+ All Categories
Home > Documents > EVALUATION OF THE TERRESTRIAL ECOSYSTEM OF THE HOUSATONIC … ·  · 2018-02-13evaluation of the...

EVALUATION OF THE TERRESTRIAL ECOSYSTEM OF THE HOUSATONIC … ·  · 2018-02-13evaluation of the...

Date post: 17-Apr-2018
Category:
Upload: buituyen
View: 216 times
Download: 1 times
Share this document with a friend
170
i *,: ,'.' EVALUATION OF THE TERRESTRIAL ECOSYSTEM OF THE HOUSATONIC RIVER VALLEY JULY 26,1994 preparedfor: The General Electric Company 100 Woodlawn Avenue Pittsfield, Massachusetts 01201 prepared by: ChemRisk® A Division of McLaren/Hart Stroudwater Crossing 1685 Congress Street Portland, Maine 04102 (207) 774-0012 in collaboration with: S.G. Martin & Associates, Inc. 7121 N. County Road 9 Wellington, CO 80549 (303) 568-9333 A Division of McLaren/Hart Environmental Engineering r fBcyc/ed paper oioH25
Transcript
  • i * , : , ' . '

    EVALUATION OF THE TERRESTRIAL ECOSYSTEM

    OF THE HOUSATONIC RIVER VALLEY

    JULY 26,1994

    preparedfor:

    The General Electric Company 100 Woodlawn Avenue

    Pittsfield, Massachusetts 01201

    prepared by:

    ChemRisk A Division of McLaren/Hart

    Stroudwater Crossing 1685 Congress Street

    Portland, Maine 04102 (207) 774-0012

    in collaboration with:

    S.G. Martin & Associates, Inc. 7121 N. County Road 9 Wellington, CO 80549

    (303) 568-9333

    A Division of McLaren/HartEnvironmental Engineering

    r fBcyc/ed paper oioH25

  • Area Environmental & Facility Programs General Electric Company 100 Woodlawn Avenue, Pittsfield, MA 01201

    July 26, 1994

    Ms. J. Lyn Cutler Section Chief, Special Projects Department of Environmental Protection 436 Dwight Street Springfield, MA 01103

    Mr. Bryan Olson U.S. Environmental Protection Agency Waste Management Division J.F. Kennedy Federal Building HRR-CAN3 Boston, MA 02203

    Re: Housatonic River (DEP # 1-0147P, EPA Permit Area 6) Terrestrial Ecosystem Assessment

    Dear Ms. Cutler and Mr. Olson:

    Enclosed for your review is a report entitled "Evaluation of the Terrestrial Ecosystem of the Housatonic River Valley," prepared by ChemRisk in collaboration with Dr. Stephen G. Martin, a terrestrial ecological expert. This report describes the methodology and results of their study of the health of terrestrial wildlife potentially exposed to PCBs in the Housatonic River floodplain.

    In consultation with ChemRisk, GE plans to use this report as one basis for the ecological risk assessment of the Housatonic River.

    We look forward to discussing this report with the DEP and EPA at our meeting on August 18. Please call me if you have any questions.

    Sincerely yours,

    Johohnn D. Ciampa Project Manager

    Enclosure

  • - 2

    cc: Alan Weinberg, DEP WERO Stephen F. Joyce, DEP Office of Commissioner Robert Bell, DEP WERO Stephen P. Winslow, DEP Susan Steenstrup, DEP WERO Carol Rowan West, DEP ORS Meg Harvey, DEP ORS Gary B. Gosbee, EPA Region I Douglas J. Luckerman, EPA Region I Ruth Bleyler, EPA Region I Ronald F. Desgroseilliers, GE Stephen C. Moore, GE Andrew J. Thomas, Jr., GE Ellen S. Ebert, ChemRisk Stephen G. Martin, Ph.D., S.G. Martin & Associates Daniel M. Woltering, ENVIRON Robert K. Goldman, Blasland, Bouck & Lee James R. Bieke, Shea & Gardner Public Information Repositories (ECL I-R-IV(A)(1))

  • EVALUATION OF THE TERRESTRIAL ECOSYSTEM

    OF THE HOUSATONIC RIVER VALLEY

    prepared for:

    The General Electric Company 100 Woodlawn Avenue

    Pittsfield, Massachusetts 01201

    prepared by:

    ChemRisk Division of McLaren/Hart

    Stroudwater Crossing 1685 Congress Street

    Portland, Maine 04102 (207) 774-0012

    July 26,1994 Reviewed and Approved By:

    Miranda HenningTask Manager

    C_3 'Date Ellen E.Ebert Project Manager

    Russell E. Keenan, Ph.D.Principal-In-Charge

    Date

    ChemRisk A Division of McLaren/Hart Environmental Engineering

  • ChemRisk - A Division of McLaren/Hart July 26,1994

    TABLE OF CONTENTS

    1.0 INTRODUCTION 1-1

    2.0 METHODS 2-1

    2.1 Ecosystem Selected for Evaluation 2-1 2.1.1 Characterization of Habitats 2-3 2.1.2 Selection of Indicator Species 2-4

    2.2 Characterization of Exposure Scenarios 2-4 2.2.1 Small Mammal Exposure Profile 2-5 2.2.2 Song Bird Exposure Profile 2-6

    2.3 Selection of Endpoints 2-7

    2.4 Selection of Study Areas 2-8 2.4.1 Flood Plain Forest Target Areas 2-8 2.4.2 Shrub Meadow Study Areas 2-10 2.4.3 Characterization of PCBs in Flood Plain Soils and River Sediments 2-11

    2.5 Ecological Assessment Methods 2-14 2.5.1 Method of Comparing Target Areas and Reference Areas 2-14 2.5.2 Method for Evaluating Avian Population Structure 2-15 2.5.3 Method for Evaluating Avian Reproductive Success 2-17 2.5.4 Method for Evaluating Small Mammal Population Structure 2-21 2.5.5 Method for Evaluating Small Mammal Age Structure 2-22 2.5.6 Method for Evaluating Small Mammal Reproductive Success . . . . 2-22

    3.0 RESULTS 3-1

    3.1 Avian Population Structure 3-2 3.1.1 Avian Population Structure in the Massachusetts Target Area 3-2 3.1.2 Avian Population Structure in the Maryland Reference Area 3-3 3.1.3 Avian Population Structure in the North Carolina Reference Area . . . 3-4 3.1.4 Summary of Findings on Avian Population Structure 3-5

    3.2 Avian Reproductive Success 3-7 3.2.1. Clutch Sizes 3-7 3.2.2 Young Hatched 3-8 3.2.3 Hatching Success 3-9 3.2.4 Summary of Findings on Avian Reproductive Success 3-9

    3.3 Small Mammal Population and Density 3-9 3.3.1 Population and Density of White-Footed Mice and Southern Red-

    Backed Voles 3-9 3.3.2 Population and Density of Shrews 3-11 3.3.3 Summary of Findings on Small Mammal Populations and Density . 3-11

  • ChemRisk - A Division of McLaren/Hart July 26, 1994

    TABLE OF CONTENTS (CONT'D)

    3.4 Small Mammal Age Structure 3-11 3.4.1. Findings on White-Footed Mice and Southern Red-Backed Voles . 3-12 3.4.2 Summary of Findings on Small Mammal Age Structure 3-12

    3.5 Small Mammal Reproductive Success 3-13 3.5.1 Reproductive Success of White-Footed Mice and Southern

    RedBacked Voles 3-13 3.5.2 Reproductive Success of Shrews 3-14 3.5.3 Summary of Findings on Small Mammal Reproductive Success ... 3-15

    3.6 Additional Ecological Data 3-15 3.6.1. Observations of Wildlife Species 3-15 3.6.2 Summary of Findings on Additional Ecological Data 3-17

    4.0 DISCUSSION 4-1

    4.1 Uncertainty Analysis 4-1

    4.2 Discussion of Results 4-4 4.2.1 Absence of a Species Normally Expected to be Present 4-4 4.2.2 Reduction of a Population or Subpopulation 4-5 4.2.3 Change in Community Structure 4-6 4.2.4 Bioaccumulation Associated with an Adverse Effect 4-7

    4.3 Conclusions 4-8

    5.0 REFERENCES 5-1

    Appendix A - Selection of Indicator Species

    Appendix B - Comparison of Target Areas and Reference Areas

    Appendix C - Results of Statistical Analyses of Avian Reproductive Data

  • ChemRisk - A Division of McLaren/Hart July 26, 1994

    LIST OF FIGURES

    Figure 2-1 Map of the Housatonic River Flood Plain Between Lenox Rd. & Woods Pond Dam 2-2a

    Figure 2-2 Summary of River Sediment and Flood Plain Soil Sampling Results & Study Area Locations 2-2b

    Figure 2-3 Summary of River Sediment and Flood Plain Soil Sampling Results & Study Area Locations 2-2c

    Figure 2-4 Summary of River Sediment and Flood Plain Soil Sampling Results & Study Area Locations 2-2d

    Figure 2-5 Summary of River Sediment and Flood Plain Soil Sampling Results & Study Area Locations 2-2e

    Figure 2-6 Avian Census Observation Route 2-15a

    Figure 3-1 Nest Heights of Avian Species Observed in Massachusetts Target Area . . . . 3-2a

    Figure 3-2 Avian Population Study: Density 3-5c

    Figure 3-3 Avian Population Study: Plot Size and Number of Species 3-6a

    Figure 3-4 Avian Population Study: Diversity 3-6b

    Figure 3-5 Avian Reproduction Evaluation: Clutch Size 3-7a

    Figure 3-6 Comparison of Observed Clutch Sizes to Ranges Reported in the Literature . 3-8a

    Figure 3-7 Avian Reporoduction Evaluation: Young Hatched 3-8b

  • ChemRisk - A Division of McLaren/Hart July 26, 1994

    LIST OF TABLES

    Table 2-1 Rationale for Selection of Indictor Species 2-4a

    Table 2-2 Ecosystem Characterization 2-4b

    Table 2-3 Exposure Analysis Summary 2-4c

    Table 2-4 Candidate Reference Areas and Rationale for Exclusion 2-9a

    Table 2-5 Summary of April 27 - June 7,1994 Flood Plain Soil Sampling and Analysis Program 2-13a

