United States
Environmental Protection
Agency
Region III
Chesapeake Bay
Program Office
Region III
Water Protection
Division
EPA 903-R-07-003
CBP/TRS 285/07
July 2007
In coordination with the Office of Water/Office of Science and Technology, Washington, D.C., and the states
of Delaware, Maryland, New York, Pennsylvania, Virginia and West Virginia and the District of Columbia
Ambient Water Quality
Criteria for Dissolved
Oxygen, Water Clarity and
Chlorophyll a for the
Chesapeake Bay and Its
Tidal Tributaries
2007 Addendum

July 2007

-------
       Ambient Water Quality Criteria

    for Dissolved Oxygen,  Water Clarity

and  Chlorophyll a for the Chesapeake Bay

            and  Its Tidal Tributaries

                  2007 Addendum


                       July 2007

              U.S. Environmental Protection Agency
                       Region III
                Chesapeake Bay Program Office
                    Annapolis, Maryland

                         and

                       Region III
                  Water Protection Division
                  Philadelphia, Pennsylvania

                    in coordination with

                     Office of Water
               Office of Science and Technology
                     Washington, D.C.

                         and

                       the states of
                Delaware, Maryland, New York,
                  Pennsylvania, Virginia and
            West Virginia and the District of Columbia

-------
                         Contents
Acknowledgments 	   v

I. Introduction  	   1
      Literature Cited	   3

II. Refinements to the Chesapeake Bay Water Quality Criteria
   Assessment Methodology  	   4
      Background	   4
      Overview of the CFD Assessment Methodology	   5
      Description and Evaluation of the CFD-Based Assessment Methodology ..   9
         Example CFD-based criteria assessment	   9
         CFD reference curves	  12
         Comparing assessment and reference curves	  15
         Development of a statistical decision-making framework	  16
         Results of the scientific evaluation  	  19
      Application of the CFD-Based Assessment Methodology	  21
         Recommendations for application of the CFD-based methodology  .  21
         Recommendations for future refinement of the
           CFD-based assessment methodology	  23
      Literature Cited	  24

III.  Application of Chesapeake Bay Water Quality Criteria
    Assessment Procedures  	  25
       Background	  25
       Assessment Units, Segmentation, and Sub-Segmentation	  26
       Data to be Used in Chesapeake Bay Criteria Assessments  	  28
       Updating the Criteria Attainment Assessment Framework  	  31
       Literature Cited	  32

IV. Refinements to the Chesapeake Bay Dissolved Oxygen Criteria
   Assessment Procedures 	  33
       Background	  33
       Temporal Periods for Assessment of Dissolved  Oxygen Criteria	  36
       Dissolved Oxygen Criteria Assessments in Shallow Waters
         Versus Open Waters	  37
       Assessment of Short Duration Dissolved Oxygen Criteria	  38
       Dissolved Oxygen Criteria Reference Curves   	  39
                                                                               Contents

-------
IV
                             Summer open-water and deep-water	  39
                                 dissolved oxygen criteria reference curves  	  42
                             Non-summer open-water dissolved oxygen criteria
                                 reference curve 	  42
                             Assessment of deep-channel instantaneous minimum
                                 dissolved oxygen criteria	  42
                           Use of Percent Saturation as Dissolved Oxygen Criteria  	  43
                           Literature Cited	  45

                   V.               to the
                      Use                          	  47
                           Background	  47
                           Shallow-Water Designated Use Attainment Assessment   	  48
                             Assessment based on the single best year ofSAV 	  50
                             Assessment based on water clarity acres	  53
                             Assessment based on CFD-based water clarity criteria attainment  .  56
                           Shallow-Water Designated Uses and SAV No-Grow Zones  	  56
                           Water Clarity Criteria Reference Curves	  57
                           Literature Cited	  59

                   VI.              a                                   	  61
                          State Water Quality Standards 	  61
                          Chlorophyll a Criteria Assessment Procedures  	  62
                          Literature Cited	  62

                   VII.
                                                              	  63
                          Design and Approach for Chesapeake Bay Shallow-Water Monitoring  .  63
                             Shallow-water monitoring design	  65
                             Continuous monitoring component  	  66
                             Water quality mapping component	  67
                          Schedule for Assessment of Shallow-Water Designated Use Habitats ..  68
                             Extending the timeframe	  69
                             Additional resources	  69
                             Assessment based on reduced monitoring  	  70
                             Segment prioritization schedule	  74
                          Dissolved Oxygen Criteria Assessments Using
                           Shallow-water Monitoring Data 	  75
                             Temporal standardization	  75
                             Scaling and interpolation issues	  77
                          Water Clarity Criteria Assessments Using
                           Shallow-water Monitoring Data 	  78
                             Analysis issues 	  79
                             Statistical modeling 	  79
                             Interpolation	  80
                          Chlorophyll a Criteria Assessments Using
                           Shallow-water Monitoring Data 	  82
                             Statistical modeling 	  83
 Contents

-------
        Modeling Approach	  84
        Analysis issues 	  84
    Literature Cited	  88

III.              for
                           	  89
    Background	  89
    Listing Category Decisions	  89
        Criteria attainment assessments	  90
        Dissolved oxygen criteria attainment assessments	  90
        Water clarity criteria attainment assessments	  91
        Chlorophyll a criteria attainment assessments  	  91
        Benthic index of biotic integrity assessments	  92
        Assessment reporting framework	  92
     Listing Decision Framework 	  95
        Segments previously listed as impaired 	  96
        Segments not previously listed as impaired	  96
        Shallow-water designated use listing decisions	  96
     Literature Cited	  97

    A, The Cumulative Frequency Diagram Method for Determining Water
       Quality Attainment: Report of the Chesapeake Bay Program
       STAC Panel to Review Chesapeake Bay Program Analytical Tools  . A-l

    B. Detailed Chesapeake Bay Water Quality Criteria Assessment
       Methodology  	B-l

    C, Evaluation of Options for Spatial Interpolation  	C-l

    D, User Guide and Documentation for the
       Chesapeake Bay Interpolator 	D-l

    E, Potential Methods for Assessing Shorter Duration
       Dissolved Oxygen Criteria	E-1

    F. Data Used in Deriving the Open-Water, Deep-Water, and
      Deep-Channel Dissolved Oxygen Criteria Summer Biological
      Reference Curves 	F-l

    G, Equations for the Open-Water, Deep-Water, and Deep-Channel
       Dissolved Oxygen Criteria Summer Biological Reference Curves . . G-l

    H, Equations for the Water Clarity Criteria Biological Reference
       Curves  	H-l

    I, Evaluation of Maryland and Virginia Chesapeake Bay Segment SAV
      Acreages from 2003 to 2005 for Prioritizing Shallow-Water Monitoring
      by Segment  	  1-1
                                                                                    Contents

-------
VI
                         J, Chesapeake Bay Estuarine Benthic Communities Assessment Protocol
                           for Maryland and Virginia 305b/303d Integrated Reports	J-l
                         K. 2006 303(d) Assessment Methods for Chesapeake Bay Benthos . . . . K-l
                         L, Addendum to the Report: Development of Diagnostic Approaches to
                           Determine Sources of Anthropogenic Stress Affecting Benthic
                           Community Condition in the Chesapeake Bay  	L-l

-------
                                                                                        VII
               Acknowledgments
This second addendum to the April 2003 Ambient Water Quality  Criteria for
Dissolved Oxygen, Water Clarity, and Chlorophyll a for Chesapeake Bay and Its
Tidal Tributaries was developed and documented through the collaborative efforts of
the  members of the Chesapeake Bay Program's Criteria Assessment Procedures
Workgroup and Water Quality Steering Committee.

PRINCIPAL AND CONTRIBUTING AUTHORS
This document resulted  from the  collaborative expertise  and talents of the
Chesapeake Bay Program's state agency, federal agency, and academic institutional
partners. The 25 principal authors (listed first) and contributing authors (listed in
alphabetical order) follow by chapter. Unless noted, author affiliations are listed
under the specific workgroup or committee acknowledgments. Chapter 1: Richard
Batiuk; Chapter 2: Steve Preston; Chapter 3: Steve Preston; Chapter 4: Richard
Batiuk, David Jasinski, Marcia Olson, and Gary Shenk; Chapter 5: Richard Batiuk;
Chapter 6: Elgin Perry, Richard Batiuk, and Larry Harding (University of Maryland
Center for Environmental Science); Chapter 7: Bruce Michael, Rick Hoffman, Mary
Ellen Ley, Ken Moore, Elgin Perry, and Mark Trice; Chapter 8: Larry Merrill, Mark
Barath, Richard Batiuk, and  Richard  Eskin;  Appendix  A:  David Secor; Mary
Christman; Frank Curriero; David Jasinski; Steve Preston; Ken Reckhow; and Mark
Trice; Appendix B: Gary Shenk; Appendix C: Steve Preston; Appendix D: Lowell
Banner (NOAA Chesapeake Bay Office), David Jasinski, and Gary Shenk; Appendix
E: Gary Shenk, Marcia Olson, and  Elgin Perry; Appendices F, G, and H:  Gary
Shenk; Appendix I: Bruce Michael; Appendix J: Mark Barath; Appendix K: Roberto
Llanso (Versar), Dan Dauer (Old Dominion University), Mike Lane (Old Dominion
University), and Jon Volstead (Versar); and Appendix L: Dan Dauer, Mike Lane, and
Roberto Llanso.

