United States Office of Research and EPA/620/R-94/026
Environmental Protection Development June 1995
Agency Washington DC 20460
Statistical Summary
EMAP-Estuaries
Virginian Province
1990 to 1993
Environmental Monitoring and
Assessment Program
-------
EPA/620/R-94/026
June 1995
Statistical Summary
EMAP-Estuaries
^ Virginian Province - 1990 to 1993
by
v
V
^ Charles J. Strobel*, Henry W. Buffum*, Sandra J. Benyi*,
Elise A. Petrocelli*, Daniel R. Reifsteckf, and Darryl J. Keith*
* U.S. Environmental Protection Agency
Narragansett, Rl 02882
f Science Applications International Corporation
Narragansett, Rl 02882
* ROW Sciences
Narragansett, Rl
Virginian Province Manager
Darryl J. Keith
EPA Project Officer
Brian Melzian
United States Environmental Protection Agency
National Health and Environmental Effects Research Laboratory
Atlantic Ecology Division
27 Tarzwell Drive
Narragansett, Rl 02882
U.S. Environmental Proteciion Agency @ Printed on Recycled Paper
Region 5.Library (PL-12J)
77 West Jackson Boulevard ISHK c.
Chicago, |L 60604 3590 ' Fl°°r
-------
ABSTRACT
Annual monitoring of indicators of the ecological condition of bays, tidal rivers, and estuaries within the
Virginian Province (Cape Cod, MA to Cape Henry, VA) was conducted by the U.S. EPA's Environmental Monitoring
and Assessment Program (EMAP) during July, August, and September, 1990-1993. Data were collected at 425
probability-based stations within the Province. Indicators monitored included water quality (temperature, salinity,
water clarity, and dissolved oxygen concentration), sediment contamination, sediment toxicity, benthic community
structure, fish community structure, and fish gross external pathology. Data are used to estimate the current status
of the ecological condition of Virginian Province estuarine resources, and provide a baseline for identifying future
trends. Cumulative distribution functions (CDFs) and bar charts are utilized to graphically display data. Estimates,
with 95% confidence intervals, are provided of the areal extent of impacted resources within the Province for those
indicators where "impacted" can be defined. Data are also presented by estuarine class (large estuaries, small
estuarine systems, and large tidal rivers).
KEY WORDS: EMAP; Environmental Monitoring and Assessment Program:, Environmental Monitoring;
Virginian Province; Indicators (biology); Estuaries; Estuarine pollution.
Page ii Statistical Summary, EMAP-E Virginian Province
-------
DISCLAIMER
Mention of trade names, products, or services does not convey, and should not be interpreted as conveying,
official EPA approval, endorsement, or recommendation.
This report represents data from the first four-year cycle of field operations of the Environmental Monitoring
and Assessment Program (EMAP). EMAP-Estuaries utilizes a probability-based scientific design necessitating
multiple years of sampling to produce estimates of the ecological condition of the Nation's estuarine resources.
This document summarizes the data collected collected in the Virginian Province from 1990 to 1993, along with
the uncertainty associated with those data. This uncertainty is expressed in the form of 95% confidence intervals
around reported values. These confidence intervals were generated from a series of equations (included in Appendix
B) based on the distribution of values across the entire Province and do not include a component for measurement
error. These equations are currently under review, and, as such, the reported uncertainties should be considered
tentative. Please note that this report contains data collected in a short index period (July to September). Appropriate
precautions should be exercised when using this information for policy, regulatory or legislative purposes.
Statistical Summary, EMAP-E Virginian Province Page iii
-------
PREFACE
Contractor support for the preparation of this document was supplied via contract number 68-C1-0005
to Science Applications International Corporation and contract number EBC682172 to ROW Sciences.
The appropriate citation for this report is:
Strobel, C.J., H.W. Buffum, S.J. Benyi, E.A. Petrocelli, D.R. Reifsteck, and D.J. Keith. 1995. Statistical Summary:
EMAP-Estuaries Virginian Province - 1990 to 1993. U. S. Environmental Protection Agency, National
Health and Environmental Effects Research Laboratory, Atlantic Ecology Division, Narragansett, RI. EPA/620/R-94/026.
This report is AED Contribution Number 1614.
Page iv Statistical Summary, EMAP-E Virginian Province
-------
ACKNOWLEDGEMENTS
This report describes the results of the first four year sampling cycle completed by EMAP-Estuaries. The
work described in this document is a culmination of the efforts and dedication of dozens of individuals. The authors
would like to take this opportunity to acknowledge the tremendous effort of all personnel involved in the EMAP-Virginian
Province effort.
First and foremost we would like to acknowledge the efforts of Dr. Fred Holland, the first Technical Director
of the Program and Steve Schimmel, the original Province Manager. Their dedication and persistence during the
early stages of the Program served as inspiration for us all and proved invaluable in making the EMAP-Virginian
Province effort succeed in meeting the highest standards set for marine environmental monitoring.
Other key researchers involved in the development of the Program and who are directly responsible for
its success include:
Richard Latimer
John Paul
Jeff Rosen
John Scott
Kevin Summers
Ray Valente
Steve Weisberg
The support staff required to complete an operation as large as EMAP-Virginian Province is extensive.
We would like to acknowledge the effort of all who were involved in supporting day-to-day operations over the
past four years, including Field Coordinators, Province Managers, the ERL-N Laboratory Director, database managers,
librarians, programmers, key administrative and support staff, and those who had to endure photocopying hundreds
upon hundreds of pages of field manuals, statistical summaries, QA Plans, etc. Those deserving special thanks
include:
Matt Aitkenhead Melissa Hughes Agnes Reilly
Jay Beaulieu Rich Jabba Linda Rogers
Jim Blake Norbert Jaworski Liz Rombauer
Marianne Burke Steve Kelly Norman Rubinstein
Dan Campbell Adam Kopacsi Peter Sampou
Sue Cielinski Carol Lee Jill Schoenherr
Jane Copeland Brian Melzian Sybil Seitzinger
Diana Coppinger Andrew Milliken Paul Selvetelli
Peter Doering Candace Oviatt Chris Smith
Craig Eller Charito Paruta Maria Tarantino
Jeff Frithsen Rose Petracca Deb Uva
Steve Hale M.J. Purick
Jim Heltshe
Statistical Summary, EMAP-E Virginian Province Page v
-------
Researchers at EPA (Cincinnati, Narragansett and Gulf Breeze), Cove Corporation, Science Applications
International Corporation, Texas A & M University, Versar, Inc, and a consortium of universities (University of
Rhode Island, Rutgers University, Philadelphia Academy of Science, and University of Maryland) contributed
significantly to this effort through the analysis of samples.
Reviewers of this document included Steve Schimmel, John Scott, John Paul, Candace Oviatt, Tom DeMoss,
John Robbie and Mike Hirshfield.
We would also like to acknowledge the assistance provided by NOAA (Milford, Narragansett, and Woods
Hole) and the University of Rhode Island in the training of field crews.
Most importantly, we would like to acknowledge the tremendous effort of all those involved in the field
effort over the past four years. Despite sea sickness, 16-hour days, equipment failure, uncooperative motel telephone
systems, and inclement weather (including two hurricanes), the six field crews successfully completed the data
collection phase to the high standards set for the Program. Without their dedication to the Program this Statistical
Summary would not be possible. Field personnel have been supplied by Science Applications International Corporation,
Versar Inc., U.S. EPA, and a consortium of universities (University of Rhode Island, Rutgers University, Philadelphia
Academy of Science, and University of Maryland), and are listed below.
Liz Abong
Jason Arruzza
Betsy Balcom
Kendal Banks
Christine Barna
Allison Brindley
Willie Burton
Liz Caporelli
Dan Card
Janis Chaillou
Sue Cielinski
Matt Cohen
Michael Cole
Diana Coppinger
Angie Correia
William Craven
Mike Crowell
Peggy Derrick
Matt DiMatteo
Craig Eller
Martin Friday
Nancy Friday
Noah Fierer
Celia Gelfman
Paul Giard
Jen Greenamoyer
John Gurley
Paula Halupa
Rick Hein
Tom Heitmuller
Meghan Hessenauer
Randy Hochberg
Marie Hodnet
Josh Holden
Charles Holloway
Christie Hurff
Rich Jabba
Fred Kelley
Russ Kelz
Mike Kenny
Kristie Killam
Tom Kirk
Laurie Kristofer
Adam Kopacsi
Damon Lear
Laura Magdeburger
Andy Matthew
Jeff McGroder
Glenn Merritt
Rich Metcalf
Andrew Milliken
John Morris
Rick Morton
Rose Newport
Donna Olejniczak
Todd Parker
Cathy Patterson
Shannon Phifer
Greg Pruitt
Dan Roelke
Mark Rousseau
John Sage
Eric Savetsky
Jill Schoenherr
Steve Serwatka
Brian Sherman
Kurt Sims
John Sirois
Carl Slack
Ward Slacum
Chris Snow
Mark Snyder
Christina Southall
Cathy Stokes
Cynthia Suchman
Maria Tarantino
Cliff Thompson
Tamara Tornatore
Steve Tulevech
Bob Wallace
Sarah Watts
Izzy Williams
Nick Wolff
Bill Yates
Madeline Young
Page vi
Statistical Summary, EMAP-E Virginian Province
-------
CONTENTS
ABSTRACT if
DISCLAIMER Hi
PREFACE iv
ACKNOWLEDGEMENTS v
CONTENTS vii
FIGURES ix
TABLES xiii
ABBREVIATIONS xiv
EXECUTIVE SUMMARY 1
1 INTRODUCTION 11
1.1 Objectives of Virginian Province Monitoring Activities 11
1.2 Program Design 12
1.3 Data Limitations 12
1.4 Purpose and Organization of this Report 12
2 OVERVIEW OF FIELD ACTIVITIES 14
3 STATISTICAL SUMMARY OF INDICATOR RESULTS 19
3.1 Biotic Condition Indicators 20
3.1.1 Benthic Index 20
3.1.2 Number of Benthic Species 22
3.1.3 Benthic Infaunal Abundance 24
3.1.4 Number of Fish Species 26
3.1.5 Total Finfish Abundance 28
3.1.6 Fish Gross External Pathology 28
3.2 Abiotic Condition Indicators 31
3.2.1 Dissolved Oxygen 31
3.2.1.1 Bottom Dissolved Oxygen 32
3.2.1.2 Dissolved Oxygen Stratification 34
3.2.2 Sediment Toxicity 35
3.2.3 Sediment Contaminants 37
3.2.3.1 Polycyclic Aromatic Hydrocarbons 38
3.2.3.2 Polychlorinated Biphenyls 41
3.2.3.3 Chlorinated Pesticides 44
3.2.3.4 Butyltins 44
3.2.3.5 Metals 45
3.2.3.6 Acid Volatile Sulfides 48
3.2.3.7 SEM/AVS Ratio 50
Statistical Summary, EMAP-E Virginian Province Page vii
-------
CONTENTS (continued)
3.2.3.8 Total Organic Carbon 51
3.2.4 Marine Debris 51
3.3 Habitat Indicators 53
3.3.1 Water Depth 53
3.3.2 Temperature 54
3.3.3 Salinity 56
3.3.4 pH 57
3.3.5 Stratification 57
3.3.6 Suspended Solids 60
3.3.7 Light Extinction 62
3.3.8 Silt-Clay Content 62
4 SUMMARY OF FINDINGS 65
4.1 Virginian Province Fact Summary 65
4.2 Findings of the Virginian Province Demonstration Project 65
5 REFERENCES 67
APPENDIX A - SAMPLING DESIGN, ECOLOGICAL INDICATORS, AND METHODS
APPENDIX B - ESTIMATION FORMULAE FOR EMAP SAMPLING IN THE LOUISIANIAN
AND VIRGINIAN PROVINCES
APPENDIX C - LINEAR REGRESSIONS OF INDIVIDUAL METALS AGAINST ALUMINUM USED IN
THE DETERMINATION OF METALS ENRICHMENT OF SEDIMENTS OF THE VIRGINIAN
PROVINCE
Page viii Statistical Summary, EMAP-E Virginian Province
-------
FIGURES
Figure 1. Percent area of the Virginian Province by estuarine class with a benthic index
value below zero 3
Figure 2. Cumulative distribution of fish abundance in numbers per standard trawl as a
percent of area in the Virginian Province 3
Figure 3. The percent of area in the large estuaries, small estuaries, and large tidal rivers
that had a low (< 2 mg/L), medium (2 to 5 mg/L), or high (>5 mg/L) oxygen
concentration in the bottom waters 4
Figure 4. Percent of area in the Virginian Province, by estuarine class, with low (<80%
of control) and very low (<60% of control) amphipod survival in sediment
toxicity tests 4
Figure 5. Cumulative distribution of combined PAHs in sediments as percent of area in
the Virginian Province 5
Figure 6. Percent area of the Virginian Province with enriched concentrations of
individual metals in sediments 6
Figure 7. The percent of area of the Virginian Province by estuarine class where
anthropogenic debris was collected in fish trawls 7
Figure 8. Cumulative distribution of water depth as a percent of area in the Virginian
Province 7
Figure 9. The percent of area of estuarine classes classified as oligohaline (<5 ppt),
mesohaline (5 to 18 ppt), and polyhaline (>18 ppt) 8
Figure 10. The percent of the area by class that had a low (<1 AOt), medium (1 to 2 AOt),
or high (>2 AC,) degree of stratification 9
Figure 11. The percent of area by estuarine class where water clarity was poor, moderate,
or good 9
Figure 12. The percent of area in the large estuaries, small estuaries, and large tidal rivers
that had a low (<20), medium (20 to 80), or high (>80) percent silt-clay in the
sediments 10
Figure 2-1. Areas of responsibility of the EMAP-VP sampling teams 15
Figure 2-2. Team 1 Base Sampling Stations 16
Figure 2-3. Team 2 Base Sampling Stations 17
Figure 2-4. Team 3 Base Sampling Stations 18
Statistical Summary, EMAP-E Virginian Province Page ix
-------
FIGURES (continued)
Figure 3-1. Example cumulative distribution of bottom dissolved oxygen concentrations as
a percent of area in the Virginian Province 19
Figure 3-2. Cumulative distribution of the benthic index as a percent of area in the
Virginian Province 21
Figure 3-3. Percent area of the Virginian Province by estuarine class with a benthic index
value below 0 22
Figure 3-4. Cumulative distribution of the mean number of infaunal benthic species per
grab as a percent of area in the Virginian Province 22
Figure 3-5. Cumulative distribution of the number of infaunal benthic species by estuarine
class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 23
Figure 3-6. Cumulative distribution of the number of infaunal benthic organisms per m2 as
a percent of area in the Virginian Province 24
Figure 3-7. Cumulative distribution of the number of infaunal benthic organisms per m2 by
class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 25
Figure 3-8. Cumulative distribution of the number of fish species per standard trawl as a
percent of area in the Virginian Province 26
Figure 3-9. Cumulative distribution of the number of fish species per trawl by estuarine
class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 27
Figure 3-10. Cumulative distribution of fish abundance in numbers per standard trawl as a
percent of area in the Virginian Province 28
Figure 3-11. Cumulative distribution of fish abundance in numbers per standard trawl by
estuarine class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 29
Figure 3-12. Cumulative distribution of bottom dissolved oxygen concentration as a percent
of area in the Virginian Province 32
Figure 3-13. The percent of area by class that had a low (< 2 mg/L), medium (2 to 5 mg/L),
or high (>5 mg/L) oxygen concentration in the bottom waters 32
Figure 3-14. Cumulative distribution of bottom oxygen concentration by estuarine class: a)
Large estuaries, b) Small estuaries, c) Large tidal rivers 33
Figure 3-15. Cumulative distribution of the D.O. concentration difference between surface
and bottom waters as a percent of area in the Virginian Province 34
Page x Statistical Summary, EMAP-E Virginian Province
-------
FIGURES (continued)
Figure 3-16. The percent of area by estuarine class that had a low (<1 mg/L), medium (1 to
5 mg/L), or high (>5 mg/L) difference in dissolved oxygen concentration
between the surface and bottom waters 35
Figure 3-17. Cumulative distribution of mean survival of amphipods in 10-day laboratory
toxicity tests (expressed as percent of control survival) 36
Figure 3-18. Percent of area in the Virginian Province, by estuarine class, with low
amphipod survival (<80% of control) in sediment toxicity tests 36
Figure 3-19. Cumulative distribution of combined PAHs in sediments as percent of area in
the Virginian Province: a) linear scale, b) log scale 40
Figure 3-20. Cumulative distribution of combined PCBs in sediments as percent of area in
the Virginian Province: a) linear scale, b) log scale 43
Figure 3-21. Cumulative distribution of tributyltin in sediments as percent of area in the
Virginian Province 45
Figure 3-22. Linear regression (with upper 95% confidence intervals) of chromium against
aluminum 47
Figure 3-23. Percent area of the Virginian Province with enriched concentrations of
individual metals in sediments 47
Figure 3-24. The cumulative distribution of the acid volatile sulfide concentration in
sediments as a percent of area in the Virginian Province 48
Figure 3-25. Cumulative distribution of the acid volatile sulfide concentration in sediments
by estuarine class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 49
Figure 3-26. Cumulative distribution of the SEM/AVS ratio in sediments as a percent of
area in the Virginian Province, 1993 only 50
Figure 3-27. The cumulative distribution of the percent total organic carbon in sediments as
a percent of area in the Virginian Province 51
Figure 3-28. Cumulative distribution of the percent total organic carbon in sediments by
estuarine class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers 52
Figure 3-29. The percent of area of the Virginian Province by estuarine class where
anthropogenic debris was collected in fish trawls 53
Figure 3-30. Cumulative distribution of water depth as a percent of area in the Virginian
Province 54
Statistical Summary, EMAP-E Virginian Province Page xi
-------
FIGURES (continued)
Figure 3-31. Cumulative distribution of bottom temperature as a percent of area in the
Virginian Province 54
Figure 3-32. Cumulative distribution of bottom temperature by estuarine class: a) Large
estuaries, b) Small estuaries c) Large tidal rivers 55
Figure 3-33. The cumulative distribution of bottom salinity as a percent of area in the
Virginian Province 56
Figure 3-34. The percent of area by estuarine class classified as oligohaline (<5 ppt),
mesohaline (5 to 18 ppt), and polyhaline (>18 ppt) 57
Figure 3-35. Cumulative distribution of bottom salinity by estuarine class: a) Large
estuaries, b) Small estuaries c) Large tidal rivers 58
Figure 3-36. Cumulative distribution of the stratified area in the Virginian Province based
on the sigma-t (as) difference between surface and bottom waters 59
Figure 3-37. The percent of the area by estuarine class that had a low (<1), medium (1 to 2),
or high (>2) degree of stratification ( A at as kg/m3) 60
Figure 3-38. The cumulative distribution of total suspended solids concentration as a percent
of area in the Virginian Province, 1991 - 1993 60
Figure 3-39. Cumulative distribution of total suspended solids concentration by estuarine
class (1991 - 1993 only): a) Large estuaries, b) Small estuaries, c) Large tidal
rivers 61
Figure 3-40. The cumulative distribution of light extinction coefficient as a percent of area
in the Virginian Province 62
Figure 3-41. The percent of area by estuarine class where water clarity was poor, moderate,
or good 63
Figure 3-42. The cumulative distribution of the percentage of silt-clay in the sediments as a
percent of area in the Virginian Province 64
Figure 3-43. The percent of area by estuarine class with a low (<20), medium (20 to 80), or
high (>80) percent silt/clay in the sediments 64
Page xii Statistical Summary, EMAP-E Virginian Province
-------
Table 3-1.
Table 3-2.
Table 3-3.
Table 3-4.
Table 3-5.
Table 3-6.
Table 3-7.
Table 3-8.
Table 4-1.
TABLES
1991 target species examined for external pathology and saved for chemical
residue analysis
Incidence of gross external pathology among fish caught in the Virginian
Province
Draft Sediment Quality Criteria values for acenaphthene, phenanthrene,
fluoranthene, and dieldrin
ER-M and ER-L guideline values for metals and organic contaminants
Range and median PAH concentrations in sediments of the Virginian Province
Range and median PCB concentrations in sediments of the Virginian Province
Range and median butyltin concentrations in sediments of the Virginian
Province
Range and median metal concentrations in sediments of the Virginian Province.
Percent area of the Virginian Province (with 95% confidence intervals) above
or below values of interest for selected indicators
30
30
37
38
39
42
44
46
66
Statistical Summary, EMAP-E Virginian Province
Page xiii
-------
ABBREVIATIONS
AED Atlantic Ecology Division, NHEERL (formerly ERL-N)
AVS Acid Volatile Sulfide
BSS Base Sampling Site
CDF Cumulative Distribution Function
DBT Dibutyltin
DO Dissolved Oxygen
dry wt Dry weight
EMAP Environmental Monitoring and Assessment Program
EMAP-E EMAP-Estuaries
ERL-N Environmental Research Laboratory, Narragansett
MET Monobutyltin
mg/L milligrams per liter = parts per million (ppm)
mg/kg milligrams per kilogram = parts per million (ppm)
kg/m3 kilograms per cubic meter
ND Not Detected
ng/g nanograms per gram = parts per billion (ppb)
NHEERL National Health and Environmental Effects Research Laboratory (U.S. EPA)
PAH Polycyclic Aromatic Hydrocarbon
PCB Polychlorinated Biphenyl
QA Quality Assurance
QC Quality Control
SEM Simultaneously Extracted Metals
SQC Sediment Quality Criteria
TBT Tributyltin
ug/g micrograms per gram = parts per million (ppm)
fj Micron
A Delta
CT, Sigma-t
%o parts per thousand (ppt)
Page xiv
Statistical Summary, EMAP-E Virginian Province
-------
EXECUTIVE SUMMARY
The EMAP Virginian Province
includes the coastal region of
the Northeast United States
from Cape Cod south to the
mouth of Chesapeake Bay. It
is composed of 23,574 km2 of
estuarine resources including
11,469 km2 in Chesapeake Bay
and 3,344 krrf in Long Island
Sound.
A total of 446 Base Sampling
Sites were identified for
sampling over the four-year
period (1990-1993).
The Environmental Monitoring and Assessment Program (EMAP) is a nationwide
program initiated by EPA's Office of Research and Development (ORD). EMAP
was developed in response to the demand for information about the degree to
which existing pollution control programs and policies protect the nation's ecological
resources.
EMAP-Estuaries (EMAP-E) represents EMAP's efforts in near-coastal environments.
These efforts are designed to provide a quantitative assessment of the regional
extent of coastal environmental problems by measuring status and change in selected
indicators of ecological condition. Specific environmental problems investigated
include:
hypoxia,
sediment contamination,
coastal eutrophication, and
habitat loss.
In 1990 EMAP-E initiated a four-year demonstration project in the estuaries
of the Virginian Province, which includes the coastal region of the Northeast United
States from Cape Cod south to the mouth of Chesapeake Bay. It is composed
of 23,574 km2 of estuarine resources including 11,469 km2in Chesapeake Bay
and 3,344 km2 in Long Island Sound.
Estuarine resources in the Virginian Province were stratified into classes by
physical dimension for the purposes of sampling and analysis. Large estuaries
in the Virginian Province were defined as those estuaries greater than 260 km2
in surface area and with aspect ratios (i.e., length/average width) of less than 18.
The areal extent of large estuaries in the Province is 16,097 km2. Large tidal
rivers were defined as that portion of the river that is tidally influenced (i.e., detectable
tide > 2.5 cm), greater than 260 km2 in surface area, and with an aspect ratio of
greater than 18. Approximately 2,602 km2 were classified as large tidal rivers.
The third class was the small estuaries and small tidal rivers which includes those
systems whose surface areas fall between 2.6 km2 and 260 km2. This class represents
4,875 km2 of the Virginian Province.
A total of 446 Base Sampling Sites (BSS: the probability-based sites used
to characterize conditions in the Province) were identified for sampling over the
four-year period (1990-1993). Of these 446 sites, 21 were deemed inaccessible
due to inadequate water depth or other logistical constraints. The 425 sites sampled
represent 94% of the estuarine surface area of the Province. All sites were sampled
by three crews during the summer index period (late July through September).
Statistical Summary, EMAP-E Virginian Province
Page 1
-------
The purpose of this report is to
provide estimates of the
ecological condition of the
estuarine resources of the
Virginian Province for the first
complete four-year cycle of
sampling
Biotic condition indicators
are characteristics of the
environment that provide
quantitative evidence of the
status of ecological resources
and biological integrity of a
sample site from which they are
collected.
The incidence of gross external
pathologies (growths, lumps,
ulcers, and fin erosion) among
fish collected in the Virginian
Province was 0.3%
The purpose of this report is to provide regional managers and administrators
with estimates of the ecological condition of the estuarine resources of the Virginian
Province for the first complete four-year cycle of sampling. A separate Assessment
Report (Paul et al., in preparation) is currently being produced to evaluate associations
between indicators as well as to evaluate the overall design of the Program. In
addition, interim reports and Statistical Summaries have been produced describing
the results of the 1990, 1991, and 1992 sampling efforts (Weisberg et al., 1993;
Schimmel et al., 1994; Strobel et al., 1994).
All EMAP-VP data used in the generation of this report were subjected to
rigorous quality assurance measures as described in the 1993 Quality Assurance
Project Plan (Valente and Strobel, 1993).
Biotic Condition Indicators
Biotic condition indicators are characteristics of the environment that provide
quantitative evidence of the status of ecological resources and biological integrity
of a sample site from which they are collected (Messer, 1990). Ecosystems with
a high degree of biotic integrity (i.e., healthy ecosystems) are composed of balanced
populations of indigenous benthic and water column organisms with species compositions,
diversity, and functional organization comparable to undisturbed habitats (Karr
and Dudley, 1981; Karr et al., 1986).
A benthic index which uses measures of organism and community condition
to evaluate the condition of the benthic assemblage was utilized in the assessment
of biological resources of the Virginian Province. The index under development
was constructed from the combined 1990 -1993 data and was developed to represent
a combination of ecological measurements that best discriminates between good
and poor ecological conditions. This index represents EMAP-E's attempt to reduce
many individual benthic indicators into a single value that has a high level of
discriminatory power between good and poor environmental conditions.
A benthic index critical value of zero was determined from the combined 1990 -
1993 Virginian Province dataset. Twenty three (± 3) percent of the bottom area
of the Virginian Province had an index value of < 0, indicating likely impacts
on the benthic community (Figure 1). The lowest incidence was found in the
large estuaries (18 ± 4%).