    Table 3-1 Summary of 1993 Avian Census Results for Massachusetts Target Area . . . . 3-2b

    Table 3-2 Summary of 1991 Avian Census Results for Maryland Reference Area . . . . 3-3a

    Table 3-3 Summary of 1991 Avian Census Results for North Carolina Reference Area 3-4a

    Table 3-4 Comparison of Avian Census Results 3-5a

    Table 3-5 Comparison of Species Densitiies in Three Reference Areas and in the Literature (territories/40ha) 3-5b

    Table 3-6 Small Mammal Population Estimates 3-10a

    Table 3-7 Summary of Flood Plain Forest Small Mammal Population Study 3-1 Ob

    Table 3-8 Summary of Shrew Population Study 3-1 la

    Table 3-9 Summary of Small Mammal Population Age Structure 3-12a

    Table 3-10 Summary of Small Mammal Reproductive Analyses 3-14a

    Table 3-11 Expected and Observed Wildlife Species of the Housatonic River Valley . . 3-16a

    Table 3-12 Rare Bird Species Observed in the Housatonic River Valley 3-16b

    Table 4-1 Summary of Potential Exposure-Effect Relationships by Measurement Endpoint and by Species 4-8a

  • ChemRisk* - A Division of McLaren/Hart July 26, 1994 Page 1-1

    EVALUATION OF THE TERRESTRIAL ECOSYSTEM OF THE HOUSATONIC RlVER VALLEY

    1.0 INTRODUCTION

    ChemRisk, the risk assessment division of McLaren/Hart Environmental Engineering, in

    collaboration with S.G. Martin & Associates, Inc., was retained by the General Electric Company

    (GE) to conduct an evaluation of the health of terrestrial wildlife populations and communities

    potentially exposed to polychlorinated biphenyls (PCBs) in soils of the Housatonic River flood plain. McLaren/Hart's ChemRisk Division is one of the largest groups of professionals dedicated

    to the use of risk assessment to understand and solve problems associated with exposure to

    chemicals released to the environment. More than 20 members of the ChemRisk staff have

    advanced training in ecotoxicology, aquatic toxicology, wildlife toxicology, ecology, marine

    biology, wetlands and environmental science; the team together has published more than 30 papers

    on ecotoxicological risk assessment in peer-reviewed journals and books. Stephen G. Martin

    holds a Ph.D. degree in Animal Ecology; he has 30 years of experience in designing, conducting,

    and evaluating research programs pertaining to environmental and anthropogenic perturbations to

    animals.

    PCBs were used as insulating liquids in select transformer applications at the GE facility in

    Pittsfield from 1932 until March 1977 (Blasland & Bouck, 1991). Before 1977, inadvertent

    releases of these materials from the facility were conveyed to the East Branch of the Housatonic

    River. Use of PCBs at the facility was discontinued in 1977.

    Flood plain soils between the GE facility and Woods Pond Dam have been investigated for the

    presence of PCBs that may have resulted from deposition of river materials on the flood plain

    during flooding events (Blasland & Bouck, 1991, 1992a, 1992c, 1993, 1994a). Data from these

    sampling efforts indicate that the current composition of PCB Aroclors in Housatonic River flood

    plain soils is approximately 99 percent Aroclor 1260, with the remainder being Aroclor 1254.

    This ecological evaluation focuses on Aroclor 1260 as the primary stressor of interest

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-2

    It has been determined that PCBs in flood plain soils are largely limited to the approximate 10-year

    flood plain between the GE facility and Woods Pond Dam (Blasland & Bouck, 1991, 1992a,

    1993, 1994a). The great majority of samples collected beyond the approximate 10-year flood plain

    contained concentrations less than 1 ppm (Blasland & Bouck, 1991, 1992a, 1993, 1994a). In

    those few areas where PCB levels greater than 1 ppm were found outside of the approximate 10

    year flood plain, these exceedances were, in general, either the result of inaccuracies in the

    topographic mapping or interpretation used in modeling the limits of the approximate 10-year flood

    plain in a couple of areas (the flood plain limits in these areas have since been corrected) or else

    were found very close to (e.g., within two feet of) the limit of the approximate 10-year flood plain

    (see Blasland, Bouck & Lee, 1994b).

    In evaluating whether the presence of PCBs in the flood plain soils has potentially stressed the

    Housatonic River valley ecosystem, a number of additional factors must be considered as potential

    stressors. Physical stressors may include both natural and anthropogenic events, such as extremes

    in natural conditions, habitat alteration, direct manipulation of wildlife and resource management

    practices. Several natural conditions may stress the Housatonic River valley ecosystem, including

    flooding, storms, and extreme temperatures. Flooding may result in drowned individual biota and

    destruction of food resources, nests, and cover. Storms, with their associated precipitation,

    winds, and extreme temperatures, can stress wildlife through starvation, freezing, drowning, and nest destruction or abandonment. Such extremes in weather are not unusual in Massachusetts.

    Alteration of habitat along the Housatonic River is most apparent within the city of Pittsfield and

    downstream to the New Lenox Road bridge. Within this reach, residential, commercial and

    industrial structures have been constructed, oxbows have been filled, native vegetation has either

    been completely removed or replaced with ornamental shrubs and grasses, and the river channel

    has been straightened for flood control purposes. In many areas upstream of the New Lenox Road

    bridge, these activities have resulted in the complete removal of suitable foraging and nesting

    habitats, thereby limiting wildlife inhabitation in these areas. Where habitats have not been entirely

    destroyed in the more developed portion of the river, they have been substantially fragmented by

    buildings, parking lots, and roads. The roads which dissect the ecosystem of the Housatonic

    River valley limit the natural (and necessary) movements of wildlife, a situation analogous to an

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-3

    island existence (Diamond and May, 1976). Ecosystems that are fragmented, like the upper

    Housatonic River valley, create small islands capable of supporting small populations of a limited

    number of species (Lovejoy and Oren, 1981).

    Downstream of the New Lenox Road-bridge, in the Housatonic Valley Wildlife Management Area

    (HVWMA), the physical stressors associated with human development are replaced by hunting,

    trapping, and resource management practices. Wildlife species legally hunted in Massachusetts

    include crow, turkey, ruffed grouse, pheasant, quail, American woodcock, various waterfowl,

    snowshoe hare, cottontail rabbit, opossum, raccoon, fox, coyote, bobcat, gray squirrel, jackrabbit,

    black bear, deer, bull frog, green frog, and snapping turtle (MDFW, 1992). Species legally

    trapped in Massachusetts include fox, coyote, skunk, opossum, weasel, bobcat, fisher, mink,

    raccoon, otter, muskrat, and beaver (MDFW, 1991). Because HVWMA is actively managed to

    promote hunting and trapping and since representatives of nearly all levels of the food web may be

    legally hunted or trapped, these activities could potentially stress the overall ecosystem, should

    overharvesting occur. In addition to these direct effects, resource management practices, such as

    mowing and herbicide application, may also stress the ecosystem by changing vegetative structure,

    thereby promoting occupancy by non-forest and exotic species, and by limiting cover for feeding,

    nesting, and rearing young. The direct effect of mowing on reduced primary productivity may also

    lead to increased competition among primary consumers for food resources. Additionally, because mowing inhibits natural succession, mowed areas are effectively maintained at stages of early

    successional development. Early successional habitats tend to be dominated by opportunistic

    species, which in turn limit species diversity and richness.

    This study was designed to evaluate the effects of PCBs on terrestrial wildlife by comparing the

    health of wildlife populations within the flood plain downstream of New Lenox Road bridge

    (hereafter referred to as the target population) to the health of reference populations that share the

    characteristics and stressors of the target population, with the exception of exposure to PCBs.

    While there are a number of methodologies for evaluating ecological effects presented in various

    guidance documents (EPA, 1989a; EPA, 1989b; EPA, 1990; EPA, 1992), the study designs,

    terminologies, and formats recommended by these sources are in many cases inconsistent and in

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-4

    other cases extremely general, making it difficult to select a single methodology that conforms to all

    available guidance and is appropriate to all sites. In designing this study, professional judgment

    has been employed in selecting a strategy that is best suited to the specific character of the

    Housatonic River flood plain target area.

    Two basic types of study methodologies are typically used to assess environmental risks. The first

    includes predictive methods, which generally require the application of mathematical models of

    food webs to simulate the transfer of a chemical from sediment, soil, or water to lower levels of the

    food web and from these lower trophic levels to higher trophic levels. The application of most

    food web models, characterized as a "bottom-up" approach, requires substantial extrapolation,

    including:

    from a species typically used in laboratory experiments (such as the rat or chicken) to a variety of species found in the natural environment;

    from observed effects in the laboratory under very controlled conditions to predicted effects in the natural environment;

    from individuals to populations; and,

    from acute effects (such as the lethal concentration to 50 percent of the population after a single exposure, the LCso) to chronic effects.

    There are several other limitations to the bottom-up approach. First, the input data on toxicity that

    are required by such models are very limited and are not compiled in any guidance document or

    database, let alone one which has received the EPA's approval. As a result, it is necessary to

    conduct a comprehensive literature review, compiling a range of toxicity criteria. These studies

    and data must then be critically evaluated for applicability to the species and stressors of interest.

    In the absence of toxicity criteria handbooks, criteria for selecting input data and the values

    themselves will vary among risk assessors.

    Second, food web models do not realistically simulate varied food sources for different species and

    for individuals within a species. Typically, modelers may input the fraction of the diet assumed to

    be derived from a limited number of locations (e.g., chemically exposed vs. unexposed) and/or

    food groups (e.g., vegetation, invertebrates, small mammals). Usually, these proportions are

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-5

    assumed to remain constant over time and to be uniform for all individuals within a species. In

    reality, the diet of most wildlife is extremely variable; food availability dictates the relative

    proportions of different prey consumed, as well as the size and location of the feeding range.

    These factors vary on a daily and seasonal basis. Because models cannot adequately simulate the

    dynamics of feeding behavior, considerable uncertainty is introduced with respect to the transfer of

    chemicals up the food web.

    Third, models cannot account for the movement patterns of free-ranging animals or the individual

    behaviors and foraging site preferences. Each of these factors substantially influences dietary

    exposure.