CRITERIA ASSESSMENT PROTOCOLS WORKGROUP

Steve Preston,  chair (U.S. Geological  Survey/Chesapeake Bay Program Office),
Harry Augustine (Virginia Department  of Environmental Quality); Mark Barath
                                                                         Acknowledgments

-------
VIM
                    (U.S. Environmental Protection Agency Region III); Thomas Barron (Pennsylvania
                    Department  of Environment); Joe  Beaman  (Maryland  Department  of the
                    Environment);  Jerusalem Bekele  (District  of Columbia  Department  of the
                    Environment); Stephen Cioccia (Virginia Department of Environmental Quality);
                    Elleanore Daub (Virginia Department of Environmental  Quality); Sherm Garrison
                    (Maryland Department of Natural Resources);  Darryl Glover (Virginia Department
                    of Environmental Quality); Peter Gold (U.S. Environmental Protection Agency
                    Region III); John Hill (Maryland Department  of the Environment); Rick Hoffman
                    (Virginia  Department of Environmental  Quality); Dave Jasinski (University of
                    Maryland Center for Environmental Science/Chesapeake  Bay Program Office); Jim
                    Keating (U.S. Environmental Protection Agency Office  of Water); Rodney Kime
                    (Pennsylvania  State Department of  the  Environment);  Larry Merrill (U.S.
                    Environmental  Protection Agency Region  III);  Bruce  Michael (Maryland
                    Department of Natural Resources); Ken Moore (Virginia  Institute of  Marine
                    Science); Shah Nawaz (District of Columbia Department of Health); Roland Owens
                    (Virginia  Department  of Environmental  Quality);  Jennifer Palmore  (Virginia
                    Department of Environmental Quality); Elgin Perry (Statistics Consultant); Charley
                    Poukish (Maryland  Department of the Environment); Matt Rowe (Maryland
                    Department of the Environment); John Schneider (Delaware Department of Natural
                    Resources and Environmental Control); Gary Shenk (U.S. Environmental Protection
                    Agency Chesapeake  Bay Program Office); Nicoline Shulterbrandt  (District of
                    Columbia  Department of Health);  Donald Smith (Virginia  Department of
                    Environmental Quality); Matt Stover (Maryland  Department of the Environment);
                    Robert Swanson (Virginia Department of Environmental Quality); Bryant Thomas
                    (Virginia Department of Environmental Quality); Mark Trice (Maryland Department
                    of Natural Resources); Michael Williams  (University  of Maryland Center for
                    Environmental  Sciences/Chesapeake Bay Program Office);  and Dave Wolanski
                    (Delaware Department of Natural Resources and  Environmental Control).



                    Diana Esher, Chair  (U.S.  Environmental  Protection Agency  Chesapeake  Bay
                    Program  Office); Richard  Batiuk (U.S. Environmental  Protection  Agency
                    Chesapeake Bay Program Office); Sheila Besse (District of Columbia Department of
                    the Environment); Bill Brannon (West Virginia  Department of Environmental
                    Protection);  Patricia Buckley  (Pennsylvania  Department of Environmental
                    Protection);  Katherine Bunting-Howarth  (Delaware  Department  of  Natural
                    Resources and Environmental Control); Jennifer  Capagnini (Delaware Department
                    of Natural  Resources  and Environmental Control); Moira Croghan  (Virginia
                    Department of Conservation and Recreation); Frank Dawson (Maryland Department
                    of Natural Resources); Rusty Diamond (Department of Environmental Protection);
                    Ron Entringer (New York Department of Environmental Conservation);  Richard
                    Eskin (Maryland  Department of the Environment); Stuart Gansell  (Pennsylvania
                    Department of Environmental Protection); Dave Goshorn (Maryland Department of
                    Natural Resources); Carlton Hay wood (Interstate Commission on the Potomac River
  Foreword

-------
                                                                                              IX
Basin; Teresa Koon (West Virginia Soil Conservation Association); Bruce Michael
(Maryland  Department of Natural Resources); Matt Monroe (West Virginia
Department  of Agriculture);  Kenn  Pattison (Pennsylvania  Department of
Environmental Protection); Alan Pollock (Virginia Department of Environmental
Quality); John Schneider  (Delaware  Department  of Natural Resources and
Environmental Control);  Rick  Shertzer  (Pennsylvania State  Department of
Environmental Protection); Tom Simpson (University of Maryland); Randolph Sovic
(West Virginia Department of Environmental Protection); Pat Stuntz (Chesapeake
Bay Commission); Ann Swanson  (Chesapeake Bay Commission);  and Robert
Yowell (Pennsylvania Department of Environmental Protection).

           Y. -                                    	rcc

The support and expert advice of all the members of the Chesapeake Bay Program's
Scientific and Technical Advisory  Committee, under the leadership  of Dr. Carl
Hershner (Virginia Institute of Marine Science) and the Executive Secretarial sup-
port of Dr.  Kevin  Sellner (Chesapeake Research Consortium) are  hereby acknowl-
edged. The  Scientific and Technical Advisory Committee convened a panel of inde-
pendent scientific  experts to provide expert advice and direction on a set of criteria
assessment  issues and procedures. The members of  the Panel for  Review of
Chesapeake Bay Program Analytical Tools were: Dr. David Secor, Chair (University
of Maryland Center for Environmental Science); Dr. Frank Curriero (Johns Hopkins
University); Dr. Mary Christman (University of Florida); and Dr. Ken Rechow (Duke
University). Elgin Perry, independent statistical consultant, provided the statistical
analysis support to the panel. Chapter 2 summarizes the panel's findings and recom-
mendations; Appendix A describes the findings and recommendations in full.
The contributions  of the independent scientific peer reviewers—selected and con-
vened by  the  Chesapeake Bay Program's  Scientific  and Technical Advisory
Committee  based on their recognized national expertise and drawn from institutions
and agencies across the country—are hereby acknowledged.

Without the efforts of the hundreds of colleagues involved in all aspects of field col-
lection,  laboratory analysis, management, and interpretation of Chesapeake Bay
Monitoring Program data over the past two decades,  these criteria could not have
been derived and the criteria assessment procedures could not have been developed.
Technical editing by Nina Fisher, independent technical editor, and document prepa-
ration by Jamie Bosiljevac,  Chesapeake Research Consortium/Chesapeake Bay
Program Office, are also acknowledged.

-------
                             chapter  |
In April 2003, the U.S. Environmental Protection Agency (EPA) published the Am-
bient Water Quality Criteria for Dissolved Oxygen, Water Clarity and Chlorophyll a
for the Chesapeake Bay and Its Tidal Tributaries (Regional Criteria Guidance) in
cooperation with and on behalf of the six watershed states—New York, Pennsyl-
vania,  Maryland,  Delaware,  Virginia, and  West  Virginia—and the  District  of
Columbia. The culmination of three years of work, the criteria document resulted
directly from the collective contributions of hundreds of regional scientists, technical
staff, and agency managers as well as the independent review by recognized scien-
tific experts across the country (U.S. EPA 2003).
In October 2004, EPA published the first addendum to the 2003 Regional Criteria
Guidance (U.S. EPA 2004). The addendum provided additional guidance on:
   • Applying the temperature-based open-water dissolved oxygen criteria required
     to protect the endangered shortnose sturgeon;
   • Assessing attainment of the instantaneous minimum and 7-day mean dissolved
     oxygen criteria using monthly mean water quality monitoring data;
   • Deriving site-specific dissolved oxygen criteria and  assessing criteria attain-
     ment of those tidal systems where the extensive adjacent tidal wetlands cause
     naturally low dissolved oxygen levels;
   • Delineating the upper and lower boundaries of the pycnocline that defines the
     vertical boundaries  distinguishing open-water, deep-water, and deep-channel
     designated uses;
   • Applying, in combination, the numerical water clarity criteria to shallow water
     habitats  and   submerged aquatic vegetation  restoration  goal acreages for
     defining  attainment  of the shallow-water bay grass designated use; and
   • Determining  where numerical chlorophyll a criteria  should  apply  to local
     Chesapeake Bay and tidal tributary waters.

From 2004 through early 2006, Delaware, Maryland, Virginia, and the District of
Columbia adopted: the EPA-published Chesapeake Bay water quality criteria for
                                                                       chapter i  »  Introduction

-------
                   dissolved oxygen, water clarity, and chlorophyll a; the EPA-recommended tidal
                   water designated uses; and the EPA-established criteria assessment procedures into
                   their respective state water quality  standards regulations. All four jurisdictions1
                   promulgated narrative chlorophyll a criteria in their standards regulations. Virginia
                   promulgated numerical segment- and season-specific chlorophyll  a criteria for the
                   tidal James River. The  District of Columbia promulgated numerical chlorophyll a
                   criteria for its reach of the tidal Potomac River and its remaining tidal waters, having
                   previously  adopted numerical chlorophyll a  criteria for protection of the tidal
                   Anacostia River.
                   The April 2003 Regional Criteria Guidance and the  October 2004  addendum docu-
                   ments published the criteria attainment assessment methods (U.S. EPA 2003, 2004).
                   These methods characterize the spatial and temporal variability of the appropriate
                   water quality parameters and provide a clear basis for deciding whether  a criterion
                   or set of criteria protecting a designated use in a specific segment of the mainstem
                   Chesapeake Bay or one of the tidal tributaries or embayments were in attainment.
                   The methods  were quite detailed; however, specific technical and procedural issues
                   remained in applying the methods as specified in the original publication by EPA
                   from April 2003. These issues required resolution  to allow Delaware,  Maryland,
                   Virginia, and the District of Columbia to assess attainment of their new Chesapeake
                   Bay water quality standards regulations fully.

