mooring systems typicall' I contribute to silt curtain ineffectiveness and failure. The recommended I mooring system (Figure 34) consists of an anchor, a chain, an anchor

-CROWN BUOY MOORING BUD/ /SILT I / CURTAIN I I I I I BOTTOM SEDIMENT I Figure 34. Recommended silt curtain mooring system rode (line or cable), and mooring and crown buoys. It is recommended I that the curtain be anchored from the section joints every 30 m in a radial pattern (Figure 31) and on both sides if the curtain is exposed I to reversing tidal currents. Half-inch (1.275-cm) polypropylene line used in conjunction with lightweight, self-burying anchors with weights of at least 4.5 kg for sandy bottom sediment and up to 34 kg for firm I mud will provide adequate holding power in most situations. However, with increasing current velocities, the anchor weights will also have to I be increased. 120. After the furled curtain has been anchored, it should be I checked to ensure that the skirt is not twisted around the flotation. 1 If this is the case, the curtain should be separated at the nearest I 96 PGE00047421 I

--^c-or. untwisted, and rejoined. The curtain in its du^lo-cd, in- I I.T^JT configuration can now be unfurled by simply cutting tie i urli v --„ o- j^ravs. If the barrier needs to be repositicned u-irin-; the I - ..ic 2n, any curtain with a long skirt depth should be rerurleo retore > oved. I Jl. Deployment model. The length of tine that a silt curtain can rain deployed in one configuration before the enclosed area nust De Uarged or the curtain moved to a new location to prevent siltation I _jng the lower edge of the curtain depends on the accumulation or tluio -side the curtain relative to the deployment geometrv, the slurr\ . ..ar~e rate, and the initial bottom gap (i.e., the distance Between • -jr skirt edge and the bottom sediment at the beginnirs or the I .rion) (Figure 35). The size of the enclosure is linitcd by the I - -CHARGE I I

///\ Js/ /s ' / S BOTTOM SEDIMENT * I R I "igure 35. Parameters affecting the schedule for moving and redeploying silt curtains I 'ai Lungth of the curtain available for the project; as the enclosed "T and bottom gap increase, the length of time before the curtain nust I -ovr'd also increases. Since it tnay be necessary to move a silt -'t.un during an operation, the following procedure can be used to I '"'OTJ i general schedule for curtain movement and redeployment. -2. fo illustrate the use of the nomograph (Figure 36) used in I • 'it-cedure, assume that approximately 975 m of curtain with a ^k 'th ^>t L.5 ra surrounds an open-water pipeline disposal operation I 97 I PGE00047422 1 I located in a quiescent nontidal environment with a uniform water depth I of 2.7 m. The circular configuration has a radius of approximately 155 m. The dredged material slurry with a solids content of 15 percent (by I weight) is discharged from a 46-cm (18-in.) pipeline at a velocity of 5.5 ra/sec. To determine when the fluid mud dredged material will build up to the lower edge of the silt curtain: I a_. Enter graph I (upper left, Figure 36) at "A" for 152 m radius. I b_. Proceed vertically to "B," the planned initial bottom gap (i.e., 1.2 m) between the silt curtain and the existing I bottom sediment. c^. Move horizontally through the right-hand axis indicating the approximate volume of the fluid mud dredged material I mound (i.e., 0.3 million cu m) to "C" (graph II). (I. Draw a vertical line from "C" through the lower axis indicating the amount of slurry pumped (i.e., 0.57 I million cu m) and into graph IV. <^. Enter graph III (lower left) at "D," the appropriate flow I velocity (i.e., 5.5 m/sec). £. Proceed vertically to the curve indicating the appropri I ate pipeline diameter (i.e., 46 cm). £. Draw a horizontal line from "E" through the right-hand discharge rate axis (i.e., at 78,974 cu m/day) and into 1 graph IV until it intersects the vertical "total volume of slurry pumped" line at "F." The length of tine before the curtain needs to be moved is estimated from the I diagonal time line that goes through "F." 123. In this example the operation can probably continue for i approximately 7.32 days before the curtain must be moved due to fluid mud accumulation up to the lower skirt edge. Figure 37 shows that the i mound will be approximately 2 m thick, under the discharge pipe and will extend radially approximately 395 m. If the configuration were semi i circular and located in a river (Figure 31B), the above procedure would be performed in the same manner using the radius of the semi i circle; however, with a semicircular configuration anchored onshore (Figure 31C) the calculated time is divided in half. Similarly, if the