    Finally, it is difficult, if not impossible, to adequately simulate the effects of both competition and

    predation in the natural environment. The fundamentals of population ecology assert that the health

    of a community is a function of all interactions within and between species (Begon and Mortimer,

    1986), as well as of physical and chemical stressors. The primary interactions within and between

    species are competition and predation. While predation is often the focus of food web models,

    competition also plays a critical role in the utilization of limited resources, such as food, water, and

    breeding territory. As such, competition directly impacts potential exposure to stressors that are

    present in these resources. By neglecting competition, food web models may substantially over-or underestimate actual exposure.

    The second type of methodology, retrospective assessment, involves field observations and the

    collection of community-level data on density, diversity, and reproductive success to determine if

    wildlife are adversely affected by the stressor. Because it is possible to focus field monitoring

    programs on those species at or near the top of the food web which are likely to face the highest

    levels of exposure, this approach may be described as "top-down."

    The "top-down" procedure ascertains whether diversity, density, and the reproductive success of

    wildlife living in contaminated areas lie within normal ranges as reported in the literature, or

    whether they compare favorably to populations inhabiting uncontaminated reference areas. By

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-6

    monitoring the actual ecosystem potentially at risk, this approach determines whether wildlife are

    healthy and whether the ecological system on which they depend is healthy and functional.

    Because of the different kinds of uncertainties associated with the "top down" and "bottom up"

    approaches, conclusions based on the two approaches may not always agree. In cases where a

    divergence exists between conclusions based on the "top down" and "bottom up" approaches,

    discrepancies likely reflect the extent of extrapolation necessary to relate the two approaches. For

    example, in a typical laboratory toxicity study, cohorts of laboratory mammals (e.g., mice, rats)

    would be administered a constant dose of a chemical in the diet, via gavage, intraperitoneally, or

    via inhalation over a relatively short period of time (days, weeks, or months). Such conditions are

    purposefully controlled in order to allow determination of a dose-response relationship. These

    same controlled conditions, however, narrow the applicability of the laboratory study's findings.

    In order to realistically apply the findings of this example laboratory study to actual conditions in

    the field, variability in the following factors would first need to be accounted for: species,

    exposure duration, no/lowest effects levels, genotype, individual/ population/community/eco

    system level effect, feeding preference, feeding behavior, grooming behavior, home range,

    metabolic rate, contaminant distribution, exposure route, and stress associated with confinement.

    Only the first three factors listed are normally accounted for in such extrapolations and, even then,

    the basis for the uncertainty factors which allow extrapolation may have little scientific foundation.

    In short, if a linkage is not apparent between results of field studies and laboratory studies or

    models, it is likely because one or more of these differences in conditions has not or cannot be

    accounted for. In the absence of a discernible relationship between field and laboratory studies,

    evaluation of a broader range of endpoints can help determine which approach provides the most

    accurate conclusions for the site in question.

    Given the high level of uncertainty associated with the "bottom-up" method, the "top-down"

    approach was selected as the more appropriate study design to be applied in this evaluation of the

    terrestrial ecosystem of the Housatonic River flood plain. This assessment focuses on potential

    impacts related to exposure to PCBs in materials deposited on the flood plain during flooding

    events. Because this investigation focuses on community structure and dynamics, the density,

    diversity, and reproductive success of two groups of animals that occupy key roles in the terrestrial

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 1-7

    ecosystem (birds and small mammals) were studied. Species density and diversity were evaluated

    using standard field censusing techniques, including mark-recapture studies for small mammals

    and plot censuses for birds. Reproductive success of small mammals was evaluated by counting

    the number of placental scars and/or embryos in female specimens. To study avian reproductive

    success, natural nests were monitored for the numbers of eggs laid and young hatched. Data

    collected for birds and small mammals inhabiting the flood plain were then compared to data for

    reference populations living in areas in which PCBs are not present and to data reported in the

    scientific literature.

  • ChemRisk* - A Division of McLaren/Hart July 26, 1994 Page 2-1

    2.0 METHODS

    To evaluate the potential effects of Aroclor 1260 on the terrestrial ecosystem of the Housatonic River valley, it was first necessary to:

    select the ecosystem to be evaluated, identify types of habitats present in that ecosystem, identify and select indicator species within those habitats, identify potential exposure scenarios, select endpoints to be evaluated, identify and select target areas and either reference areas or literature studies, demonstrate comparability of target areas and reference areas, identify study methodologies, identify statistical and analytical methods with which to compare the ecological health of target areas and reference areas, and

    perform the field studies.

    The methodologies used in evaluating each of these initial steps are discussed in the following

    sections.

    2.1 Ecosystem Selected for Evaluation

    The reach of the Housatonic River with the highest concentrations of PCBs in flood plain soils is

    the 10-year flood plain between the GE facility and Woods Pond dam (Blasland & Bouck, 1992a,

    1993, 1994). From the GE facility to the New Lenox Road bridge, habitats are either small or

    have been substantially altered and/or fragmented. This would confound any evaluation of

    ecological communities within this reach. Between the New Lenox Road bridge and Woods Pond,

    there are more widespread and unfragmented wildlife habitats. Although other physical stressors

    (such as hunting, trapping, and natural resource management practices) may be present

    downstream of the New Lenox Road bridge, these stressors should not confound the assessment,

    provided they are also present in reference areas that are evaluated and that species that are hunted

    or trapped are not included as indicator species.

    Areas of the Housatonic River valley downstream of Woods Pond dam are less suitable for

    evaluating the effects of PCBs on ecological receptors for several reasons. First, the PCB

    concentrations in both sediments and flood plain soils are substantially lower downstream of

    Woods Pond dam than they are upstream, with concentrations detected in flood plain soils

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-2

    downstream of the dam averaging below 1 ppm (Blasland & Bouck, 1991). Accordingly, the

    potential for PCB exposure along this stretch of the river would likewise be substantially lower.

    Moreover, the habitat in this portion of the flood plain is less suitable for wildlife than the habitat

    between the New Lenox Road bridge and Woods Pond. In virtually all areas downstream of

    Woods Pond dam where the flood plain is broad, land use is dominated by agriculture. Regular

    plowing, tilling, and harvesting practices on agricultural land strongly affect the natural ecosystem,

    rendering most agricultural lands suitable habitat only for opportunistic and tolerant species. In

    other areas downstream of Woods Pond dam, the flood plain is so narrow as to be practically

    nonexistent; PCBs would not likely be deposited on these steep banks, which are characteristic of

    scouring environments.

    For these reasons, the present evaluation focuses on the ecosystem located within the 10-year flood

    plain of the Housatonic River between the New Lenox Road bridge and Woods Pond dam, as

    illustrated in Figure 2-1. The concentrations of PCBs in flood plain soils within this reach range

    from below detection to around 100 ppm, as illustrated in Figures 2-2 through 2-5. Given the

    elevated concentrations of PCBs in flood plain soils, coupled with the widespread availability of

    desirable wildlife habitat along this stretch of the Housatonic River, this study area allows for the

    assessment of potential effects under worst-case (maximum) exposure conditions. If adverse

    effects are not observed in this ecosystem, then they would not be expected to occur where the

    potential for exposure is lower.

    The flood plain south of New Lenox Road to Woods Pond dam is relatively wide. Approximately

    one-half mile south of New Lenox Road, the flood plain along the east bank of the river is

    confined by October Mountain, while the west bank has a relatively flat topography resulting in an

    extended flood plain (Blasland & Bouck, 1991). Within this reach, the Housatonic River is

    associated with substantial wetland flood plains and backwater areas. Here, the river is slow-

    mixing and meandering as it travels through relatively flat areas of local topography, and

    deposition of sediment within the backwater areas is pronounced (Stewart Laboratories, 1982).

    Because the mesic flood plains between New Lenox Road and Woods Pond dam also provide the

    largest intact wildlife habitat upstream of Woods Pond dam, they were targeted for ecological

    study.

  • CtemRisk-A DMston of McLaren/Hart JULY29, 1994 PAGE 2-2A

    SEE FIGURES 2-2 THROUGH 2-5 FOR DETAIL mi >

    1200' 1200'

    APPROXIMATE SCALE

    24-00

    LIMIT OF APPROXIMATE FLOOD PLAIN

    EDGE OF WATER

    ROADWA-T OR TRAIL

    1 THE LIMIT OF FLOOD PLAIN REPRESENTS THE APPROXIMATE 10-YEAR FLOOD PLAIN

    Original includes color coding.

    BLASLAND, BOUCK & LEE, INC. ENGINEERS & SCIENTISTS

    GENERAL ELECTRIC COMPANY PITTSFIELD, MASSACHUSETTS

    MAP OF THE HOUSATONIC RIVER FIGURE FLOOD PLAIN BETWEEN

    7/94 54-AK NES LENOX RD. & WOODS POND DAM 1019741R/CHEMRSK/10197G01 DWG

    2-1

  • Ctt*nRtek-A DMaton of McLaren/Hart JULY 29.1994 PAQE2-2B

    1988-1989 PARCEL 29-1 SAMPLING LOCATION

    PCB CONCENTRATION (PPM DRV WEIGHT) 0-4 DEPTH

    PC8 CONCENTRATION (PPM DRY WEIGHT) 4"-8" DEPTH

    RESULTS Of DUPLICATE ANALYSIS

    FLOOD PLAIN TRANSECT LINE

    TRANSECT SAMPLE LOCATION (1990-1992)

    TRANSECT SAMPLE LOCATION NAME PCB CONCENTRATION (PPM) 0"-6' DEPTH PCB CONCENTRATION (PPM) 6--12" DEPTH

    ROOD PLAIN SIM EVALUATION SAMPLE LOCATION (AUGUST 1992)

    DECKER CANOE LAUNCH STM EVALUATION SAMPLE LOCATION (JULY 1993)

    SAMPLE IDENTIFICATION

    PCB CONCENTRATION (PPM DRV WEIGHT) a'-e' DEPTH SEDIMENT SAMPLING LOCATION (1981-1982) r / SEDIMENT SAMPLING LOCATION (1990-1991)

    FLOOD PLAIN COMPOSITE SAMPLING LOCATION (APRIL & MAY 1994)

    LIMIT OF APPROXIMATE IDYEAR FLOOD PLAIN

    NOTES.