                   This second addendum documents the revised, refined, and new criteria assessment
                   methods for the published Chesapeake  Bay dissolved oxygen, water clarity, and
                   chlorophyll a criteria.
                      •  Chapter 2 documents refinements to and recommendations for further devel-
                        opment  of the spatial interpolation and  statistical aspects  of the  overall
                        Chesapeake  Bay water quality criteria attainment assessment methodology.
                      •  Chapter 3  documents the resolution of and recommended procedures for
                        addressing  a series  of overarching Chesapeake Bay  water quality  criteria
                        assessment issues.
                      •  Chapter 4  documents  refinements and  additions to  the  procedures for
                        assessing the previously published Chesapeake Bay dissolved oxygen criteria.
                      •  Chapter 5  documents  refinements and  additions to  the  procedures for
                        assessing the previously published Chesapeake Bay water clarity criteria and
                        determining attainment of the shallow-water bay grass designated use.
                      •  Chapter 6  documents  refinements and  additions to  the  procedures for
                        assessing attainment of state-adopted numerical concentration-based chloro-
                        phyll a criteria.
                   'References throughout the text to "states" or "jurisdictions" means a collective reference to the states
                   of Delaware and Maryland, the Commonwealth of Virginia, and the District of Columbia. All four have
                   Chesapeake Bay tidal waters within their jurisdictional boundaries.
chapter

-------
   •  Chapter 7 documents new recommended methodologies and procedures for
     using shallow-water monitoring data in assessing attainment of Chesapeake
     Bay water quality criteria and tidal water designated uses.
   •  Chapter 8 documents a recommended 303(d) list decision-making framework
     for assessment of Chesapeake Bay and its tidal tributaries and embayments.

This document represents the second formal addendum to the 2003 Chesapeake Bay
water quality criteria document; as such, readers should regard the sections in this
document as new or replacement chapters and appendices to the original published
report. The criteria attainment assessment procedures published in this  addendum
replace and otherwise supercede similar criteria assessment procedures originally
published in the 2003 Regional Criteria Guidance and 2004 addendum (U.S. EPA
2003, 2004). Publication of future addendums by EPA on behalf of the Chesapeake
Bay Program watershed jurisdictional partners is likely  as  continued  scientific
research and management  applications reveal  new insights  and  knowledge that
should be incorporated into revisions of state water quality standards regulations in
upcoming triennial reviews.
U.S. Environmental Protection Agency. 2003. Ambient Water Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a for Chesapeake Bay and Its Tidal Tributaries. EPA
903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.

U.S. Environmental Protection Agency. 2004. Ambient Water Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a for Chesapeake Bay and Its Tidal Tributaries -
2004 Addendum. EPA 903-R-04-005. Region III  Chesapeake Bay  Program Office,
Annapolis, MD.
                                                                        chapter i  »  Introduction

-------
                                         chapter||
                 Refinements  to  the  Chesapeake
                       Bay  Water  Quality  Criteria
                       Assessment  Methodology
                                         BACKGROUND
                The Chesapeake Bay water quality criteria were designed to protect the ecological
                integrity of the Bay's tidal waters. To ensure that the criteria are being attained and
                the Chesapeake Bay ecosystem is, in fact, protected, adequate means to measure and
                evaluate water quality relative to the criteria must exist. The Bay is a highly diverse
                and variable system; these characteristics make precise assessment of water quality
                criteria attainment difficult. Thus, it is critical to design both a data collection system
                and a data analysis methodology carefully to make the best use of existing resources
                and provide the best possible assessment of water quality criteria attainment. Such a
                design  can inform stakeholders about the status of impairments and whether the
                impairments have been removed once management actions have resulted in the
                achievement of the desired restoration goals.
                To address the need for enhanced water quality criteria assessments brought on by
                the states' adoption of new Chesapeake Bay water quality standards, the Chesapeake
                Bay Program1 redesigned its tidal monitoring network to provide a framework for
                interpreting the data. To the extent possible (within funding constraints), existing
                monitoring programs were either enhanced to support  criteria assessment or new
                monitoring programs were established to address monitoring gaps. Given the diver-
                sity  of  tidal habitats throughout the  Bay,  establishing a comprehensive  tidal
                monitoring network required different types of monitoring.
                lrThe Chesapeake Bay Program, formed in 1983 by the first Chesapeake Bay agreement, is a unique
                regional partnership guiding the restoration of the Chesapeake Bay and its tidal tributaries. On water
                quality issues, the Chesapeake  Bay Program partners include Delaware, Maryland, New York,
                Pennsylvania, Virginia, West Virginia, the District of Columbia, the Chesapeake Bay Commission, the
                U.S. Environmental Protection Agency, over 20 other federal agencies, academic institutions, local gov-
                ernments, and citizen groups.
chapter ii  •  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
Developing a methodology for assessing criteria attainment using these data was
also critical. Ideally the criteria assessment methodology would prove useful in
several ways:  1) it could be applied consistently for many water quality criteria
components; 2) it would provide a common framework for assessing data collected
over multiple scales; 3)  it would provide a basis for using as much of the informa-
tion contained in the collected data as possible; 4) it would provide a clear basis for
making decisions on criteria attainment; and 5) it would provide diagnostic infor-
mation  regarding  the spatial and temporal  patterns of criteria violations.  The
cumulative frequency diagram (CFD) approach, described in the original  2003
Chesapeake Bay water quality criteria document, was designed with many of these
objectives in mind (U.S. EPA 2003a).
                 OF THE

The original 2003 Chesapeake Bay water-quality criteria document fully describes
the CFD methodology (Chapter 6, pages 154-178), but is summarized briefly here
(U.S.  EPA 2003a).  Criteria assessment using the CFD methodology is based on
interpolation within a spatially defined grid. Described later in this chapter, this grid-
based interpolation provides  the  spatial framework for use of all of the data. It
weights each data location according to the amount of area (or volume) it represents.
Water quality parameter levels in all interpolator grid cells are estimated based on
interpolation algorithms, providing a complete "map" of water quality throughout
the assessed area (Figure II-1). Water quality parameter levels in each grid cell are
compared to the applicable criteria levels to establish an estimate of the spatial extent
of criteria exceedance (non-attainment). Aggregating the total amount of space (area
or volume) in which the criteria are exceeded provides a basis for estimating the
percentage of the spatial assessment unit (designated use within a segment) in which
the criteria were exceeded for that monitoring  cruise.  These measures of criteria
exceedance are then compiled over the entire assessment period to develop a cumu-
lative frequency diagram, or CFD. The CFD is a well-known and well-established
statistical procedure commonly used to  describe hydrologic and environmental data
(Helsel and Hirsch 1992).

The CFD assessment methodology evolved from the need to allow for variability in
water quality parameters due  to unusual events. For the water quality parameter to
be assessed, a criterion threshold is established; when the threshold is exceeded, the
system is considered impaired. All water quality parameters, however, are inherently
variable  in space and time. Because of this variability, it is unlikely that even a
healthy Chesapeake Bay ecosystem will attain the threshold absolutely in all places
and at all times.

Spatially, small regions may persistently exceed the criteria's threshold due to poor
flushing or other natural conditions. Such areas should not automatically lead to the
assumption that the entire assessment unit is impaired. Similar logic applies in the
        chapter ii  «  Refinements to the Chesapeake Bay 'Water Quality Criteria Assessment Methodology

-------
6
                    Figure 11-1. Example of interpolation of Chesapeake Bay water quality data.
                   temporal dimension. Water quality in a large area of a segment may exceed the
                   criteria's threshold for a short time. If this degradation proves infrequent and short-
                   lived with the segment quickly returning to a healthy state, this situation does not
                   represent an impairment of the ecologically defined designated use of the segment.

                   Recognition that ephemeral exceedances of the criterion's threshold in time or space
                   do not represent persistent impairment of the segment's designated use ultimately
                   led to the development  of a criteria assessment methodology that deems  such
                   exceedances as acceptable. Persistent, widespread criteria exceedance, however, is
                   considered an impairment of the segment's designated use (U.S. EPA 2003a).