A "standard" fish trawl (trawling at a specified speed for a specified time)
was performed at each station to collect information on the distribution and abundance
of fish. Because many factors influence fish abundance, poor catch may not be
an indication of degraded conditions, but simply the natural habitat. Catches of
<10 fish/trawl (catch per unit effort) occurred at stations representing approximately
36 ± 3% of the Province (Figure 2).
The incidence of gross external pathologies (growths, lumps, ulcers, and fin
erosion) among fish collected in the Virginian Province was 0.3%. Of the over
16,000 fish examined, 55 were identified as having one or more of these conditions.
Page 2
Statistical Summary, EMAP-E Virginian Province
-------
Twenty three (± 3) percent of
the bottom area of the Virginian
Province had an index value of
< 0, indicating likely impacts on
the benthic community.
Abiotic condition indicators
quantify the levels of stresses
to which organisms are
exposed.
50 -i
40-
30
0)
2 20
o
rx
10-
0-
Province Large Small Tidal
Figure 1. Percent area of the Virginian Province by estuarine class
with a benthic index value below 0. (Error bars represent 95%
confidence intervals).
200 400 600 800 1000
Number of Fish per Trawl
1200
Figure 2. Cumulative distribution of fish abundance in numbers per
standard trawl as a percent of area in the Virginian Province. (Dashed
lines are the 95% confidence intervals).
Abiotic Condition Indicators
Abiotic condition indicators historically have been the mainstay of environmental
monitoring programs, because these indicators quantify the levels of stresses to
which organisms are exposed.
One potential stress to aquatic organisms is a low concentration of dissolved
oxygen (DO). Two and 5 mg/L are values employed by EMAP to define severe
and moderate hypoxia, respectively. Approximately 25 ± 3% of the sampled area
of the Province contains bottom waters with DO concentrations less than or equal
Statistical Summary, EMAP-E Virginian Province
Page 3
-------
Approximately 25 ± 3% of the
sampled area of the Province
contains bottom waters with DO
concentrations less than or
equal to 5 mg/L.
Approximately 5 ± 2% of the
sampled area exhibited bottom
DO conditions <2.0 mg/L.
Approximately 10 ± 2% of the
sampled area of the Virginian
Province contained sediments
which were toxic to the
amphipod Ampelisca abdita
during 10-day exposures.
90 i
<2
2 to 5
>5
Province Large Small Tidal
Figure 3. The percent of area by class that had a low (< 2 mg/L),
medium (2 to 5 mg/L), or high (>5 mg/L) oxygen concentration in the
bottom waters. (Error bars represent 95% confidence intervals).
to 5 mg/L (Figure 3). "Bottom" is defined as one meter above the sediment-water
interface. Approximately 5 ± 2% of the sampled area exhibited bottom DO conditions
<2.0 mg/L. Dissolved oxygen conditions <2.0 mg/1 were evident in all three classes
of estuaries (Figure 3).
In addition to measuring contaminants in sediments, acute toxicity tests were
performed on sediments collected at each site to determine if they were toxic to
the tube-dwelling amphipod, Ampelisca abdita. Sediments were classified as toxic
if amphipod survival in the test sediment was less than 80% of that in the control
sediment and statistically different from control survival. Approximately 10 ±
2% of the sampled area of the Virginian Province contained sediments which were
toxic to the amphipod during 10-day exposures (Figure 4). Sediments were highly
toxic (i.e., survival < 60% of control) in 2 ± 1% of the area of the Province.
Province Large Small Tidal
Figure 4. Percent of area in the Virginian Province, by estuarine
class, with low (<80% of control) or very low (<60% of control)
amphipod survival in sediment toxicity tests. (Error bars represent
95% confidence intervals).
Page 4
Statistical Summary, EMAP-E Virginian Province
-------
75% of the area of the Province
contained sediments with
concentrations of PAHs <1,200
ng/g, with a maximum
measured concentration at any
station of 80,100 ng/g.
Draft Sediment Quality Criteria
for PAHs were exceeded at
only one station within the
Province.
Stations representing only 1 ±
1% of the area of the Province
exceeded any Long and
Morgan ER-M value for PAHs.
100
80
as
£ 60
20-
0
20 40 60 80
Combined PAHs (ng/g dry wt x 1,000)
100
Figure 5. Cumulative distribution of combined PAHs in sediments as a
percent of area in the Virginian Province. (Dashed lines are the 95%
confidence intervals).
Sediments collected at each station were analyzed for both organic contaminants
and metals. Because of the complex nature of sediment geochemistry, the ecological
impact of elevated contaminant levels is not well understood. Although no definitive
statement can be made estimating the overall aerial extent of sediment contamination
in the Virginian Province, the results of several different approaches are presented
and discussed.
Figure 5 shows the distribution of the sum of measured polycyclic aromatic
hydrocarbons (PAHs) in the Virginian Province. The complete list of analytes
included in this summation can be found in Section 3. The 75th percentile for total
PAHs was approximately 1,200 ng/g (i.e., 75% of the area of the Province contained
sediments with concentrations of PAHs <1,200 ng/g), with a maximum measured
concentration at any station of 80,100 ng/g.
Draft EPA Sediment Quality Criteria (SQC) are currently available for the
PAHs acenaphthene, phenanthrene, and fluoranthene; and the pesticide dieldrin.
Draft PAH SQC were exceeded at only one small estuary station within the Province
(ca. 0.07% of the area).
Stations representing only 1 ± 1% of the area of the Province exceeded any
ER-M (Effects Range-Median from Long et a/., 1995) value for PAHs.
The extent to which polluting activities have affected concentrations of metals
in sediments is complicated by the natural variation of metals in sediments. Crustal
aluminum concentrations are generally many orders of magnitude higher than
anthropogenic inputs; therefore, aluminum can be used to "normalize" for differing
crustal abundances of trace metals. The process utilized was inefficient for several
metals (i.e., r2 for the regression < 0.4), but performed well for As, Cr, Fe, Hg,
Mn, Ni, Sb, and Zn. The percent area of the Virginian Province with sediments
enriched by metals pertains only to the metals mentioned above. Figure 6 presents
the results of this normalization. The metal exhibiting the greatest extent of enrichment
is manganese. Approximately 46 ± 5% of the area of the Province showed enrichment
of sediments with at least one metal. Thirty seven (± 7), 64 ± 3, and 69 ± 16
Statistical Summary, EMAP-E Virginian Province
Page 5
-------
Approximately 46 ± 5% of the
area of the Province showed
enrichment of sediments with at
least one metal. Note that
enrichment does not imply
potential ecological effects.
Stations representing only 4 ±
2% of the area of the Province
exceeded any Long and
Morgan ER-M value for metals.
40-
30-
I
5
20-
10-
As Cr Fe Hg Mn Ni Sb Zn
Figure 6. Percent area of the Virginian Province with enriched
concentrations of individual metals in sediments. (Error bars represent
95% confidence intervals).
percent of the large estuary, small estuary, and large tidal river class areas sampled
contained sediments with metals concentrations exceeding predicted background
levels. This only shows the percent of the Province with elevated concentrations
of metals, and does not indicate the magnitude of enrichment; therefore, this does
not imply concentrations are elevated to the point where biological effects might
be expected. As shown below, sediment from only a fraction of this area contains
concentrations of metals high enough to result in ecological effects.
Stations representing only 4 ± 2% of the area of the Province exceeded any
ER-M (Effects Range-Median from Long et al., 1995) value for metals. It should
be noted that earlier EMAP-E documents utilized the Long and Morgan (1990)
values. ER-M and ER-L values have subsequently been updated (Long et al.,
1995) and it is these newer values that are used in this report. The major difference
is an increase in the ER-M values for metals, resulting in a significant reduction
in the percent area of the Province in exceedence.
Presence of marine debris in fish trawls was documented by field crews as
being encountered at stations representing 20 ± 3% of the Virginian Province
area (Figure 7). The small estuary class had the largest percent area (35 ± 9%)
where trash was found.
Page 6
Statistical Summary, EMAP-E Virginian Province
-------
Presence of marine debris in
fish trawls was documented by
field crews as being
encountered at stations
representing 20 ± 3% of the
Virginian Province area.
Habitat indicators describe the
natural physical and chemical
conditions of the sites sampled.
CO
Province Large Small Tidal
Figure 7. The percent of area of the Virginian Province by estuarine
class where anthropogenic trash was collected in fish trawls. (Error
bars represent 95% confidence intervals).
Habitat Characterization
Habitat indicators describe the natural physical and chemical conditions of
the sites sampled. These parameters are important modifying factors controlling
both abiotic and biotic condition indicators.
Figure 8 shows the distribution of water depth in the Virginian Province.
The area shallower than 2 m is underestimated because this was the minimum
depth sampled.
Based on the sampling design where a single station represents a statistical
area (e.g., 70 km2 for large estuary sites), 6% of the area of the Province could
not be sampled due to inadequate water depth or inaccessibility.
20 30
Water Depth (meters)
40
50
Figure 8. Cumulative distribution of water depth as a percent of area
in the Virginian Province. (Dashed lines are the 95% confidence
intervals).
Statistical Summary, EMAP-E Virginian Province
Page?
-------
90
80
70
! 60
<
"o 50
<5
40
30
20
10-
0
<5
5 to 18
Province Large Small Tidal
Figure 9. The percent of area by estuarine class classified as
oligohaline (<5%o), mesohaline (5 to 18%o), or polyhaline (>18%<>).
(Error bars represent 95% confidence intervals).
Bottom water temperatures in the Virginian Province ranged from 12°C to
30°C during the summer sampling season.
Figure 9 illustrates the distribution of oligohaline (<5%o salinity), mesohaline
(5-18%o), and polyhaline (>18%c) water in the Virginian Province and by class.
Vertical density differences (a function of both salinity and temperature) in
the waters of the Virginian Province can be large enough to result in a reduction
in mixing between surface and bottom waters, potentially allowing the bottom
waters to become hypoxic. Degree of stratification in the Virginian Province
was measured as the delta (A) ot, which is the difference in a, (sigma-t, a density
measurement) between surface and bottom waters. Approximately 72 ±3% of
the Province area had a Ac, of <1 unit; thus the majority of the water in the Virginian
Province was well-mixed (Figure 10). Only 13 ± 3% of the Province area was
strongly stratified (Aat >2).
Water clarity was determined from light extinction coefficients, which describe
the attenuation of light as it passes vertically through the water column. We are
defining low water quality as water in which a diver would not be able to see
his/her hand when held at arms length (i.e., only 10% of incident sunlight reaches
a depth of one meter; light attenuation coefficient > 2.303). Moderate water clarity,
in terms of human vision, is defined as water in which a wader would not be able
to see his/her feet in waist-deep water (i.e., only 25% of incident sunlight reaches
a depth of one meter; light attenuation coefficient > 1.387). Water clarity was
good in 81 ± 3% of the area of the Virginian Province (Figure 11). Water of
low clarity was found in 6 ± 2% of the Province and an additional 13 ± 2% had
water of moderate clarity.
The silt-clay (mud) content of sediments (the fraction <63|u particle diameter)
is an important factor determining the composition of the biological community
at a site, and is therefore important in the assessment of the benthic community.
The distribution of mud (>80% silt-clay) vs sand (<20% silt-clay) is illustrated
in Figure 12.
Page 8
Statistical Summary, EMAP-E Virginian Province
-------
The majority of the water in the
Virginian Province was well-
mixed. Only 13 ± 3% of the
Province area was strongly
stratified.
Water clarity was good in 81 ±
3% of the area of the Virginian
Province Water of low clarity
was found in 6 ± 2% of the
Province and an additional 13 ±
2% had water of moderate
clarity.
100-1
80
(0
CD
< 60-
c
£> 40
CD
Q.
20-
1 to 2
>2
Province
Large
Small
Tidal
Figure 10. The percent of the area by estuarine class that had a low (<1),
medium (1 to 2), or high (>2) degree of stratification (A o, as kg/m3). (Error
bars represent 95% confidence intervals).
100-1
80-
E 60-
5
D
5 40-
20-
ifa
Low
Moderate
Good
Province Large Small Tidal
Figure 11. The percent of area by estuarine class where water clarity
was poor, moderate, or good. (Error bars represent 95% confidence
intervals).
Statistical Summary, EMAP-E Virginian Province
Page 9
-------
70-,
Province Large Small Tidal
Figure 12. The percent of area by estuarine class with a low (<20),
medium (20 to 80), or high (>80) percent silt/clay in the sediments.
(Error bars represent 95% confidence intervals).
Page 10
Statistical Summary, EMAP-E Virginian Province
-------
SECTION 1
INTRODUCTION
The Environmental Monitoring and Assessment
Program (EMAP) is a nationwide program initiated by
EPA's Office of Research and Development (ORD).
EMAP was developed in response to the need to imple-
ment a monitoring program that contributes to com-
parative ecological risk assessment and decisions related
to environmental protection and management. EMAP
is an integrated federal program; ORD is coordinating
the planning and implementation of EMAP with other
federal agencies including the Agricultural Research
Service (ARS), the Bureau of Land Management (BLM),
the U.S. Fish and Wildlife Service (FWS), the Forest
Service (FS), the U.S. Geological Survey (USGS), and
the National Oceanic and Atmospheric Administration
(NOAA). These other agencies and offices participate
in the collection and analysis of EMAP data and will
use these data to guide their policy decisions as
appropriate.
EMAP-Estuaries (EMAP-E) represents one portion
of EMAP's efforts in near-coastal environments. These
efforts are designed to provide a quantitative assessment
of the regional extent of coastal environmental problems
by measuring status and change in selected ecological
condition indicators to address specific environmental
problems including:
hypoxia,
sediment contamination,
coastal eutrophication, and
habitat loss.
EMAP-E initiated a four-year Demonstration Project
in 1990 in the estuaries of the Virginian Province (i.e.,
estuaries, bays and sounds between Cape Cod, MA and
Cape Henry, VA: Holland, 1990; Weisberg et al.,
1993). One of the objectives of the Demonstration Project
was to test the EMAP design, logistical approach and
various ecological condition indicators.
1.1 Objectives of Virginian Province Moni-
toring Activities
The objectives of the EMAP-Estuaries monitoring
program, as described in the EMAP-Estuaries Program
Plan (Holland 1990), are as follows:
Provide a quantitative assessment of the regional
extent of coastal environmental problems by measuring
pollution exposure and ecological condition,
Measure changes in the regional extent of environmental
problems for the nation's estuarine and coastal
ecosystems,
Identify and evaluate associations between the
ecological condition of the nation's estuarine and
coastal ecosystems and pollutant exposure, as well
as other factors known to affect ecological condition
(e.g., climatic conditions, land use patterns), and
Assess the effectiveness of pollution control actions
and environmental policies on a regional scale (i.e.,
large estuaries like Chesapeake Bay) and nationally.
Additional objectives of the Virginian Province monitoring
program were to:
obtain data on Virginian Province-specific variability
in ecological indicators; and
Statistical Summary, EMAP-E Virginian Province
Page 11
-------
develop and refine assessment procedures for
determining the ecological status of estuaries and
apply these procedures to establish baseline
conditions in the Virginian Province.
Data collected in the first EMAP cycle (1990 - 1993)
will be used to establish baseline conditions in the
Virginian Province for future trends analyses.
As part of establishing baseline conditions in the
Virginian Province, several assessment questions relating
to ecological conditions were addressed. Among these
questions are:
What proportion of the bottom waters of the
estuaries of the Virginian Province experience
hypoxia (e.g., dissolved oxygen concentrations
< 2 or 5 mg/L)?
What proportion of the estuarine sediments of the
Virginian Province have a benthic community
structure indicative of polluted environments?
What is the incidence of gross external pathologies
among fish species in the Virginian Province?
What proportion of estuarine sediments in the
Virginian Province contain elevated levels of
anthropogenic chemical contaminants?
What proportion of estuarine sediments in the
Virginian Province contain anthropogenic marine
debris?
A simple classification scheme based on the physical
dimensions of an estuary was used to develop three classes
of estuaries -- large estuaries, large tidal rivers, and small
estuaries/small tidal rivers. Large estuaries in the Virginian
Province were defined as those estuaries greater than
260 km2 in surface area and with aspect ratios (i.e.,
length/average width) of less than 18. Large tidal rivers
were defined as that portion of the river that is tidally
influenced (i.e., detectable tide > 2.5 cm), greater than
260 km2 in surface area, and with an aspect ratio of greater
than 18. Small estuaries and small tidal rivers were designated
as those systems whose surface areas fell between 2.6
km2 and 260 km2. These criteria resulted in the identification
of 12 large estuaries; 5 large tidal rivers; and 144 small
estuaries/small tidal rivers.
1.3 Data Limitations
EMAP is designed to provide data on a regional scale.
This design creates a limitation for those interested in
smaller scale studies. For example, each of the 144 small
systems (e.g., Raritan Bay or the Elizabeth River) is
represented by a single station, the location of which
is randomly selected. The assumption is made that this
station is representative of an area of the Province equal
to the area of that system. In total, these stations are
expected to provide an accurate portrayal of conditions
in small systems across the Province; however, the design,
at its current scale, does not allow for the study of conditions
in individual small systems. The reader should consult
Appendix A and the Near Coastal Program Plan (Holland,
1990) for additional information on the statistical design.
1.2 Program Design
Sample collection in the Virginian Province focused
on ecological indicators (described in Holland, 1990 and
Appendix A) during the index sampling period, the
period when many estuarine responses to anthropogenic
and natural stresses are anticipated to be most severe.
This index period was identified in 1990 based on
dissolved oxygen conditions in the Province to be the
end of July through the end of September. The
sampling design combines the strengths of systematic
and random sampling with an understanding of estuarine
ecosystems in order to provide a probability-based
estimate of estuarine status in the Virginian Province.
1.4 Purpose and Organization of This Report
This Statistical Summary is meant to provide large
quantities of information without including extensive
interpretation of these data. The purpose of this report
is to provide estimates of the ecological condition of
the estuarine resources of the Virginian Province based
solely on selected individual indicators. These estimates
are based on samples collected during a four-year sampling
period, 1990-1993. A separate interpretative Assessment
Report is currently being produced to evaluate associations
between indicators as well as to evaluate the overall design
of the Program (Paul et al, in preparation). This report
will provide more detail on the condition of ecological
resources (in addition to Province-wide estimates, it will
Page 12
Statistical Summary, EMAP-E Virginian Province
-------
present results for the major estuarine watersheds in the
Province), evaluate associations between ecological
condition and other indicators, and attempt an evaluation
of the effectiveness of the program in meeting its
objectives. In addition, interim reports and Statistical
Summaries have been produced describing the results
of the 1990, 1991, and 1992 sampling efforts (Weisberg
et al., 1993; Schimmel et al, 1994; Strobel et al., 1994).
This report is organized into sections addressing the
objectives and results of the Virginian Province
monitoring program. Section 1 describes the objectives
of the Program and limitations on the use of the data
presented in this report.
Section 2 provides a brief overview of the sampling
effort, including providing maps showing all sampling
locations.
Section 3 is the statistical summary of the data
collected during the survey. Also included in this
section is pertinent Quality Assurance/Quality Control
information.
Section 4 summarizes the findings of the monitoring
program in the Virginian Province.
Section 5 lists the references cited in this report.
Appendix A provides an overview of the sampling
design used for base-level monitoring, as well as details
concerning special studies conducted to assess spatial
variability. This appendix also describes the selected
indicators used in this study.
Appendix B presents the equations utilized in the
construction of Cumulative Distribution Functions
(CDFs) and the associated confidence intervals.
Appendix C presents the plots of the regressions of
individual metals concentrations in sediments against
aluminum concentrations used in the determination of
areal extent of metals enrichment.
Statistical Summary, EMAP-E Virginian Province Page 13
-------
SECTION 2
OVERVIEW OF FIELD ACTIVITIES
The Virginian Province includes the coastal region
of the northeast United States from Cape Cod south to
the mouth of Chesapeake Bay. It is composed of
23,574 km2 of estuarine resources including 11,469 km2
in Chesapeake Bay and 3,344 km2in Long Island Sound.
The Virginian Province survey was conducted
annually during late July through early September from
1990 to 1993. A probability-based sampling design was
used to sample major estuarine resources proportionately
(Overton etal, 1991; Stevens etal, 1991). This design
makes it possible to estimate the proportion or amount
of area in the Virginian Province having defined
environmental conditions.
Four hundred and forty six stations in the Virginian
Province, located between Cape Cod (MA) and Cape
Henry (VA), were scheduled for sampling over four
years.
Sample collection in the Virginian Province focused
on ecological indicators during the index sampling
period, when responses of estuarine resources to
anthropogenic and natural stresses are anticipated to be
most severe (e.g., high temperatures, low dissolved
oxygen). The basic sampling design provides a
probability-based estimate of estuarine status in the
Virginian Province. Additional sites were also sampled
to collect information for specific hypothesis testing and
other specific study objectives (Schimmel, 1990; Strobel
et at., 1992). A more detailed discussion of the
indicators and sampling methods can be found in
Appendix A.
Base Sampling Sites (BSS) are the probability-based
sites which form the core of the EMAP-E monitoring
design for all provinces, including the Virginian
Province. Data collected from these sites are the basis
of this statistical summary. Four hundred and twenty
five BSS were sampled during the index period over
four years. A total of 446 were scheduled for sampling;
however 21 were deemed unsampleable due to inadequate
water depth or inaccessibility.
Stations sampled each year were divided among three
sampling teams, each covering a specific area of
responsibility (Figure 2-1). Each team was comprised
of two, four-person alternating crews which sampled
for five or six consecutive days (depending on the year).
During this period, the crew was assigned responsibility
for sampling a cluster of stations. The order in which
clusters were to be sampled was randomized to assure
stations were not uniformly sampled across the Province
in a North-South series. Each Base Sampling site was
visited once during the index period. Figures 2-2, 2-3,
and 2-4 present maps of all the Base Sampling Sites
scheduled for sampling in the Virginian Province
monitoring program.
Each year prior to sampling all crew members were
required to attend an intensive 4-6 week training course
covering all aspects of sampling. Crews were also
required to participate in one week of "dry runs" prior
to sampling.
The Program was successful in its attempt to collect
large amounts of information and samples over a
relatively short time period. The overall effectiveness
of the sampling plan is reflected in the high percentage
of stations for which usable data were obtained for the
variety of parameters measured. As stated above, 21
stations could not be sampled. These stations represent
only six percent of the area of the Province. Although
the remaining stations (425) were successfully sampled,
not all stations were sampled for all parameters. The
number of stations with valid data are included in the
discussion of each indicator in Section 3.
Page 14
Statistical Summary, EMAP-E Virginian Province
-------
Narragansett,
SAMPLING TE
Edistonf NJ*
SAMPLING TEAM 2
SAMPLING TEAM 3
Figure 2-1. Areas oi Responsibility of the EMAP-VP Sampling Teams.
Statistical Summary, EMAP-E Virginian Province
Page 15
-------
in
c
g
«
CO
O)
c
CO
co
CD
E
CO
to
(V
3
O)
Page 16
Statistical Summary, EMAP-E Virginian Province
-------
1990 Stations
A 1991 Stations
1992 Stations
1993 Stations
Figure 2-3. Team 2 Base Sampling Stations.
Statistical Summary, EMAP-E Virginian Province
Page 17
-------
Stations
1991 Stations
1992 Stations
1993 Stations
Figure 2-4. Team 3 Base Sampling Stations.
Page 18
Statistical Summary, EMAP-E Virginian Province
-------
SECTION 3
STATISTICAL SUMMARY OF INDICATOR RESULTS
The EMAP indicator strategy includes four types
of ecological indicators: Biotic Condition, Abiotic
Condition, Habitat, and Stressor. In this section the
statistical results of the Virginian Province Survey are
described for each indicator with discussions categorized
by major indicator type. Stressor data are not collected
as part of the field effort; therefore, they are not
discussed in this report. The following discussion is
organized by indicator type. Indicators will be briefly
described, and in most cases the Cumulative Distribution
Function (CDF) will be shown to present the frequency
of occurrence of observations within the Province. Bar
graphs are also presented, where appropriate, to
delineate the proportions of the Province or estuarine
class resources that are impacted, or falling above or
below values of interest. Methods are only briefly
discussed for individual indicators. A more thorough
discussion can be found in Appendix A.
The term "impacted" is used throughout this
document. EMAP is using this terminology when
scientific data are available to distinguish between good
and poor ecological conditions. As an example many
states consider a dissolved oxygen concentration below
2 mg/L to be deleterious to aquatic life. We are
therefore defining the portion of the Virginian Province
with a dissolved oxygen concentration below 2 mg/L
to be "impacted". It is important to note that available
criteria for defining "impacted" conditions do not exist
for all EMAP indicators of ecological condition.
As described in Section 1.4, this is not intended as
an interpretative report. Results are summarized and
reported in this section with only limited interpretation
provided. Interpretation and conclusions will be
provided in a separate interpretative Assessment Report
(Paul el al., in preparation).
CDFs display the full distribution of the values ob-
served for an indicator plotted against the cumulative
100
80
60
40
20
2468
Bottom Dissolved Oxygen (mg/L)
10
Figure 1. Example cumulative distribution of bottom
dissolved oxygen as a percent of area. (Dashed lines are the
95% confidence intervals).
percentage of area in the class or Province. They provide
information on both central tendency (e.g., median) and
the range of values in one easily interpreted graphical
format (Holland, 1990). For example, Figure 3-1 shows
the cumulative distribution function of instantaneous
bottom dissolved oxygen (DO) concentrations for the
Virginian Province.
The x-axis represents observed DO concentrations
ranging from 0 to 10 mg/L. The y-axis represents the
cumulative percentage of estuarine area within the Virginian
Province. The dotted lines represent the 95% confidence
intervals for the CDF. The CDF provides the reader
with a powerful tool to evaluate the extent of conditions
of any indicator within the Province or class. For example,
the reader could be interested in the portion of area within
the Province that was characterized by a DO concentration
of 2 mg/L or less, a potential biological criterion. This
concentration intersects with the cumulative area in the
Province at 5 ± 2%. The reader might also be interested
in a state regulatory criterion of 5 mg/L, and the CDF
shows that 25 ± 3% of the estuarine bottoms waters had
DO concentrations below these levels. From a positive
viewpoint, the reader may be interested in the amount
of area above 7 mg/L (e.g., as a criterion for fish farming)
Statistical Summary, EMAP-E Virginian Province
Page 19
-------
and the CDF shows that approximately 23 ± 3% of the
bottom waters in the Province were observed to be
above 7 mg/L DO (i.e., 77 ± 3% < 7 mg/L).
Criteria values for the assessment of impacted
versus non-impacted areas are often subjective at best.