    1 THE BASE MAP FEATURES PRESENTED ON THIS FIGURE WERE PHOTOGRAMMETRICALl'i MAPPED FROM APRIL 1990 AERIAL PHOTOGRAPHS (AREA NORTH OF NEW LENOX ROAD) AND FROM APRIL 1983 AERIAL PHOTOGRAPHS (AREA SOUTH OF NEW LENOX ROAD

    2 THE LIMIT OF FLOOD PLAIN REPRESENTS THE APPROXIMATE 1O-YEAR FLOOO PLAIN DELINEATION OF 10-YEAR FLOOD PLAIN IS BASED ON HEC-2 HYDRAULIC MODELING PERFORMED BY BLASLAND & BOUCK ENGINEERS P C (1991) AND AVAILABLE TOPOGRAPHIC MAPPING

    3 TAX ASSESSORS PARCEL IDENTIFICATION NUMBERS AND BOUNDARY INFORMATION OBTAINED FROM CITl OF PITTSFIELD S TAX ASSESSORS OFFICE AND IS CURRENT THROUGH DECEMBER 31 1991

    4 ALL SAMPLE LOCATIONS ARE APPROXIMATE

    5 DATA QUALIFIERS REFERENCE NO NOT DETECTED J ESTIMATED VALUE BELOW OUANT1TAT10N LIMIT fSHRUB MEADOW

    DECKER CANOE LAUNCH AREA APPROXIMATE SCALE

    BLASLAND, BOUCK & LEE, INC ENGINEERS & SCIENTISTS

    GENERAL ELECTRIC COMPANY PITTSFIELD MASSACHUSETTS

    CONTINUED ON SHEET 2-3

    SUMMARY OF RIVER SEDIMENT X 10197G03 DWG AND FLOOD PLAIN SOIL SAMPLING LA ONOFFREF. -MAP_NUM -PROPERTY -ROW GC-OBG 7/94 54- RLP NES

    RESULTS & STUDY AREA LOCATIONS 10I9741R/10197GO.'!

  • ChemR&F-A Division of McLaren/Hart JULY 29, 1994 PAQE 2-2C

    CONTINUED ON SHEET 2-2

    SHRUB MEADOW

    TARGET ARE

    SEDIMENT SAMPLING LOCATION (1981-1982)

    SEDIMENT SAMPLING LOCATION (1990-1991)

    FLOOD PLAIN COMPOSITE SAMPLING LOCATION (APRIL In MAY 1994)

    FLOOD PLAIN SAMPLING LOCATION (JUNE 1991)

    PC9 CONCENTRATION (PPM DRY WEIGHT) 0"-6" DEPTH

    NEST LOCATION

    FLOOD PLAIN LIMIT OF APPROXIMATE 10 -YR FLOOD PLAIN

    FOREST TARGET AREA EDGE OF WATER PAVED ROADWAY

    UNPAVED ROADWA1 OR TRAIL

    RAILROAD

    VEGETATION

    THE BASE MAP FEATURES PRESENTED ON THIS FICURF WERE PHOTOGRAMMETRICALLY MAPPED FROM APRIL 1983 AERIAL PHOTOGRAPHS

    THE LIMIT OF FLOOD PLAIN REPRESENTS THE APPROXIMATE 10-YEAR FLOOD PLAIN DELINEATION OF 10-YEAR FLOOD PLAIN IS BASED ON HEC-2 HYDRAULIC MODELING PERFORMED B~i BLASLAND & BOUCK ENGINEERS P C (I9S1) AND AVAILABLE TOPOGRAPHIC MAPPING

    TAX ASSESSORS PARCEL IDENTIFICATION NUMBERS AND BOUNDARY INFORMATION OBTAINED FROM CITY OF PITT5FIE1D S TAX ASSESSORS OFFICE AND IS CURRENT THROUGH DECEMBER 31 1991

    4 ALL SAMPLE LOCATIONS ARE APPROXIMATE

    5 DATA QUALIFIERS ND NOT DETECTED J ESTIMATED VALUE BEIOW QUANflTATION LIMIT

    APPROXIMATE SCALE

    BLASLAND, BOUCK & LEE, INC. ENGINEERS & SCIENTISTS

    CONTINUED ON SHEET 2-4 GENERAL ELECTRIC COMPANY PITTSFIELD, MASSACHUSETTS

    SUMMARY OF RIVER SEDIMENT FIGURE X 10197G04 DWG LA ON-' OFF>REF MAP_NUM -PROPERT1 -ROW GE> 06O 7/94 54 RLP NES

    AND FLOOD PLAIN SOIL SAMPLING RESULTS & STUDY AREA LOCATIONS 2-3

    1019741R/10197G04DVK:

  • ChemfVsk'-A DMslon of McLaren/Haft JULY29. 1994 PAGE2-2D

    CONTINUED ON SHEET 2-3

    FLOOD PLAIN TRANSECT LINE

    TRANSECT SAMPLE LOCATION (1990-1992)

    TRANSECT SAMPLE LOCATION NAME PCB CONCENTRATION (PPM) 0"-6" DEPTH PCB CONCENTRATION (PPM) 6"-12" DEPTH PCB CONCENTRATION (PPM) 12"-18" DEPTH PCB CONCENTRATION (PPM) 18"-24" DEPTH

    SEDIMENT SAMPLING LOCATION (1931-1932)

    SEDIMENT SAMPLING LOCATION (1990-1991)

    FLOOD PLAIN SAMPLING LOCATION (JUNE 1994)

    PCB CONCENTRATION (PPM DRY WEIGHT! 0"-6" DEPTH

    NEST LOCATION

    LIMIT OF APPROXIMATE 10-YEAR FLOOD PLAIN

    EDGE OF WATER

    PAVED ROADWAY

    UNPAVED ROADWA't OR TRAIL

    RAILROAD

    VEGETATION

    MOTtS

    1 THE BASE MAP FEATURES PRESENTED ON THIS FIGURE WERE PHOTOGRAMMETRICALLY MAPPED FROM APRIL 198"* AERIAL PHOTOGRAPHS

    ? THE LIMIT OF FLOOD PLAIN REPRESENTS THE APPROXIMATE 10-YEAR FLOOD PLAIN DELINEATION OF 10-YEAR FLOOD PLAIN IS BASED ON HEC-2 HYDRAULIC MODELING PERFORMED BY BLASLAND & BOUCK ENGINEERS, PC (1991) AND AVAILABLE TOPOGRAPHIC MAPPING

    3 TAX ASSESSORS PARCEL IDENTIFICATION NUMBERS AND BOUNDARY INFORMATION OBTAINED FROM CITY OF PITTSFIELD S TAX ASSESSORS OFFICF AND IS CURRENT THROUGH DECEMBER 31 1991

    4 ALL SAMPLE LOCATIONS ARE APPROXIMATE

    5 DATA QUALIFIERS ND NOT DETECTED

    Original includes color coding.

    APPROXIMATE SCALE

    BLASLAND, BOUCK & LEE, INC. ENGINEERS & SCIENTISTS

    GENERAL ELECTRIC COMPANY PITTSFIELD, MASSACHUSETTS

    SUMMARY OF RIVER SEDIMENT FIGURE AND FLOOD PLAIN SOIL SAMPLING

    RESULTS & STUDY AREA LOCATIONS 2-4 X 10197GOSDWG LA ON-> OFF-.REF- MAP_NUM.'PROPERTY -ROW.GE- OBG> 7/94 54- RLP, NES 1019741R/10197G05 DWG

  • ChemRlsk^A DMslon of McLaren/Hart JULY29. 1994 PAGE 2-2E

    1-3A (008 PPM)

    1-36 (1 2 PPM)

    1-3C

    (029 PPM)

    A FLOOD PLAIN STM EVALUATION SAMPLE LOCATION (AUGUST 1992)

    1-3A SAMPLE IDENTIFICATION

    ^0 06 PPM) PCS CONCENTRAT10N-(PPM DRY WEIGHT) 0"-6" DEPTH

    SEDIMENT SAMPLING LOCATION IB (1981-1982)

    SEDIMENT SAMPLING LOCATION a (1990-1991) FLOOD PLAIN SAMPLING LOCATION (JUNE 1994)

    PCB CONCENTRATION (PPM OR) WEIGHT) (46) 0"-6 DEPTH

    NEST LOCATION

    * LIMIT OF APPROXIMATE 10-YEAR FLOOD PLAIN

    EDGE OF WATER

    PAVED ROADWAY

    UNPAVED ROADWAt OR TRAIL

    RAILROAD

    VEGETATION

    NOTES.

    1 THE BASE MAP FEATURES PRESENTED ON THIS FIGURE WERE PHOTOGRAMMETRICALL1 MAPPED FROM APRIL 1983 AERIAL PHOTOGRAPHS

    2 THE LIMIT OF FLOOD PLAIN REPRESENTS THE APPROXIMATE 10-YEAR FLOOD PLAIN DELINEATION OF 10-YEAR FLOOD PLAIN IS BASED ON HEC-2 HYDRAULIC MODELING PERFORMED BY BLASLAND * BOUCK ENGINEERS P C (1991) AND AVAILABLE TOPOGRAPHIC MAPPING

    3 TAX ASSESSORS PARCEL IDENTIFICATION NUMBERS AND BOUNDARY INFORMATION OBTAINED FROM CITY OF PITTSFIELD'S TAX ASSESSORS OFFICE AND IS CURRENT THROUGH DECEMBER 31 1991

    4 AIL SAMPIE LOCATIONS ARE APPROXIMATE

    .

    Original includes color coding.

    200 0 200 4-00

    APPROXIMATE SCALE

    BLASLAND, BOUCK & LEE, INC. ENGINEERS & SCIENTISTS

    GENERAL ELECTRIC COMPANY PFTSFIELD, MASSACHUSETTS

    SUMMARY OF RIVER SEDIMENT FIGURE AND FLOOD PLAIN SOIL SAMPLING

    RESULTS & STUDY AREA LOCATIONS 2-5 X 101Q7G06DWG LON--OFF--REF- -MAP_NUM -PROPERTY.-ROW GE- OBG"* 94 ?4- RLP NES I0174IR/10197G060WG

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-3

    2.1.1 Characterization of Habitats

    Based on Land Use and Property Ownership Maps (Blasland & Bouck, 1992b), National

    Wetlands Inventory maps, and field observations, much of the land between New Lenox Road and

    Woods Pond dam that is potentially affected by a 10-year flood event is comprised of flood plain

    forests, shrub meadows, and cattail marshes.