                   The criteria assessment methodology determines how much  of the spatial  assess-
                   ment unit is not in compliance with the criteria (percent of space) for each moment
                   in time. In the second step of the methodology, a determination is made of how often
                   (percent of time) a segment is out of attainment by more than a fixed percent of
 chapter ii  •  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
space. The results of these queries can be presented in graphical form with percent
of time plotted against percent of space.
Figure 11-2 illustrates a typical CFD based on 12 measures of spatial extent of criteria
exceedance over time. In general, if a segment is in attainment with the criterion,
then one expects a high frequency of dates for  which the percent out of attainment
is low. In this case, the CFD should descend rapidly from the upper left corner, pass
not far from the lower left corner, and then proceed to the lower right corner. The line
in Figure II-2 shows the typical hyperbolic shape commonly observed using the CFD
to assess water quality criteria in the Chesapeake Bay. The closer the CFD  curve
comes to the origin (lower  left corner), the  better the attainment of the assessed
segment. A curve that is  far from the origin indicates  that a larger percent of space
in the segment is out of attainment and the probability of use impairment increases.
The CFD  methodology  offers  many  advantages over  other  criteria assessment
approaches. Through interpolation, it provides a method for using data collected in
areas surrounding the area of interest  (the spatial assessment unit). This factor is
important since the sample size of observations  within a spatial assessment unit may
not be sufficient to determine the area (or volume) of exceedance within the unit
accurately. The  method  also weights  the data  collected from a given location
according to the amount of area (or volume) that the location represents. This capa-
bility is important because data may be collected from locations that do not represent
    1.0

    0.9

    0.8

    0.7
 ~  0.6
 "o
 o  0.5
 o
 §• 0.4
 Q.
    0.3

    0.2

    0.1

    O.Ol
      0.0     0.1    0.2     0.3     0.4     0.5     0.6    0.7
                                  Proportion of Space
                                                            0.8
                                                                   0.9
                                                                          1.0
      II-2. A water quality criteria attainment assessment cumulative frequency diagram
(CFD) based on 12 measures of the spatial extent of criteria exceedance over time.

-------
                   all areas of the spatial assessment unit; providing equal weight to such data could
                   bias the assessments.
                   A second advantage is that the CFD incorporates the spatial-temporal pattern of
                   criteria exceedance into the assessment. The shape of the curve offers information on
                   patterns of exceedance in space and time. Such information may prove helpful in
                   understanding the causes of impairments (see page 162 in U.S. EPA 2003a).
                   A third advantage is that it bases the assessment on biologically determined patterns
                   of allowable criteria exceedance. Reference curves are ideally developed in the same
                   way as assessment curves and should reflect the degree of criteria exceedance that
                   can be withstood by the  ecological  communities without impairing the designated
                   use. Thus, comparison of the assessment curve to the reference curve ensures that
                   any allowable criteria exceedances do not occur in a spatial or temporal pattern that
                   could, in reality, represent impairment at the scale of the entire assessment unit (see
                   pages 162-178 in U.S. EPA 2003a). Local persistent effects could still have high
                   impairment.
                   Finally, the combined elements of the CFD criteria assessment methodology fully
                   and effectively address all five factors used to determine attainment of designated
                   uses: magnitude, duration, frequency, space, and time. After conducting a national
                   review of TMDL programs, the National Research Council (2001) concluded that
                   establishing these conditions is  crucial for  successful application of state water
                   quality standards.
                   The CFD methodology is a new and innovative method of water quality criteria
                   assessment, representing an improvement over methods  used in other parts of the
                   country (STAC 2006). The standard practice for assessing compliance with water
                   quality criteria throughout the United States is by sampling monthly at a fixed set of
                   stations and  gauging compliance strictly from a count of exceedances of those
                   samples. Sampling stations are typically located for convenience (e.g., accessibility).
                   Consequently, reluctance to re-evaluate and change location (so as to maintain a time
                   series at a fixed point) is common; no consideration is given to the representative-
                   ness of the sample for the space/time not sampled.
                   Most  assessments are based simply on EPA assessment guidance in which all
                   samples in a given area were compiled; attainment was assumed if no more than 10
                   percent of the samples exceeded the standard (U.S. EPA 1997). In this approach, all
                   samples are assumed to be fully representative of the specified space and time and
                   are simply combined as if they were random samples from a uniform population.
                   This approach was necessary in the  past because the technology did not exist for a
                   more rigorous method of data analysis; however, it neglected spatial and temporal
                   patterns in the criteria measures. The CFD approach was designed to characterize
                   these spatial and temporal patterns and weight samples more accurately based on the
                   amount of space or time that they actually represent.
chapter

-------
The CFD methodology was first applied in the Chesapeake Bay for the most recent
303(d) listing cycle, completed in the spring of 2006 and based on data from 2002
through 2004. The  CFDs were developed  and used primarily for  the dissolved
oxygen open-water and deep-water 30-day mean criteria because insufficient data
and data analysis techniques existed to assess the higher-frequency dissolved oxygen
criteria components. Similarly, the water clarity criteria were not assessed based on
the CFD because few tidal systems had sufficient shallow-water monitoring data for
an assessment.

In fall 2005, the  Chesapeake Bay Program's  Scientific and Technical Advisory
Committee (STAC)  established a scientific panel to review  and refine  the CFD
assessment methodology. Nationally recognized academic experts in spatial  and
environmental statistics made up  the panel.  The STAC-convened panel concluded
that the CFD approach is both feasible and innovative, qualifies as the best available
science, and represents an improvement over criteria assessment methods used in the
past (STAC 2006).
The panel also recognized, however, that the approach remains in the early stages of
management application. It stated that the CFD approach deserves further directed
study and analysis to evaluate the bias and imprecision that can occur due to limita-
tions  in available  data  and  in current  interpolation and CFD algorithms (STAC
2006). This chapter provides guidance for criteria assessment application, summa-
rizes findings from the  CFD evaluations, and offers recommendations for further
refinement of the CFD assessment methodology. Appendix A provides a  complete
copy of the scientific panel's final report.
  DESCRIPTION  AND  EVALUATION OF  THE  CFD-BASED
                 ASSESSMENT METHODOLOGY

The methodology for estimating the CFD is most easily described as a series of eight
steps as shown in Table II-1. These steps, described below, provide a framework for
considering the process and are elucidated by a simple example. More detailed
discussions of each step follow later in this chapter.

EXAMPLE CFD-BASED CRITERIA ASSESSMENT

To illustrate the CFD criteria assessment methodology, a simple theoretical example
based on a small data set can prove useful. Assume a segment for which the inter-
polation grid is 4 cells  by 4 cells. In reality, the number of grid cells is much larger
(hundreds to thousands), but this small grid is illustrative.  Also assume that data
were collected on five distinct dates, and that each date is representative of the appro-
priate time scale (in an actual application, data would be collected over many more
dates). The criterion threshold for this fictitious water quality parameter is 3.
        chapter ii  •  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
10
                            -1  Steps for constructing and assessing criteria attainment using cumulative
                               frequency diagrams (CFDs).

                       1.  Collect data from a spatial network of locations on several dates during the
                          assessment period.

                       2.  For each date, interpolate the data spatially over the entire system to obtain esti-
                          mates of water quality using a two- or three-dimensional grid of interpolation
                          cells.

                       3.  Aggregate interpolations to the appropriate temporal scale (e.g., if evaluating the
                          30-day mean, take the average of all interpolations for each date in the month).

                       4.  For each interpolator cell, assess whether the applicable criterion is exceeded.

                       5.  For each assessment unit, compute the percentage of interpolator cells that exceed
                          the criterion as an estimate of the percent of area (or volume) within the spatial
                          assessment unit that exceeds the criterion.

                       6.  Rank the percent of area estimates for the set of all sample days in the  assessment
                          period from largest to smallest and sequentially assign to these ranked percents a
                          value that estimates percent of time. Add the end points of (100%, 0%) and (0%,
                          100%).

                       7.  Plot the paired percent of area (or volume) and percent of time data on a graph
                          with the percent of area on the x-axis and percent of time on the y-axis. The
                          resultant plot is the assessment cumulative frequency diagram or CFD.

                       8.  Compare the assessment CFD (from step 7) to the appropriate reference CFD. If at
                          any point the assessment CFD exceeds the reference CFD (i.e., a given level of
                          spatial noncompliance occurs more often than allowed for a given amount of
                          time), then the criterion is in non-attainment. Consequently, the segment fails to
                          meet that designated use.
                     An illustration  of the eight steps for computing the  CFD for these simplified
                     constraints is shown on the facing page. The three columns show the first three steps.
                     Column 1 provides fictional data  for five dates  for five fixed  locations in a two-
                     dimensional grid. Column 2 shows a fictional interpolation of these data to cover the
                     entire grid. Column 3 gives the compliance status of each cell in the grid with 1 indi-
                     cating non-attainment and 0 signifying attainment.