Indeed, many of the criteria values used in this
document, though based on reasonable scientific
judgement, are debatable. The CDF allows the user to
select his/her own criterion value and re-evaluate the
proportion of area in the Virginian Province which is
considered impacted.
Equations used in the generation of the CDFs and
associated confidence intervals are provided in Appendix
B. These equations were provided by EMAP-Estuaries,
and estimate confidence intervals based solely on the
distribution of values across the Province, regardless
of measurement error. The reader should note that the
equations for large estuaries and large tidal rivers differ
from those used for generating single-year estimates in
the 1991 and 1992 Statistical Summaries. The original
equations were designed solely for use with single-year
data and are not appropriate for use in this report. It
should be noted that these equations are still under
review and may be further refined in the future to
address additional measurements of variability.
Comparing results generated using the "old" vs "new"
equations shows only small changes resulting.
3.1 BIOTIC CONDITION INDICATORS
Biotic condition indicators (previously termed
response indicators) are characteristics of the environ-
ment that provide quantitative evidence of the status of
ecological resources and the biological integrity of the
sample site from which they are collected (Messer,
1990). Ecosystems with a high degree of biotic integrity
(i.e., "healthy" ecosystems) are composed of balanced
populations of indigenous benthic and water column
organisms with species compositions, diversity, and
functional organization comparable to undisturbed
habitats (Karr and Dudley, 1981; Karr et al., 1986).
Biotic condition indicators measured include measures
of both fish and benthic community structure.
3.1.1 Benthic Index
Condition of the benthic community was used as
an indicator because previous studies have suggested
that they are sensitive to pollution exposure (Pearson
and Rosenberg, 1978; Boesch and Rosenberg, 1981).
They also integrate responses to exposure over relatively
long periods of time. One reason for their sensitivity
to pollutant exposure is that benthic organisms live in
and on the sediments, a medium that accumulates
environmental contaminants over time (Schubel and Carter,
1984; Nixon et al., 1986). The sedentary nature of many
benthic invertebrates also may maximize their exposure
to pollutants which accumulate in sediments.
Three 440cm2 grab samples were collected at each
station and sieved through a 0.5 mm sieve. Organisms
and debris collected on the sieve were preserved in 10%
formalin and returned to the laboratory for sorting,
identification, enumeration, and biomass determination.
A benthic index which uses measures of community
condition to evaluate the condition of the benthic assemblage
was utilized in the assessment of biological resources
of the Virginian Province. The index under development
was determined from data collected in all four years of
sampling and was constructed to represent a combination
of ecological measurements that best discriminates between
good and poor ecological conditions. The index represents
EMAP-E's attempt to reduce many individual indicators
into a single value that has a high level of discriminatory
power between good and poor environmental conditions.
The reader should note that this index is different from
the one used in earlier EMAP-VP reports. This index
was determined using all four years of data as part of
the assessment exercise currently underway.
The process for developing an index of benthic community
condition has been documented for the 1990 (Weisberg
et al., 1993) and, separately, for the 1990-91 (Schimmel
et al., 1994) data sets. This process entails several discrete
steps: identification of a set of benthic parameters to
define conditions that include components of faunal and
functional diversity and structure; determination of the
statistical relationships between these benthic parameters
and habitat variables; normalization of those benthic parameters
that are strongly associated with habitat condition;
identification of a test data set that clearly distinguishes
relatively pristine sites from those exhibiting toxic
contamination, hypoxia, or both; and application of
Page 20
Statistical Summary, EMAP-E Virginian Province
-------
discriminant analysis to the test data set to determine
those benthic parameters whose variation is most closely
associated with differences in reference and impacted
condition. This same process was used with the 1990-93
data set.
The development of this index is detailed in the
EMAP-Virginian Province Four-Year Assessment Report
(Paul et al., in preparation) and Appendix A of this
document.
The three benthic parameters of the index were:
salinity-normalized expected Gleason's D (Washington,
1984) for infaunal and epifaunal species, salinity-
normalized expected number of tubificids, and
abundance of spionids. The richness measure is
associated with reference conditions (positive
contribution) and the latter two measure are associated
with impacted conditions (negative contribution).This
index results in a slightly better classification efficiency
(ca. 90% for the test data set) than the index utilized
in the 1991 and 1992 Statistical Summaries (ca. 85%).
This index also performs much better for low salinity
waters than the previous index.
The discriminant score calculation normalizes the
individual parameters based on the mean and standard
deviation for the parameter in the test data set. The
critical value for discriminating between reference and
impacted sites was determined to be zero using the
following equation:
Benthic Index Score =
1.389 (pet expect Gleason - 51.5) / 28.4
-0.651 (normalized tubificid abundance - 28.2) / 119.5
- 0.375 (spionid abundance - 20.0) /45.4
Where:
Percent Expected Gleason diversity index value =
Gleason / (4.283 - 0.498*bottom salinity
+ 0.0542 * bottom salinity2
- 0.00103* bottom salinity3) * 100
Normalized Tubificid Abundance =
Tubificids - 500 * e
-15*bottom salinity
Twenty three (±3) percent of the bottom area of
the Virginian Province sampled had an index value of
< 0, indicating likely impacts on the benthic community
(Figure 3-2). The percent area classified as impacted
among the three classes of estuaries are 18 ± 4 %, 35
± 6 %, and 33 ± 14% for large estuaries, small estuarine
systems, and large tidal rivers, respectively (Figure 3-3).
100
Benthic Index Score
Figure 3-2. Cumulative distribution of the benthic index as a percent of area in the Virginian Province. (Dashed lines
are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 21
-------
50-
40-
30-
0)
Q.
20-
10-
Province Large Small Tidal
Figure 3-3. Percent area of the Virginian Province by estuarine class wilh a benthic
index value below 0. (Error bars represent 95% confidence intervals).
Quality Assurance
Valid benthic index data exist for 401 of the 446
Base Sampling Sites in the database (90%). The 401
sites sampled represent 88% of the area of the Virginian
Province. The current benthic index database contains
no QA qualifier flags, indicating that all the data are
valid and need no qualification. Additional QA
information can be found in the 1990-1993 Virginian
Province Quality Assurance Report (Strobel et at.,
1995).
3.1.2 Number of Benthic Species
Number of infaunal benthic species has been used
to characterize the environmental condition of estuarine
habitats for specific salinity and grain size conditions.
The mean number of species from three replicate 440
cm2 grabs collected at each station resulted in numbers
of infaunal benthic species ranging from 0 to 68 (Figure
3-4), with the maximum number of species per grab being
68, 52, and 25 in the large estuaries, small estuaries,
and large tidal rivers respectively (Figure 3-5).
10 20 30 40 50 60
Benthic Species (mean number per grab)
70
Figure 3-4. Cumulative distribution of the mean number of infaunal benthic species per grab as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 22
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
b) Small Estuaries
10
20
c) Large Tidal Rivers
10 20 30 40 50 60
Benthic Species (mean number per grab)
70
Figure 3-5. Cumulative distribution of the number of infaunal benthic species per grab by estuarine class: a) Large estuaries,
b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 23
-------
Because community composition is strongly
influenced by factors other than environmental "health"
(e.g., salinity and grain size), we cannot infer that a low
number of species necessarily represents an impacted
community. However, the CDFs presented provide
baseline information that can be useful in assessing
future trends in community structure.
Quality Assurance
Valid benthic data exist for 404 of the 446 Base
Sampling Sites in the database (91%). The 404 sites
sampled represent 89% of the area of the Virginian
Province. The current benthic database contains no QA
qualifier flags, indicating that all the data are valid and
need no qualification. As part of laboratory QA the
analytical laboratory is required to resort, re-identify,
and recount 10% of the samples. All reanalysis results
were within 10% (QA control limit) of the original data.
Additional QA information can be found in the 1990-
1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.1.3 Benthic Infaunal Abundance
Abundant benthic organisms, particularly in com-
munities characterized by multiple species and feeding
types, suggest a productive estuarine environment.
Infaunal abundances ranged from 0 to over 138,000
organisms per square meter (Figure 3-6). Using <200
organisms per square meter (8.8 per grab) and <500
organisms per square meter (22 per grab) as indicators
of low and moderate abundances, respectively, 7 ± 2%
of the Virginian Province had low abundances, and 9
± 2% had low to moderate abundances. Because of natural
variation in benthic populations and modifying factors
such as salinity and grain size, low abundance, as defined
above, does not necessarily imply that the community
is impacted; however, this information can be useful in
detecting trends.
The percent area of low abundance was low in all
three estuarine classes. Five ± 2 , 8 ± 3, and 15 ± 9
percent of the area of large estuaries, small estuaries,
and large tidal rivers, respectively, exhibited benthic
abundances of < 200 organisms per square meter (Figure
3-7). The highest number of individuals (138,674 per
m2) was found in the large estuary class, with maximums
of 76,530 and 20,591 found in the small estuary and large
tidal river classes, respectively.
Quality Assurance
Valid benthic data exist for 404 of the 446 Base Sampling
Sites in the database (91%). The 404 sites sampled represent
89% of the area of the Virginian Province. The current
benthic database contains no QA qualifier flags, indicating
that all the data are valid and need no qualification. As
part of laboratory QA the analytical laboratory is required
to resort, re-identify, and recount 10% of the samples.
All reanalysis results were within 10% (QA control limit)
of the original data. Additional QA information can be
found in the 1990-1993 Virginian Province Quality Assurance
Report (Strobel et al, 1995).
100
80
CO
£ 60
^ 40
20
7
/
i
20 40 60 80 100 120
Total Benthic Abundance (#/m2 x 1,000)
140
Figure 3-6. Cumulative distribution of the total number of infaunal benthic organisms per m2 as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 24
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
b) Small Estuaries
c) Large Tidal Rivers
Figure 3-7. Cumulative
100 -
80-
CB
2 60 -
s?
40
20-
0.
₯
i
i
0 20 40 60 80 100 120 140
100 i. .1.1-
80
CO
2> 60
^0
40
.
r
-
. .
20 |
o-
i i i i i i i
0 20 40 60 80 100 120 140
100
80
CO
£ 60 -
s?
40 -
20-
0.
' "*"^ -*
'
1
|:
F
0 20 40 60 80 100 120 140
Total Benthic Abundance (#/m2 x 1 ,000)
distribution of the total number of infaunal benthic organisms per m2 by estuarine class: a) Large
estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary,
EMAP-E Virginian Province Page 25
-------
3.1.4 Number of Fish Species
Fish were collected by trawling with a 15 m, high-
rise otter trawl with a 2.5-cm mesh cod end. The net
was towed for 10 (± 2) minutes against the tide (if
significant tidal current existed) between 0.5 and 1.5
m/s (1-3 knots). All fish caught in the trawl were
identified to species and counted; up to 30 fish of a
species from each collection were measured to the
nearest millimeter.
Zero to 17 species offish were collected from single
standardized trawls performed at each base station in
the Virginian Province (Figure 3-8). A total of 104 spe-
cies were collected in standard trawls throughout the
Province over four years.
Fish catch can be affected by many variables
including salinity, habitat, and migrations; therefore, a
critical value for the number of species that must be
caught in a net for the area to be considered "healthy"
is not available. We can only report the incidence of
high vs low catches. Low catch does not imply that the
area is impacted in reference to this indicator. However,
as described above for benthic indicators, these data can
be useful in detecting future trends in fish community
structure on a provincial scale. Attempts to develop a
fish index, similar to the benthic index described above,
have not been successful to date. However, EMAP's
efforts to develop such an index are continuing.
Two or fewer species were caught in standard trawls
in 30 ± 3% of the Virginian Province. Alternatively,
at least five fish species were collected throughout 35
± 4% of the sampled area of the Province. No fish were
collected at 26 stations, representing 7 ± 2% of the area
of the Province. The areas producing no fish catch were
located primarily in large estuaries (9 ± 3% of the area;
Figure 3-9). Fish were collected at all but eight small
estuary stations (96 ± 3% of the area) and at all but two
large tidal river stations (96 ± 4% of the area: Figure
3-9).
Quality Assurance
Valid fish species composition data exist for 390
of the 446 Base Sampling Sites in the database (87%).
The 390 sites sampled represent 87% of the area of the
Virginian Province. The current fish community database
contains no QA qualifier flags, indicating that all the
data are valid and need no qualification.
Each crew was required to preserve the first one or
two (depending on the year) individuals of each species
collected. These fish were shipped to an analytical laboratory
for taxonomic verification by a fisheries expert.
Three types of errors were detected: misspelled or
incomplete species names (in the database), misidentifications,
and fish that could not be identified in the field. Errors
falling into the first category were easily detected, corrected
in the database, and documented.
0 2 4 6 8 10 12 14
Number of Fish Species per Trawl
Figure 3-8. Cumulative distribution of the number of fish species per standard trawl as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 26
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
b) Small Estuaries
c) Large Tidal Rivers
4 6 8 10 12 14
Number of Fish Species per Trawl
Figure 3-9. Cumulative distribution of the number of fish species per standard trawl as a percent of area by estuarine class:
a) Large estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 27
-------
The second type of error was misidentifications.
In all cases the crew identified a closely-related species,
such as longspine porgy instead of scup, or brown
bullhead catfish instead of the yellow bullhead.
Unidentified fish were sent in as either unknowns
or partial unknowns (e.g., herring uncl.). Most mis-
identified or partially identified individuals were
juveniles.
A total of 43,049 fish, representing 112 species,
were collected in all trawls (both standard and non-
standard) from Base Sampling Sites during the Project.
The percentage of errors detected was less than one
percent and all except the incomplete identification of
individuals of five species were corrected in the
database. Additional QA information can be found in
the 1990-1993 Virginian Province Quality Assurance
Report (Strobel et al., 1995).
3.1.5 Total Finfish Abundance
Abundant nektonic organisms, especially in
communities characterized by multiple species and
feeding types, suggest a stable and productive food web.
Finfish abundance in standard trawls ranged from 0 to
1,244 fish per trawl throughout the Province (Figure 3-
10). A total of 35,489 fish were collected in standard
trawls conducted at Base Sampling Sites in 1990-1993.
Figure 3-11 illustrates fish abundance by system class.
Low fish catches (<10 fish per trawl) were experienced
in 38 ± 5%, 37 ± 9%, and 21 ± 8% of the area in large
estuaries, small estuaries, and large tidal rivers, respectively.
Overall, low catches were experienced in 36 ± 3% of
the area in the Virginian Province. As with the fish species
indicator, only high versus low catches are reported with
no inference made as to the quality of the area relative
to this indicator.
Quality Assurance
Valid fish abundance data exist for 390 of the 446
Base Sampling Sites in the database (87%). The 390
sites sampled represent 87% of the area of the Virginian
Province. The current fish community database contains
no QA qualifier flags. Additional QA information can
be found in the 1990-1993 Virginian Province Quality
Assurance Report (Strobel et al., 1995).
3.1.6 Fish Gross External Pathology
Field crews examined the first 30 individuals of each
fish species for evidence of external pathology (growths,
lumps, ulcers, and fin erosion). In 1991 crews were only
required to examine ten target species (Table 3-1).
100 -
200 400 600 800 1000
Number of Fish per Trawl
1200
Figure 3-10. Cumulative distribution of fish abundance in numbers per standard trawl as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 28
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
200
400
600
800
1000
1200
b) Small Estuaries
40
20
n -
1 , 1 : 1 . 1 . 1 \ .
200
400
600
800
1000
1200
c) Large Tidal Rivers
200 400 600 800 1000
Number of Fish per Trawl
1200
Figure 3-11. Cumulative distribution of fish abundance in numbers per standard trawl as a percent of area by estuarine class:
a) Large estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 29
-------
The results of these examinations are presented in
Table 3-2. The overall incidence of gross external
pathologies in examined fish was approximately 0.3%,
or three per thousand. This value was fairly consistent
between years. Species dwelling or feeding on the
bottom had the highest incidence of occurrence.
Table 3-1. 1991 target species examined for external
pathology and saved for chemical residue analysis.
Common Name
Atlantic Croaker
Bluefish
Channel Catfish
Scup
Spot
Summer Flounder
Weakfish
White Catfish
White Perch
Winter Flounder
Scientific Name
Micropogonias undulatus
Pomatomus saltatrix
Ictalurus punctatus
Stenotomus chrysops
Leiostomus xanthurus
Paralichthys dentatus
Cynoscion regalis
Ameiurus catus
Morone americana
Pleuronectes americanus
Quality Assurance
Valid fish pathology data exist for 390 of the 446
Base Sampling Sites in the database (87%). The 390
sites sampled represent 87% of the area of the Virginian
Province. The current fish community database contains
numerous QA qualifier flags describing the results of
quality assurance exercises.
Crews were required to preserve all fish with a pathology
and a subset of pathology-free fish for examination by
an expert pathologist. However, in 1990 to 1992 fish
were also collected for contaminant analyses (although
only those collected in 1991 were analyzed) which took
priority over pathology QA. Therefore, a fish with a
noted pathology could have been processed for chemical
analysis rather than pathology QA. Because of this, pathology
results reported for 1990 to 1992 were based on the crews'
observations. The 1991 and 1992 Statistical Summaries
report high rates of "false positive" based on the laboratory
review of preserved specimens.
Table 3-2. Incidence of gross external pathology among fish caught in the Virginian Province (standard
trawls only).
Lumps
Growths
Ulcers
Fin Rot
Total3
All Species (1991 excluded)
Frequency 6 10 20 19 55
Total # Fish Examined 16,884 16,884 16,884 16,884 16,884
Percent Incidence 0.04% 0.06% 0.12% 0.11% 0.33%
Target Species Only (see Table 3-1)
Frequency 6 6 18 11 40
Total # Fish Examined 11,845 11,845 11,845 11,845 11,845
Percent Incidence 0.05% 0.05 0.15 0.09 0.34
a "Total" need not equal the sum of lumps, growths, ulcers and fin rot if multiple pathologies were found on a
single fish.
Page 30
Statistical Summary, EMAP-E Virginian Province
-------
Thousands of specimens were sent in "blind" each
year, meaning the pathologist did not know which fish
were identified by the crew as having a pathology or
being pathology-free. Because of the quantity of
specimens the pathologist needed to examine and the
effect of preservation on the condition of these
pathologies, we began to feel that the errors may be due
to the QA process and not the crews' observations. In
1993 fish were no longer collected for chemical
analysis. Crews were instructed to send in all fish
observed to have a pathology (as well as a subset of
pathology-free fish). Unlike previous years, these fish
were identified for the pathologist as "pathology fish"
or "reference fish". Because all fish suspected of having
a pathology were sent in for verification, 1993 results
are based on the pathologist's observations. It should
be noted that whenever the pathologist disagreed with
the crew's classification, a second pathologist also
examined the fish.
As mentioned above, the incidence of pathology in
the Virginian Province was similar among years,
suggesting that the 1990-1992 field observations
reported in earlier EMAP reports were valid. Neverthe-
less, pathology results should be used with caution.
Additional QA information can be found in the 1990-
1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.2 ABIOTIC CONDITION INDICATORS
Abiotic condition indicators (previously termed
exposure indicators) provide information on the potential
exposure of organisms to environmental stresses, and
have historically been the mainstay of environmental
monitoring programs. Indicators of exposure measured
during the Virginian Province Survey were dissolved
oxygen concentration, sediment toxicity (Ampelisca
abdita), sediment contaminants, and marine debris.
3.2.1 Dissolved Oxygen
Dissolved oxygen (DO) is critically important to
aquatic systems because it is a fundamental requirement
of fish, shellfish and other aquatic biota. DO was
measured in two ways over the four years: instantaneous
point measurements (vertical profiles), and continuous
measurements (from deployed instruments) at base
stations for a minimum of 24 hours (1991 only). "Bottom"
relative to dissolved oxygen and other water quality measure-
ments is defined as one meter above the sediment/water
interface.
Vertical profiles of dissolved oxygen and other water
quality parameters were obtained using a SeaBird SeaLogger
CTD. DO data included in this report are point measurements
from this profile taken one meter above the sediment/water
interface. As a QA measure, and backup to the CTD,
beginning in 1991 additional bottom measurements of
DO were obtained at every station using a YSI model
58 DO meter.
In addition to single point measurements of DO at
a station at a specific time, in 1991 continuous bottom
measurements of DO were made for a minimum of 24
hours using a Hydrolab DataSonde 3 datalogger deployed
one meter off the bottom at base stations. Measurements
were taken every 15 minutes until the unit was retrieved.
Continuous DO measurements should provide a more
complete picture of the dissolved oxygen conditions at
a station (i.e., by monitoring the periods when benthic
and water column respiration is higher) than instantaneous
measurements. Minimum DO concentrations, as determined
from the full Hydrolab data set from each base station
over the entire Province, ranged from 0.0 to 8.3 mg/L.
The 1991 data show that approximately 8 ± 7% of the
sampled area of the Province experienced DO concentrations
as low as 2 mg/L over the 24 hour period of deployment,
compared to an estimate of 5 ± 5% for instantaneous
measurements.
The percent area classified as impacted based on a
value of <2 mg/L in the Virginian Province calculated
from continuous and instantaneous DO measurements
do not differ significantly. Data collected during the
1990 Demonstration Project show that temporal variability
in DO concentration has a weaker diurnal component
than is present in other regions (i.e., the Gulf of Mexico),
and that a much longer time series is required to "better"
classify a station as impacted than is attained using a
simple point measurement. This, in addition to the logistics
and additional cost involved in the relatively short-term
deployment of the DataSondes, resulted in continuous
measurements being discontinued after 1991.
Statistical Summary, EMAP-E Virginian Province
Page 31
-------
3.2.1.1 Bottom Dissolved Oxygen
Data collected from 1990 to 1993 indicate that
approximately 25 ±3% of the sampled area of the
Province contains bottom waters with a dissolved
oxygen concentration less than or equal to 5 mg/L
(Figure 3-12). Approximately 5 ± 2% of the Province
exhibited bottom DO conditions <2 mg/L, defined by
EMAP-E as severely hypoxic.
Dissolved oxygen conditions <2 mg/L were evident
in all three estuary classes (Figures 3-13 and 3-14).
Approximately 6 ± 2%, 0.2 ± 1.3%, and 10 ± 6% of the
areas of large estuaries, small estuaries, and large tidal
rivers, respectively, contained measured concentrations
of bottom DO of < 2 mg/L. Most of the severely hypoxic
water in the large tidal river class was found in the lower
Potomac River in 1990. Twenty eight ±5%, 21 ±11%,
and 18 ± 6% of the area of large estuaries, small estuaries,
and large tidal rivers, respectively, contained measured
dissolved oxygen concentrations <5 mg/L in bottom waters.
Quality Assurance
Valid bottom DO data exist for 420 of the 446 Base
Sampling Sites in the database (94%). The 420 sites
sampled represent 92% of the area of the Virginian Province.
The process of quality assuring water quality data is lengthy,
and includes a review of the CTD profiles and QA data,
100 -
4 6
Bottom Dissolved Oxygen (mg/L)
10
Figure 3-12. Cumulative distribution of bottom dissolved oxygen concentration as a percent of area in the Virginian
Province. (Dashed lines are the 95% confidence intervals).
90
80
70
CO
S> 60
o 50
| 40
^ 30
20
10
0
<2
2 to 5
>5
Province Large Small Tidal
Figure 3-13. The percent of area by class that had a low (< 2 mg/L), medium (2 to 5 mg/L), or high (>5 mg/L)
oxygen concentration in the bottom waters. (Error bars represent 95% confidence intervals).
Page 32
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
b) Small Estuaries
10
c) Large Tidal Rivers
2468
Bottom Dissolved Oxygen (mg/L)
10
Figure 3-14. Cumulative distribution of bottom dissolved oxygen concentration as a percent of area by estuarine class: a)
Large estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 33
-------
automated checks, splitting of profiles into multiple
files, and determining the appropriate values to report
for surface and bottom measurements. Bottom values
reported are generally the mean of the measurements
recorded during the bottom soak portion of the profile.
The current water quality database contains numerous
QA qualifier flags resulting from this process.
The CTDs utilized for the measurement of dissolved
oxygen were calibrated prior to the start of the sampling
season. In addition, a calibration check was performed
at each station by comparing measurements to those
obtained using an air-calibrated YSI meter. Acceptable
precision was ± 0.5 mg/L.
When multiple visits were made to a station, the
data utilized in the generation of this report are from
the first visit with valid surface AND bottom DO
measurements. When valid measurements from the CTD
were not available, the YSI measurement was utilized.
This allows for the calculation of surface-bottom
differences (i.e., stratification) as discussed in the
following section. Additional QA information can be
found in the 1990-1993 Virginian Province Quality
Assurance Report (Strobel et al., 1995).
3.2.1.2 Dissolved Oxygen Stratification
The difference between surface and bottom DO
concentrations measured at base sampling stations is
illustrated in Figure 3-15. Differences between bottom
and surface DO were less than 1 mg/L in 58 ± 4% of
the area of the Province. Approximately 8 ± 2% of the
area of the Province showed differences greater than 5
mg/L.
Figure 3-16 illustrates DO differences by estuarine
class. Most of the highly stratified area was found in
the large estuaries (8 ± 3% of the area with a difference
exceeding 5 mg/L), with the largest A DO measured being
8.7 mg/L.
Quality Assurance
Four hundred and ten of the 446 Base Sampling Sites
in the database (92%) have both valid surface and bottom
dissolved oxygen data collected on the same visit, allowing
the calculation of DO stratification. The 410 sites sampled
represent 91% of the area of the Virginian Province.
The process of quality assuring water quality data is lengthy,
and includes a review of the CTD profiles and QA data,
automated checks, splitting of profiles into multiple files,
and determining the appropriate values to report for surface
and bottom measurements. The current water quality
database contains numerous QA qualifier flags resulting
from this process.
When multiple visits were made to a station, the data
utilized in the generation of this report are from the first
visit with valid surface AND bottom DO measurements.
When valid measurements from the CTD were not available,
the YSI measurement was utilized. Multiple measurement
100 -
02468
Dissolved Oxygen Difference (mg/L)
Figure 3-15. Cumulative distribution of the D.O concentration difference between surface and bottom waters as a
percent of area in the Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 34
Statistical Summary, EMAP-E Virginian Province
-------
90 -i
80
70-
S 60-
<
o 50-
8 40^
CD
30-
20-
10-
1 to 5
>5
0
Province Large Small Tidal
Figure 3-16. The percent of area by estuarine class that had a low (<1 mg/L), medium (1 to 5 mg/L), or high (>5 mg/L)
difference in D.O. concentration between the surface and bottom waters. (Error bars are 95% confidence intervals).
types (e.g., YSI and CTD) were never mixed in the
determination of DO stratification. Additional QA
information can be found in the 1990-1993 Virginian
Province Quality Assurance Report (Strobel et al.,
1995).