    The Housatonic flood plain forest is dominated by mixed hardwood trees with scattered pine. It is

    intermittently flooded and has well developed ground, shrub, and tree strata. Ground cover is

    prevalent, consisting mainly of ferns. The shrub stratum, which covers approximately one-third of

    the flood plain forest, is relatively diverse and is dominated by arrowood and sugar maple. The

    tree stratum is well developed both in terms of density and diversity. The diversity is typical of a

    forest controlled by flooding; while many species of trees (balsam poplar, black oak, black cherry,

    white and yellow birch, ironwood, etc.) are represented, one species dominates (sugar maple).

    The ground, shrub, and tree strata together comprise a diverse community providing high quality

    habitat for song birds and small mammals.

    Shrub meadows along the Housatonic River between New Lenox Road and Woods Pond dam are

    early successional ecosystems dominated by grass-forb-shrub communities. Within the region of

    interest, most shrub meadows are actively managed for wildlife and hunting by mowing and other

    practices. Vegetation is abundant, even in mowed areas. The ground cover is comprised of

    grasses and forbs, interspersed with occasional emergent shrubs.

    As their name would suggest, cattail marshes are dominated by cattails, although limited numbers

    of red osier dogwood, willow, and silver maple are interspersed among the cattails. Arrowheads,

    water hyacinths, and lily pads are generally present in the backwaters of the river along the cattail

    marshes. Duck blinds and nest boxes have been established in many of these marsh areas.

    Additionally, because the cattail stands are of limited size, are frequendy inundated with water, and

    are dominated by a single plant species, they do not support extensive or diverse terrestrial wildlife

    populations.

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-4

    2.1.2 Selection of Indicator Species

    Upon completion of a literature review, indicator species were selected based on feeding habits,

    territory sizes, densities, trophic levels, and extent of human interference associated with each of

    the individual species expected in the ecosystem. While large mammals, raptors, and piscivorous

    species were considered in the selection of indicator species, there were a number of reasons why

    they were not determined to be optimal indicator species for this evaluation. A complete review of

    the rationale behind the selection of target species is provided in Appendix A and summarized in

    Table 2-1.

    The indicator species selected for evaluation included the Northern short-tailed shrew, white-footed

    mouse, Southern red-backed vole, Eastern phoebe, barn swallow, wood thrush, American robin,

    yellow warbler, American redstart, rose-breasted grosbeak, and red-winged blackbird.

    As shown in Table 2-2, each of the indicator species shares several characteristics which directly

    influence their potential for exposure to PCBs in flood plain soils. Specifically, all of the indicator

    species have home ranges substantially smaller than the 10-year flood plain, ensuring that the

    majority of the area where they forage is within the study area. Second, with few exceptions, the

    indicator species are secondary consumers, indicating that exposures are likely to occur through

    foodweb transfer, as well as through direct contact and ingestion of soils during grooming.

    Potential for exposures through direct contact varies among indicator species; those that nest and forage on or near the ground, such as the small mammals, wood thrush and red-winged blackbird,

    clearly have increased potential for exposure through direct contact and inhalation. In contrast,

    most of the activities of the American redstart are conducted in the upper canopy, where exposures

    are limited to food web transfer as it feeds on flying insects that metamorphosed from larval stages

    living in flood plain soil or river sediments.

    2.2 Characterization of Exposure Scenarios

    The potential exposure routes, frequencies, and timings for the indicator species used in this

    ecological evaluation are summarized in Table 2-3. For all species, the spatial scale of their

    primary activities is closely related to the spatial scale of PCBs in flood plain soils. That is,

    depending upon the species, certain or all activities such as foraging, incubating, preening or

  • ChemRisk A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1.

    Species Considered Advantages

    Large Mammals

    White-tailed Range includes the Berkshires Deer Social value (hunting,

    aesthetics)

    Black Bear Range includes the Berkshires Social value (hunting, aesthetics)

    Moose Range includes the Berkshires Social value (aesthetics)

    Rationale for Selection of Indicator Species

    Selected as an Indicator

    Disadvantages Species?

    Diet primarily consists of vegetation, so opportunity would NO be limited for exposure through the food web Large feeding territories and home ranges expand beyond the target area, suggesting it would be difficult to distinguish between target and reference populations Limited potential for exposure due to large feeding territories and home ranges Sample sizes would be limited due to low population densities typical of large mammals Preferred habitat is limited within the 10-year flood plain of the Housatonic River valley

    Diet primarily consists of vegetation, so opportunity NO would be limited for exposure through the food web Large feeding territories and home ranges expand beyond target area, suggesting it would be difficult to distinguish between target and reference populations Limited potential for exposure due to large feeding territories and home ranges Sample sizes would be limited due to low population densities typical of large mammals Preferred habitat is limited within the 10-year flood plain of the Housatonic River valley

    Diet primarily consists of vegetation, so opportunity NO would be limited for exposure through the food web Large feeding territories and home ranges expand beyond the target area, suggesting it would be difficult to distinguish between target and reference populations Limited potential for exposure due to large feeding territories and home ranges Sample sizes would be limited due to low population densities typical of large mammals Preferred habitat is limited within the 10-year flood plain of the Housatonic River valley

    Page 1 of 8

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page2-4a

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Selected as Species

    Considered Advantages Disadvantages an Indicator

    Species?

    Furbearers

    Marten Because diet consists of small Current distribution is restricted to north of the St. NO mammals, birds, and insects, opportunity for dietary exposure through food chair transfer is high Home ranges are normally small; therefore individuals are naturally restricted to the reference or target area Easily trapped for study Social value (trapping)

    Lawrence River. Therefore, it is unlikely that populations exist in the Berkshires Preferred habitat is upland, characterized by spruce, balsam, and hemlock; in contrast, the habitat of the Housatonic River valley is comprised of bottomlands and cattail marshes

    Fisher Social value (trapping) Extirpation of fishers from New England has been reported. Presence in the Berkshires is doubtful

    NO

    Limited potential for exposure due to large home ranges Large home ranges (12.8 to 24 km in diameter) which expand beyond the target area would make it difficult to distinguish between target and reference populations

    Long-tailedWeasel

    Diet primarily consists of small mammals, increasing potential

    exposure through food web transfer

    Social value (trapping)

    The grassy and eroded shores of the Housatonic River differfrom the rocky shores preferred by long-tailed weasels. Due to habitat unsuitability, populations of sufficient size for quantitative study are not expected in the Housatonic River valley Limited potential for exposure due to large home ranges

    NO

    Ermine

    Diet consists primarily of small mammals, increasing potential exposure through food web transfer

    Social value (trapping)

    Preferred habitat (rock piles or other heavy cover close to awatercourse) is not present in the Housatonic River valley. Due to habitat unsuitability, populations of sufficient size for quantitative study are not expected in the Housatonic River valley Distinguishing between reference and target population would be difficult, given ermines' propensity to travel great distances to feed

    NO

    Page 2 of 8

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Species Considered Advantages

    Furbearers (cont'd)

    Mink Diet includes fish, frogs, and aquatic insects, increasing opportunity for accumulation through the food web

    Considered highly sensitive to PCBs

    River Otter Relatively small but stable populations inhabit most New England states

    Diet includes fish, increasing opportunity for accumulation through the food web

    Selected as an Indicator

    Disadvantages Species?

    The mink's solitary lifestyle would make it difficult to NO locate a target population of adequate size to render statistically meaningful results Extensive home ranges expand beyond the target area, suggesting it would be impossible to distinguish between target and reference populations, due to normal low population densities, populations of sufficient size for quantitative study are not expected in the Housatonic River valley Mink are subject to a wide range of stressors (trapping, habitat destruction, and other human disturbances), the effects of which would confound observations of population effects

    Active trapping status in MA would confound observations NO of population effects in the field Feed in both contaminated and uncontaminated areas within a large home range; thus, potential exposure is limited Extensive home ranges expanding beyond the study area would make it impossible to distinguish between target and reference populations Sample sizes would be limited due to naturally low populations densities

    Page 3 of 8

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Selected as Species an Indicator

    Considered Advantages Disadvantages Species?

    Small Mammals

    White- Diet includes insects, increasing Diet also consists of seeds, acoms, nut, fruits, and green YES footed opportunity for accumulation of plants, reducing opportunity for dietary exposure Mouse contaminant through the food

    web Commonly found in woodlands

    and thickets like those of the flood plain forests along the Housatonic River

    Adequate sample sizes; easily studied due to ease of trapping and high population densities

    Close contact with soil increases exposure through dermal contact, inhalation, and incidental ingestion

    Small home ranges ensure that individuals are naturally restricted to target or reference areas

    Well researched species, permitting comparisons with literature reports

    Shrew Insectivorous feeding behavior Mark and recapture population studies are not feasible, due to YES Species increases opportunity for extremely rapid metabolism, which causes the animals to

    exposure through the food web starve to death within hours of being trapped Commonly found in woodlands

    and thickets like those of the flood plain forests along the Housatonic River

    Close proximity to soil increases exposure through dermal contact, inhalation, and incidental ingestion

    Page 4 of 8

  • ChemRiskJuly 26, 1994 Page 2-4a

    A Division of McLaren/Hart

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Selected as an Indicator

    Species?