                     In this hypothetical example, the assessment curve is clearly greater than the refer-
                     ence curve and in non-attainment of the criterion,  therefore, the designated use is not
                     met. EPA recommends that any exceedance of the attainment CFD above the refer-
                     ence CFD should be considered non-attainment of the criterion and, consequently,
                     the designated use.
  chapter ii  »  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                                        11
  Step 1. Collect data at
  known locations.
Step 2. Interpolate the
data to grid cells.
Steps 3-4. Determine
attainment status of
each cell.
  Date 1
Date 1
Date 1
3


2





5


3


1
Date 2
1


1





3


1


1
Date 3
4


1





2


2


1
Date 4
1


4





2


4


1
DateS
1


1





2


3


1
3
4
3
2
4
4
3
3
5
5
4
3
3
2
1
1
                               Date 2
                               Date 3
                               Date 4
                               DateS
1
2
1
1
2
2
3
1
3
3
2
1
1
2
1
1
4
3
2
1
3
2
2
1
2
2
1
1
2
1
1
1
1
2
3
4
2
2
3
3
3
2
2
1
4
3
1
1
1
2
1
1
2
2
1
1
3
2
1
1
3
2
1
1
                              Date 3
                              Date 4
                              DateS
1
1
1
0
1
1
1
1
1
1
1
1
1
0
0
0
Date 2
0
0
0
0
0
0
1
0
1
1
0
0
0
0
0
0
1
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
1
1
1
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
  Step 5: Determine
  percent attainment by date.
Sample date
Date 1
Date 2
DateS
Date 4
DateS
Percent
space
75.00%
18.75%
18.75%
43.75%
12.50%
Step 6. Rank the percent of
space values and assign percent
of time as (100*R/(N+1)), where
R is rank and N is sample size.
Sample date

Date 1
Date 4
Date 2
Date 3
DateS

Ranked
percent
space
100%
75.00%
43.75%
18.75%
18.75%
12.50%
0%
Percent time
0%
16.67%
33.33%
50.00%
66.67%
83.33%
100%
          Steps 7 and 8. Figure II-3 illustrates the plot of this
          theoretical assessment CFD and the comparison to a
          hypothetical reference curve. In this hypothetical ex-
          ample, the assessment area shows non-attainment. For a
          percent area of 18.75, the allowable frequency on the
          reference curve is about 17 percent. That is, 18.75 per-
          cent of the segment area should not be out of attainment
          more that 17 percent of the time. For Date 3, the esti-
          mated frequency of 18.75 percent of segment area in
          non-attainment is 66.67 percent. Thus the frequency of
          18.75 percent of space out of attainment exceeds the 17
          percent allowed. The reference curve is exceeded for
          dates 4 and 1 as well.2
                                         2In this cumulative distribution framework, the actual date is
                                         not relevant. One should not infer that non-attainment occurred
                                         on that date if the data point associated with a date falls above
                                         the reference. The date is used here as a label for each coordi-
                                         nate pair.
         chapter ii  •  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
12
                               0%    10%   20%   30%    40%   50%   60%   70%    80%   90%   100%
                                                Percent of Area Exceeding the Criterion
                           II-3. Graphical representation of the CFD from the above theoretical example
                    assessment curve (blue) with a hypothetical reference curve (black).
                    CFD
                    Two approaches are feasible in defining the reference curves proposed for use in the
                    CFD assessment methodology. One is biologically based and identifies appropriate
                    regions of the Bay, its tidal tributaries, and its embayments that have healthy biolog-
                    ical indicators and are in attainment of their designated use (U.S. EPA 2003a). The
                    CFDs are developed for these areas in the same way that assessment CFDs would be
                    developed elsewhere. Curves generated  for biologically healthy tidal areas are
                    considered "reference" curves.
                    For example, healthy benthic indices of biotic integrity (IBI) scores might be used as
                    indicators of adequate bottom dissolved oxygen (Weisberg et al.  1997; U.S. EPA
                    2003a). Thus, after stratifying by salinity zone and perhaps other factors, a series of
                    dissolved oxygen reference CFD curves could be developed from the existing moni-
                    toring database. One  advantage  of this approach is that each biological reference
                    curve could be tailored to each designated-use-based criteria component. This tech-
                    nique tailors the pattern of criteria exceedance that the part of the Bay ecosystem
                    could tolerate and remain healthy to the protected species and biological communi-
                    ties and the specific criterion component. Thus, each reference curve may have a
                    somewhat different shape (see pages 168-177 in U.S. EPA 2003a).
  chapter

-------
                                                                                                  13
In some cases, development of a biologically-based reference curve is not possible
due to lack of data describing the health of the relevant species or biological commu-
nities. Such  cases require a different approach. The  EPA recommends use of a
default reference curve in situations for which a biologically based reference curve
remains unavailable. This default reference curve is defined as a hyperbolic curve
that encompasses no more than 10 percent of the area of the CFD graph (percent of
space multiplied by percent of time) (see page 174 in U.S. EPA 2003a) (Figure II-4).
The default reference curve has the following important properties:  1) the plot is
symmetric about the 1:1 line; 2) the plot is hyperbolic; 3) the total area under the
100%
"o
o
5
o .0
S0* ~ 7n°/
c 0
8 °!
i- £ Rn%
Q_ W
_ "C
^ Q)
~ g OU/o
'o X
giu
Q- 0> 4QO/ _
w E 40/0
(0 3
o> o
jZ ^S
Iz ®
o < 20%
c
0>
o
Q.











0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Area/Volume Exceeding the Criterion
          Default reference curve for application in the attainment assessment of
Chesapeake Bay water quality criteria for which biologically based reference curves
cannot be derived.
curve equals 10 percent; and 4) the ends of the curve pass through x- and y-axis
intercepts (100, 0) and (0, 100), respectively.
Figure II-4 is defined by the equation:

                         (x + b)(y + b) = a
where:  b = 0.0429945 and a = b2 + b.
Equation 1

-------
14
                    No specific theoretical basis underlies this definition of the default reference curve,
                    but the definition does provide equal weight to exceedances occurring in either space
                    or time. This approach is appropriate since  no information exists to indicate that
                    either time or space should take precedence. Selection of the 10 percent value is
                    based on its consistency with past national  EPA guidance (U.S. EPA 1997). The
                    default reference curve is hyperbolic, making it similar in  shape to biologically
                    based reference curves. In fact, the shape of the default reference curve is quite
                    similar to some of the established biologically based reference curves, such as the
                    30-day mean open-water dissolved oxygen reference curve (Figure II-5).

"o
o
o
8,1
IB 70/0
80
0)0) 60 ^
Q. £
~° «
o> TJ ^n% -
_ -a au/0
o a>
S S
a 1

o
Q.










0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume Exceeding the Criterion
                               Biological reference curve for 30-day mean open-water dissolved oxygen
                    criterion applied for assessment during the summer months (June-September) only.

                    A default reference curve, defined as a hyperbolic curve encompassing no more than
                    10 percent frequency exceedances, was also considered. Such a curve is based on a
                    simple model:
y = u + a; + bj
                                                                                       Equation 2
                    where a is temporal term with variance 2a and b is spatial term with 2b. The vari-
                    ance of Xy is  $2a + $2b = $2. The standard deviation of Xy is  sqrt(2) = $. Ten
                    percent of the xy- should fall above u +  1.2815  * $ where  1.2815 is  the 90th
                    percentile of the standard normal distribution. Thus, assuming normality, a popula-
                    tion with equal spatial and temporal variance and a mean that is 1.2815 *   below
                    the threshold criterion should have an exceedance rate of 10 percent over space and
                    time. Figure II-6 shows the CFD for the 10 percent frequency exceedance default
                    reference curve in black.
  chapter

-------
                                                                                                  15
Also plotted on this same axis in blue in Figure II-6 is a default reference curve
based on 10 percent of the area of the percent space x percent time (the default refer-
ence curve described previously and illustrated in Figure II-4). This evaluation was
undertaken given an approach  to  deriving and assessing attainment of numerical
chlorophyll a criteria is based largely on thresholds that should rarely be exceeded
in healthy populations (e.g., the 90th percentile). These two curves are very close in
shape, further supporting the use of the default reference based on a 10 percent area
under the curve. The EPA recommends use of the default reference curve, illustrated
in Figure III-4  and defined by  Equation 1,  when an  applicable  biologically-based
reference curve is not available.
             0.1     0.2     0.3    0.4     0.5     0.6    0.7    0.8     0.9    1.0
      II-6. Comparison of hyperbolic curves based on 10 percent of area under the curve
(blue) and 10 percent frequency exceedance (black).
              <  ',

Reference curves are more or less continuously defined while  assessment curves
have relatively few discrete measures.  Biological reference curves can contain
hundreds of points; the default reference curve has an infinite number of points. By
contrast, curves for three-year assessments of summer (June-September) monthly
means will  have 12  data points  with the  curve defined by linear interpolation
between neighboring points. For this reason, it is possible for portions of the assess-
ment curve  to be above the reference  curve  even  without any measured point
exceeding the reference curve.  This situation  becomes  more comprehensible by
        chapter in  «  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
16
                    understanding that reference curves typically have positive curvature and that this
                    curvature can dip below the line between consecutive points  on the  assessment
                    curve, causing a spurious, non-allowable exceedance.

                    To address this problem, the EPA recommends that reference curves be evaluated
                    only at the temporal axis points in the assessment curve as illustrated in Figure II-7.
                    For non-continuous biological reference  curves, the points should be interpolated
                    from neighboring points. Appendix B provides a detailed description of the complete
                    Chesapeake Bay water quality criteria attainment assessment methodology.
Dissolved Oxygen DW Monthly new curve discrete points CB7PH 2002—2004
1.0


0.9
0.8

0.7(

E 0.6'
F
0 <
c 0.5
lo,

0.3
0.2

0.1
3 ja j a

D




\
|V.
\
r \
i \
V X.

- """"""•SN
Ox s%.!tv
^-V 	
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Fraction of Space

-O- assessment curve
""""' reference curve

-

-

_

-
_


_

-

0.9
0.8

0.7(

E 0.61
F
0 <
c 0.5
~
I 0.4
0.3

0.2
0.1
0.0







-\
\
- \
$ \
\ V,
- to- "\

- """V
- s X '***'»,
e- - - a s**'*^.^^^
0.8 0.9 1.0 Fraction of Space

-O- assessment curve
,„„,,,, reference curve
-
_

_

_

_

-
_

-
_
— r~— ._

















0.8 0.9 1.0

        II-7. The graph on the left (A) shows spurious non-attainment as the reference curve passes below the
 assessment curve between points. The graph on the right (B) shows attainment as the reference and assessment
 curves are evaluated at the same temporal axis points.
                                      OF A

                    A statistical framework for making decisions on water quality criteria attainment
                    based on the CFD methodology would yield additional information on the certainty
                    of the attainment decisions. It would also help direct appropriate monitoring strate-
                    gies to reduce uncertainties.  However,  many theoretical obstacles remain in
                    developing such a  framework. The CFD methodology is a new and innovative
                    approach to water quality criteria assessment. The relatively recent application of
                    this methodology to criteria assessment suggests that conducting further evaluations
                    and making improvements  should prove  constructive.  The  following section
                    discusses the steps in applying the CFD methodology.