3.2.2 Sediment Toxicity
Sediment toxicity tests were performed on the
composite sample of surficial sediments (top two cm)
collected from each sampling site. Solid-phase sediment
toxicity tests (Swartz et al., 1985) with the tube-
dwelling amphipod, Ampelisca abdita, were conducted
according to procedures described in U.S. EPA/ACE
(1991) and ASTM (1991). Sediments were classified
as toxic if amphipod survival in the test sediment was
less than 80% of that in the control (a.k.a. "reference")
sediment and significantly different from the control.
The relative health of test organisms was determined
via the use of reference toxicant tests as described
below. Approximately 10 ± 2% of the sampled area of
the Virginian Province contained toxic sediments (Figure
3-17). However, only 2± 1% of the area had sediments
where survival was below 60% of control survival (i.e.,
sediments were highly toxic). The estuarine class with
the largest proportion of toxic sediments was the small
estuary class (12 ± 6%), with the large estuaries and
large tidal river classes exhibiting a lesser extent of
toxicity (10 ± 3% and 3 ± 4%, respectively: Figure 3-
18). Highly toxic sediments were evenly distributed among
the three classes (2 ± 1%, 3 ± 4%, and 1 + 1% of the
area of large estuaries, small estuaries, and large tidal
rivers, respectively).
Quality Assurance
Valid sediment toxicity data exist for 373 of the 446
Base Sampling Sites in the database (84%). The 373
sites sampled represent 82% of the area of the Virginian
Province.
As per the QA Project Plan (Valente and Strobel,
1993), the laboratory was required to maintain a control
chart for toxicity testing using a reference toxicant. The
laboratory used SDS (sodium dodecyl sulfate) or calcium
chloride (1990 only) as their reference material, running
a standard 48-hour water-only toxicity test with the toxicant
whenever EMAP samples were run. The control chart
shows that the LC50 for SDS ranged from <2.57 to 11.2
mg/L, with all but one value falling within two standard
deviations of the mean as required in the QA Plan. In
1990 the LC50 for calcium chloride ranged from 0.4
to 1.2 mg/L, with all values falling within two standard
deviations of the mean.
The LC50 of one reference toxicity test conducted
in 1992 (<2.57 mg/L) fell outside of two standard deviations
of the mean. Results of the failed reference test, as well
Statistical Summary, EMAP-E Virginian Province
Page 35
-------
co
100
80 -
60 --
40 -
20 -
0 --
0
20 40 60 80
Mean Amphipod Survival (% of control)
100
Figure 3-17. Cumulative distribution of mean survival of amphipods in 10-day laboratory toxicity tests (expressed as
percent of control survival). (Dashed lines are the 95% confidence intervals).
20 -,
15-
03
" 10-
c
o>
o
Q>
Q_
0-
< 60%
< 80%
Province Large Small Tidal
Figure 3-18. Percent of area in the Virginian Province, by estuarine class, with low (<80% of control) or very low (<60% of
control) amphipod survival in sediment toxicity tests. (Error bars represent 95% confidence intervals).
as the results of all tests included in that batch, were
examined. No anomalies were noted and no re-testing
was performed.
In 1991 several tests failed to meet EMAP QA
requirements for control organism survival. Of the 19
tests conducted, three exhibited control organism
survival less than the required 85% (this was following
repeating all tests that failed on the first attempt). These
tests were "flagged" in the database (ST-L code) and
were not included in the data set utilized to generate this
statistical summary. Numerous other data qualifier codes
exist in the sediment toxicity database; however, they
indicate minor deviations from standards (e.g., fewer
than five replicates, or fewer than 20 organisms per replicate).
Those data were included in the generation of this report.
Page 36
Statistical Summary, EMAP-E Virginian Province
-------
Additional QA information can be found in the
1990-1993 Virginian Province Quality Assurance Report
(Strobel et al, 1995).
3.2.3 Sediment Contaminants
A wide variety of contaminants have been released
to marine systems due to human activities. Some of
these compounds and elements have properties which
cause them to associate with particulate material, and
many of these chemicals are also persistent in the
environment. Contaminants with this combination of
properties can accumulate to high concentrations in
sediments and may become available to aquatic
organisms. The organic compounds measured included
selected polycyclic aromatic hydrocarbons (PAHs),
polychlorinated biphenyl (PCB) congeners, chlorinated
pesticides, butyltins and several metals. Because of the
complex nature of sediment geochemistry, and possible
additive, synergistic, and antagonistic interactions
among multiple pollutants, the ecological impact of
elevated contaminant levels is not well understood.
Several strategies for estimating biological effects from
contaminant concentrations which will be discussed are
the EPA Sediment Quality Criteria approach, the Long
and Morgan approach, and the SEM/AVS
(simultaneously extracted metals/acid volatile sulfides)
approach for metals. Because these various techniques
result in different estimates, definitive estimates of
percent area of the Province with overall contaminant
concentrations high enough to cause ecological impacts
cannot be provided. However, the data collected will
form a baseline for monitoring trends in sediment
contamination and are extremely valuable in that
respect.
EPA is currently in the process of establishing
Sediment Quality Criteria (SQC). Draft SQC are
presently available for four of the analytes EMAP-VP
is measuring: Acenaphthene, phenanthrene, fluoran-
thene, and dieldrin (U.S. EPA, 1993a-d). SQC are ex-
pressed as fjg analyte / g organic carbon; therefore,
concentrations must first be normalized for the organic
carbon content of the sediment. Only those sediments
with organic carbon concentrations >0.2% can be
examined using this approach. Separate SQC values
have been established for freshwater and saltwater
sediments.
SQC values for the four analytes measured are listed
in Table 3-3, along with the upper and lower bounds.
It is important to note that these values are still in draft
form and are subject to change as the documents proceed
through the peer review process.
Table 3-3. U.S. EPA draft Sediment Quality Criteria for analytes
measured. Freshwater (F), Saltwater (S), and
upper and lowerconfidence intervals are included.
All values are ug/g organic carbon.'
Analyte
Acenaphthene
Phenanthrene
Fluoranthene
Dieldrin
F/S
F
S
F
S
F
S
F
S
SQC
130
230
180
240
620
300
11
20
Upper
SQC
280
500
390
510
1300
640
24
44
Lower
SQC
62
110
85
110
290
140
5.2
9.5
In the approach proposed by Long and Morgan (1990),
a database has been accumulated from the scientific literature
comparing concentrations of analytes to ecological or
toxic effects. Based on this review, effect levels have
been established. The ER-M (Effects Range-Median)
criterion is concentration representing the median, or
50th percentile of the biological effects data reviewed.
The lower ER-L (Effects Range-Low) value represents
the 10th percentile. Concentrations below the ER-L value
are rarely expected to elicit effects. ER-M and ER-L
values as modified by Long et al. (1995), are presented
in Table 3-4. It should be noted that the values used
in earlier EMAP-E documents were the Long and Morgan
(1990) values. ER-M and ER-L values have subsequently
been updated (Long et al., 1995) and these newer values
are used in this report. The major difference is an increase
in the ER-M values for metals, resulting in a significant
reduction in the percent area of the Province in exceedence.
Sediments collected at stations representing approximately
6 ± 2% of the Province exceeded the ER-M value for
any analyte, either inorganic or organic. Exceedences
Statistical Summary, EMAP-E Virginian Province
Page 37
-------
were measured in 5 ± 2%, 5 ± 2%, and 14 ± 6% of the
area of large estuaries, small estuaries, and large tidal
rivers, respectively. The maximum number of
exceedences at any station was 13 (in one small
estuary), with most stations exceeding for only one or
two analytes.
Table 3-4. ER-M and ER-L guideline values for
metals and organic contaminants. Values
are from Long et al., 1995.
Analyte
ER-M
ER-L
Metals (ua/a or pom dry weight sediment)
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Silver
Zinc
70
9.6
370
270
218
0.71
51.6
3.7
410
8.2
1.2
81
34
46.7
0.15
20.9
1.0
150
Organic contaminants (no/a or oob dry weight sediment)
Acenaphthene
Acenaphthylene
Anthracene
Fluorene
2-methyl naphthalene
Naphthalene
Phenanthrene
Benz(a)anthracene
Benzo(a)pyrene
Chrysene
Dibenzo(a,h)anthracene
Fluoranthene
Pyrene
Total PAHs
p,p'-DDE
Total DDT
Total PCBs
500
640
1,100
540
670
2,100
1,500
1,600
1,600
2,800
260
5,100
2,600
44,792
27
46.1
180
16
44
85.3
19
70
160
240
261
430
384
63.4
600
665
1,700
2.2
1.58
22.7
Using ER-L values, sediments collected at stations
representing approximately 47 ± 4% of the Province exceeded
the criterion for at least one analyte, with the maximum
number of exceedences being for 23 analytes.
Another approach, specific to divalent metals, is to
compare the amount of simultaneously-extracted metal
(SEM) in the sediment v/ith the amount of acid volatile
sulfide (AVS). The SEM/AVS ratio is the molar ratio
of the sum of cationic metals to AVS. Samples with
a value less than one contain more AVS than metal, suggesting
that all the metal would be bound by the AVS (1:1 binding
ratio) and none would be bioavailable. Therefore, if
the SEM/AVS ratio is < 1, metal-related toxicity or effects
should not exist. If the ratio is greater than one, some
metal may be bioavailable, depending on the availability
of other potential binding agents such as total organic
carbon.
Where appropriate, all three of these approaches are
presented. It should be noted that no single approach
has gained wide acceptance throughout the scientific
community. The results of the available analyses are presented
without endorsement of any single method for associating
ecological effects with sediment contaminant levels.
3.2.3.1 Polycyclic Aromatic Hydrocarbons
Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous
in marine sediments (Laflamme and Kites, 1978). These
compounds are widespread because of the large number
and variety of PAH sources including oil spills, natural
oil seeps, forest fires, automobile exhaust, domestic heating,
power plants and other combustion processes. With
the exception of specific oil releases, the majority of
PAHs found in marine sediments are believed to originate
from combustion processes (Windsor and Hites, 1979).
PAH concentrations tend to correlate with the degree
and urbanization or industrialization and, therefore, these
compounds are often considered to be indicators of anthropo-
genic activity.
Range and median concentrations for PAHs measured
in the Virginian Province are listed in Table 3-5. Combined
PAH values reported in this table reflect the summation
of the concentrations of all of the PAH compounds that
were measured. This summation is not listed as "total"
PAH because only a select list of PAHs were measured
and many other PAH compounds could be found in these
sediments. Combined PAH concentrations for low level
Page 38
Statistical Summary, EMAP-E Virginian Province
-------
Table 3-5. Range and median PAH concentrations in sediments of the Virginian Province (1991 to 1993
only).
Analyte (weight
a* b
Concentration (ng/g dry weight)
MIN
MAX
Median
Median
Detection Limit0
Acenaphthene (L)
Acenaphthlylene (L)
Anthracene (H)
Benz(a)anthracene (H)
Benzo(b+k)fluoranthene (H)
Benzo(g,h,i)perylene (H)
Benz(a)pyrene (H)
Benz(e)pyrene (H)
Biphenyl (L)
Chrysene (H)
Dibenz(a,h)anthracene (H)
Fluoranthene (H)
Fluorene (L)
lndeno(1,2,3-c,d)pyrene (H)
Naphthalene (L)
1-methylnaphthalene (L)
2-methylnaphthalene (L)
2,6-dimethylnaphthalene (L)
2,3,5-trimethylnaphthalene (L)
Perylene (H)
Phenanthrene (H)
1-methylphenanthrene (H)
Pyrene (H)
Combined PAHs
Letter in parenthesis indicates
f~\ »-> f\ c* + r» i i /-\ « \mno r\\s i"»l i i /-\f\t~t fwf-\rv
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
high molecular
^ +K^\r»^-» o i-» o l\ /o t*n
1,770
209
2,620
4,000
6,550
2,060
2,970
2,510
292
6,590
544
19,900
2,320
2,240
1,500
477
1,120
489
184
2,020
11,800
983
14,900
80,100
weight compound
r^ A o o + o +*-\ r-J !»-> + 1-» *
ND
ND
ND
23.2
68.9
25.7
25.4
26.6
ND
33.1
ND
51.9
ND
28.5
15.5
ND
13.0
ND
ND
38.8
38.4
ND
55.3
562
(H) or low
i + f\\tt +l-i i r% i
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
na
molecular weight compound (L).
contaminated by a chip of PAHs blown out of a smokestack of a passing ship.
For each "not detected" the laboratory supplied a detection limit. This value is the median of these values
for each analyte.
na = not applicable
ND = not detected
Statistical Summary, EMAP-E Virginian Province
Page 39
-------
samples are artificially low because analytes that were
not detected (ND) were assigned a value of zero for
calculation of the Combined concentration. Combined
PAH concentrations (Table 3-5) showed a large range
(ND - 80,100 ng/g) with a median concentration of 562
ng/g in Virginian Province sediments.
This large range of PAH concentrations can be seen
in the cumulative distribution of combined PAHs shown
in Figure 3-19. This figure shows that the sediments
of the vast majority of the area of the Province contain
low concentrations of PAHs. The 75th percentile for
total PAHs was approximately 1,200 ng/g dry weight.
Figure 3-19b is the CDF plotted on a log scale to better
illustrate the distribution of concentrations at the lower
end of the scale.
It should be noted that one large estuary station was
eliminated from this analysis. The station, with a concentration
of PAHs of 141,000 ng/g located near a shipping channel
at the mouth of Chesapeake Bay in a sandy environment.
Sediments from this station did not show any toxicity,
analytes other than PAHs were not elevated, and the benthic
community was not indicative of a degraded environment.
All evidence suggests that this exceedence was an artifact,
possibly due to a "chip" of material dislodged from the
smokestack of a passing ship. Inclusion of this station
in the analyses has little effect on the overall distribution
of PAH contaminants across the Province.
40 -
20 -
n
i i i i i i i i i i i i i i , i , , , i
20,000 40,000 60,000
Combined PAHs (ng/g dry wt)
80,000
100,000
100 1,000
Combined PAHs (ng/g dry wt)
10,000
100,000
Figure 3-19a&b. Cumulative distribution of combined PAHs in sediments as a percent of area in the Virginian Province
(1991-1993): a) Linear scale, b) Logarithmic scale. (Dashed lines are the 95% confidence intervals).
Page 40
Statistical Summary, EMAP-E Virginian Province
-------
As discussed above, draft Sediment Quality Criteria
are available for three PAHs: acenaphthene, phenan-
threne, and fluoranthene. The SQCs (see Table 3-3) for
freshwater and saltwater sediments were exceeded at
only one small estuary station visited over the three
years, representing 0.07 % of the area of the Province.
This exceedence was for fluoranthene and phenanthrene.
Applying the more conservative Lower SQC values in
Table 3-3 increases the percentage only slightly (0.4%
of the area in exceedence). It is important to note that
these estimates were based on only those sediments with
a total organic carbon content of >0.2% (72 ± 3% of
the area of the Province). For the purpose of this
exercise, those stations excluded were treated statistical-
ly as missing values.
Stations representing only 1 ± 1 % of the Province
exceeded any ER-M value for PAHs, with the highest
incidence found in the large tidal rivers (3 ± 4% of the
area).
Using the lower ER-L values, sediments collected
at stations representing approximately 24 ± 4% of the
Province exceeded the criterion for at least one PAH.
Petroleum and combustion-type PAH sources contain
very different PAH compound distributions. Because
of this, the distributions of PAHs in a sample can
provide information on the relative importance of petro-
leum versus combustion PAH sources (Lake et al.,
1979). Petroleum products contain relatively large
amounts of lower molecular weight compounds relative
to combustion sources which are dominated by higher
molecular weight compounds (listed in Table 3-5).
Examination of the distribution of PAHs in samples
reveals that high molecular weight compounds dominate
in almost all samples, indicating that combustion is the
major source of PAHs in Virginian Province sediments.
Quality Assurance
Serious analytical problems were encountered by
the laboratory analyzing the 1990 sediment chemistry
samples. As a result of these problems, the data
generated did not meet the rigid Quality Assurance
requirements specified by EMAP, and the 1990 PAH
results were deemed of unacceptable quality for use in
this assessment. Therefore, the above estimates are
based on only three years of data (1991 - 1993), and
sediment PAH data exist for only 292 of the 446 Base
Sampling Sites in the database (65%). The 292 sites
sampled represent 65% of the area of the Virginian Province.
Although data were collected in only three of the four
years of sampling, as discussed in Appendix B, stations
sampled each year are assumed to represent conditions
within the Province as a whole, with the four-year estimate
being the mean of individual yearly estimates.
The current PAH database contains three QA qualifier
flags, two related to non-detects and the detection limit,
and one cautioning the user that the data are suspect and
should be used with caution (the SC-C code). These
data are generally included in this report because, although
the exact concentration may be questionable, it does indicate
whether the concentration was high or low. In general,
all data collected in a given year for an individual analyte
were "flagged" if there appeared to be a consistent bias
in the results of analysis of QC samples (i.e., Standard
Reference Materials or SRMs) for that analyte. These
data are utilized in this report because not including these
analytes in the "Combined PAH" calculation would result
in an artificially low value. The number of analytes with
the SC-C code varied between years, but was generally
low (e.g., the only PAH flagged in 1993 was chrysene
and this was because the mean percent recovery for chrysene
from the SRM was 135.5% [acceptable range 70-130%]).
Additional QA information can be found in the 1990-1993
Virginian Province Quality Assurance Report (Strobel
etal., 1995).
3.2.3.2 Polychlorinated Biphenyls
Environmental measures of PCBs have been conducted
using a variety of techniques including their measurement
as industrial mixtures (e.g., Aroclors) (Hutzinger et al.,
1974), by level of chlorination (Gebhart et al., 1985)
and as individual congeners (Mullin et al., 1984; Schantz
et al., 1990). Each of these techniques has both positive
and negative aspects based on the specific application
for which the PCB data are needed. For this study, PCBs
were measured as a series of 18 selected congeners (Table
3-6). These congeners were selected to produce data
consistent with the National Oceanographic and Atmospheric
Administration's National Status and Trends Program.
The congeners included on this list are some of the more
abundant chlorobiphenyls found in environmental samples
as well as some (congeners 105 and 118) that are considered
to have a high potential for toxicity (McFarland and Clarke,
1989).
Statistical Summary, EMAP-E Virginian Province
Page 41
-------
The PCB congeners measured are identified based
on the numbering convention proposed by Ballschmiter
and Zell (1980). Concentration ranges and median
values measured for the individual congeners are listed
in Table 3-6. Also included in this table is a summation
of the measured congeners referred to as Combined
PCBs. This term was used instead of "total" PCBs to
differentiate it from measurements of all of the PCBs
in a sample. Combined PCB concentrations for low
level samples are artificially low because congeners that
were not detected were assigned a value of zero for
calculation of the combined concentration. Combined
PCB concentrations ranged from the detection limit to
1,040 ng/g dry weight with a median concentration of
4.61 ng/g. The cumulative distribution of combined
PCBs in the Virginian Province is shown in Figure 3-20.
This plot shows that low concentrations of PCBs were
found in the majority of the area of the Province. PCBs
were not detected in 37 ± 4% of the area of the Province.
The 75th percentile for total PCBs was approximately
10 ng/g dry weight. Figure 3-20b is the CDF plotted
on a log scale to better illustrate the distribution of concentra-
tions at the lower end of the scale.
Stations representing only 2 ± 1% of the Province
exceeded the ER-M value for total PCBs, with the highest
incidence found in the large tidal rivers (10 ± 3% of the
area). The relatively high value for large tidal rivers
is due to PCB contamination of the Hudson River.
Table 3-6. Range and median PCB concentrations in sediments of the Virginian Province (1991 to 1993 only).
Analyte
MIN
Concentration (ng/g dry weight)
MAX
Median
Median
Detection Limit3
PCBS
PCB18
PCB28
PCB44
PCB52
PCB66
PCB101
PCB105
PCB118
PCB128
PCB138
PCB153
PCB170
PCB180
PCB187
PCB195
PCB206
PCB209
Combined PCBs
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
84.5
76.1
346
72.6
107
152
57.8
34.8
58.8
134
97.0
83.8
49.4
82.0
30.2
10.2
21.6
29.4
1,040
ND
ND
0.490
ND
ND
0.348
0.296
ND
0.346
ND
0.446
0.526
ND
ND
ND
ND
ND
ND
4.61
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
na
a For each "not detected" the laboratory supplied a detection limit. This value is the median of these values
for each analyte.
na = not applicable
ND = not detected
Page 42
Statistical Summary, EMAP-E Virginian Province
-------
500 1000
Combined PCBs (ng/g dry wt)
1500
100 -
1 10 100
Combined PCBs (ng/g dry wt)
1,000
Figure 3-20a&b. Cumulative distribution of combined PCBs in sediments as a percent of area in the Virginian Province
(1991-1993): a) Linear scale, b) Logarithmic scale. (Dashed lines are the 95% confidence intervals).
Quality Assurance
Serious analytical problems were encountered by
the laboratory analyzing the 1990 sediment chemistry
samples. As a result of these problems, the data
generated did not meet the rigid Quality Assurance
requirements specified by EMAP, and the 1990 PCB
results were deemed of unacceptable quality for use in
this assessment. Therefore, the above estimates are
based on only three years of data (1991 - 1993), and
sediment PCB data exist for only 293 of the 446 Base
Sampling Sites in the database (66%). The 293 sites
sampled represent 66% of the area of the Virginian
Province. Although data were collected in only three
of the four years of sampling, as discussed in Appendix
B, stations sampled each year are assumed to represent
conditions within the Province as a whole, with the four-year
estimate being the mean of individual yearly estimates.
The current PCB database contains four QA qualifier
flags, two related to non-detects and the detection limit,
and two cautioning the user that the data are suspect and
should be used with caution (the SC-C and SC-D code).
These data are generally included in this report because,
although the exact concentration may be questionable,
it does indicate whether the concentration was high or
Statistical Summary, EMAP-E Virginian Province
Page 43
-------
low. In general, all data collected in a given year for
an individual analyte were flagged with the SC-C code
if there appeared to be a consistent bias in the results
of analysis of QC samples (i.e., Standard Reference
Materials or SRMs) for that analyte. These data were
utilized in the generation of this report because not
including these analytes in the "Combined PCB"
calculation would result in an artificially low value.
The number of analytes with the SC-C code varied
between years, but was generally low (e.g., only two
PCB congeners were flagged in 1993). PCB
concentrations were "verified" by dual column
confirmation, meaning the concentration was measured
twice using two different gas chromatograph columns.
The SC-D code was applied if the two measurements
differed by greater than a factor of three. In these cases
the lower value was reported and assigned the SC-D
code. Additional QA information can be found in the
1990-1993 Virginian Province Quality Assurance Report
(Strobel et al, 1995).
3.2.3.3 Chlorinated Pesticides
In addition to PCBs, several other chlorinated com-
pounds were monitored in the sediments of the Virginian
Province. The analysis of chlorinated pesticides in
marine sediments is difficult because of the extremely
low concentrations generally present. This, coupled
with EMAP's rigorous QA Program, resulted in the need
to qualify all pesticide results as being of unknown or
questionable quality. Therefore, results of these
analyses are not presented in this report and cannot be
used in our assessment of conditions within the
Province.
3.2.3.4 Butyltins
Until its recent ban for most uses (Huggett et al.,
1992), tributlytin (TBT) was used in many boat anti-
fouling paint formulations. As a result of this usage,
TBT and its breakdown products, dibutyltin (DBT) and
monobutyltin (MET) have subsequently been detected
in many harbors (Seligman et al., 1989). The presence
of TBT in aquatic systems has generated considerable
concern because of the potent effects of this compound
on some species (Rexrode, 1987; Heard et al., 1989).
Tributlytin can be rapidly converted to DBT and MET
in the water column but may be relatively resistant to
degradation in marine sediments (Adelman et al., 1990).
The concentrations of butyltin compounds in this report
are reported as nanograms of the respective butyltin ion
per gram of dry sediment. Caution should be used when
comparing TBT concentrations among studies because
of the different ways that it is reported (e.g., sometimes
reported as ng tin /g sediment).
The maximum TBT concentration observed was 764
ng/g; DBT and MET levels were generally lower than
those of TBT (Table 3-7). Figure 3-21 shows the cumulative
distribution of TBT in sediments as a percent of area
in the Virginian Province. TBT was not detected (detection
limit of approximately 12 ng/g) in 33 ± 3% of the area
of the Province and 19 ±3% of the area contained sediments
with TBT concentrations of less than 25 ng/g. Concentrations
exceeding 100 ng/g were detected at stations representing
4 ± 1 % of the area of the Province.
Table 3-7. Range and median butyltin concentrations in sediments of the Virginian Province.
Analyte
Concentration (ng ion /g dry weight)
MIN
MAX
Median
Median
Detection Limit3
Monobutyltin (MBT+3)
Dibutyltin (DBT+2)
Tributyltin (TBT+)
ND
ND
ND
196
108
764
ND
ND
23.0
17.8
9.8
12.2
a For each "not detected" the laboratory supplied a detection limit. This value is the median of these values
for each analyte.
ND = not detected
Page 44
Statistical Summary, EMAP-E Virginian Province
-------
100 200 300 400 500 600
Tributyltin (ng TBT + /g dry wt)
700 800
Figure 3-21. Cumulative distribution of tributyltin in sediments as a percent of area in the Virginian Province. (Dashed
lines are the 95% confidence intervals).
Quality Assurance
TBT data exist for 392 of the 446 Base Sampling
Sites in the database (88%). The 392 sites sampled
represent 87% of the area of the Virginian Province.
The current TBT database contains three QA qualifier
flags, two related to non-detects and the detection limit,
and one cautioning the user that the data are suspect and
should be used with caution (the SC-C code). These
data are included in this report because, although the
exact concentration may be questionable, it does indicate
whether the concentration was high or low. In general,
all data collected in a given year for an individual
analyte were flagged with the SC-C code if there
appeared to be a consistent bias in the results of analysis
of QC samples (i.e., Standard Reference Materials or
SRMs) for that analyte. Because of the analytical
difficulties involved in analyzing for butyltins, all
butyltin data have been assigned the SC-C code
indicating that they should be considered estimates and
should be used with caution. This code was applied
mainly because of the relatively high detection limit (see
Table 3-7) relative to expected environmental concentra-
tions. Percent recoveries for the SRM also generally
fell marginally outside of the acceptable range (80-
120%). Additional QA information can be found in the
1990-1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.2.3.5 Metals
The median and range of metals concentrations measured
are listed in Table 3-8. Elemental concentrations in sediments
are highly variable, due not only to contaminant inputs,
but to natural differences in sediment types as well. Several
approaches have been used to normalize sediment metals
concentrations for variations due to sediment type differences.