    YES

    NO

    Species Considered Advantages

    Small Mammals (cont'd)

    Southern Commonly found in forests Red-backed similar to those in the flood Vole plains along the Housatonic

    River Owls, hawks, foxes, bobcats,

    mustelids, and raccoons feed on meadow voles; therefore, alterations in vole populations could potentially affect all species that prey on them

    Small home ranges naturally restrict individuals to contaminated or uncontaminated areas

    Adequate sample sizes easily studied, due to ease of trapping and naturally high population densities

    Close contact with soil increases exposure through dermal contact, inhalation, and incidental ingestion

    Easily studied concurrently with other small mammals

    Avian Species

    Belted Piscivorous feeding behavior Kingfisher increases opportunity for

    exposure through the food web Social value (aesthetics, bird

    watching)

    Disadvantages

    Diet consists of vegetative material, thus reducing opportunity for dietary exposure

    Solitary behavior would make it difficult to locate a target population of adequate size to render statistically meaningful results Woody roots, which are common in the Housatonic study area, impede nest excavation; therefore, belted kingfisher would not breed along the Housatonic River Masses of pondweed and water buttercup (like those found on the reaches of the Housatonic River) are a deterrent to foraging

    Page 5 of 8

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Species Selected as i Considered Advantages Disadvantages Indicator Spec

    Avian Species (cont'd)

    Osprey Piscivorous feeding behavior Ospreys are migratory species in the Berkshires. In MA, they NO increases opportunity for only breed near the coast (Veil and Peterson, 1993) exposure through the food web Large feeding ranges may expand far beyond the target area,

    Social value (aesthetics, bird suggesting it would be difficult to distinguish between target watching) and reference populations

    Clear waters are preferred for foraging over those with extensive emergent and submergent vegetation, like the Housatonic River.

    Bald Eagle Piscivorous feeding behavior Does not nest in the region NO increases opportunity for Sample sizes would be limited due to naturally low population exposure through the food web densities Social value (aesthetics, bird Clear waters are preferred for foraging over those with extensive watching) emergent and submergent vegetation, like the Housatonic River

    Open canopy and old growth forests, which are not generally present along the Housatonic, are preferred habitat

    Human disturbances that are common in the Housatonic River valley may deter bald eagles from inhabiting the region

    Studying an endangered species, particularly one known to avoid human disturbance, might disturb the birds to the extent that behavior and reproductive success might be adversely affected

    Huge feeding areas make it impossible to differentiate between target and reference populations

    Other Birds High trophic level enhances Sample sizes expected to be limited due to naturally low NO of Prey opportunities for exposure population densities

    through food chain transfer Large feeding territories and home ranges expand beyond the Social value (aesthetics, bird study area; therefore, extent of exposure to chemicals within the watching) Housatonic flood plain is unknown

    Great Blue Piscivorous feeding behavior Great blue herons do not nest in the Housatonic River valley; NO Heron increases opportunity for nesting restricted to more remote areas of the Berkshires (Keefe,

    exposure through the food web personal communication, 1993) Inspection of the nests for Large feeding territories that expand beyond the target area reproductive success could be would make it difficult to distinguish between reference and easily accomplished, since the target populations birds nest in dense colonies Open canopies and exposed limbs generally used for nesting are Social value (aesthetics, bird rare along the Housatonic River. Therefore, breeding watching) populations sufficient for a quantitative study are not expected

    (and are not present) Human disturbances that are common on the Housatonic River

    valley may deter great blue herons from nesting in the region. Therefore, sufficient populations for a quantitative study would be difficult to locate

    Page 6 of 8

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1. Rationale

    Species Considered Advantages

    Avian Species (cont'd)

    Other Colonial Nesting Piscivorous

    Ring-necked Pheasant

    Wood Duck

    Mallard, Canada Goose, American Black Duck

    Piscivorous feeding behavior increases opportunity for exposure through the food web Inspection of the nests for reproductive success could be easily accomplished, since the birds nest in dense colonies Social value (aesthetics, bird watching)

    Adequate sample sizes could be easily obtained due to naturally high population densities Small territory sizes naturally restrict individuals to either reference or target area Social value (hunting)

    Do not travel far from waterbodies for feeding; thus potential for exposure via incidental ingestion and dermal contact is increased Social value (hunting)

    Commonly breed in marshes along Housatonic River Social value (hunting)

    for Selection of Indicator Species (cont'd)

    Selected as an Indicator

    Disadvantages Species?

    According to maps prepared by Veil and Peterson (1993) and NO field observations, very few colonial nesting birds breed in the Berkshires

    Populations are actively managed through hunting and NO restocking. Therefore, the effects being studies may be masked or exaggerated by the management practices According to MA atlas (Veit and Petersen, 1993), there are few breeding pairs in western MA. It would be difficult to locate a breeding population of adequate size

    Diet is largely herbivorous, limiting potential for exposure NO via food web transfer Habitat in the target area does not provide sufficient natural nesting cavities; therefore, a substantial breeding wood duck population could not be supported naturally The populations are actively managed through habitat management and hunting; therefore, it would be impossible to distinguish population effects potentially caused by PCBs from those associated with hunting and poor natural nesting opportunities.

    Forage over wide areas, traveling to upland meadows, NO farmlands, and parks; therefore, exposure to PCBs would be low Diet is entirely herbivorous, limiting potential for exposure via food web transfer Mallards and American black duck populations are hunted. Therefore, it would be impossible to distinguish population effects potentially caused by PCBs from those associated with hunting

    Page 7 of 8

  • ChemRisk A Division of McLaren/Hart July 26, 1994 Page 2-4a

    Table 2-1. Rationale for Selection of Indicator Species (cont'd)

    Selected as Species an Indicator

    Considered Advantages Disadvantages Species?

    Avian Species (cont'd)

    Insectivo- Insectivorous feeding behavior Passerines occupy middle trophic levels. Organisms higher YES rous increases opportunity for on the food web may provide better indications of the effects Passerines exposure through the food web of PCBs on the ecosystem if they could be studied in

    Small home ranges and feeding sufficient numbers territories are within the target area, making it easy to distinguish between target and reference populations

    Adequate sample sizes are easily monitored, due to high population densities

    Breeding populations inhabit the Housatonic River valley

    Social value (aesthetics, bird watching)

    Note: All sources of information are detailed in Appendix A.

    Page 8 of 8

  • Indicator Species

    Small Mammals White-footed mouse

    Masked shrew

    Northern short-tailed shrew

    Southern red-backed vole

    Preferred Habitat

    Interiors and edges of deciduous, mixed, and coniferous forestsb

    Damp deciduous and coniferous woodlands

    with grasses, rocks, logs or stumps15

    Open habitats, such as banks of streams and meadows*5

    Cool moist deciduous. mixed, or coniferous forests among mossy rocks, logs, or tree rootsb

    ^ f ol^-f*pI

    > 5 i 3

    Table 2-2. Ecosystem Characterization o

    Foraging Migratory Territory/Home Nest Height Height

    Preferred Food Reproductive Cycle Timing

    Cycle Timing in Massachusetts8

    Range Size (ha)

    (m above ground)

    (m above ground)

    Seeds, acorns, nuts, Breed in late February to nm 0.06 - 0.22 for adult 0-2e of fruits, and tender green November15 malesb

    plantsb Young born March to 0.02 - 0.15 for adult December15 femalesb

    Litters per year not specified15

    Insects, worms, Breed in late April to late 0.04b O15

    spiders, snails, and September or October15

    slugs15 Young born in late April to September or October15

    Up to 3 litters per year15

    Insects, plants, worms Breed March to September15 nm 0.2-0.5lb Oc Ob

    snails, and small Young born April to vertebrates15 September15

    2 to 3 litters per year15

    Mainly green Breed from mid -January to nm 0.10-1.44b 0 d3 vegetation15 late November15

    Page 1 of 3

  • Table 2-2. Ecosytem Characterization (continued)

    Preferred Habitat

    Birds American redstart

    American robin

    Barn swallow

    Eastern phoebe

    Breeds in orchards, saplings bordering pastures, second-growth deciduous woodlands; in shade trees and second-growth maples; also in willow and alder thickets bordering ponds and streams13

    Breed in open woodlands, clearings, fields, orchards, and woodland edgesb

    Breed on rural farmlands13

    Breed in woodlands cliffs, ravines, agricultural and suburban areas, often near streams'3

    Preferred Food

    Insects, such ascaterpillars, bugs, flies, moths, small grasshoppers, beetles, and wasps'3

    Wild and cultivated fruits, insects, and earthworms'3

    (60% vegetable matter and 40% animal)b

    Flying insects and occasionally fruitb

    Flying insects and occasionally small fruitb

    Reproductive Cycle Timing

    Egg dates are May 21 to June 21d

    Incubation period is 12 to 14 daysb

    1 brood per year13

    Egg dates are April 12-25 to July 2-3d

    Incubation period is 1 1 to 14 days'3

    2 broods per year*3

    Egg dates are May 1 1 to August 3b

    Incubation period is 15 days'3

    1 or 2 broods per year*3

    Egg dates are April 27 to August 15d

    Incubation period is 15 to 17 days'3

    2 broods per year13

    Migratory Cycle Timing

    in Massachusettsa

    Arrive mid-May1'

    Leave late August and September*1

    Arrivemid-April*1

    Leave late October*1

    Arrivemid-April*1

    Leave late September to early October*1

    Arrive earlyMarchd

    Leave late September to early October"1

    Territory/Home Range

    Size (ha)

    0.1-0.4b

    0.1-0.3b

    Restricted to nest siteb

    0.3-1.45 per pair*3

    Foraging Nest Height Height

    (m above (m above ground) ground)

    1.2-9.1b 1.5-15.2b

    1.5 4.5, 0-3e

    up to 21.3b

    1.5-15.2b 1.5- 16e

    0.8 to 6.1b 0.8-76

    Chem

    Risk

    A D

    ivision of McL

    aren/E

    uly 26, 1994 'age 2-4b

    Page 2 of 3

    http:0.3-1.45http:0.1-0.3bhttp:0.1-0.4b

  • IIP

    Table 2-2. Ecosytem Characterization (continued)

    Preferred Habitat

    Red-winged blackbird

    Rose-breasted grosbeak

    Wood thrush

    Yellow warbler

    Preferred Food

    Breed in marshes, Insects, weed seeds, swamps, wet meadows and grainb

    ponds, dry fields. Prefer wetlands with extensive growth of cattails, bulrushes, sedges, and reedb

    Breed on the edges of Insects and spiders moist deciduous (50%) , seeds, and second-growth woods, fruits'' in thickets, suburban

    trees, and old orchards, as well as wooded borders

    of swamps and streams'5

    Breed in lowland Insects (62%) and deciduous or mixed fruits (38%)b

    forests; shady, cool, mature upland forests,

    often near a swamp, pond, stream, or lake; require abundant under growth*5

    Breed in farmlands, Insects, caterpillars, orchards, roadsides, and cankerworms, beetles, along streams and weevils, plant lice, lakesb and grasshoppers