                          1 —

                    One of the advantages of the CFD approach is that it can accommodate a variety of
                    input data and still arrive at the same assessment endpoint. Data collection methods
                    currently in place include: fixed-station data, cruise track data, continuous moni-
                    toring  data, aircraft flight path  data, and satellite  imagery data.  Because of the
  chapter ii  »  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                                   17
interpolation step, all of these data can be used with varying degrees of success to
estimate the total spatial distribution (to the limit of interpolator pixel size) of a water
quality parameter.
Interpolation can place data collected at various spatial densities on a common footing.
On the one hand, this capability is advantageous because data collected at different
spatial densities are available for the criteria assessment process. On the other hand, it
can be misleading to accept interpolated surfaces from different data sources as equiv-
alent  without qualifying each interpolation with  a measure of the estimation  error
associated with each data type. Clearly, an interpolation based on hundreds of points per
segment (such as cruise track  data) more  accurately reflects the true non-attainment
percentage when compared to an interpolation based on two or three points per segment
(such as a fixed-station data). Of the various types of interpolation algorithms available
and reviewed, kriging is best positioned to address this issue (STAC 2006). Kriging
offers advantages over inverse distance weighting in that it provides the best assessment
of the estimation error associated with interpolation, but has not been implemented to
date. Other methods, such as interpolating polynomials, splines, and locally weighted
regression methods, should also be explored.

      3—                          of
Depending on the interpolation  method and the  statistics  available, it may  be
possible to calculate the probability of exceedance of the temporal mean at  each
point given the likely variance and the value(s) observed during the period. This step
is necessary to calculate probabilities in the following  step.

      4-	
Determining the percent attainment of each grid cell from each interpolation seems
simple.  If the estimated value for a grid cell is above  (chlorophyll  a) or below
(dissolved oxygen,  water clarity) the criterion, then that cell is not in attainment.
While interpolation allows for standardization of many types of data, pointwise
attainment determination allows for standardization of many criteria. Because attain-
ment is determined at moments in time and points in space, it is possible to vary the
criterion in time and space. If different  levels of a water quality constituent are
acceptable in different seasons, then the criterion  can vary seasonally. It is possible
to  implement different criteria over space for a segment that bridges, for example,
oligohaline and mesohaline salinity regimes. It might even be possible to let the
criterion be a continuous function of some ancillary variable such as temperature or
salinity, although this situation requires that such data exist for every interpolator
cell. The only requirement is that the final attainment determination be "yes" or "no"
for each interpolator cell.

        chapter in  »  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
18
                    Currently, limited pointwise  attainment determination compliance has  been im-
                    plemented. For example, the open-water 30-day mean dissolved oxygen criterion is
                    5 mgxliter1, except when the  ambient salinity drops below 0.5 psu and the criterion
                    becomes 5.5 mgxliter1 (U.S. EPA 2003a). During the summer months,  the open-
                    water designated use boundaries are selected based on local density conditions
                    reflecting stratification of the  water column.

                    Even the simplicity of this concept diminishes when examining interpolation error.
                    Consider the assessment of two interpolator cells from  an interpolation based on
                    cruise track data. While both  interpolations could have the same value, each could
                    have a different level of error. Such different levels of error could mean  that these
                    were different probabilities that the criteria were actually exceeded. For the simple
                    assessment of non-attainment, however, they count the same. Thus, one advantage of
                    a statistical framework is that it accounts for different levels of error throughout the
                    interpolation grid and these error levels could be incorporated into a single overall
                    assessment of attainment.

                    Stt;  ;  ~\                         in

                    Computing a percentage should also be simple. The estimate is simply 100 times the
                    number of cells not in attainment divided by the total number of cells. As a rule, the
                    uncertainty of a binary process can be modeled using a binomial distribution. The
                    issue of uncertainty described in step 3  propagates into  computing the percent of
                    attainment for a segment. In addition, estimated values for interpolator  cells have a
                    complex dependence structure,  ruling out a simple binomial  model. The rules
                    governing the uncertainty  of this step  are  also complex.  The mathematics  for
                    modeling this propagation of error are feasible, but have not yet been developed.

                               •            "ne

                    While the CFD's percent-of-space coordinate provides a simple interpretation of the
                    percent of the  spatial assessment unit that is out of attainment on a given date, the
                    percent-of-time coordinate is  not simply the percent of time  out of attainment at a
                    given point. Instead this coordinate is interpreted similarly to that of a cumulative
                    distribution function; it represents the percent of time that  the associated spatial
                    percent of non-attainment is exceeded. For example, if the (percent space, percent
                    time) coordinates for a point  on the CFD are (90, 10), the spatial percent of non-
                    attainment is greater than or equal to 90 percent about 10 percent of the time.
                    This step is very similar to computing an empirical distribution function, which is an
                    estimator of a cumulative distribution function. This similarity brings to mind
                    statistical inference  tools associated with empirical  distribution  functions—the
                    Kolmogorov-Smirnov, Shapiro-Wilk, Anderson-Darling, or Cramer-von Mises — as
                    candidates  for inference about the  CFD (STAC 2006). These procedures  model
                    uncertainty as a function of sample size only (in this case, the number of sample
                    dates). Since they do not account for uncertainty associated with the number of
  chapter ii » Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                                19
samples collected in space (i.e., number of sampling stations), this indicates that they
are to provide a statistical framework that truly accounts for error in Chesapeake Bay
water quality criteria assessments.

       7      8—

When comparing the assessment curve to the reference curve, the issue of uncer-
tainty becomes most important. The preceding discussion clearly  indicates that
uncertainty in the assessment curve represents an accumulation of uncertainty gener-
ated in and propagated  through the preceding steps. If the  reference curve  is
biologically based, it is derived under the same system of error propagation. Devel-
oping the statistical algorithms to quantify this uncertainty poses a challenge.
Even if the uncertainty can be properly quantified, the issue of who gets the benefit
of doubt due to this uncertainty can prove difficult to resolve.

This problem of uncertainty in the regulatory process is widespread and not limited
to the CFD approach. Nonetheless, it must be dealt with. One option is to require that
the assessment curve be  significantly above the reference curve to establish non-
attainment. This option protects the regulated party from being deemed out  of
attainment due to random effects. If assessment CFD curves are not accurately deter-
mined, however,  it could lead to poor protection of environmental  health and
designated uses. A second option is to require that the assessment curve be signifi-
cantly below the reference curve to establish attainment. This option results in strong
protection of the environmental resource, but could lead to the regulated party imple-
menting unnecessary and expensive management actions.
Some compromise between these extremes is needed. The simplest compromise is
to ignore variability and compare the assessment curve to the reference curve.  As
long as unbiased estimation is implemented for both the assessment curve and the
reference curve, this third option will result in roughly equal numbers of false posi-
tive (declaring non-attainment when, in fact, compliance exists) and false negative
(declaring  attainment  when,  in  fact, non-attainment exists)  results.  This last
approach is balanced and the one  currently recommended by  EPA.  Under this
approach,  however, no  mechanism exists to motivate error reduction by improving
the data sets on which the criteria assessments are based.