The approach taken in the 1991 and 1992 Virginian Province
Statistical Summaries (Schimmel et al., 1994; Strobel
et al., 1994) was to normalize against aluminum. Determina-
tion of metal-aluminum relationships in background sediments
enables estimation of the extent of enrichment of metals
in sediments. A more detailed description of the metal-
aluminum normalization is described in Appendix A.
Figure 3-22 presents an example of a metal regression
plot (for Cr). The predicted metal-aluminum relationship
(solid line) is obtained from the regression, along with
the upper bound of the 95% confidence interval for predicted
values (dashed line). Values above the upper bound are
greater than expected (i.e., enriched) based on the aluminum
concentration measured in the sediment. This "excess"
metal is derived from additional sources other than crustal
background sediment, presumably, although not necessarily,
from anthropogenic activity. As described in Appendix
A, data were first statistically screened to generate a set
of unenriched stations which could be used to predict
the crustal metal-to-aluminum relationship. This process
was inefficient for several metals, but performed well
Statistical Summary, EMAP-E Virginian Province
Page 45
-------
Table 3-8. Range and median metal concentrations in sediments of the Virginian Province.
Analyte
MIN
Concentration (pg/g dry weight)
MAX
Median
Median
Detection Limit3
Major
Aluminum
Iron
Manganese
Trace
1,760
653
11.6
98,300
64,700
6,430
43,400
22,600
392
na
na
na
Antimony
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Tin
Zinc
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
152°
36.2
7.99
856
1,500
13,600b
3.27
136
6.87
9.69
80.1
865
0.285
5.79
0,219
41 5
16.0
24.8
0.055
15,4
0.331
0.047
2.19
81.5
0.051
0.98
0.031
1.80
2.35
1.80
0.004
1.70
0.110
0.007
0.120
1.80
a For each "not detected" the laboratory supplied a detection limit. This value is the median of these values
for each analyte.
b Lead shot is suspected as the cause of this elevated concentration. An elevated antimony level was also
detected in this sample, and antimony is a hardener used in lead shot. Excluding this station results in a
maximum antimony concentration of 49.1 ug/g and a maximum lead concentration of 323 ug/g.
na = not applicable
ND = not detected
for As, Cr, Fe, Hg, Mn, Ni, Sb, and Zn. The percent
area of the Virginian Province with sediments enriched
by metals pertains only to the metals mentioned above.
Regressions and regression parameters for these metals
are presented in Appendix C. Coefficients (r2) ranged
from 0.39 for antimony to 0.89 for iron.
As can be seen in the regression figures presented
in Appendix C, the magnitude of enrichment for
individual metals varies from being generally less than
a factor of two to greater than an order of magnitude.
Often a given station exhibits substantial enrichment of
more than one metal. The spatial extent of sediments
with elevated concentrations of metals can be estimated
once stations with enriched metal concentrations are identified
(Figure 3-23). For several metals, the proportion of the
Province in which metals concentrations are enriched
is substantial, e.g., Hg and Mn. One 1992 station in
Chesapeake Bay exhibited sediment concentrations of
both Pb and Sb several orders of magnitude higher than
Page 46
Statistical Summary, EMAP-E Virginian Province
-------
200
3 data points excluded:
341,348, & 865 pgCr
O
Figure 3-22. Linear regression of Chromium against aluminum. (Dashed line is the upper 95% confidence interval).
NOTE: Three data points were excluded for clarity.
As Cr Fe Hg Mn Ni Sb Zn
Figure 3-23. Percent area of the Virginian Province with enriched concentrations of individual metals in sediments.
(Error bars represent 95% confidence intervals).
any other station. This is likely due to lead shot
(presumably from duck hunters) included in the sample.
The co-occurrence of lead and antimony (Sb is a
hardener used in lead shot) at this station supports this
hypothesis.
Approximately 46 +5% of the area of the Province
showed enrichment of sediments with at least one metal.
Thirty seven (± 7), 64 ± 3, and 69 ± 16 percent of the
large estuary, small estuary, and large tidal river class
areas sampled contained sediments with metals concentrations
exceeding predicted background levels. Although a significant
proportion of the Province contains sediments with potentially
enriched levels of metals, this does not imply ecological
impact. The level of enrichment is generally low, and
most of the metals present are likely bound by AVS,
making them biologically unavailable (see Section 3.2.3.7).
Statistical Summary, EMAP-E Virginian Province
Page 47
-------
Stations representing approximately 4 ± 2% of the
Province exceeded at least one ER-M value for metals,
with the incidence similar across classes.
Using the more protective ER-L values, sediments
collected at stations representing approximately 41 ±
5% of the Province exceeded the criterion for at least
one metal. It should be noted, as discussed earlier, that
most of the metals found in the sediments are likely
bound by AVS or TOC and may not be biologically
available.
As stated earlier, the values used in earlier EMAP-E
documents were the Long and Morgan (1990) values.
ER-M and ER-L values have subsequently been updated
(Long et al., 1995) and it is these newer values that are
used in this report. The major difference is an increase
in the ER-M values for metals, resulting in a significant
reduction in the percent area of the Province in
exceedence.
Quality Assurance
Sediment metals data exist for 396 of the 446 Base
Sampling Sites in the database (89%). The 396 sites
sampled represent 88% of the area of the Virginian
Province. The current sediment metals database
contains three QA qualifier flags, two related to non-
detects and the detection limit, and one cautioning the
user that the data are suspect and should be used with
caution (the SC-C code). These data are included in
this report because, although the exact concentration
may be questionable, it does! indicate whether the concentration
was high or low. In general, all data collected in a given
year for an individual analyte were flagged with the SC-C
code if there appeared to be a consistent bias in the results
of analysis of QC samples (i.e., Standard Reference Materials
or SRMs) for that analyte, The number of analytes flagged
with the SC-C code varied by year, but ranged from one
to four. Additional QA information can be found in the
1990-1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.2.3.6 Acid Volatile Sulfides
Acid volatile sulfides are defined as the fraction of
sulfide in the sediments that can be extracted with cold
hydrochloric acid. They exist in sediments mainly as
iron monosulfide complexes, and are important in determining
the biological availability of a number of cationic metals,
primarily zinc, lead, copper, nickel, and cadmium. Acid
volatile sulfides measured in sediments of the Virginian
Province ranged from ND to 5,010 mg/kg dry weight
sediment. The CDFs of percent area as a function of
AVS concentration is shown in Figures 3-24 and 3-25.
Quality Assurance
Virginian Province AVS data exist for 288 of the
446 Base Sampling Sites in the database (65%). The
288 sites sampled represent 65% of the area of the Virginian
Province. The collection of AVS data did not begin until
1991. Although data were collected in only three of the
1000 2000 3000 4000
Acid Volatile Sulfides (mg/kg)
5000
Figure 3-24. Cumulative distribution of the acid volatile sulfide concentration in sediments as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Page 48
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
40 j
20-
n -
, 1 1 . 1 , 1 , 1
1000
2000
3000
4000
5000
b) Small Estuaries
1000
2000
3000
4000
5000
c) Large Tidal Rivers
1000 2000 3000 4000
Acid Volatile Sulfides (mg/kg)
5000
Figure 3-25. Cumulative distribution of the acid volatile sulfide concentration in sediments as a percent of area by estuarine
class: a) Large estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 49
-------
four years of sampling, as discussed in Appendix B,
stations sampled each year are assumed to represent
conditions within the Province as a whole, with the four-
year estimate being the mean of individual yearly
estimates.
AVS data collected in 1991, although included in
this assessment, should be used with caution. Because
of the collection method employed, samples may have
been partially oxidized in the process of sample
collection, transport, and analysis. Sediments collected
for chemical analysis (including AVS) were a composite
of the surficial layer from multiple grabs. The sediments
were thoroughly mixed to produce this homogenate.
It is possible that this mixing process resulted in the
oxidation of some of the AVS, reducing the measured
concentrations. Comparison of 1991 results with those
of 1992 and 1993 does not support this hypothesis;
nevertheless we feel the data should be used with
caution. The collection methodology was changed in
1992 to eliminate this problem.
Additional QA information can be found in the
1990-1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.2.3.7 SEM/AVS Ratio
As stated above, AVS exist in sediments mainly as
iron monosulfide complexes, and are important in
determining the biological availability of a number of
cationic metals, primarily zinc, lead, copper, nickel, and
cadmium. In 1993 simultaneously-extracted metals (SEM)
were measured in addition to total metals (described below).
SEM are extracted from sediments as part of the AVS
extraction, and represent the fraction which is potentially
bioavailable. One factor controlling the bioavailability
of metals is AVS (DiToro et al., 1991). The SEM/AVS
ratio is the molar ratio of the sum of cationic metals to
AVS. Samples with a value less than one contain more
AVS than metals, suggesting that all the metals would
be bound by the AVS (1:1 binding ratio) and none would
be bioavailable. Therefore if the SEM/AVS ratio is <
1, metal-related loxicity or effects should not exist. If
the ratio is greater than one, some metal may be bioavailable;
however, it is likely that other sediment components such
as TOC may bind-up some or all of the residual SEM.
The ratio of SEM to AVS for 1993 Virginian Province
data is shown in Figure 3-26. Approximately 74 ± 7%
of the area of the Virginian Province contains sediments
with an SEM/AVS ratio < 1, indicating that any metals
present would be unavailable and no toxic effects would
be expected. Only 3 + 3% of the area contains sediments
with a ratio exceeding 10.
Quality Assurance
Virginian Province SEM/AVS data exist for 98 of
the 446 Base Sampling Sites in the database (22%). The
98 sites sampled represent 25% of the area of the Virginian
100 -
80
CO
£ 60
4
-------
Province. SEM data were collected only in 1993.
Although data were collected in only one of the four
years of sampling, as discussed in Appendix B, stations
sampled each year are assumed to represent conditions
within the Province as a whole, with the four-year
estimate being the mean of individual yearly estimates.
Additional QA information can be found in the 1990-
1993 Virginian Province Quality Assurance Report
(Strobel et al., 1995).
3.2.3.8 Total Organic Carbon
Organic carbon, as measured by EMAP in the sedi-
ments, includes all forms of carbon except carbonate.
Organic carbon accumulates in sediments of the marine
environment as a function of the proximity and magni-
tude of the various sources of organic matter and the
physical, and biological factors that influence erosion
and deposition. Organic matter is an important modifier
of the physical and chemical conditions in the benthic
ecosystem and serves as the primary source of food for
the bottom fauna. As discussed earlier, organic carbon
also plays a critical role in the geochemistry of organic
contaminants in sediments.
The organic carbon content measured in sediments
of the Virginian Province ranged from 0 to 7.01% by
weight. The CDF of percent area as a function of the
total organic carbon present in the sediments for all
estuaries is shown in Figure 3-27. The pattern is largely
determined by the large estuaries (Figure 3-28) which
account for the largest part of the Province area.
Approximately 28 ± 3% of the area of the Province contained
TOC concentrations < 0.2%, resulting in the exclusion
of this area from analyses using Sediment Quality Criteria.
Quality Assurance
Total organic carbon data exist for 396 of the 446
Base Sampling Sites in the database (89%). The 396
sites sampled represent 88% of the area of the Virginian
Province. The current TOC database contains three QA
qualifier flags, two related to non-detects and the detection
limit, and one cautioning the user that the data are suspect
and should be used with caution (the SC-C code). No
TOC data have been assigned the SC-C code. Additional
QA information can be found in the 1990-1993 Virginian
Province Quality Assurance Report (Strobel et al., 1995).
3.2.4 Marine Debris
Anthropogenic debris is perhaps the most obvious
sign of human use and environmental degradation. The
presence of anthropogenic debris in the field of view
or the inconvenience caused when it fouls a boat propeller
or fishing line can diminish the recreational value of
the estuarine environment. "Trash" is most likely to
be found in large tidal rivers and small estuaries where
human settlement and recreational activities are most
intense.
100
234567
Total Organic Carbon (% dry wt)
Figure 3-27. Cumulative distribution of the percent total organic carbon in sediments as a percent of area in the
Virginian Province. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 51
-------
a) Large Estuaries
b) Small Estuaries
c) Large Tidal Rivers
2345
Total Organic Carbon (% dry wt)
Figure 3-28. Cumulative distribution of the percent total organic carbon in sediments by estuarine class: a) Large estuaries,
b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Page 52
Statistical Summary, EMAP-E Virginian Province
-------
"5
0)
Q_
50 -i
40-
30
Province Large Small Tidal
Figure 3-29. The percent of area of the Virginian Province by estuarine class where anthropogenic trash was collected
in fish trawls. (Error bars represent 95% confidence intervals).
The debris collected in bottom trawls was examined
as an indicator of environmental degradation in the
Virginian Province. This indicator was selected over
floating debris because it was felt it was less subjective,
i.e., the collection and observation methodology was
quantitative. Debris was found on the bottom of
approximately 20 ± 3% of the Virginian Province area
sampled (Figure 3-29). The small estuary class had the
largest percent area (35 ± 9%) where trash was found.
Trash was found in 15 ± 4% of the area of the large
estuaries and 28 ± 13% of the area of large tidal rivers.
Quality Assurance
Valid debris data exist for 392 of the 446 Base
Sampling Sites in the database (88%). The 392 sites
sampled represent 88% of the area of the Virginian
Province. Additional QA information can be found in
the 1990-1993 Virginian Province Quality Assurance
Report (Strobel et al, 1995).
3.3 Habitat Indicators
Habitat indicators describe the natural physical and
chemical conditions of the sites sampled in the Virginian
Province study.
3.3.1 Water Depth
The depth distribution in the Virginian Province,
as determined from the boat's fathometer, is shown in
Figure 3-30. The area shallower than 2 m is underestimated
because this is the minimum depth sampled. Based on
the sampling design where a single station represents
a given area, 8% of the area of large estuaries was
unsampleable due to inadequate water depth or inaccessibility.
Small estuaries were considered unsampleable if the water
depth did not exceed 2 m anywhere in the system. Such
systems account for approximately 1.3% of the area of
small systems in the Virginian Province. Overall, 6%
of the area of the Province was deemed unsampleable
due to water depth or inaccessibility.
It should be noted that these data are not normalized
for stage of the tide. The tidal range at EMAP stations
varies from a couple of centimeters in tidal river stations
to approximately two meters at some of the northern
stations.
Statistical Summary, EMAP-E Virginian Province
Page 53
-------
100
10 20 30
Water Depth (meters)
40
50
Figure 3-30. Cumulative distribution of water depth as a percent of area in the Virginian Province. (Dashed lines are
the 95% confidence intervals).
Quality Assurance
Valid water depth data exist for 425 of the 446 Base
Sampling Sites in the database (95%: all stations
visited). The 425 sites sampled represent 94% of the
area of the Virginian Province. Additional QA
information can be found in the 1990-1993 Virginian
Province Quality Assurance Report (Strobel et al,
1995).
3.3.2 Temperature
Bottom water temperature in the Virginian Province
ranged from 12°C to 30°C during the summer sampling
period. The cumulative distribution function of bottom
temperature is shown in Figure 3-31. The lowest bottom
temperatures measured in the Province occurred in a small
estuary at the eastern end of Cape Cod, MA.
Bottom temperature in the small estuaries ranged
from 12°C to 29°C (Figure 3-32b). Large tidal rivers
had a steep CDF (Figure 3-32c) and exhibited the smallest
temperature range (23°C to 30°C).
12 14
16 18 20 22 24
Bottom Temperature (°C)
26 28
30
Figure 3-31. Cumulative distribution of bottom water temperature as a percent of area in the Virginian Province.
(Dashed lines are the 95% confidence intervals).
Page 54
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
10 12 14 16 18 20 22 24
b) Small Estuaries
10 12 14
c) Large Tidal Rivers
100
80
60
40
20 -
0
10 12 14 16 18 20 22 24
Bottom Temperature (°C)
26 28
30
Figure 3-32. Cumulative distribution of bottom water temperature as a percent of area by estuarine class: a) Large estuaries,
b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 55
-------
Quality Assurance
Valid temperature data exist for 415 of the 446 Base
Sampling Sites in the database (93%). The 415 sites
sampled represent 92% of the area of the Virginian
Province. The process of quality assuring water quality
data is lengthy, and includes a review of the CTD
profiles and QA data, automated checks, splitting of
profiles into multiple files, and determining the
appropriate values to report for surface and bottom
measurements. The current water quality database
contains numerous QA qualifier flags resulting from this
process. Additional QA information can be found in
the 1990-1993 Virginian Province Quality Assurance
Report (Strobel et al, 1995).
3.3.3 Salinity
Salinity is determined by freshwater discharge and
seawater intrusion. Salinity in the broad sounds of the
northern extent of the Province is, in general, higher
than salinity in the coastal plain estuaries south of the
Hudson River. The CDF for bottom salinity (Figure 3-
33) reflects the different salinity characteristics of the
large estuarine systems (Figure 3-34).
The CDF for small estuaries (Figure 3-35) is
dominated by small systems in the Chesapeake Bay
which account for most of the area between 12 and
20%c. The low salinity tail of the CDF is due to the
contribution of small river systems, whereas the high
salinity component is due to embayments supplied with
high salinity waters from the northern sounds. The range
of salinities was greatest in small estuaries (0.1 to 32%c),
with the ranges for large estuaries and large tidal rivers
being 4 to 33 and 0.1 to 22%0) respectively (Figure 3-35).
The data show 21 ± 10% of the large tidal river area
to be fresh water (salinity < 0.5%c). Large tidal rivers
also contain the largest oligohaline area (35 ± 10% <
5%c) compared to 9 ± 3% for small estuaries and 1 ±
1% for the large estuaries (Figure 3-35).
Quality Assurance
Valid salinity data exist for 420 of the 446 Base Sampling
Sites in the database (94%). The 420 sites sampled represent
92% of the area of the Virginian Province. The process
of quality assuring water quality data is lengthy, and
includes a review of the CTD profiles and QA data, automated
checks, splitting of profiles into multiple files, and determining
the appropriate values to report for surface and bottom
measurements. The current water quality database contains
numerous QA qualifier flags resulting from this process.
Additional QA information can be found in the 1990-1993
Virginian Province Quality Assurance Report (Strobel
et al., 1995).
100 -
10
15 20
Bottom Salinity (ppt)
25
Figure 3-33. Cumulative distribution of bottom water salinity as a percent of area in the Virginian Province. (Dashed
lines are the 95% confidence intervals).
Page 56
Statistical Summary, EMAP-E Virginian Province
-------
Province Large Small Tidal
Figure 3-34. The percent of area by estuarine class classified as oligohaline (<5%o), mesohaline (5 to 18%o), or
polyhaline (>18%0). (Error bars represent 95% confidence intervals).
3.3.4 pH
The negative log of the hydrogen ion concentration,
or pH, of estuarine and coastal waters, similar to
salinity, depends on the mixing of sea water and fresh
water from land drainage. Sea water is well-buffered,
with its pH usually falling between 8.1 and 8.4. The
pH of fresh water runoff depends upon the characteris-
tics of the land drained and can be quite variable.
The measured pH of Virginian Province estuaries
ranged from 6.3 to 9.2. The range for large estuaries
was 7.1 to 8.6. This range also accounted for 97 ± 1%
of the area of the Province. The lowest pH values
occurred in the freshwater portions of the large tidal
rivers with the lowest values measured in the
Rappahannock River. High pH values were generally
associated with sea water inflow; however, the highest
pH value was found in the upper Potomac River near
Washington DC, and the second highest in upper
Chesapeake Bay (Northeast River). High pH values can
be caused by algal blooms resulting from eutrophication.
Quality Assurance
Valid pH data exist for 410 of the 446 Base Sampling
Sites in the database (92%). The 410 sites sampled represent
92% of the area of the Virginian Province. The process
of quality assuring water quality data is lengthy, and
includes a review of the CTD profiles and QA data, automated
checks, splitting of profiles into multiple files, and determining
the appropriate values to report for surface and bottom
measurements. The current water quality database contains
numerous QA qualifier flags resulting from this process.
Additional QA information can be found in the 1990-1993
Virginian Province Quality Assurance Report (Strobel
et a/., 1995).
3.3.5 Stratification
Vertical density differences (i.e., stratification), if
large enough, can result in a reduction of mixing between
surface and bottom waters, potentially allowing the bottom
waters to become hypoxic. Stratification may also create
conditions that enhance phytoplankton growth, which
might ultimately result in increased biomass settling to
the bottom contributing an additional biological oxygen
demand in the stratified environment.
Statistical Summary, EMAP-E Virginian Province
Page 57
-------
a) Large Estuaries
b) Small Estuaries
c) Large Tidal Rivers
10 15 20 25
Bottom Salinity (ppt)
Figure 3-35. Cumulative distribution of bottom water salinity as a percent of area by estuarine class: a) Large estuaries, b)
Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Page 58
Statistical Summary, EMAP-E Virginian Province
-------
Fresh water runoff can be an important factor in this
process because it both provides low density water
which helps to maintain stratification and often carries
high nutrient concentrations which support plant growth.
Stratification may also be caused by warming of the
surface waters, especially where salinity is uniform.
The development of stratification depends not only on
the magnitude of the density difference between surface
and bottom waters, but also on the depth of those waters
and the physical energy available for mixing. It should
be noted that stratification is affected by many factors
including stage of the tide and recent rainfall events.
The data presented here have not been normalized or
adjusted for any such factors.
Stratification in the Virginian Province is shown as
a CDF of Aat, which is the 2). The bar chart
for stratification by class (Figure 3-37) shows that small
estuaries were least stratified (2 ± 2% with Aat >2) and
best mixed (95 ± 3% with A0, < 1.0). Large estuaries
had the greatest range of Aat(0 to 9.5).
Quality Assurance
Valid surface and bottom salinity and temperature
(and therefore, stratification) data exist for 413 of the
446 Base Sampling Sites in the database (93%). The
413 sites sampled represent 92% of the area of the Virginian
Province. The process of quality assuring water quality
data is lengthy, and includes a review of the CTD profiles
and QA data, automated checks, splitting of profiles into
multiple files, and determining the appropriate values
to report for surface and bottom measurements. The
current water quality database contains numerous QA
qualifier flags resulting from this process.
When multiple visits were made to a station, the data
utilized in the generation of this report are from the first
visit with valid surface AND bottom density measurements.
Additional QA information can be found in the 1990-1993
Virginian Province Quality Assurance Report (Strobel
et al, 1995).
10
Ac.
Figure 3-36. Cumulative distribution of the surface to bottom sigma-t density difference as a percent of area in the Virginian
Province. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 59
-------
100n
Province
Large
Small
Tidal
Figure 3-37. The percent of the area by estuarine class that had a low (<1), medium (1 to 2), or high (>2)
degree of stratification (A o, as kg/m3). (Error bars represent 95% confidence intervals).
3.3.6 Suspended Solids
The amount of suspended matter in the water is
dependent on the physical and biological conditions at
the site. Both the concentration and composition (i.e.,
size distribution and organic vs inorganic origin) of
suspended material affects light extinction and water
clarity and thus the productive and aesthetic qualities
of the water.
The data presented in this section represent surface
values only. Suspended solids concentrations in the waters
of the Virginian Province ranged from 2.7 to 71 mg/L
(Figure 3-38). The relative condition of Virginian Province
waters in large estuary, small estuary, and large tidal
river classes are similar (Figure 3-39).
100
0 20 40 60
Total Suspended Solids (mg/L)
Figure 3-38. Cumulative distribution of total suspended solids concentration as a percent of area in the Virginian
Province, 1991-1993. (Dashed lines are the 95% confidence intervals).
Page 60
Statistical Summary, EMAP-E Virginian Province
-------
a) Large Estuaries
40
60
80
b) Small Estuaries
20
40
60
80
c) Large Tidal Rivers
20 40 60
Total Suspended Solids (mg/L)
80
Figure 3-39. Cumulative distribution of total suspended solids concentration as a percent of area by estuarine class (1991 -1993
only): a) Large estuaries, b) Small estuaries, c) Large tidal rivers. (Dashed lines are the 95% confidence intervals).
Statistical Summary, EMAP-E Virginian Province
Page 61
-------
Quality Assurance
Valid suspended solids data exist for 298 of the 446
Base Sampling Sites in the database (67%). The 298
sites sampled represent 67% of the area of the Virginian
Province. Samples for total suspended solids were
collected at all Base Sampling Sites beginning in 1991.
In 1990 they were collected only at Indicator Testing
and Evaluation Sites, and those data are not included
in this report.
In 1992 three samples were accidentally destroyed
when dust fell in the dried sample pan (and there was
insufficient sample for reanalysis). Overall in 1992,
44% of the TSS results were flagged in the database as
being of questionable quality because of inadequate QA.
Additional QA information can be found in the 1990-
1993 Virginian Province Quality Assurance Report
(Strobel et al, 1995).
3.3.7 Light Extinction
The light extinction coefficient is a measure of the
attenuation of sunlight in the sea. It is the natural
logarithm of the ratio of the intensity of light of a
specified wavelength on a horizontal surface to the
intensity of the same wavelength light on a horizontal
surface 1 m deeper. The extinction coefficient of
photosynthetically active radiation (PAR) was calculated
from depth and PAR measurements made with the
SeaBird CTD. The extinction coefficient is an important
measure of the light available for photosynthesis and
of the aesthetic qualities of the water for human use.
We are defining low water clarity as water in which
a diver would not be able to see his/her hand when held
at arms length. This corresponds to an attenuation coefficient
>2.303 which is equivalent to the transmission of 10%
of the light incident on the surface to a depth of 1 m.
Good water clarity corresponds to an extinction coefficient
of <1.387, which is equivalent to the transmission of
25% of the light incident on the water surface to a depth
of 1 m. In terms of human vision, a wader in water of
good clarity would be able to see his/her feet in waist-deep
water.
Water clarity was good in 81 ± 3% of the sampled
area of the Virginian Province (Figure 3-40). Water of
low clarity was found in 6 ± 2% of the Province and
an additional 13 ± 2% of the Province had water of moderate
clarity. Thus, in 19 ± 2% of the waters in the Virginian
Province waders would not be able to see their toes in
waist deep water. Water of low clarity was found in
3 ± 2% of the large estuarine area, 11 ± 11% of the small
estuarine area, and in 15 ± 13% of the large tidal river
area (Figure 3-41). These differences in water clarity
may be due to fundamental differences in the dynamic
properties of the classes as well as differences in the
intensity of human use. Large estuaries had the greatest
percent area of high water clarity (91 ± 3%).