    (94% animal, 6%

    vegetable)*3

    Reproductive Cycle Timing

    Egg dates are May 10 to July 18C

    Incubation period is 10 to 12 days'5

    2 to 3 broods per year*1

    Egg dates are May 23 to July 15d

    Incubation period is 12 to 14 days'5

    1 or 2 broods per yearb

    Egg dates are May 25 to June 26d

    Incubation period is 12 to 14 days'5

    2 broods per year15

    Egg dates are May 20 to June 2Qd

    Incubation period is 10 to 11 daysb

    1 brood per year13

    Migratory

    Cycle Timing

    in Massachusetts3

    Arrive midto late Marchd

    Leave late October*1

    Arrive

    mid to late

    Mayd

    Leave September"1

    Arrive late April

    early Mayd

    Leave late August and

    September4

    Arrive late Aprild

    Leave late July and early August"1

    Territory/Home Range

    Size (ha)

    0.02 0.5b

    NA

    - 0.6-2.8b

    NA

    Nest Height(m aboveground)

    0-4.3, typicallyless than 1.8b

    1.8-7.9b

    1.5-15.2,typically 1.5-3.7b

    0.6-4.6,typically 0.9 - 2.4b

    Division of M

    Foraging Height

    (m above ground)

    FOb i e*

    0-30+b

    0-2b

    1.2-12.2b

    0.9-2.4b

    a. nm = non-migratory b. DeGraaf and Rudis, 1986 c. Godin, 1977 d. Veil and Peterson, 1993 e. Based on best estimate Page 3 of 3 NA = not available

    http:0.9-2.4bhttp:1.2-12.2bhttp:1.5-3.7bhttp:1.8-7.9bhttp:0.6-2.8b

  • ChemRisk - A Division of McLaren/Hart July 26,1994 Page 2-4c

    Table 2-3. Exposure Analysis Summary Approximate %

    Activities of Day Potential Exposure during which Associated Timing of

    Indicator Species Routes8 with Exposure Exposure is Possibleb Exposure Small Mammals Northern short-tailed FWT>II>IN/DC Feeding, foraging, burrowing, 100% FWT, II: year-round

    shrew grooming, and parental carec DC, IN: April to November

    Masked shrew FWT>II>IN/DC Feeding, foraging, burrowing, 100% FWT, II: year-round grooming, and parental care DC, IN:

    April to November

    White-footed mouse H>IN/DCFWT Feeding, foraging, burrowing, 50% FWT, II: year-round grooming, and parental care DC, IN: April to November

    Southern red-backed II>IN/DCFWT Feeding, foraging, burrowing, 50% FWT, II: year-round vole grooming, and parental care DC, IN: April to November

    Birds American redstart FWTII Feeding, foraging, and preening 58% FWT, II: mid-May to late

    August or September

    American robin FWT>II>IN/DC Feeding, foraging, nest building, 55% FWT, II, IN, DC: ground gleaning, and preening mid-April to late October

    Barn swallow FWT>II>IN/DC Feeding and nest building 55% FWT: mid-April to late September or early October II, IN, DC: mid-to late April

    Eastern phoebe FWT>II>IN/DC Feeding and nest building 55% FWT: early March to late September II, IN, DC: March

    Red-winged FWT>n>IN/DC Feeding, foraging, ground 55% FWT, II, IN, DC:

    blackbirdd gleaning, and preening mid-to late March to late October

    Rose-breasted FWT>II>IN/DC Feeding, foraging, gleaning, 58% FWT, n, IN, DC: grosbeak and preening mid-to late May to September

    Wood thrush FWT>II>IN/DC Feeding, foraging, ground 58% FWT, II, IN, DC: gleaning, and preening late April-early May to late

    August-September

    Yellow warbler FWT>II Feeding, foraging, and branch 55% FWT, II: late April to gleaning late July or early August

    For all species, primary potential stressor is assumed to be PCBs and medium of interest is floodplain soil, except as noted a. Exposure routes denoted, in order of importance, as foodweb transfer (FWT), dermal contact (DC), incidental ingestion (II), and inhalation (IN) b. Approximate percent of day during which exposure is possible was calculated based on the following assumptions:

    daylight averages approximately 14 hours in summer (58% of the day) daylight averages approximately 13 hours when including spring, summer, and/or early fall (55% of the day) daylight averages approximately 12 hours when including spring, summer, and late fall (50% of the day)

    c. Parental care involves adult organisms feeding, incubating, and cleaning their offspring d. Red-winged blackbird may contact floodplain soil, sediment, and surface water

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-5

    grooming, and sleeping occur in close contact with flood plain soils containing PCBs.

    Additionally, the relatively small home ranges of all indicator species limit the activities of the

    individuals studied to within the 10-year flood plain.

    The temporal scale of the presence of PCBs coincides closely with both the life and reproductive

    cycles of these animals. While PCBs are assumed present in the flood plain soils year-round,

    contact is substantially limited when soils are frozen or covered with snow, as in winter. Exposure

    is also limited during the winter months due to the behavior of the indicator species (e.g, migration

    of songbirds).

    The following exposure profiles describe the potential for exposure of small mammals and

    songbirds (passerines) to PCBs in flood plain soils. Potential routes of exposure to PCBs for

    small mammal and passerines include: (1) food web transfer via ingestion of organisms that

    bioaccumulate PCBs from soil; (2) incidental ingestion of soil particles contaminated with PCBs;

    (3) inhalation of volatilized PCBs (expected to be a very minor pathway); and (4) dermal

    absorption via contact with exposed body surfaces. The activities associated with each pathway,

    the relative importance of each pathway, and the timing of exposures are summarized. Factors

    that affect the rates of uptake associated with each pathway, including duration and frequency of

    exposure, relative area of exposed body surfaces, and relative absorption at point of contact, are

    also discussed.

    2.2.1 Small Mammal Exposure Profile

    Potential exposure of small mammals to PCBs may occur during a variety of daily activities, such

    as feeding, foraging, burrowing, grooming, nest building, and parental care (e.g., nursing,

    feeding, grooming). Factors that affect the magnitude of a small mammal's exposure over a

    lifetime include: food habits (insectivorous vs. herbivorous); winter status (hibernating vs. non-

    hibernating); and duration of daily activity (nocturnal vs. diurnal). Although some mammals

    hibernate in the winter, shrews, voles, and white-footed mice are active all year (Godin, 1977) so

    that exposure frequency may be as high as twelve months per year for some activities. However,

    in western Massachusetts, the ground is normally frozen between November and April, thus

    limiting the potential for exposure during this period.

  • ChemRisk - A Division of McLaren/Hart July 26,1994 Page 2-6

    Because shrews are expected to experience relatively high frequency of contact and exposure point

    concentrations when they feed on insects, it is likely that the food web transfer largely determines

    potential exposures for them. Incidental ingestion is likely the second most important pathway for

    shrews because they groom frequently and are in close contact with soil as they burrow and travel.

    Although shrews may inhale PCBs, exposures resulting from this route are expected to be

    insignificant. Exposure through dermal contact should also be relatively low since their fur should

    substantially impede direct contact with soil and dermal absorption.

    For voles and mice, exposure to PCBs via incidental ingestion may be greater than exposure via

    food web transfer, since a large portion of their diet is vegetation. Uptake of PCBs by plants has

    been shown by several researchers to be negligible (Babish et al., 1979; Chou et al., 1978; Iwata

    and Gunther, 1976; Miyazaki et al., 1975; Fries and Marrow, 1981; O'Connor et al., 1990; Davis

    et al., 1981; Strek et al., 1981; Sawhney and Hankin, 1985). During the winter, when mice and

    voles feed to a greater extent on insect larvae and earthworms, food web transfer may become a

    more important route of exposure. During the rest of the year, ingestion of particulate-bound

    PCBs while feeding on vegetation, foraging in leaf litter, and grooming is likely to be the primary

    route of potential exposure to PCBs. As with shrews, inhalation and dermal contact are likely

    minor contributors to overall exposure.

    2.2.2 Song Bird Exposure Profile

    Song birds have the greatest potential for PCB exposure from flood plain soils during feeding,

    foraging, nest building, roosting, and preening. Factors that determine the magnitude of exposure

    over a lifetime include food habits (herbivorous vs. insectivorous), nesting and feeding locations

    (ground vs. canopy), and exposure duration. Throughout their breeding season in Massachusetts

    (March to October), birds may potentially be exposed to PCBs in flood plain soils during daylight

    hours when they are active. Birds are active from just before dawn to dusk (Brooke and Birkhead,

    1991).

    The potential magnitude of exposure is largely determined by food web transfer for all passerines

    studied in this assessment The primary source of uptake for insectivorous birds are the organisms

    on which they feed that bioaccumulate PCBs. Incidental ingestion, as it occurs during foraging,

    feeding, preening, and nest building, is thought to be the second most significant exposure

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-7

    pathway for these birds. Birds that feed and/or nest on the ground (e.g. American robin, red-

    winged blackbird, rose-breasted grosbeak, and wood thrush) likely have higher exposure levels

    via incidental ingestion than do those that nest and/or feed above ground (e.g. American redstart,

    barn swallow, Eastern Phoebe, and yellow warbler). Birds, such as American robin, barn

    swallow and Eastern phoebe, that use mud to build nests, may also have increased exposure via

    incidental ingestion of sediment

    Although inhalation and dermal contact may occur while birds forage, build nests, or roost on the

    ground, the inhalation pathway is expected to be less significant than food web transfer and

    incidental ingestion for several reasons. For American redstarts and yellow warblers, exposure via

    inhalation is assumed to be negligible because they feed, forage, and nest in the canopy, and

    because concentrations in the ambient air are expected to be very low relative to concentrations in

    soil or in food sources. Exposure via dermal contact should also be low, because adult birds are

    covered with feathers and there is little opportunity for contact of contaminated soil with exposed

    skin. While the feet and legs may contact flood plain soil, they are covered with keratinized scales

    that are relatively impermeable (Welty, 1975).