           OF TIIF
Beginning  in fall 2005, the Chesapeake Bay Program's  Scientific and Technical
Advisory Committee (STAC) appointed a panel of scientists to evaluate and refine
the CFD water quality criteria assessment methodology. Evaluations included tests
on the effects of: 1) sample densities in time and space; 2) varying levels of attain-
ment; and 3)  varying  degrees of spatial  and temporal covariance. Appendix  A
provides a complete copy of the panel's final report while Appendix C offers a narra-
tive evaluation of the options for spatial interpolation.
        chapter in  »  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
20
                    In general, the STAC panel analysis and review indicated that the CFD approach can
                    combine spatial and temporal data to support inferences on attainment or exceedance
                    of the water quality criteria (STAC 2006). The panel viewed the CFD approach as
                    innovative — one that has general application in water quality criteria assessments.
                    In comparison to other jurisdictional authorities, the Chesapeake Bay Program has
                    taken a lead in monitoring and assessment based upon  scientific design (designated
                    uses) and emphasis on statistical  evidence. Advancement in the CFD approach
                    should provide an important precedent for states outside the Chesapeake Bay region.
                    Because the CFD is both feasible and innovative, the panel felt that it qualifies as the
                    best available approach. On the other hand, the panel recognized that the approach
                    remains nascent and deserves further directed study and analyses to evaluate the bias
                    and imprecision that can occur due to small sample densities, non-independence in
                    temporal trends, and inadequate spatial interpolations.
                    The panel found that the CFD approach in its current form is feasible, but requires
                    additional research to further refine and strengthen it as a statistical tool. The CFD
                    builds on important statistical theory related to cumulative distribution functions; as
                    such, its statistical  properties can be simulated and deduced.  In its analyses,  the
                    STAC  panel showed that constructing  confidence ellipses that support inferences
                    related to threshold curves or other tests  of spatial and temporal compliance  are
                    feasible. Understanding fundamental properties of how the CFD represents likely
                    covariances of attainment in time and space and how temporal and spatial correla-
                    tions interact  with sample  size  effects require  additional research.  Further,
                    researchers must also analyze biases across regions and designated-use segments.
                    The panel expects that two to three years  of directed research and development are
                    necessary to identify and measure potential sources of bias and imprecision  for
                    criteria attainment determinations.
                    In the near future, the panel foresees that the CFD approach will prove particularly
                    powerful when linked to continuous spatial data streams through the  cruise-track
                    monitoring program, and when able to utilize continuous temporal data generated
                    through further deployment of remote sensing platforms  in the  Chesapeake Bay
                    (e.g., Chesapeake Bay Observing System). These data  sets will allow greater preci-
                    sion and accuracy in both threshold and attainment determinations made using  the
                    CFD approach.
                    The  STAC panel concluded that success of the CFD-based assessment rests upon
                    decision rules related to the  biological reference curves.  These curves represent
                    desired segment-designated use water quality  outcomes  and reflect sources of
                    acceptable natural  variability (STAC 2006). The reference and attainment curves
                    should follow the same general approach in derivation: water quality data collection,
                    spatial interpolation, comparison to biologically  based water quality criteria, and
                    combination of space-time attainment data through a CFD. Therefore, the biological
                    reference curves allow implementation of a tolerance threshold presuming the data
                    used to derive the reference curve were sampled similarly to the assessment curve.
  chapter

-------
                                                                                              21
That is, the reference curve defines the degree to which criteria violations can be
tolerated without resulting in impairment of the designated use.

Bias and uncertainty are driven in CFD curves by sample densities in time and space.
Therefore, the STAC panel advised that similar sample densities be used in the deri-
vation of assessment and reference curves. As such densities are not always feasible,
additional analytical methods  are needed to weight sampling densities equally
between attainment and reference curves.
              APPLICATION  OF  THE  CFD-BASED
                 ASSESSMENT  METHODOLOGY

RECOMMENDATIONS FOR APPLICATION OF THE CFD-BASED
METHODOLOGY

As  stated above,  the CFD-based  water quality  criteria assessment methodology
offers the potential for significant  benefit  in accurately assessing Chesapeake Bay
water quality criteria attainment. As the STAC CFD Review Panel has indicated,
however, that the  methodology is  new and additional evaluations and refinements
should be performed (STAC 2006) (Appendix A). The EPA agrees with the panel's
conclusions, strongly supports the  findings that further research is needed, and will
support those efforts in whatever way possible in the coming years. In the meantime,
the EPA recommends the following approach in undertaking Chesapeake Bay water
quality criteria assessments.

As  described above,  the Chesapeake Bay Program collects data at two different
scales for water quality criteria attainment assessment. In each case, the  design of
data collection program focuses on assessments at a specific scale. The fixed-station
data are designed for segment and baywide  assessments and the shallow-water moni-
toring  data  are designed to assess the  small  tidal  tributaries  and the  Bay's
shallow-water habitats. Given the different scales, separate interpolations  are likely
necessary using the  most appropriate  interpolation algorithm. The STAC CFD
Review Panel evaluated  two possible  options  for  spatial interpolation, recom-
mending kriging  as  the  better of the two alternatives (STAC 2006).   Kriging,
however, has not  been fully developed for application in Chesapeake Bay water
quality criteria attainment assessment.

Until kriging is fully developed as an option for whole-Bay assessment based on the
fixed-station data,  the EPA recommends that spatial interpolations continue using the
current Chesapeake Bay Program's inverse distance weighting (IDW) algorithm-
based interpolator (Appendix D).  Spatial interpolation of the fixed-station data for
assessment of criteria attainment  in the mainstem Bay and major tidal tributaries
requires several specific capabilities including: 1) the data must be interpolated in
three-dimensions (i.e., with depth); 2) the data must be interpolated into the tidal
tributaries and around bends in these tidal rivers; and 3) the interpolation needs to be
        chapter ii •  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
22
                    automated to complete large number of criteria assessments efficiently and routinely.
                    These capabilities are  not currently  available using  a kriging algorithm, but the
                    Chesapeake Bay Program IDW interpolator is designed with these capabilities in
                    mind. Thus, the EPA recommends that large-scale interpolations (segment, baywide)
                    continue to be based on the fixed-stations data be performed using the Chesapeake
                    Bay Program IDW interpolator. As kriging is developed further for use, this option
                    may be recommended in the future.

                    For the  criteria assessment of small  tidal tributaries  and the Bay's shallow-water
                    habitats based on data from the shallow-water monitoring program, the EPA recom-
                    mends implementation of a  kriging algorithm, where possible. The shallow-water
                    monitoring  program yields  data to  assess criteria attainment in relatively  few
                    systems at any one time. Thus, it is possible to provide the more focused evaluations
                    of individual interpolations  that  kriging requires.  Furthermore, the intensive data
                    collection provided by  the shallow-water monitoring program is particularly
                    conducive to detailed statistical analysis. To utilize the data's information fully, a
                    more thorough statistical interpolation procedure, such as kriging, should be imple-
                    mented. The shallow-water systems are highly dynamic and thus better characterized
                    by more intensive data collection combined with a more rigorous statistical interpo-
                    lation algorithm.  For  these  reasons,  the EPA recommends  that  kriging  be
                    implemented, where possible, for criteria assessment based on shallow-water moni-
                    toring data.

                    Given the recommendation above, the EPA further advises that the states develop the
                    expertise to perform spatial interpolation based on statistical  methods. Assessment
                    of the shallow waters will largely fall to the states,  with some support from the
                    Chesapeake Bay Program Office. Guidelines  are  being  developed for the use of
                    kriging in shallow-water criteria assessment. The procedure is detailed, however, and
                    requires expertise in geographic information systems, spatial statistics, and computer
                    programming. Questions remain about how best to implement kriging as an option
                    for spatial interpolation. The EPA plans to provide support through the Chesapeake
                    Bay Program Office  to ensure  that spatial interpolations based on kriging are
                    performed consistently for all shallow waters of the Bay when practical.

                    In general, most of the tidal waters of the Chesapeake Bay mainstem and major trib-
                    utaries  remain impaired. This  judgment was confirmed  by  the  assessments
                    performed during the 2006 303(d) listing cycle and by listing decisions made prior
                    to that time. The 2006 assessments indicated that many of the assessment units were
                    far out of attainment with little need  to confirm the conclusions through statistical
                    analysis. As restoration efforts proceed and  more Bay tidal waters approach attain-
                    ment of their designated uses, then statistical procedures  may become important to
                    ensure that waters are properly removed from the 303(d) list as soon as possible.
                    Given that it may require several years for the Bay to respond to management
                    actions,  there is ample time to conduct the studies necessary to develop the required
                    statistical decision-making framework based on the CFD. The EPA recommends that
                    assessment of criteria attainment continue as in 2006 when the decision rule was that
  chapter

-------
                                                                                               23
any criterion exceedance greater than that of the appropriate reference curve indi-
cates non-attainment of that criterion and, therefore, the designated use.