23456
Light Extinction Coefficient
Figure 3-40. Cumulative distribution of light extinction coefficient as a percent of area in the Virginian Province.
(Dashed lines are the 95% confidence intervals).
Page 62
Statistical Summary,, EMAP-E Virginian Province
-------
100 -
80-
03
< GO-
'S
c
1 40-
Q_
20-
0-
Low
Moderate
Good
Province Large Small Tidal
Figure 3-41. The percent of area by estuarine class where water clarity was poor, moderate, or good. (Error bars
represent 95% confidence intervals).
Quality Assurance
Valid water clarity data exist for 408 of the 446
Base Sampling Sites in the database (91%). The 408
sites sampled represent 91% of the area of the Virginian
Province. The process of quality assuring water quality
data is lengthy, and includes a review of the CTD
profiles and QA data, automated checks, splitting of
profiles into multiple files, and determining the
appropriate values to report for surface and bottom
measurements. The current water quality database
contains numerous QA qualifier flags resulting from this
process.
When multiple visits were made to a station, the
data utilized in the generation of this report are from
the first visit with a valid PAR profile. Additional QA
information can be found in the 1990-1993 Virginian
Province Quality Assurance Report (Strobel et al.,
1995).
3.3.8 Percent Silt-Clay Content
The silt-clay (mud) content of sediments (the
fraction <63|j) is an important factor determining the
composition of the biological community at a site, and
is therefore important in the assessment of the benthic
community. Percent mud is also useful when examining
sediment chemistry data because the available surface
area for sorption of contaminants is partially a function
of grain size, with fine-grained sediments (i.e., mud)
generally being more susceptible to contamination than
sands exposed to the same overlying water.
All silt-clay results presented in this report are for
the surficial sediments (0-2 cm) collected as part of the
chemistry/toxicity homogenate.
The CDF of silt-clay content for the Virginian Province
is shown in Figure 3-42. Forty-six (± 4) percent of the
area contained sandy sediments (<20% silt-clay), and
26 ± 3% of the area had muddy sediments (>80% silt-clay).
The sediment size distribution in large estuaries was dominated
by sands, in small estuaries by muds, and in tidal rivers
it was variable (Figure 3-43).
Sediment size distribution is primarily a result of
the different physical characteristics of the separate system
classes. For example, small systems are often estuaries,
bays, tidal creeks and rivers with low flow rates, which
result in high deposition rates of fine-grained material.
The large area of sandy sediments found in the large
estuaries of the Virginian Province are most likely the
result of either the winnowing of sediments or the transport
of marine sands. The mouth of the Chesapeake Bay is
an example of the latter where sands are carried in from
the ocean (Hobbs et al., 1992). Long Island Sound is
Statistical Summary, EMAP-E Virginian Province
Page 63
-------
100 -f
40 60
Percent Silt/Clay
80
100
Figure 3-42. Cumulative distribution of the percentage of silt/clay in sediments as a percent of area in the Virginian
Province. (Dashed lines are the 95% confidence intervals).
<20
20 to 80
>80
Province Large Small Tidal
Figure 3-43. The percent of area by estuarine class with a low (<20), medium (20 to 80), or high (>80) percent silt/clay
in the sediments. (Error bars represent 95% confidence intervals).
an example of a system where the coarser sediments at
the entrance are mainly a result of strong tidal currents
transporting away the fine fraction (winnowing), leaving
behind the coarser sands and gravel (Akapati, 1974;
Gordon, 1980).
Quality Assurance
Valid sediment grain size data exist for 394 of the
446 Base Sampling Sites in the database (88%). The
394 sites sampled represent 88% of the area of the Virginian
Province. The current sediment grain size database contains
no QA qualifier flags. Additional QA information can
be found in the 1990-1993 Virginian Province Quality
Assurance Report (Strobel et al., 1995).
Page 64
Statistical Summary, EMAP-E Virginian Province
-------
SECTION 4
SUMMARY OF FINDINGS
Thousands of pieces of information on the condition
of estuarine resources in the Virginian Province were
collected and analyzed. The major findings of the four-
year Virginian Province Demonstration Project are
highlighted in this section.
4.1 Virginian Province Fact Summary
The Virginian Province includes the coastal
region of the Northeast United States from Cape
Cod south to the mouth of Chesapeake Bay.
It is composed of 23,574 km2 of estuarine
resources including 11,469 km2 in Chesapeake
Bay and 3,344 km2 in Long Island Sound.
Estuarine resources in the Virginian Province
were stratified into classes for purposes of
sampling and analysis. The classes and their
areal extent are as follows: Large estuaries,
16,097 km2; small estuaries, 4,875 km2; and
large tidal rivers, 2,602 km2.
The large estuary class includes Chesapeake Bay,
Delaware Bay, Long Island Sound, Block Island
Sound, Buzzards Bay, Narragansett Bay, and
Nantucket Sound.
The large tidal river class includes the James,
Rappahannock, Potomac, Delaware, and Hudson
Rivers.
The small estuary class includes 144 estuarine
systems of various types between 2.6 and 260
km2 in area.
4.2 Findings of the Virginian Province
Demonstration Project
Of the 446 Base Sampling Sites initially selected,
all but 21 (4.7%) were successfully sampled.
The majority of the data collected at these
stations met the quality control standards set by
the Program. Overall six percent of the area of
the Province could not be sampled because of
inadequate water depth (< 2m) or inaccessibility.
Because all samples could not be collected from
all sampled sites (e.g., sediment could not be
collected when the bottom was rocky), this value
is higher for selected indicators.
A benthic index was developed to discriminate
between good and poor environmental conditions.
Based on this index, approximately 25 ± 3% of
the Province area could be classified as potential-
ly degraded relative to the benthic community.
Bottom dissolved oxygen concentrations <2 mg/L
were measured at stations representing 5 ± 2%
of the Province area. Concentrations <5 mg/L
were measured in 25 ±3% of the area of the
Province.
Draft EPA Sediment Quality Criteria (SQC) are
currently available for four of the analytes
EMAP measures in sediments: acenaphthene,
phenanthrene, fluoranthene, and dieldrin. SQC
were exceeded (for PAHs) at three small estuary
stations representing 0.07 percent of the area of
the Virginian Province. The SQC for dieldrin
was not exceeded at any station.
Statistical Summary, EMAP-E Virginian Province
Page 65
-------
Sediments collected from stations representing
approximately 46 ± 5% of the Province area
were determined to contain elevated levels of
metals; however, metals concentrations high
enough to potentially result in ecological effects
were measured in only 4 ± 2% of the area of
the Province based on exceedence of ER-M
values.
Table 4-1 summarizes the data presented in
Section 3 for selected Biotic Condition, Abiotic
Condition, and Habitat indicators.
Table 4-1. Percent area of the Virginian Province (with 95% confidence intervals) above or below values of
interest for selected indicators.
Percent area
Estuarine Condition
Province
Large
Large Tidal
Estuary River
Small
Estuary
Benthic Index
<0
23 ±3
18 ±4
33 ± 14
35 ±6
Total Benthic Abundance
<200 / m2
<500 / m2
Bottom DO
<2 mg/L
<5 mg/L
Sediment Toxicity
(% control survival)
<80%
<60%
7 ±2
9 ±2
5±2
25 ± 3
1 0 ± 2
2 ± 1
5±2
7 ± 3
6 ±2
28 ±5
1 0 ± 3
2 ± 1
15 ± 9
24 ± 11
10 ±6
18 ±6
3 ± 4
1 ± 1
8 ± 3
10 ±3
0.2 ± 1.3
21 ± 11
12 ± 6
3 ±4
Enriched metals
any metal
above background 46 ± 5 37 ± 7
Marine Debris
presence 20 ± 3 15 ± 4
Salinity
Polyhaline(>18%o) 66 ± 3 78 ± 4
Mesohaline (5 to 18%0) 28 ± 3 21 ± 4
Oligohaline (< 5%«) 6 ± 1 1 ± 1
69 ± 16
28 ± 13
7 ±4
58 ±4
35 ± 10
64 ±3
35 ± 9
57 ± 10
34 ± 10
9 ± 3
Page 66
Statistical Summary, EMAP-E Virginian Province
-------
SECTION 5
REFERENCES
Adelman, D., K.R. Hinga and M.E.Q. Pilson. 1990. Biogeochemistry of butyltins in an enclosed marine
ecosystem. Environ. Sci. Technol. 24: 1027-1032.
Akapati, B.N. 1974. Mineral composition and sediments in eastern Long Island Sound. Maritime Seds. 10:
19-30.
ASTM (American Society of Testing and Materials). 1991. Standard guide for conducting 10-day static
sediment toxicity tests with marine and estuarine amphipods. Annual Book of ASTM Standards Volume
11.04:1052-1015.
Ballschmiter , K., and M. Zell. 1980. Analysis of polychlorinated biphenyls (PCBs) by glass capillary gas
chromatography. Fresenius Z. Anal. Chem. 302: 20-31.
Bilyard, G.R. 1987. The value of benthic infauna in marine pollution monitoring studies. Mar. Poll. Bull.
18:581-585.
Boesch, D.F. and R. Rosenberg. 1981. Response to stress in marine benthic communities. In: G.W. Barret
and R. Rosenberg, eds., pp. 179-200. Stress Effects on Natural Ecosystems. New York: John Wiley and
Sons.
Cochran, W. G. 1977. Sampling Techniques. 3rd edition. John Wiley, New York.
DiToro, D.M., C.S. Zarba, D.J. Hansen, W.J. Berry, R.C. Swartz, C.E. Cowan, S.P. Pavlou, H.E. Allen, N.A.
Thomas, and P.R. Paquin. 1991. Technical basis for establishing sediment quality criteria for nonionic
organic chemicals using equilibrium partitioning. Environ. Toxicol. and Chem. 10:1541-1583.
Forstner, U. and G.T.W. Wittmann. 1981. Metal pollution in the aquatic environment. 2nd revised edition.
New York: Springer-Verlag.
Gebhart, J.E., T.L. Hayes, A.L. Alford-Stevens and W.L. Budde. 1985. Mass spectrometric determination of
polychlorinated biphenyls as isomer groups. Anal. Chem. 57: 2458-2463.
Gordon, R.B. 1980. The sedimentary system of Long Island Sound. Advances in Geophysics 22: 1-39.
Statistical Summary, EMAP-E Virginian Province Page 67
-------
Heard, C.S., W.W. Walker and W.E. Hawkins. 1989. Aquatic lexicological effects of organotins: An
overview. Proceedings, pp. 554-563. Oceans '89 Conference and Exposition on Science and Engineering.
Washington, DC: Institute of Electrical and Electronics Engineers.
Hinga, K.R. 1988. Seasonal predictions for pollutant scavenging in two coastal environments using a model
calibration based upon thorium scavenging. Mar. Environ. Res. 26:97-112.
Hobbs, C.H., III, J.P. Halka, R.T. Kerhin, and M.J. Carron. 1992. Chesapeake Bay sediment budget. J.
Coast. Res. 8(2): 292 - 300.
Holland, A.F., ed. 1990. Near Coastal Program Plan for 1990: Estuaries. EPA 600/4-900/033.
Narragansett, RI: U.S. Environmental Protection Agency, Environmental Research Laboratory, Office of
Research and Development.
Holland, A.F., A.T. Shaughnessy, L.C. Scott, V.A. Dickens, J.A. Ranasinghe, and J.K. Summers. 1988.
Progress report: Long-term benthic monitoring and assessment program for the Maryland portion of
Chesapeake Bay (July 1986-October 1987). PPRP-LTB/EST-88-1. Prepared for Maryland Power Plant
Research Program and Maryland Department of the Environment, Office of Environmental Programs.
Columbia, MD: Versar, Inc., ESM Operations.
Holland, A.F., A.T. Shaughnessy, L. C. Scott, V.A. Dickens, J. Gerritsen, and J.A. Ranasinghe. 1989.
Long-term benthic monitoring and assesssment program for the Maryland portion of Chesapeake Bay:
Intepretive report. Columbia, MD: Versar, Inc. for Maryland Department of Natural Resources, Power
Plant Research Program. CBRM-LTB/EST-2.
Huggett, R.J., M.A. Unger, P.P. Seligman and A.O. Valkirs. 1992. The marine biocide tributyltin. Environ.
Sci. Technol. 26: 232-237.
Hunsaker, C. and D. Carpenter, eds. 1990. Ecological indicators for the Environmental Monitoring and
Assessment Program. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
Research and Development. EPA 600/3-90/060.
Hutzinger, O, S. Safe and V. Zitko. 1974. The Chemistry of PCBs. Cleveland, OH: CRC Press. 269pp.
Karr, J.R., and D.R. Dudley. 1981. Ecological perspective on water quality goals. Environ. Manage. 5:55-
68.
Karr, J.R., K.D. Fausch, P.L. Angermeier, P.R. Yant, and I.J. Schlosser. 1986. Assessing biological
integrity in running waters: a method and its rationale. Special Publication 5. Champaign, II: Illinois
Natural History Survey.
Kemp, W.M. and W.R. Boynton. 1980. Influence of biological and physical processes on dissolved oxygen
dynamics in an estuarine system: Implications for measurements of community metabolism. Estuarine
and Coastal Mar. Sci. 11:407-431.
Knapp, C.M., D.R. Marmoreck, J.P. Baker, K.W. Thornton, J.M. Klopateck, and D.F. Charles. 1990. The
indicator development strategy for the Environmental Monitoring Assessment Program. Washington, DC:
U.S. EPA Office of Research and Development, EPA 600/3-91/023.
Page 68 Statistical Summary, EMAP-E Virginian Province
-------
Laflamme, R.E. and R.A. Kites. 1978. The global distribution of polycyclic aromatic hydrocarbons in recent
sediments. Geochimica et Cosmochimica Acta 42: 289-303.
Lake, J.L., C. Norwood, C. Dimock and R. Bowen. 1979. Origins of polycyclic aromatic hydrocarbons in
estuarine sediments. Geochimica et Cosmochimica Acta 43: 1847-1854.
Long, E.R., D.D. MacDonald, S.L. Smith, and F.D. Calder. 1995. Incidence of adverse biological effects
within ranges of chemical concentrations in marine and estuarine sediments. Environ. Mgt. 19(1): 81-97.
Long, E.R. and L.G. Morgan. 1990. The potential for biological effects of sediment-sorbed contaminants
tested in the National Status and Trends Program. NOAA Technical Memorandum NOS OMA 52.
Rockville, MD: US Departement of Commerce, National Oceanic and Atmospheric Administration,
National Ocean Service.
McFarland, V.A. and J.U. Clarke. 1989. Environmental occurrence, abundance, and potential toxicity of
polychlorinated biphenyl congeners: Considerations for a congener-specific analysis. Environ. Health
Perspectives 81: 225-239.
Messer, J.J. 1990. EMAP Indicator Concepts. In: C.T. Hunsaker and D.E. Carpenter, eds., Ecological
Indicators for the Environmental Monitoring and Asssessment Program. EPA 600/3-90/060. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and Development.
Mullin, M.D., C.M. Pochini, S. McCrindle, M. Romkes, S.H. Safe and L.M. Safe. 1984. High-resolution
PCB analysis: Synthesis and chromatographic properties of all 209 PCB congeners. Environ. Sci. Technol.
18: 468-476.
Nixon, S.W., C.D. Hunt, and B.L. Nowicki. 1986. The retention of nutrients (C,N,P), heavy metals (Mn,
Cd, Pb, Cu), and petroleum hydrocarbons in Narragansett Bay. In: P. Lasserre and J.M. Martin, eds., pp.
99-122. Biogeochemical Processes at the Land-sea Boundary. New York: Elsevier.
Overton, W.S., D.L. Stevens and D. White. 1991. Design Report for EMAP, Environmental Monitoring and
Assessment Program. Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research
Laboratory.
Paul, J.F., J.H. Gentile, S.C. Schimmel, K.J. Scott, and D.E. Campbell, (in preparation). Assessment of
Estuarine Conditions in the Virginian Province Using 1990-93 EMAP Data. Narragansett, RI: U.S.
Environmental Protection Agency, Office of Research and Development.
Pearson, T.H. and R. Rosenberg. 1978. Macrobenthic succession in relation to organic enrichment and
pollution of the marine environment. Oceanogr. Mar. Biol. Ann. Rev. 16:229-311.
Plumb, R.H. 1981. Procedure for handling and chemical analysis of sediment and water samples. Technical
Report EPA/CE-81-1. Prepared for the U.S. Environmental Protection Agency/Corps of Engineers
Technical Committee on Criteria Dredge and Fill Material. Vicksburg, MS: Environmental Laboratory,
U.S. Army Waterways Experiment Station.
Reifsteck, D.M., C.J. Strobel, and D.J. Keith. 1993. EMAP-Estuaries 1993 Virginian Province Field
Operations and Safety Manual. Narragansett, RI: U.S. Environmental Protection Agency, Office of
Research and Development, June 1993.
Statistical Summary, EMAP-E Virginian Province Page 69
-------
Rexrode, M. 1987. Ecotoxicity of tributyltin. Proceedings, pp. 554-563. Oceans '87 Conference and
Exposition on Science and Engineering. Washington, DC: Institute of Electrical and Electronics
Engineers.
Rhoads, D.C. 1974. Organism-sediment relations on the muddy sea floor. Oceanogr. Mar. Biol. A. Rev.
12:263-300.
Rhoads, D.C., P.L. McCall, and J.Y. Yingst. 1978. Disturbance and production on the estuarine sea floor.
Amer. Sci. 66:577-586.
Rosen, J.S., J. Beaulieu, M. Hughes, H. Buffum, J. Copeland, R. Valente, J. Paul, F. Holland, S. Schimmel,
C. Strobel, K. Summers, K.J. Scott, and J. Parker. 1990. Data base management system for coastal
demonstration project. Narragansett, RI: EPA Office of Research and Development. (Internal report).
Schantz, M.M., B.A. Benner, Jr., S.N. Chesler, B.J. Koster, K.E. Hehn, S.F. Stone, W.R. Kelly, R. Zeisler
and S.A. Wise. 1990. Preparation and analysis of a marine sediment reference material for the
determination of trace organic constituents. Fresenius J. Anal. Chem. 338: 501-514.
Schimmel, S.C. 1990. Implementation Plan for Environmental Monitoring and Assessment Program Near
Coastal Demonstration Project. Narragansett, RI: U.S. Environmental Protection Agency, Environmental
Research Laboratory, Office of Research and Development.
Schimmel, S.C., B.D. Melzian, D.E. Campbell, C.J. Strobel, S.J. Benyi, J.S. Rosen, and H.W. Buffum. 1994.
Statistical Summary: EMAP-Estuaries Virginian Province - 1991. EPA/620/R-94/005 Narragansett, RI:
U.S. Environmental Protection Agency, Environmental Research Laboratory, Office of Research and
Development.
Schropp, S.J., F.G. Lewis, H.L. Windom, J.D. Ryan, F.D. Calder, and L.C. Bumey. 1990. Interpretation of
metal concentrations in estuarine sediments of Florida using aluminum as a reference element. Estuaries
13:227-235.
Schubel, J.R. and H.H. Carter. 1984. The estuary as a filter for the fine-grained suspended sediment. In:
V.S. Kennedy, ed., pp. 81-104. The Estuary as a Filter. Orlando, FL: Academic Press.
Seligman, P.P., J.G. Grovhoug, A.O. Valkirs, P.M. Stang, R.Fransham, M.O. Stallard, B. Davidson and R.F.
Lee. 1989. Distribution and fate of tributyltin in the United States marine environment. Applied
Organometalic Chem. 3: 31-47.
Stevens, D.L., Jr., A.R. Olsen, D. White. 1991. Environmental Monitoring and Assessment Program --
integrated sampling design. Draft report. Corvallis, OR: Environmental Research Laboratory, U.S.
Environmental Protection Agency.
Strobel, C.J., S.J. Benyi, D.J. Keith, H.W. Buffum and E.A. Petrocelli. 1994. Statistical Summary: EMAP-
Estuaries Virginian Province - 1992. EPA/620/R-94/019 Narragansett, RI: U.S. Environmental
Protection Agency, Environmental Research Laboratory, Office of Research and Development.
Strobel, C.J., Valente, R.M., and D.J. Keith. 1995. Quality Assurance Report: EMAP-Virginian Province,
1990-1993. Narragansett, RI: U.S. Environmental Protection Agency, Environmental Research
Laboratory, Office of Research and Development (in review).
Page 70 Statistical Summary, EMAP-E Virginian Province
-------
Strobel, C.J., D.M. Reifsteck, and S.C Schimmel. 1992. Environmental Monitoring and Assessment Program
EMAP'-Estuaries, Virginian Province Logistics Plan for 1992. Narragansett, RI: U.S. Environmental
Protection Agency, Environmental Research Laboratory, Office of Research and Development, January
1992.
Swartz R.C., W.A. DeBen, J.K. Jones, J.O. Lamberson, and F.A. Cole. 1985. Phoxocephalid amphipod
bioassay for marine sediment toxicity. In: R.D. Cardwell, R. Purdy, and R.C. Banner, eds., pp. 284-307.
Aquatic Toxicology and Hazard Assessment: Seventh Symposium. Philadelphia, PA: American Society for
Testing and Materials.
Terrell, T.T. 1979. Physical regionalization of coastal ecosystems of the United States and its territories.
FWS/OBS-79/80. Office of Biological Services, U.S. Fish and Wildlife Service.
Turekian, K.K. 1977. The fate of metals in the oceans. Geochimica et Cosmochimica Acta 41:1139-1144.
U.S. EPA. 1989. Briefing Report to the EPA Science Advisory Board on the equilibrium portioning
approach to generating sediment quality criteria. EPA 440/5-89-002. Washington, DC: U.S. EPA,
Criteria and Standards Division.
U.S. EPA/ACE. 1991. Evaluation of dredged material proposed for ocean disposal (Testing manual).
Prepared by the U.S. Environmental Protection Agency, Office of Marine and Estuarine Protection and
Department of the Army, United States Army Corps of Engineers, February 1991.
U.S. EPA. 1993a. Proposed Sediment Quality Criteria for the Protection of Benthic Organisms:
Acenaphthene. Washington DC: U.S. Environmental Protection Agency, Office of Science and
Technology. In Review.
U.S. EPA. 1993b. Proposed Sediment Quality Criteria for the Protection of Benthic Organisms:
Phenanthrene. Washington DC: U.S. Environmental Protection Agency, Office of Science and
Technology. In Review.
U.S. EPA. 1993c. Proposed Sediment Quality Criteria for the Protection of Benthic Organisms:
Fluoranthene. Washington DC: U.S. Environmental Protection Agency, Office of Science and
Technology. In Review.
U.S. EPA. 1993d. Proposed Sediment Quality Criteria for the Protection of Benthic Organisms: Dieldrin.
Washington DC: U.S. Environmental Protection Agency, Office of Science and Technology. In Review.
U.S. EPA. 1995. EM AP-Estuaries Laboratory Methods Manual, Volume I - Biological and Physical
Analyses. Cincinnati, OH: U.S. Environmental Protection Agency, Environmental Monitoring Systems
Laboratory, Office of Research and Development.
Valente, R., and C.J. Strobel. 1993. EM AP-Estuaries Virginian Province 1993 Quality Assurance Project
Plan. Narragansett, RI: U.S. Environmental Protection Agency, Office of Research and Development,
Environmental Research Laboratory. July 1993.
Warwick, R.M. 1980. Population dynamics and secondary production in benthos. In: K.R. Tenore and B.C.
Coull, eds., Marine Benthic Dynamics, Belle W. Baruch Library in Science Number 11. Columbia, SC:
University of South Carolina Press.
Statistical Summary, EMAP-E Virginian Province Page 71
-------
Washington, H.G. 1984. Diversity, biotic and similarity indices, a review with special relevance to aquatic
ecosystems. Water Resources, 18( 6): 653-694.
Weisberg, S.B., J.B. Frithsen, A.F. Holland, J.F. Paul, K.J. Scott, J.K. Summers, H.T. Wilson, R.M.Valente,
D.G. Heimbuch, J. Gerritsen, S.C. Schimmel, and R.W. Latimer. 1993. EMAP-Estuaries, Virginian
Province 1990 Demonstration Project Report. EPA/620/R-93/006. Narragansett, RI: U.S. Environmental
Protection Agency, Environmental Research Laboratory, Office of Research and Development.
Welsh, B.L. and F.C. Eller. 1991. Mechanisms controlling summertime oxygen depletion in Western Long
Island Sound. Estuaries 14:265-278.
Windom, H.L., S.J. Schropp, F.D. Calder, J.D. Ryan, R.G. Smith, L.C. Bumey, F.G. Lewis, and C.H.
Ralinson. 1989. Natural trace metal concentrations in estuarine and coastal marine sediments of the
southeastern United States. Environ. Sci. Technol. 23:314-320.
Windsor, J.G., Jr. and R.A. Hites. 1979. Polycyclic aromatic hydrocarbons in Gulf of Maine sediments and
Nova Scotia soils. Geochimica et Cosmochimica Acta 43: 27-33.
Page 72 Statistical Summary, EMAP-E Virginian Province
-------
APPENDIX A
SAMPLING DESIGN, ECOLOGICAL INDICATORS,
AND METHODS
A.I Region and Estuarine Classification
EMAP-E monitoring is conducted on regional and
national scales. Standardized methods are employed,
and the entire Virginian Province is sampled
synoptically within a defined "index" time period to
ensure comparability of data within and among sampling
years. EMAP-E identified boundaries for 12 estuarine
regions (Holland, 1990) based on biogeographic
provinces defined previously by NOAA and the U.S.
Fish and Wildlife Service using major climatic zones
and prevailing major ocean currents (Terrell, 1979)
(Figure A-l). The 1990-1993 Virginian Province
Demonstration Project included the estuarine resources
located along the irregular coastline of the mid-Atlantic
coast between Cape Cod, MA and Cape Henry, VA,
including: Buzzards Bay, Narragansett Bay, Long Island
Sound, New York/New Jersey Harbors, Delaware Bay,
and Chesapeake Bay. Five major rivers within the
Province were monitored: the Hudson, the Delaware,
the Rappahannock, the Potomac, and the James.
A review of the literature identified potential
classification variables that reduced within-class
variability. These variables included physical attributes
(salinity, sediment type, depth), and extent of pollutant
loadings. The use of salinity, sediment type, and
pollutant loadings as classification variables (i.e., a
priori strata) would result in the definition of classes
for which areal extents could vary dramatically from
year-to-year or even over the index sampling period of
EMAP-E. This stratification process requires
establishment of a sampling frame prior to sampling;
thus misclassification of sample sites within a class
should be minimized. Stratification by sediment type,
depth, or salinity was considered to be difficult because
detailed maps of sediment and water column characteristics
were not available or are often unreliable for much of
the Virginian Province. These attributes were not used.