~~i 98 i PGE00047423 I I I I I I \ I I~~

~~~LO* VELOCITY M /SEC TOTAL VOLUME OF "5LUHOY PUMPED 10 I 5-0- 6 i ^ j ~ TT I~~

~~! l I f . I ll I i i~~

'.'oleograph depicting the relationship among different parameters that affect the redeployment schedule :"or i silt curtains during an operation. It is assumed that the dredged material slurry is 15 percent solids by weight, the fluid mud is 25 percent solids by weight, i .;nd the fluid mud rcound has a slope of 1:200. »y_"v" rarer to example in text (paragraph 122) i 99 i PGE00047424 I I I MAXIMUM HEIGHT, H OF MOUND, M UNDER DISCHARGE I 10 15 ^0 21 10 IS I I I I

100 ^oo 3 CO 40O SCO »OO 600 I MAXIMUM RADIUS ,R Or MOUND, M I Figure 37. Dimensions of a fluid mud mound with a slope of 1:200 I curtain is deployed in a square configuration with sides of length L, assume that the curtain is circular or elliptical in shape with radius of L/2. I 124. As pointed out previously, this procedure can be used to calculate an approximate schedule for moving silt curtains. Because of I the varying characteristics of an operation (i.e., slurry density, pumping time, etc.) and the settling/consolidation characteristics of I the fluid mud, there may be some variability associated with the rates of dredged material accumulation. However, this model does provide a I conservative time framework (i.e., a shorter length of time between curtain movements than might be necessary) for planning the silt curtain operation. Additional experience should indicate possible modifications I for improving the accuracy of this procedure. 125. Maintenance. To maximize the effectiveness of a silt curtain I operation, maintenance is extremely important. This entails moving the I 100 I PGE00047425 I I

• -_n '-/ay : rorn the turbidity source just before the liui^ -iu ! '' or ^:ors tnc lower edge of the skirt, replacing; worn or Kro-:on snchor -.:ia maintaining the, integrity of the curtain by I . _:cf3 ar.c/or tears in the curtain fabric. Tears in tne flotation _ .^~ :_m be repaired in the water with a hand type pop rivet sun. | . -. :<> tears in the skirt may be repaired on land with a vim'l/nvlon . ".1 cit and VIXYLFIX or PVC glue. Because extensively torn sections • >t oe returned to the manufacturer for refurbishing, one or two soare •cions should be purchased for immediate substitution in the field, ar -maintenance will not only decrease the curtain's effectiveness I ..'ticular operation, but also increase the cost of reconaition- _ curtain for reuse. I -~.. -'ecoverv. After the operation has been completed, the cur '.ould be refurled, the anchor/mooring system recovered, and the -r. returned to the launching site for cleaning, repacking, and • If properly stored in a location that is uncxposed to the I :tj, the curtain can be maintained in its existing condition tor -"._ .ears and reused on subsequent operations. •

Flocculant Injection i I -7. It may be possible under certain conditions to marginally mm -e cne settling velocity of the small percentage of dredged | **. slurry that is suspended in the water column during an open loeline disposal operation by injecting polyelectrolytes (floe- B "• 1 into the dredge pipeline before the slurry is discharged, r, the practicality of this technique is probably limited, at I -•Je to the variability in the solids concentration of the slurry, - ri solids concentrations that must be treated, as well ,>s the hi"h • na r.any logistical problems'associated with handling, mxuij;, and * "- n-4 rlocculants into the slurry. Therefore, the use or floe cu -> reduce dredged material dispersion at open-water pipolino I ---1 operations is not recommended. I

~~101 •~~

~~PGE00047426~~