    2.3 Selection of Endpoints

    The assessment endpoints selected for evaluation in this study included:

    absence of species or taxonomic family or order normally expected to be present in a particular habitat;

    reduction of a population or subpopulation due to the stressor of interest; adverse change in the structure of a community; and bioaccumulation that results in an adverse effect.

    In order to evaluate these assessment endpoints, a number of field investigations were undertaken

    in an effort to evaluate the following measurement endpoints:

    diversity of avian species in the flood plain forest; density of avian populations in the flood plain forest; reproductive success of avian populations in all habitats; small mammal population structure in the flood plain forest and shrub meadow; age structure of small mammal populations in the shrub meadow; and reproductive success of small mammals in the shrub meadow and flood plain forest.

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-8

    2.4 Selection of Study Areas

    In order to evaluate the effects of the exposure scenarios, it was first necessary to identify

    appropriate target areas within the potentially affected area, as well as appropriate points of

    comparison, or reference areas. While the identification of target sites within the contaminated area

    was relatively straightforward, it was considerably more difficult to locate areas outside of the contaminated area that could serve as points of reference for observations made with respect to

    ecosystem health. Potential reference sites were evaluated by studying National Wetlands

    Inventory maps, U.S. Geological Survey topographic maps, and aerial photographs, as well as

    through field reconnaissance via helicopter, automobile, foot, and canoe, and through consultation

    with personnel at the Western District of the Commonwealth of Massachusetts Division of Fish

    and Wildlife.

    Specific study areas were defined for each measurement endpoint except avian reproductive

    success. Nests that were monitored for this endpoint were differentiated by their location within or

    outside of the 10-year flood plain, rather than by localized study area or habitat type. For example,

    target nests were typically found along the river banks, under the New Lenox Road bridge, or in

    the cattail stands in the river. Reference nests were located well outside of the 10-year flood plain,

    including along October Mountain Road, in the reference shrub meadow, and in the Crane Farm

    barn in Dawson. Feeding preferences and territory sizes of individual species monitored were

    evaluated in order to verify that reference birds would not be expected to visit the target area and vice versa.

    In addition, data from the literature (Ehrlich et al., 1988) on normal clutch sizes for each avian

    indicator species were also compared to observations of target nests. The target and reference

    areas used for all measurement endpoints other than avian reproductive success are described

    below by habitat type.

    2.4.1 Flood Plain Forest Target Areas

    The flood plain forest target area is located south of the New Lenox Road bridge in an area of the

    flood plain forest accessible from the Lenox Hunt Club (Figure 2-3). The area is 5.85 hectare in

    size and homogenous in habitat. Its north, south, east, and west coordinates are: 509527.383N;

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-9

    508486.377N; 128973.445E; and 127857.700E, respectively. This target area was selected from

    among several other tracts of flood plain forest for several reasons. This site is the largest intact,

    undisturbed flood plain forest between New Lenox Road bridge and Woods Pond dam, the reach

    of the Housatonic River with both high concentrations of PCBs and suitable wildlife habitat. Other

    forested areas within this reach of the river are smaller, have a greater proportion of edge habitat,

    and are less accessible than the tract selected as a target area. Furthermore, potential exposure of

    wildlife to PCBs at the locality selected as a target area appeared greater than at other candidate

    sites, based on general patterns of river sediment and flood plain soil analytical data available prior

    to the study area selection and on visual evidence of flooding.

    After a great deal of investigation of the area within a 50 mile (or greater) radius of Pittsfield, it was

    determined that suitable reference sites were not available for the flood plain forest. The absence of

    a suitable reference site is in part a reflection of the target area's location within a HVWMA. This

    area differs substantially from land uses of surrounding areas, where residential and agricultural

    land uses predominate. In most locations outside of HVWMA, human development, agricultural

    activity, and industrialization have drastically impacted the land, rendering it unsuitable for

    comparison to the target site. In such areas it would not be possible to distinguish between those

    human influences either unrelated to or related to historic PCB releases. While the Housatonic

    River upstream of Pittsfield, tributaries to the Housatonic River, and other river valleys (including

    the Hoosic and the Konkapot River valleys) were all considered for use as reference sites, no areas

    of comparable size, land use, or habitat type were located close enough to the target site to allow simultaneous studies to be conducted. Several candidate reference locations are noted and reasons

    for their exclusion are summarized in Table 2-4.

    Because an adequate reference flood plain forest could not be located, a comprehensive literature

    search was conducted in an effort to identify a suitable reference study for comparison to the avian

    population study. Because passerines are highly sensitive to subtle variations in habitat as they

    establish breeding territories, the most important criteria for selecting reference studies were habitat

    type (i.e., flood plain forest), habitat quality (e.g., forest maturity, isolation from human influence,

    size), and habitat structure (e.g., flooding regime, development of strata, canopy cover, dominance

    and abundance of plant species). A broad array of sources was searched, including reports of state

    and federal agencies and nongovernmental organizations, Master's and Doctoral dissertations, and

    scientific journals.

  • *05 ITable 2-4. Candidate Reference Areas and Rationale for Exclusion

    Location of Candidate

    Peru State Wildlife Management Area, Peru

    Woodlands west of railroad tracks and treatment site, Lenox

    Middlefield Slate Forest, Middlefield

    Hinsdale Flats State Wildlife Management Area, Hinsdale

    October Mountain State Forest

    Konkapot River Valley, Clayton

    Konkapot River Valley, Hartsville

    Hoosic River Valley, Adams

    American Legion Park, Dalton

    Description

    Woodlands adjacent to wetlands between North Road and Mongue Road; access via FraryRoad

    Limited flood plain forest bordered by railroad tracks and transmission lines

    Wetlands W of Middlefield Road

    E of Rt 8 and S of Middlefield Road; access via trail just W of Fassell Road intersection

    Wetland SW of Ashley Brook and Ashley Lake, Washington, MA

    On CT, MA border

    Immediately S of Lake Buell

    EofMt. Greylock

    Wooded private property E of American Legion Park

    Rationale for Exclusion as Reference Area for Avian or Small Mammal Census

    Habitat is structurally different from target site as a result of infrequent flooding and higher elevation (588 m)a

    Patches of flood plain forest are fragmented and very limited in size; forest very close to (and in some places, within) 10-year flood plain

    Elevation greater than treatment (504 m) with poor access; would require extreme effort to access area; habitat is structurally different from target site

    Structurally different from and higher elevation (442 m) than target site; understory very developed and affected by calcareous geology and soils

    Elevation higher than target site (588 m); abundant spruce; habitat structurally different from target site

    River is much smaller than Housatonic; steep valleys preclude extensive flood plain forest development

    River is much smaller than Housatonic; steep valleys preclude extensive flood plain forest development

    Steep valleys preclude extensive flood plain forest development

    Fragmented and small in size

    a. Elevation of treatment flood plain forest target site is 294 m.

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-10

    Based on the criteria described above, two series of breeding bird censuses were selected as the

    most suitable reference studies. These are breeding bird plot censuses that were conducted since

    1981, as part of the Breeding Bird Census in a flood plain forest in Maryland (Criswell and

    Gauthey, 1983; Van Velzen and Van Velzen, 1985, 1986, 1987, 1988; Gauthey 1984, 1989,

    1990, 1991, 1992) and in a flood plain forest in North Carolina (Crotteau et al., 1978, 1979;

    Christensen et al., 1980,1982; Hall et al., 1984; Van Velzen and Van Velzen, 1985; Hall, 1990;

    Mueller and Hall, 1991; Mueller and Mueller, 1992; Mueller, 1993). The annual Breeding Bird

    Census is a nationwide effort jointly sponsored by the Association of Field Ornithologists and

    Cornell Laboratory of Ornithology. It includes detailed information on the habitats of the

    individual study plots, which facilitates comparisons with outside databases. The Breeding Bird

    Census shares objectives and methodologies virtually identical to those of the avian census

    conducted at the Housatonic flood plain forest. No comparable reference sites closer to the

    Housatonic River could be identified in the scientific literature. Although it would clearly be

    desirable to use reference sites located closer to the target area, these two series of censuses were

    determined to be appropriate reference studies because they are very similar to the target area with

    respect to habitat type, quality, and structure. A detailed comparison of the target area and

    reference area habitats is presented in Appendix B.

    The scientific literature was also searched for reference studies for observations of white-footed

    mice and Southern red-backed vole populations in the flood plain forest. A 27-month study

    conducted by Batzli (1977) in Illinois on white-footed mice and a 3-year study by Miller and Getz

    (1977) in Connecticut on white-footed mice and Southern red-backed voles were chosen as

    reference studies for the small mammal studies. These studies had objectives similar to this

    analysis, adequate interpretations of findings for comparisons of results, and were conducted in

    flood plain forest habitats similar to the Housatonic River Valley flood plain.

    2.4.2 Shrub Meadow Study Areas

    Suitable shrub meadow target and reference areas were both located south of the New Lenox Road

    bridge at sites accessible from the Lenox Hunt Club (Figures 2-2 and 2-3). The shrub meadow

    target area is located within the approximate 10-year flood plain, just north of the flood plain forest

    target area. The reference shrub meadow is located outside of the 10-year flood plain, directly

    northeast of the Lenox Hunt Club and east of the railroad tracks. The north, south, east and west

  • ChemRisk - A Division of McLaren/Hart July 26, 1994 Page 2-11

    coordinates of the target shrub meadow are: 509691.593N; 509481.927N; 129057.165E; and

    128543.098E, respectively. The north, south, east, and west coordinates of the reference shrub meadow are: 511313.026N; 510884.591N; 128170.279E; and 127999.333E, respectively.

    2.4.3 Characterization of PCBs in Flood Plain Soils and River Sediments

    To assess the concentrations of PCBs to which the receptors in the areas under study could be

    exposed, the results of prior sampling in and near those areas were reviewed. This evaluation

    included not only the PCB concentrations detected in flood plain soils in and near the areas under

    study, but also the PCB concentrations in nearby river sediments. PCBs in river sediments affect

    the total exposures of insectivorous small mammals and birds via foodweb transfer, since a portion

    of the diet of these organisms is comprised of insects and insect larvae that mature in river water

    and sediment. Additionally, because


Recommended