A simple classification scheme based on the physical
dimensions of an estuary was used to develop three classes -
- large estuaries, large tidal rivers, and small estuaries/small
tidal rivers. Large estuaries in the Virginian Province
were defined as those estuaries greater than 260 km2 in
surface area and with aspect ratios (i.e., length/average
width) of less than 18. Large tidal rivers were defined
as that portion of the river that is tidally influenced (i.e.,
detectable tide > 2.5 cm), greater than 260 km2, and with
an aspect ratio of greater than 18. Small estuaries and
small tidal rivers were designated as those systems whose
surface areas fell between 2.6 km2 and 260 km2. These
designations excluded estuarine water bodies less than
2.6 km2 in surface area. These resources were included
in the sampling frame by incorporating them into the
adjacent water body, but they were not sampled separately.
Application of the classification scheme based upon
geometric dimensions (criteria unlikely to change in reasonable
time frames) to the Virginian Province estuarine resources
resulted in the identification of 12 large estuaries; 5 large
tidal rivers; and 144 small estuaries / small tidal rivers.
Statistical Summary, EMAP-E Virginian Province
Page A - 1
-------
Columbian
Virginian
Acadian
Carolinian
West Indian
Figure A-1. EMAP-Estuaries biogeographical provinces.
A.2 Sampling Design
Sample collection in the Virginian Province focused
on ecological indicators (see Section A.3) during the
index sampling period (late July through September),
the period when many estuarine responses to
anthropogenic and natural stresses are anticipated to be
most severe. The sampling design employed combines
the strengths of systematic and random sampling with
an understanding of estuarine ecosystems in order to
provide a probability-based estimate of estuarine status
in the Virginian Province. In addition, some special-
study sites were sampled to collect information for spe-
cific hypothesis testing and other specific study
objectives. This resulted in sampling seven types of
sampling sites (stations) for the Virginian Province
survey.
Base Sampling Sites (BSS) are the probability-based
sites which form the core of the EMAP-E monitoring
design for all provinces, including the Virginian
Province. Data collected from these sites are the basis
of this preliminary status assessment. There were 446
BSS to be sampled during the four-year project.
Index Sites (IND) were a continuation of a special
study initiated in 1990. They are associated with the
base sampling sites in small estuarine systems and
large tidal rivers and are located in depositional
environments where there is a high probability of sediment
contamination or low dissolved oxygen conditions.
A total of 86 IND siles were monitored.
Long-term Trend Sites (LTT) were a select number
of 1990 BSS that were revisited in 1991 through 1993.
They were sampled each year to investigate the within-
station annual variability. Twelve LTT sites were
monitored (only 11 in 1991).
Long-term Trends Spatial Transect (LTS) sites were
located along a transect originating at selected large-estuary
LTT stations. Twelve (12) LTS sites were associated
Page A - 2
Statistical Summary, EMAP-E Virginian Province
-------
with four transects in 1991 and were all located in
the Chesapeake Bay. Three LTS sites were placed
along each transect at 0.25,0.5, and 1.0 statute miles
from the associated LTT station to evaluate the
spatial variability within a sampling cell. LTS sites
were monitored only in 1991.
Indicator Testing and Evaluation (ITE) sites were
sampled to determine the reliability, sensitivity, and
replicability of indicator responses for discriminating
between sites with "known" environmental
conditions. These sites were selected on the basis
of historical information concerning dissolved oxygen
concentration and sediment contamination. These
sites were used to develop indices and test the
discriminatory power of specific indicators. The
number of ITE sites varied with year with some being
revisited over multiple years. A total of 22 sampling
events at ITE stations were conducted over the four-
year period.
Replicate sampling sites (REP) were randomly
located in small estuarine systems. Replicate stations
were designed to provide information on within-
system spatial variability. Twenty-nine REP sites
were sampled as part of the Demonstration Project.
Supplemental sites (SUPP) were sites located in the
Delaware Bay (large estuary) in 1990. These sites
were selected using the same design used to locate
large estuary sites, but on a smaller scale. The
purpose of supplemental sites was to investigate the
effect of scale on the design. A total of 24
supplemental sites were sampled in 1990.
A.3 Indicators
EMAP monitoring focuses on indicators of biological
response to stress, and uses measures of exposure to
stress or contamination as a means for interpreting that
response. Traditionally, estuarine monitoring has
focused on measures of exposure (e.g., concentrations
of contaminants in sediments) and attempted to infer
ecological impacts based on laboratory bioassays. The
advantage of the ecologically-based approach
emphasized in EMAP is that it can be applied to
situations where multiple stressors exist, and where
natural processes cannot be modeled easily. This is
certainly the case in estuarine systems, which are subject
to an array of anthropogenic inputs and exhibit a great
biotic diversity and complex physical, chemical, and biological
interactions.
The implementation plan for the Virginian Province
(Schimmel, 1990) listed three general indicator categories
for the Demonstration Project: core, developmental and
research. Table A-l lists the EMAP-Virginian Province
indicators.
Table A-1. Ecological indicators used in the Virginian
Province Survey.
Category
Core
Developmental
Research
Indicator
Benthic Species Composition & Biomass
Habitat Indicators
(Salinity, pH, Temperature,Water
Depth, % Silt-Clay)
Sediment Contaminants
Sediment Toxicity
Dissolved Oxygen Concentration
Gross Pathology of Fish
Marine Debris
Fish Community Composition & Lengths
Water Clarity
Histopathology of Fish
Fish Biomarkers (1990 and 1991)
Contaminants is Fish Tissue (1991 only)
A.4 Indicator Sampling Methods
The EMAP indicator strategy involves four types of
ecological indicators (Hunsaker and Carpenter, 1990;
Knapp et al., 1990): Biotic condition, abiotic condition,
habitat, and stressor. Biotic condition indicators are ecological
characteristics that integrate the responses of living resources
to specific or multiple pollutants and other stresses, and
are used by EMAP to assess overall estuarine condition.
Abiotic condition indicators quantify pollutant exposure
and habitat degradation and are used mainly to identify
associations between stresses on the environment and
degradation in biotic condition indicators. Habitat indicators
provide basic information about the natural environmental
gradients. Stressor indicators are used to quantify pollution
inputs or stresses and identify the probable sources of
pollution exposure. Tables A-2 and A-3 list individual
indicators.
Statistical Summary, EMAP-E Virginian Province
Page A - 3
-------
Table A-2. Ecological indicators categorized as biotic
condition, abiotic condition, and habitat indicators.
Table A-3. Subcomponents of ecological indicators.
Indicator Type
Biotic Condition
Abiotic Condition
Habitat
Indicator
Benthio Community Composition
Benthic Abundance
Benthic Biomass
Fish Community Composition
Fish Lengths
Pathology in Fish
Sediment Contaminants
Sediment Toxicity
Dissolved Oxygen Concentrations
Marine Debris
Water Clarity
Salinity
Temperature
Percent Silt-Clay
PH
Water Depth
Descriptions of the methods used for individual
indicators have been taken from the Near Coastal
Program Plan (Holland, 1990), the Virginian Province
Implementation Plan (Schimmel, 1990), the 1993
Virginian Province Field Operations and Safety Manual
(Reifsteck et al., 1993), and the EMAP-E Laboratory
Methods Manual (USEPA, 1995).
A.4.1 Biotic Condition Indicators
A.4.1.1 Benthos
Benthic invertebrate assemblages are composed of
diverse taxa with a variety of reproductive modes,
feeding guilds, life history characteristics, and
physiological tolerances to environmental conditions
(Warwick, 1980; Bilyard, 1987). As a result, benthic
populations respond to changes in conditions, both
natural and anthropogenic, in a variety of ways (Pearson
and Rosenberg, 1978; Rhoads et al., 1978; Boesch and
Rosenberg, 1981). Responses of some benthic
organisms indicate changes in water quality while others
indicate changes in sediment quality. Most benthic
organisms have limited mobility. They are not as able
to avoid exposure to pollution stress as many other
estuarine organisms (e.g., fish). Benthic communities
have proven to be a reasonable and effective indicator
of the extent and magnitude of pollution impacts in
Primary Indicator Subcomponents
Benthos
Fish
Total abundance
Species composition
Species diversity
Abundance by species
Percentage by taxonomic group
Biomass
Biomass by taxonomic group
Total abundance
Species composition
Species diversity
Abundance by species
Percentage by taxonomic group
Mean length by species
Gross Pathology Type of disorder
Dissolved Instantaneous at sampling
Oxygen Continuous for 24-hr (15-min intervals)
Sediment Toxicity Ampelisca abdita 10-day test
Sediment 23 polycyclic aromatic hydrocarbons
Contaminants 15 metals
15 pesticides
18 PCB congeners
Butyltins
Sediment Percent silt-clay
Characters Acid Volatile Sulfides (AVS: 1991 -1993)
Total organic carbon (TOC)
estuarine environments (Bilyard, 1987; Holland et al.,
1988 and 1989).
Benthic samples for evaluation of species composition,
abundance, and biomass were collected at all sampling
sites. Samples were collected with a Young-modified
van Veen grab that samples a surface area of 440 cm2.
Three (3) grabs were collected at each base, index, or
long-term site. A small core (60 cc) was taken from
each grab for sediment characterization. The remaining
sample was sieved through a 0.5 mm screen using a backwash
technique that minimized damage to soft-bodied animals.
Samples were preserved in 10% formalin-rose bengal
solution and stored for at least 30 days prior to processing
to assure proper fixation.
In the laboratory, macrobenthos were transferred from
formalin to an ethanol solution and sorted, identified
to lowest practical taxonomic level, and counted. Biomass
was measured for key taxa and all other taxa were grouped
according to taxonomic type (e.g., polychaetes, amphipods,
Page A - 4
Statistical Summary, EMAP-E Virginian Province
-------
decapods). Shell-free dry weight was determined using
an analytical balance with an accuracy of 0.1 mg after
drying at 60°C. Large bivalves were shucked prior to
determining biomass. Smaller shells were removed by
acidification using a 10% HC1 solution.
A.4.1.2 Fish
There are several advantages to using fish as a
potential indicator of estuarine condition. Because of
their dominant position at the upper end of the estuarine
food web, fish responses integrate many short-term and
small-scale environmental perturbations. Fish are
known to respond to most environmental problems of
concern in estuaries, including eutrophication, habitat
modification, and pathogenic or toxic contamination.
Fish were collected by trawling with a 15 m, high-
rise otter trawl with a 2.5-cm mesh cod end. The net
was towed for 10 minutes against the tide (if significant
tidal current existed) between 0.5 and 1.5 m/s (1-3
knots). All fish caught in the trawl were identified to
species and counted; up to 30 fish of a species from
each collection were measured to the nearest millimeter.
Individuals collected in standard trawls were
inspected for gross external pathological disorders at
all stations where fish were collected. This inspection
included checking body surface and fins for lumps,
growths, ulcers and fin erosion. In 1991 only target
species (Table A-4) were examined. Specimens with
observed gross pathologies were preserved in Dietrich's
solution for subsequent laboratory verification and
histological examination. At indicator testing sites, all
specimens exhibiting gross pathologies, and up to 25
pathology-free specimens of each target species (and
10 of non-target species), were preserved for quality
control checks of field observations. These fish also
received histopathological examinations related to liver
lesions, spleen macrophage aggregates, and gill or
kidney disfunction (research indicator).
Table A-4. 1991 target species examined for external
pathology and saved for chemical residue analysis.
Common Name
Atlantic Croaker
Bluefish
Channel Catfish
Scup
Spot
Summer Flounder
Weakfish
White Catfish
White Perch
Winter Flounder
Scientific Name
Micropogonias undulatus
Pomatomus saltatrix
Ictalurus punctatus
Stenotomus chrysops
Leiostomus xanthurus
Paralichthys dentatus
Cynoscion regalis
Ameiurus catus
Morone americana
Pleuronectes americanus
A.4.2 Abiotic Condition Indicators
A.4.2.1 Sediment Collection Procedures
Sediments were collected using the same Young-modified
van Veen grab used for benthic invertebrate sampling.
The top 2 cm of 6 to 10 grabs were placed in a mixing
bowl and homogenized. After approximately 2,000 cc
of sediment were collected and completely homogenized,
the sediment was distributed among containers for sediment
characterization, sediment chemistry, and sediment toxicity
testing.
A.4.2.2 Sediment Characterization
The physical characteristics of estuarine sediments
(e.g., grain size) and certain chemical aspects of sediments
(e.g., acid volatile sulfide [AVS] content, total organic
carbon [TOC] content) influence the distribution of benthic
fauna and the accumulation of contaminants in sediments
(Rhoads, 1974; Plumb, 1981; DiToro et al, 1991). Sediment
silt-clay content was determined to help interpret biotic
condition indicator data and sediment contaminant
concentrations. AVS and TOC were collected not only
as interpretive aids but also as potential covariates for
toxic contaminant concentrations.
Subsamples from each benthic grab and contaminant/
toxicity homogenate were retained for grain size determination.
A sample for determination of AVS content was removed
from the homogenate (1991) or directly from each grab
being composited (1992-1993). Samples were shipped
Statistical Summary, EMAP-E Virginian Province
Page A - 5
-------
on ice to their respective processing laboratory. Samples
for the determination of silt/clay content were sieved
using a 63um mesh sieve. Both an aliquot of the filtrate
and the fraction retained on the sieve were dried in an
oven at 60°C and weighed to calculate the proportion
of silt/clay in the sample.
The AVS collection method employed in the 1991
Survey permitted the potential release of sulfides when
the materials where processed on-board the sampling
vessel and in subsequent shipping. The sample was
collected from a homogenized composite (i.e., allowing
maximal exposure to oxygen). As a result, the accuracy
of the 1991 AVS measurements could be in doubt
although the precision may remain reliable as all
samples were treated similarly. Modifications to the
collection methods were incorporated into the 1992
sampling program to prevent a recurrence of these
problems. Beginning in 1992 a two-cm deep core was
removed from each grab included in the homogenate.
These "plugs" were placed in the AVS container without
mixing, thereby reducing the oxidation of the sample.
The container was filled to the top to further reduce the
probability of oxidation.
A.4.2.3 Sediment Contaminants
Metals, organic chemicals, and fine-grained sediments
entering estuaries from freshwater inflows, point sources
of pollution, and various non-point sources including
atmospheric deposition, generally are retained within
estuaries and accumulate in the sediments (Turekian,
1977; Forstner and Wittmann, 1981; Schubel and Carter,
1984; Nixon et al., 1986; Hinga, 1988). Samples were
collected from a homogenate created during sampling
by combining the top 2 cm of sediment from 6-10
sediment grabs. The sediment was placed in clean glass
jars with teflon liners or polypropylene containers (for
organics and metals analyses, respectively), shipped on
ice, and stored frozen in the laboratory prior to analysis
for contaminants. Sediments were analyzed for the
NOAA Status and Trends suite of contaminants (Table
A-5).
A.4.2.4 Sediment Toxicity
Sediment toxicity testing is the most direct measure
available for determining the toxicity of contaminants
in sediments to indigenous biota. It improves upon
direct measurement of sediment contaminants because
many contaminants are tightly bound to sediment particles
or are chemically complexed and, therefore, are not
biologically available (U.S. EPA, 1989). Sediment toxicity
testing, however, cannot be used to replace direct measurement
of the concentrations of contaminants in sediment because
such measurements are an important part of interpreting
the results of toxicity tests.
Toxicity tests were performed on the composite sediment
samples from each station. Tests were conducted using
the standard 10-day acute test method (Swartz et al.,
1985; ASTM 1991) and the: tube-dwelling amphipod Ampelisca
abdita.
A.4.2.5 Dissolved Oxygen
Dissolved oxygen (DO) is a fundamental requirement
for maintenance of populations of benthos, fish, shellfish,
and other estuarine biota. DO concentrations are affected
by environmental stresses, such as point and non-point
discharges of nutrients or oxygen-demanding materials
(e.g., particulates, dissolved organic matter). In addition,
stresses that occur in conjunction with low DO concentrations
may be even more detrimental to biota (e.g., exposure
to hydrogen sulfide, decreased resistance to disease and
contaminants). DO levels are highly variable over time,
fluctuating widely due to tidal action, wind stress, and
biological activity (Kemp and Boynton, 1980; Welsh
and Eller, 1991).
Dissolved oxygen was sampled in three ways during
the Virginian Province survey: 1) instantaneous water
column profiles using a SeaBird model SEE 25 CTD,
(2) point-in-time bottom oxygen conditions with a YSI
(model 58) oxygen meter and the SeaBird CTD, and 3)
continuous 24-72 nr measurements of bottom concentrations
using a Hydrolab DataSonde 3 data logging array (1991
only). The first two measurements were taken at all sites,
and the continuous measurements were taken at base
stations (BSS) only, and only in 1991.
The Hydrolab DataSonde 3 data logger deployed at
each 1991 Base site for 24-72 hours collected continuous
DO data at 15-min intervals. The DataSonde 3 also collected
salinity, temperature, water depth, and pH data. The
instruments were calibrated prior to every deployment,
and were checked on-board ship immediately prior to
deployment and following retrieval by comparison to
the YSI oxygen meter. These instruments were deployed
Page A - 6
Statistical Summary, EMAP-E Virginian Province
-------
Table A-5. EMAP Virginian Province: Sediment Chemistry Analytes
Analyte Code
Definition
TOC
AG
AL
AS
CD
CR
CU
FE
HG
MN
Nl
PB
SB
SE
SN
ZN
PCB8
PCB18
PCB28
PCB44
PCB52
PCB66
PCB101
PCB105
PCB118
PCB128
PCB138
PCB153
PCB170
PCB180
PCB187
PCB195
PCB206
PCB209
MBT
DBT
TBT
OPDDE
PPDDE
OPDDD
PPDDD
OPDDT
PPDDT
Total Organic Carbon Concentration in ug/g Dry Weight
Silver Concentration in ug/g Dry Weight
Aluminum Concentration in ug/g Dry Weight
Arsenic Concentration in ug/g Dry Weight
Cadmium Concentration in ug/g Dry Weight
Chromium Concentration in ug/g Dry Weight
Copper Concentration in ug/g Dry Weight
Iron Concentration in ug/g Dry Weight
Mercury Concentration in ug/g Dry Weight
Manganese Concentration in ug/g Dry Weight
Nickel Concentration in ug/g Dry Weight
Lead Concentration in ug/g Dry Weight
Antimony Concentration in ug/g Dry Weight
Selenium Concentration in ug/g Dry Weight
Tin Concentration in ug/g Dry Weight
Zinc Concentration in ug/g Dry Weight
2,4'-dichlorobiphenyl in ng/gram
2,2',5-trichlorobiphenyl in ng/gram
2,4,4'-trichlorobiphenyl in ng/gram
2,2',3,5'-tetrachlorobiphenyl in ng/gram
2,2',5,5'-tetrachlorobiphenyl in ng/gram
2,3',4,4'-tetrachlorobiphenyl in ng/gram
3,3',4,4',5-pentachlorobiphenyl in ng/gram
2,2',4,4',5-pentachlorobiphenyl in ng/gram
2,3,3',4,4'-pentachlorobiphenyl in ng/gram
2,21>3,3',4,4'-hexachlorobiphenyl in ng/gram
2,2',3,4,4',5'-hexachlorobiphenyl in ng/gram
2,2',4,4',5,5'-hexachlorobiphenyl in ng/gram
2,2',3,3',4,4',5-heptachlorobiphenyl in ng/gram
2,2',3,4,4',5,5'-heptachlorobiphenyl in ng/gram
2,2',3,4',5,5',6-heptachlorobiphenyl in ng/gram
2,2',3,3',4,4',5,6-octachlorobiphenyl in ng/gram
2,2',3,3',4,4',5,5',6-nonachlorobiphenyl in ng/gram
Decachlorobiphenyl in ng/gram
Mono-butyl Tin in ng/gram
Di-butyl Tin in ng/gram
Tri-butyl Tin in ng/gram
2,4'-DDE DDT and metabolites in ng/gram
4,4'-DDE DDT and metabolites in ng/gram
2,4'-DDD DDT and metabolites in ng/gram
4,4'-DDD DDT and metabolites in ng/gram
2,4'-DDT DDT and metabolites in ng/gram
2,4'-DDT DDT and metabolites in ng/gram
(Continued)
Statistical Summary, EMAP-E Virginian Province
Page A - 7
-------
Table A-5 (continued).
Analyte Code
Definition
ALDRIN
ALPHACHL
TNONCHL
DIELDRIN
HEPTACHL
HEPTAEPO
HEXACHL
LINDANE
MIREX
NAPH
MENAP2
MENAP1
BIPHENYL
DIMETH
ACENTHY
ACENTHE
TRIMETH
FLUORENE
PHENANTH
ANTHRA
MEPHEN1
FLUORANT
PYRENE
BENANTH
CHRYSENE
BENZOBFL
BENZOKFL
BENAPY
BENEPY
PERYLENE
INDENO
DIBENZ
BENZOP
SAND_PC
SICL PC
Aldrin in ng/gram
Alpha-Chlordane in ng/gram
Trans-Nonachlor in ng/gram
Dieldrin in ng/gram
Heptachlor in ng/gram
Heptachlor epoxide in ng/gram
Hexachlorobenzene in ng/gram
Lindane (gamma-BHC) in ng/gram
Mirex in ng/gram
Naphthalene in ng/gram
2-methylnaphthalene in ng/gram
1-methylnaphthalene in ng/gram
Biphenyl in ng/gram
2,6-dimethylnaphthalene in ng/gram
Acenaphthlylene in ng/gram
Acenaphthene in ng/gram
2,3,5-trimethylnaphthalene in ng/gram
Fluorene in ng/gram
Phenanthrene in ng/gram
Anthracene in ng/gram
1-methylphenanthrene in ng/gram
Fluoranthene in ng/gram
Pyrene in ng/gram
Benz(a)anthracene in ng/gram
Chrysene in ng/gram
Benzo(b)fluoranthene in ng/gram
Benzo(k)fluoranthene in ng/gram
Benzo(a)pyrene in ng/gram
Benzo(e)pyrene in ng/gram
Perylene in ng/gram
ldeno(1,2,3-c,d)pyrene in ng/gram
Dibenz(a,h)anthracene in ng/gram
Benzo(g,h,i)perylene in ng/gram
Sand Content (%)
Silt-Clay Content (%)
Page A - 8
Statistical Summary, EMAP-E Virginian Province
-------
approximately 1 m from the bottom. The stored data
were downloaded to a computer and the unit was
serviced and recalibrated for subsequent deployment at
another site. Dataloggers were also deployed in 1990
for the entire summer (ca. 60 days) at selected sites
(Long-Term Dissolved Oxygen [LTDO] sites) to
evaluate variability. Units deployed in 1990 were set
to log at 30-minute intervals. The results of this special
study are discussed in the 1990 Demonstration Project
Report (Weisberg et al, 1993).
Water column profiles for dissolved oxygen were
collected at each station using a SeaBird SBE-25
SeaLogger CTD. This unit was equipped with probes
to measure salinity, temperature, depth, pH, DO, light
transmission, fluorescence, and PAR. The unit was
equilibrated at the sea surface, and then lowered through
the water column at ca. 1A m/s until it reached a depth
of one meter above the bottom where it was allowed
to equilibrate. It was then returned to the surface and
all CTD data were downloaded to an on-board computer
for review and storage. If the cast appeared unusual or
failed QC it was repeated. Beginning in 1991 a bottom
water sample was collected using a Go-Flo water
sampling bottle, and the dissolved oxygen concentration
of the sample determined with a YSI Model 58 DO
meter. This measurement served as a check on the CTD
probe as well as a back-up in case the CTD failed.
A.4.2.6 Marine Debris
The kinds and amounts of floating and submerged
(i.e., collected in trawls) marine debris were noted at
all stations. Debris was categorized as paper, plastics,
metal, glass, wood, and other wastes. Only debris of
anthropogenic origin was included. Wastes that were
comprised of composited materials (e.g., metal, wood,
and plastic) were categorized based on their dominant
component.
A.4.3 Habitat Indicators
Habitat indicators provide basic information about
the natural environmental setting. Habitat indicator data
discussed in this report include water depth, salinity,
temperature, pH, water clarity, and sediment silt/clay
content.
All water quality measurements were made using
the Seabird model SEE 25 CTD (described earlier). This
unit was equipped with probes to measure salinity, temperature,
depth, pH, DO, light transmission, fluorescence, and PAR.
Measurements of water clarity are incorporated into
the CTD casts that were performed at each station. Included
in the CTD instrumentation package are a SeaTech
transmissometer and a Biospherical PAR (Photosynthetically
Active Radiation) sensor. As the CTD is lowered through
the water column, transmissivity and PAR data are continually
logged.
Surficial water samples were collected at all stations
for determination of Total Suspended Solids (TSS). Samples
were refrigerated, returned to the laboratory, filtered through
a glass-fiber filter, dried and weighed.
Sediment silt/clay content was measured on samples
taken from the surficial sediment (top two cm) homogenate
from which chemistry and toxicity samples were also
removed.
A.5 Data Collection and Sample Tracking
Each field crew was supplied with two portable computers
and appropriate software to facilitate electronic recording
of the data, data transfer, and sample tracking. All samples,
shipments, and equipment were labelled with bar-coded
labels to facilitate sample tracking and reduce transcription
errors. Field computers were equipped with bar code
readers to record sample identification numbers. Receiving
laboratories were also equipped with bar code readers
to facilitate the receiving process and to rapidly convey
information concerning lost or damaged shipments.
Copies of all data entered into the field computer
were stored on the hard disk and copied to diskettes.
Information on the hard disk was transferred daily via
modem to the Information Management Center at AED-
Narragansett (RI). Backup diskettes and hard-copy data
sheets were shipped weekly to the Center.
All transferred data were examined within 24-48 hours
of collection by EMAP-E personnel. Errors were brought
to the attention of the field crews for correction and
resampling, if required. All electronic data were checked
against paper data forms for verification. Further information
on the details of the Near Coastal data management systems
are presented in Rosen et al. (1990).
Statistical Summary, EMAP-E Virginian Province
Page A - 9
-------
A.6 Analytical Methods For This Statistical
Summary
Three types of analyses were conducted for this
report: 1) direct descriptions of measured indicators, 2)
development of modified or adjusted indicators (e.g.,
metal contaminants in sediments), and 3) development
of indices based on directly measured indicators. These
analyses are documented in a Virginian Province 1990
Demonstration Report (Weisberg et al., 1993) and
Appendix C of this document.
A.6.1 Cumulative Distribution Functions
(CDFs)
All ecological indicators collected during the
Virginian Province survey were characterized using
Cumulative Distribution Functions (CDFs). These
functions describe the full distribution of indicators in
relation to their areal extent within the Province. All
observations are weighted based upon surface area
associated with each sampling site. The area associated
with each sampling unit in large estuaries was equal to
the hexagonal spaces created by the EMAP grid (70
km2). For tidal river and small estuary classes, the area
associated with each station was determined using the
ARC/INFO data model which produces areal and
perimeter estimates. For the tidal river class
ARC/INFO was used to delineate the extent of 25 km-
long segments beginning at the mouth of the river on
a 1:100,000 digital line graph. The area of a large tidal
river station is equal to the area of the segment
containing the station. For small estuarine systems, the
station area is equal to the area of the system in which
it was randomly located. The total areas associated with
the three classes is: large estuaries - 16,097 km2; large
tidal rivers - 2,602 km2; and small estuaries - 4,875 km2.
To generate estimates across classes (strata), weights
for stations within each class were adjusted so that the
total of the weights for that class was equal to the total
area represented by all stations (including unsampleable
stations) within that class. The equations used in the
generation of CDFs are described in Appendix B.
A. 6.2 Adjustment To Known Covariates
In several cases, variability in observed indicators
might reflect relationships to known habitat or control
variables. Examples of these relationships are: variation
in estuarine biota resulting from sampling throughout
the salinity gradient; variation in sediment toxicity tests
with different mortalities associated with the controls;
and variation in sediment metals observed at a site resulting
from variations in the amount of natural crustal materials
at the site. In all these cases, the observed data must
be adjusted in order to construct CDFs or to compare
observations from different locations.
A.6.2.1 Adjustment for Natural Habitat Gradients
Estuarine biota are largely controlled by their
environmental settings, both natural and anthropogenic.
Natural gradients, particularly in salinity and silt-clay
content, are common in estuaries. Many estuarine organisms
may represent overlapping discrete distributions along
these gradients. Thus, normalization of ecological measures
over habitat gradients is a common tool used to interpret
information when such normalization is necessary. Such
relationships were examined; however, no data were
normalized in the production of this report.
A.6.2.2 Adjustment for Experimental Controls
Estimates of the area in the Virginian Province containing
toxic sediments were based on the results of toxicity
tests using the amphipod, Ampelisca abdita. For this
summary, a relative measure of toxicity was created to
facilitate comparisons between sites over a series of bioassays.
This adjustment is necessary because control mortalities
vary among test series. Sediments were determined to
be toxic if: survival of the test organism in test sediments
was less than or equal to 80% of the survival observed
in clean, control sediments; survival in test and control
sediments were significantly different (p < 0.05); and
survival in control sediments was > 85%. This results
in an adjustment to the observed survival rates in test
sediments that accounts for variability due to differences
in the controls for individual bioassays. These criteria
are consistent with those established in U.S. EPA/ACE
(1991).
Page A - 10
Statistical Summary, EMAP-E Virginian Province
-------
A.6.2.3 Adjustment for Natural Crustal
Properties
The extent to which anthropogenic activities have
affected concentrations of metals in sediments is
complicated by the natural variation of concentrations
due to differing particle size distributions in sediments.
Because of surface adsorptive and complexation
processes, fine-grained sediments will naturally have
higher trace metal concentrations than coarse sediments.
In some studies, e.g., the National Status and Trends
program, reported concentrations are adjusted for this
variation by normalizing the concentrations by the fine-
grained fraction determined separately. As an alterative
to actual size-fractionation measurements, a number of
authors (Windom et al., 1989; and Schropp et al., 1990)
have determined relationships between sediment
concentrations of trace metals and other elements
indicative of fine-grained crustally-derived material,
e.g., aluminum, iron and manganese. The most
commonly used of these indicator elements is aluminum,
due to its large natural abundance, freedom from
common anthropogenic contaminant sources and
significant correlation with both the fine-grained fraction
and trace metal concentrations in clean, un-impacted
sediments. The correlation between aluminum and trace
metals in fine-grained sedimentary material has a
geochemical basis related to the composition of crustal
material from which the fine particles are derived and
the natural adsorption and complexation processes
occurring during "weathering" of the crustal material.
Once background sediment metal-aluminum relation-
ships have been determined, concentrations of metals
expected from background material can be subtracted
from total metal concentrations, allowing residual,
presumably anthropogenic, contributions to be assessed.
Background metal-aluminum relationships are
derived by linear regression of sediment concentrations
of each element against aluminum concentrations in the
same sediment. Some investigators have used log-
transformed metals concentrations in the regression
analyses. Such transformations do not improve
correlation of the metals-aluminum concentrations of
this data set. Furthermore, linear regressions provide
direct correlation with the physical mixing and
geochemical factors noted above which affect the overall
concentration of metals in sediments. This correlation
is lost when the concentrations are transformed. Conse-
quently, no data transformations were performed prior
to regression analysis.
Use of linear regression to determine metal-aluminum
relationships in background sedimentary material can
only be successful if the sediments do not include contributions
from sources other than natural background sediments.
The data sets used in this study were statistically screened
to eliminate samples which might contain additional source
materials. This was accomplished by performing linear
regressions of concentrations of aluminum against each
metal. The residuals (the differences between the measured
concentrations and those predicted from the regression)
were then tested for normal distribution. If the residuals
were found not to be normally distributed, samples which
had studentized residual values greater than 2 were eliminated
from the data set. Regression of the reduced data set
was repeated and the residuals tested again for normal
distribution. This process was repeated for each metal
until residuals from the regressions were all normally
distributed, at which point the remaining samples were
assumed to represent natural, background sediments.
The regression relationships derived for the background
sediments were then applied to the original data set. Samples
with trace metal concentrations exceeding the upper 95%
confidence limit for that metal's regression against aluminum
were designated as enriched. It should be noted that no
assessment was made as to the magnitude of enriched;
metal concentrations might be only slightly above the
95% confidence limit or might exceed the limit by factors
of 10-100. The categorization "enriched" was applied
to any sediment with a metal concentration higher than
that expected from the background sediment aluminum
metal relationship at the 95% confidence level.
A. 6.3 The Benthic Index
A benthic index, which uses individual measures
of the benthic community, was utilized to report on the
condition of the benthic biological resources of the Virginian
Province. The index, as used in this report, was developed
from a subset of the data collected over all four years
of sampling and was constructed to represent a combination
of individual benthic measures that best discriminates
between good and poor benthic conditions. This current
index is EMAP's continued attempt to reduce many individual
indicators into a single value that has a high level of
discriminatory power between good and poor ecological
conditions. The reader should note that the index as
Statistical Summary, EMAP-E Virginian Province
Page A - 11
-------
used in this report is a revision to prior ones used in
earlier EMAP-VP reports. It has always been the intent
of the program to continually revise the benthic index
as more data became available (Weisberg et al., 1993),
and the current index represents the effort using four
years of available EMAP data (Paul et al., in
preparation).
The process for developing an index of benthic
community condition has been documented for the 1990
(Weisberg et al., 1993) and, separately, for the 1990-91
(Schimmel et al., 1994) data sets. This process entails
several discrete steps: identification of a set of benthic
parameters to define conditions that include components
of faunal and functional diversity and structure;
determination of the statistical relationships between
these benthic parameters and habitat variables;
normalization of those benthic parameters that are
strongly associated with habitat condition; identification
of a test data set that clearly distinguishes relatively
pristine sites from those exhibiting toxic contamination,
hypoxia, or both; and application of discriminant
analysis to the test data set to determine those benthic
parameters whose variation is most closely associated
with differences in reference and impacted condition.
This same process was used with the 1990-93 data set.
The benthic index developed using the 1990-91 data
set suffered from poor representation of impacted and
reference conditions in low salinity (< 5 ppt) in the test
data set. This benthic index was highly correlated with
salinity and appeared to misclassify good sites in the
oligohaline and impacted sites in the meso- and
polyhaline. This led to the refinement of the candidate
benthic parameters, utilization of more stringent criteria
for assignment of sites to the reference and impacted
categories in the test data set, and revision of the
benthic index based upon the four-year data set (1990-
93). Statistical analyses indicated that most measures
of diversity and abundance of low salinity tubificids
were highly correlated with salinity and required
normalization (Weisberg et al., 1993).
The test data set was constructed to contain an equal
number of impacted and reference sites, and equal
number of sites exhibiting each condition in each of the
three salinity zones (< 5, 5-18, > 18 ppt), and a relative
balance of muddy and sandy sites in each
condition/salinity category. For a site to be included in
the reference condition test data set, all of the following
were met: bottom dissolved oxygen > 7 mg/1; no more
than three sediment contaminant concentrations exceeding
Long et al. (1995) ER-L values, and none exceeding
ER-M concentration; and survival in a sediment toxicity
test 90% or better and not significantly different from
control survival. Ten sites in each of the oligo-, meso-,
and polyhaline zones were selected. Thirty impacted
sites were selected based on criteria that included: low
bottom dissolved oxygen (< 2 mg/1); low survival in the
toxicity test and multiple concentration exceedances of
ER-M values. The test data set contained 60 cases; 30
were categorized as impacted and 30 were reference.
The discriminant analyses identified a series of highly
correlated benthic indices that correctly classified reference
and impacted sites in the test data set. The benthic parameter
common to the candidate indices that accounted for the
greatest degree of variability was a measure of species
richness, Gleason's D (Washington, 1984). The benthic
index that was chosen (1) maximized classification efficiency
using the test data set (goal of ca. 90% correct classification),
(2) provided a good degree of cross-validation with the
test data set (goal of ca. 80% cross-validation), (3) produced
a good classification efficiency with a validation data
set (goal of ca. 80% correct classification), and (4) had
the individual parameters contribute to the overall score
consistently with our understanding of benthic communities.
The stations that met the reference and impacted site
criteria, but were not used in the test data set for the
discriminant analysis, were used as a validation data set
(52 cases).
The three benthic parameters of the index were: salinity-
normalized expected Gleason's D for infaunal and epifaunal
species; salinity-normalized expected number of tubificids
and abundance of spionids. The richness measure is associated
with reference conditions (positive contribution) and the
latter two measure are associated with impacted conditions
(negative contribution).
The discriminant score calculation normalizes the
individual parameters based on the mean and standard
deviation for the parameter in the test data set. The critical
value for discriminating between reference and impacted
sites was determined to be zero using the following equation:
Page A - 12
Statistical Summary, EMAP-E Virginian Province
-------
Benthic Index Score =
1.389 (pet expect Gleason - 51.5) / 28.4
-0.651 (normalized tubifield abundance - 28.2) / 119.5
- 0.375 (spionid abundance - 20.0) / 45.4
Where:
Percent Expected Gleason diversity index value =
Gbason / (4.283 0.498*bottom salinity
+ 0.0542 * bottom salinity2
- 0.00103* bottom salinity3) * 100
Normalized Tubificid Abundance =
Tubificids - 500 * e
-15*bottom salinity
Statistical Summary, EMAP-E Virginian Province
Page A - 13
-------
APPENDIX B
ESTIMATION FORMULAE FOR EMAP SAMPLING
IN THE LOUISIANIAN AND VIRGINIAN PROVINCES
Acknowledgements: The equations described in this section for large estuary, large tidal river, and
whole province estimates were formulated by Douglas Heimbuch and Harold Wilson of Coastal Environmental
Services Inc., Linthicum, MD and Stephen Weisberg of Versar Inc., Columbia, MD
B.I INTRODUCTION
This appendix describes the equations used for making the four-year estimates of the areal extent of conditions
of interest (and estimates of variances for these estimates) reported in this document. Equations were formulated
using data from the first four years of EMAP-Estuaries (EMAP-E) sampling conducted in the Virginian and Louisianian
Provinces. The recommended methods were chosen to be consistent with the sampling designs employed in each
of the estuarine system classes (large estuaries, large tidal rivers, and small estuaries). This appendix describes
the generic equations for each class, followed by specific instructions for application to Virginian Province data.
The equations and associated SAS programs were provided to the EMAP-Virginian Province team by EMAP-Estuaries.
The reader should note that the large estuary and large tidal river equations differ from those used for generating
single-year estimates in the 1991 and 1992 Statistical Summaries. These equations represent a refinement of the
earlier equations. It should be noted that these equations are still under review and may be further refined in the
future to address additional measurements of variability. Any alterations to the equations should result in only
minor changes. A comparison of the confidence intervals generated using the "old" equations with those reported
in this section shows only small changes result.
The 95% confidence intervals reported in this document are calculated as 1.96 times the standard error of
the estimate, with the standard error (of the estimate) being the square root of the variance of the estimate.
B.2 LARGE SYSTEM RECOMMENDED METHODS
The recommended method for large system estimation is based on a sampling design in which sampling
stations are selected within hexagons of a randomly overlaid grid. If a station within a hexagon is on land, then
no sample is obtained from that hexagon. The estimated subnominal (i.e., impacted) area for one year of the survey
is based on Horvitz-Thompson estimation methods and is given by:
Statistical Summary, EMAP-E Virginian Province Page B-l
-------
where,
Y,= the estimated subnominal area in year t
A= the known total area of large systems in the province
n= the total number of hexagons subject to sampling in the province
zti= the response from hexagon i in year t (=1 if subnominal, 0 otherwise)
x(/= 1 if a sample is obtained from hexagon /', 0 otherwise.
The indicator variable Xti can also be defined in a manner to estimate the subnominal area of a particular
subpopulation of the province (e.g., Delaware Bay as a subpopulation within the Virginian Province). In this case,
Xti would be defined as 1 if the sample was obtained in the subpopulation of interest, and zero otherwise. The
total area used in this calculation (A) would be the known area of the subpopulation of interest.
The variance of the estimated subnominal area is estimated using a formula based on the Yates-Grundy formula
for the variance of the Horvitz-Thompson estimator, and the Taylor series expansion formula for the variance of
a quotient:
v£r(Y)=var A
N}(AN
i N/4
)(
where,
var(N) +v&r(D)^ _2cov(N,D)
N2 D2 ND
cdv(N,D) = W
n n
[ which is equivalent to, cov(N,D) =£ £ W,/x,,-x,;) (xt/ztl-xtjzt)
1-1 f>l
Page B-2 Statistical Summary, EMAP-E Virginian Province
-------
n
Vtl= the proportion of random placements of the grid which result in points / and y lying in the same hexagon.
Two stations can be jointly selected for sampling in the same year only if they are located in separate hexagons.
The probability of joint inclusion is therefore related to the complement of the probability that the random placement
of the grid causes the two points to be included in the same hexagon. For stations that are sufficiently distant,
the probability of being included in the same hexagon is equal to zero.
The recommended method for estimating the subnominal area is based on the ratio of two random quantities:
the number of actual samples associated with a subnominal response, and the total number of actual samples.
The total number of actual samples is a random quantity due to the edge effect of hexagons that include both land
and water area. An alternative method could be used for estimation that is based only on the product of the number
of hexagons with a subnominal response and the area of a hexagon. However, this method could produce estimates
of subnominal area that are greater than the total known area due to a greater than expected number of sample
stations falling in water. The ratio method is recommended to insure that estimates of subnominal area are not
adversely affected by actual sample size in this manner.
To estimate the four-year average subnominal area in large systems, the average of the annual estimates
is calculated:
where,
Y= the estimated average response over four years.
The estimated variance of the four-year average is given by:
The recommended method for estimating the four-year average subnominal area employs years as independent
strata. An alternative method could have been used which combines data from all years into one procedure without
stratifying by year. In this method, years that received more effort due to random sample sizes would receive
more weight in the calculation of the four year average. By treating years as strata, each year receives equal weight
in the estimation of the four-year average. The recommended method also insures that the year to year variation
in the response variable will not affect the variance of the four year average estimate. This is consistent with
the view of interannual variability as a fixed effect (rather than a random effect) when characterizing a specified
set of years.
Statistical Summary, EMAP-E Virginian Province Page B-3
-------
Application to the Virginian Province
The recommended method for estimating the subnominal area for one year can be applied to large system
data from the Virginian Province. However, sampling stations in the Virginian Province were not randomly selected
within hexagons, but were obtained at the center point of each hexagon. For this reason, the recommended method
for the estimation of the variance of the subnominal area estimate in one year cannot be directly applied. An approximate
estimate of this variance can be calculated by redefining Vtj as :
VtJ= for i and; in adjacent hexagons, 0 otherwise,
where (3 is the proportion of all pairs of samples that are from adjacent hexagons. This approximation is based
on the following relationships:
and,
where,
n= number of hexagons in the grid
a= area of each hexagon (in appropriate units)
7i, = joint inclusion probability for station locations / and j.
Therefore,
- 1 1
na(na-1)P(1 -V) + na(na-1)(1 -P) = n(n-1)
32 r, &2
where,
V= average of non-zero Vy values
P = proportion of i,j pairs that are no farther apart than the distance between centers of adjacent hexagons,
and,
V= for large "a" (relative to the size of a sampling station).
pn
Page B-4 Statistical Summary, EMAP-E Virginian Province
-------
The recommended method for estimating the variance of the four-year average can be applied to the approximate
variances of the one year estimates. This approach does not account for any gain in precision that may be caused
by the four-year interpenetrating design which was implemented in the Virginian Province. The potential gain
in precision is due to negative covariance among the annual estimates. The magnitude of the negative covariance
depends on the degree of spatial autocorrelation of distances less than the size of a hexagonal cell. Analyses of
the data suggest that spatial autocorrelation in the response variables is insignificant at distances as small as 2.5
km. Therefore, little increase in precision is anticipated and the recommended method is likely to provide an adequate
approximation.
B.3 TIDAL RIVER RECOMMENDED METHODS
The recommended methods for tidal rivers estimation is based on a stratified random sampling design. Each
river is stratified into areas of river length equal to 25 kilometers. In each year, at least one sample is obtained
in each stratum with some strata providing two samples (Lousinian Province only: see notes on Application to
Virginian Province). Each 25-km segment was divided into four subsegments, with one being sampled each year.
The statistical area applied to each station was equal to the area of the subsegment the station resided in plus the
area of the next three upstream subsegments. The recommended method for estimating subnominal area for tidal
rivers is:
Yt=A^
where,
Yt= estimated subnominal area in year t
A= the total known area of tidal river systems in the province
n= the number of sampled tidal river strata in the province
W,= the area of stratum /
mu
XX
zs = = the average response in year t and stratum /
mi,
ztlj= the response in year t, stratum /', sample j (1 if subnominal, 0 otherwise)
m,= the number of observations in year rand stratum /.
The recommended method can be applied to estimate the subnominal area in a particular tidal river within the
province. In this application, only data from the strata of interest would be utilized, and the total area (A) would
be that of only the selected tidal river.
Statistical Summary, EMAP-E Virginian Province Page B-5
-------
The variance of the estimated subnominal area is calculated as:
A2S2
var(Y)~
where,
var{Y)= the estimated variance of the subnominal area estimate in year f,
S2 = -^^ (the pooled estimate of within stratum variance)
EK-D
n'= the number of strata with replicate samples (mti > 2).
The variance of the estimated subnominal area is based on the estimate of within-strata variance pooled across
strata which contain at least two observations. The estimation of variance requires that at least one stratum contains
two or more observations (see note on application to Virginian Province).
The estimate of the four-year average subnominal area is calculated as the average of the annual estimates:
Y ' \*- Y
where,
Y= the estimated four-year average subnominal area.
The estimated variance of the four-year average estimate is:
16t.
Application to the Virginian Province
The recommended method for estimating tidal river subnominal area can be applied to Virginian Province
data. However, the area subject to sampling in each tidal river changed over the four-year period because the
statistical area applied to each station was equal to the area of the subsegment the station resided in plus the area
of the next three upstream subsegments. Stations in the first subsegment were sampled in 1990, third subsegment
in 1991, second subsegment in 1992, and fourth subsegment in 1993. Therefore, in year 1 the reach from 0 km
to 125 km was subject to sampling, in year 2 the reach from 12 km to 125 km was subject to sampling, in year
3 the reach from 6.25 km to 125 km, and in year 4 the reach from 18.75 km to 125 km was subject to sampling.
Page B-6 Statistical Summary, EMAP-E Virginian Province
-------
The tidal river sampling in the Virginian Province consisted of one sample per stratum. Since replicate observations
were not obtained in any stratum, the approximate estimate of within-stratum variance applied to Louisianian Province
data can also be applied to data from the Virginian Province. In this application, approximations of within-stratum
variances can be calculated separately for each tidal river.
The recommended methods for the estimate of the four-year average subnominal area and corresponding
variance can be directly applied to tidal river data from the Virginian Province. This approach does not account
for any gain in precision that may be caused by the four-year interpenetrating design which was implemented in
the Virginian Province. Similarly to the approach for large systems, the potential gain in precision is a function
of the degree of spatial autocorrelation in the response variables. Since studies have suggested that the degree
of spatial autocorrelation is small, little increase in precision from the interpenetrating design is anticipated, and
the recommended approach is likely to produce an adequate approximation.
B.4 SMALL SYSTEM RECOMMENDED METHODS
For small estuarine systems, estimates of CDFs and associated variances were computed based on a random
selection of small systems within the Province, with replicate samples taken from a subset of the selected systems
(Cochran, 1977). Unlike large estuaries and large tidal rivers, only a portion of the area of this class is sampled
each year; therefore, a single four-year estimate is produced from the entire dataset as opposed to pooling individual
yearly estimates. This method is directly applicable to Virginian Province data without modification The resulting
CDF estimate is:
P - H
r
where,
PSx = CDF estimate for value x
m, = number of samples at small system /
A, = area of small system I
= /1 if
t 0 o
response is less than x
otherwise
n = number of small systems sampled
Since replicate samples were only obtained at a subset of the sampled small estuarine systems, the formula
for the estimated variance taken from Cochran (1977 eq. 11.30) was modified to produce the following estimate
of the approximate mean squared error (MSE) of the CDF estimate:
Statistical Summary, EMAP-E Virginian Province Page B-7
-------
A>
n n-1 n*,,,
where,
n* = number small systems with replicate samples
O2 -
o2, -
A = the total area of small systems in the Province (4,875 km2)
N = number small systems in Province (144)
B.5 ENTIRE PROVINCE RECOMMENDED METHODS
The recommended estimate for the subnominal area for the entire province (i.e., across all system classes)
is the sum of the subnominal area estimates of the large systems, tidal rivers, and small systems:
/=Us
where,
Ot= the estimated subnominal area for the entire province in year t
Ytl= the estimated subnominal area in year f for system class /, (/'=large, tidal, small).
The estimated variance for the subnominal area in the entire province is the sum of the component variances:
I-/./.S
where,
var(U) = the estimated variance of the subnominal area in the entire province,
var(Y,)= the estimated variance of the subnominal area in year f, system class /'.
The recommended methods for estimation in the entire province are based on the assumption of the system
classes as being independent strata. The methods can be directly applied to one-year and four-year average estimates,
and to data from both the Louisianian and Virginian Provinces.
Page B-8 Statistical Summary, EMAP-E Virginian Province
-------
APPENDIX C
LINEAR REGRESSIONS OF INDIVIDUAL METALS AGAINST
ALUMINUM USED IN THE DETERMINATION OF METALS
ENRICHMENT OF SEDIMENTS OF THE VIRGINIAN PROVINCE
As discussed in Section 3.2.3.5, concentrations of
individual metals were normalized against the crustal
element aluminum in an attempt to provide a basis for
estimating the areal extent of enrichment of these metals
in Virginian Province sediments. The method utilized
is described in Appendix A (Section A.6.2.3). For each
metal, a regression and an upper 95% confidence
interval was determined and plotted (Figures C-l to C-
8). Stations with concentrations falling above the upper
95% confidence interval were classified as enriched for
that metal. This process was inefficient for several
metals, but performed well for As, Cr, Fe, Hg, Mn, Ni,
Sb, and Zn. Regressions and regression parameters
(slope, intercept, and correlation coefficient: Table C-l)
for only these metals are included in this report.
Statistical Summary, EMAP-E Virginian Province Page C - 1
-------
0
Figure C-1. Linear regression of Arsenic against aluminum. (Dashed line is the upper 95% confidence
interval).
200
150
100
\
o
3 data points excluded:
341,348, & 865
8
10
Al (%)
Figure C-2. Linear regression of Chromium against aluminum. (Dashed line is the upper 95%
confidence interval). NOTE: Three data points were excluded for clarity.
Page C - 2
Statistical Summary, EMAP-E Virginian Province
-------
Figure C-3. Linear regression of Iron against aluminum. (Dashed line is the upper 95% confidence
interval).
1t
1 data point
excluded: 3.27 ug
Hg
0
Figure C-4. Linear regression of Mercury against aluminum. (Dashed line is the upper 95% confidence
interval). NOTE: One data point was excluded for clarity.
Statistical Summary, EMAP-E Virginian Province
Page C - 3
-------
3000 T
2000
1000
3 data points excluded: 5550,
5850, & 6430 pg Mn
Figure C-5. Linear regression of Manganese against aluminum. (Dashed line is the upper 95%
confidence interval). NOTE: Three data points were excluded for clarity.
Figure C-6. Linear regression of Nickel against aluminum. (Dashed line is the upper 95% confidence
interval).
Page C - 4
Statistical Summary, EMAP-E Virginian Province
-------
2 data P°ints
excluded: 49 & 152
0
Figure C-7. Linear regression of Antimony against aluminum. (Dashed line is the upper 95% confidence
interval). NOTE: Two data points were excluded for clarity.
750
500
N 250 +
*«b * J*»J1»'.* *»^
. ..-3 «n*»*« .
0
4 6
Al (%)
8
10
Figure C-8. Linear regression of Zinc against aluminum. (Dashed line is the upper 95% confidence
interval).
Statistical Summary, EMAP-E Virginian Province
Page C - 5
-------
Table C-1. Metal-aluminum regression parameters obtained from Virginian Province sediment data (m = slope,
b = intercept, r2 = correlation coefficient).
Element
Regression parameters
m
As
Cr
Fe
Hg
Mn
Ni
Sb
Zn
1.06
9.64
5,581
0.010
54.22
4.66
0.006
21.83
1.28
-1.55
-953
0.002
69.05
-3.40
0.006
-5.43
0.49
0.82
0.89
0.48
0.74
0.76
0.39
0.69
Page C - 6
Statistical Summary, EMAP-E Virginian Province
U.S. GOVERNMENT PRINTING OFFICE: 1995-6 50-006/ 22053
------- |