                         FOR                          OF
THE

As part of its conclusions, the STAC CFD Review Panel identified several critical
remaining issues requiring resolution in the near future (STAC 2006). The EPA
agrees with the recommendations for future development and advises that the Chesa-
peake  Bay  Program partners  ensure  that  the work is completed in a timely,
appropriate manner.
The following list identifies some of the critical aspects requiring further research as
recommended by STAC (2006).  See Appendix A for additional details.
  1. Effects of Sampling Density on CFD Results. The CFD is a special case of
     an unbiased estimator for a cumulative distribution function of a population.
     Like the cumulative distribution function, the CFD is a function of the mean
     and the variance of the population under assessment. The better the mean and
     variance are characterized with sample data, the more accurate the shape of the
     CFD. As the sampling density increases, the estimated CFD begins to approach
     the true CFD. If the sampling density is low, however, then sampling error
     could  become important with the potential to affect the shape of the CFD and
     ultimately the accuracy of the compliance assessment. Furthermore, the poten-
     tial for sample size to affect  the shape could create an attainment assessment
     bias if the reference curve  and assessment  curve  are based  on  different
     sampling densities. Conditional simulation methods developed by the STAC
     panel  show promise in resolving these issues and mitigating potential biases
     caused by sample size differences.
  2. Choice of Interpolation  Method, The STAC panel's research considered
     several interpolation methods and outlined the features  of each (Table  C-l in
     Appendix C). These features illustrated tradeoffs between ease of implementa-
     tion and  maximizing information garnered from the data.  Further work is
     needed to compare the features to the requirements of wide-scale implementa-
     tion of Chesapeake Bay criteria assessment procedures and to formulate a plan
     for tractable implementation that results in credible assessments. One strategy
     is to implement easily performed analysis (e.g., IDW) as a screening tool to
     identify cases for which attainment/non-attainment is clear, and then  imple-
     ment  more  labor-intensive  methods  (e.g., kriging)  for  cases in  which
     attainment is more difficult to resolve.
  3. Three-Dimensional  Interpolation. Assessments  of  the  dissolved oxygen
     criteria attainment requires three-dimensional interpolation. The field of three-
     dimensional  interpolation, however, is not  as highly  developed as that  of
     two-dimensional interpolation. Efforts are needed to evaluate research in three-
     dimensional interpolation further and to seek additional outside scientific input
        chapter ii  * Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
24
                          and review to implement the best available technology for this aspect of criteria
                          assessment.
                       4. High-Density Temporal Data. As currently formulated, criteria assessment
                          for  most of the Bay's open waters are based on "snapshots"  in time of the
                          spatial extent of criteria exceedance  estimated  through interpolation.  Data
                          collected for use in interpolation span  several days given the large area being
                          sampled. New technologies  should soon be capable of producing high-density
                          data in both space and time. Interpolation should accommodate data collected
                          densely in space. It is unclear, however, how the CFD process will accommo-
                          date data that are densely clustered in time. Further work is needed to evaluate
                          methods to fully utilize the temporally intensive data currently being collected.
                       5. Implementation and  Review. As a rule of thumb, the best test of any new
                          procedure is putting it to  work with  stakeholder involvement. Through its
                          Criteria Assessment Protocols Workgroup, the  Chesapeake Bay Program has
                          already established a forum for resolving the details of CFD implementation.
                          At appropriate intervals in this process, however, the Chesapeake Bay Program
                          should seek  independent scientific and  technical review of the implementation
                          status of the assessment methodology.
                     Helsel, D.R. and R.M. Hirsch.1992. Statistical Methods in Water Resources. Studies in Envi-
                     ronmental Science #49. Elsevier Science Publishers, Amsterdam, Netherlands. 552 pp.
                     Scientific and Technical Advisory  Committee (STAC).  2006. The  Cumulative Frequency
                     Diagram Method for Determining Water Quality Attainment: Report of the Chesapeake Bay
                     Program STAC Panel to Review of Chesapeake Bay Analytical Tools. STAC Publication 06-
                     003.  9 October 2006.  Chesapeake Bay Program  Scientific and Technical Advisory
                     Committee. Chesapeake Research Consortium,  Edgewater, MD.
                     U.S. Environmental Protection Agency (U.S. EPA). 1997. Guidelines for Preparation of the
                     Comprehensive State  Water Quality Assessments  (305(b) reports) and Electronic  Updates.
                     Assessment and Watershed Protection Division, Office of Wetlands, Oceans and Watersheds,
                     Office of Water, U.S. Environmental Protection Agency, Washington, D.C.
                     U.S. Environmental Protection Agency. 2003a. Ambient Water Quality Criteria for Dissolved
                     Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal Tributaries.
                     EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.
                     U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
                     Bay Designated Uses and Attainability. EPA  903-R-03-004. Region III  Chesapeake  Bay
                     Program Office Annapolis, MD.
                     Weisberg, S.B., J.A. Ranasinghe, D.M. Dauer, L.C. Schaffner, RJ. Diaz, and J.B.  Frithsen.
                     1997. An estuarine benthic index of biotic integrity (B-IBI) for Chesapeake Bay. Estuaries
                     20:149-158.
  chapter ii  »  Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                    25
                        chapter|||
  Application  of Chesapeake Bay
           Water  Quality  Criteria
         Assessment  Procedures
                         BACKGROUND

Beginning in the late 1990s and continuing through 2003, the Chesapeake Bay
Program partners developed new Chesapeake Bay water quality criteria designed
specifically to protect the ecological health of the Bay (U.S. EPA 2003a). Delaware,
Maryland, and Virginia, along with the District of Columbia, then adopted these
criteria and new tidal water designated uses into their water quality standards regu-
lations. The states' new Chesapeake Bay water quality standards were applied for the
first time in each state's 2006 Clean Water Act 303(d) listing cycle.
The four jurisdictions also adopted criteria assessment methods — published by EPA
in 2003 and in a 2004 addendum — into state water quality standards regulations (U.S.
EPA 2003 a, 2004a). The methods characterize the spatial and temporal variability of
the appropriate water quality parameters, while providing a clear basis for determining
whether a portion of the Bay's tidal waters reached attainment of the applicable desig-
nated  use.  Despite  the methods' detail, technical limitations remained for  their
complete application. This chapter and those that follow address many of the  prior
technical limitations. Continued efforts to develop further refinements to the criteria
assessment methodology in specific areas, however, will likely remain.

In addition to the technical limitations, obstacles related to the states' transition from
an old set of water quality standards to the newer, more detailed Chesapeake Bay
water quality standards also existed. Differences occurred in the spatial extent of past
listing/delisting decisions. New water quality criteria components  also exist that
have never been previously assessed. Furthermore, the mechanisms and processes
used to report listings in the past required updating to allow reporting based on the
states' new Chesapeake Bay water quality standards regulations. As with the tech-
nical limitations referenced above, an ongoing effort to refine and update the
methodology for making future listing decisions based on the new Chesapeake Bay
water quality standards will also be required (see Chapter 8 for further details).
            chapter iii  •  Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
26
                    To assess attainment of the Chesapeake Bay water quality criteria, the spatial and
                    temporal extent over which they apply must be  defined. The temporal extent is
                    defined implicitly for each component of the states' Chesapeake Bay water quality
                    standards. Described on page 150 in the 2003 EPA Chesapeake Bay water quality
                    criteria document (U.S. EPA 2003a) and adopted into the jurisdictions' water quality
                    standards regulations, the spatial extent is defined by the intersection  of a Chesa-
                    peake  Bay  Program  segment  (U.S. EPA 2004b,  2005a) and  each  tidal water
                    designated use (U.S. EPA 2003b, 2004c). The spatial units defined by this intersec-
                    tion are referred to as "spatial assessment units."  The intent is for each unit to be
                    assessed and listed independently on each jurisdiction's 303(d) list (part 1 through
                    part 5) (see Chapter 8 for further details).

                    The scale of the Chesapeake Bay spatial assessment units is large, with selection
                    based specifically on conditions in the Bay and on the factors affecting  these condi-
                    tions. The Chesapeake Bay Program segments themselves were  based on salinity
                    regimes, circulation patterns,  and other natural physical features, but are generally
                    reflective of variations in water quality  conditions and living resource communities
                    (U.S. EPA 2004b, 2005a). Thus, these segments serve as appropriate spatial units for
                    measuring the scope of water quality impairments in the Chesapeake Bay, its tidal
                    tributaries, and its embayments. They also work  at a logical scale for developing
                    necessary management  plans (TMDLs). Many of the water quality impairments
                    currently extend over large areas of the Bay and its tidal tributaries, so performing
                    assessments and reporting on these impairments at the segment scale are both appro-
                    priate. Developing management plans at this scale  is also appropriate since multiple
                    jurisdictions often contribute to impairments.
                    Even though the scale of the spatial assessment units is suitable, in many cases it
                    varied from the scale of past tidal water quality criteria attainment assessments. The
                    change in scale introduced several challenges to the states as they implemented the
                    new Chesapeake Bay water quality criteria and tidal water designated uses. Bound-
                    aries of some previously established state assessment units were moved  or shrunk to
                    address the spatial variability in some state water quality standards assessment meas-
                    ures. Furthermore, management decisions (e.g., listing certain waters as impaired,
                    developing TMDLs) had  already been made based on the previously established
                    assessment units and were being implemented at the time the new Chesapeake Bay
                    water quality standards were adopted into state regulation. Thus, it was  necessary to
                    establish procedures for transitioning to new spatial assessment units  and relating
                    prior management decisions to new assessments that were sometimes  defined at a
                    different spatial scale.

                    In general, the states could address the differences in boundary locations by making
                    small adjustments to state-defined spatial units. Primarily, adjustments  consisted of
                    small changes in the boundaries of the previously state-defined assessment units to

  chapter iii  *  Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
                                                                                                  27
make them coincident  with the boundaries of the
larger  Chesapeake  Bay Program segments.  This
way, the smaller assessment units nest within the
larger ones and the larger-scale assessment results
can be attributed to each of the smaller units within.
The approach  allows  states to remain  consistent
with previous listing decisions while accounting for
the  broader  designated-use-segment-assessment
results on their 303(d) lists.

In some cases, adoption of the new Chesapeake Bay
spatial assessment units represented a less detailed
and possibly less precise assessment of water-quality
criteria attainment. For  example, Figure III-l  illus-
trates Chesapeake Bay  Program  segment CB7PH,
which covers the southeastern portion of the main-
stem Chesapeake within Virginia.  As is typical in
most of the Bay, the shoreline is extremely complex
with many small tidal  rivers, creeks, and embay-
ments. These smaller tidal habitats may have different
water quality than the mainstem Bay  section of the
segment due  to different circulation patterns or land
uses or pollution sources that dominate local water
quality conditions. These smaller tidal habitats may
even have monitoring information that demonstrates
the differences in water quality conditions. In such a
case, it may make sense to separate the smaller tidal
river, creek, or  embayment from the main assess-
ment unit by subdividing it to create a new smaller
spatial unit for separate assessment. Thus, the states have the option to "sub-segment"
larger units to characterize conditions in specific parts of the Bay, its tidal tributaries,
and embayments more precisely.
Allowing jurisdictions to subdivide the larger segments is consistent with national
EPA guidance and with EPA-published Chesapeake Bay water quality criteria assess-
ment guidance, which both provide specific considerations for sub-segmenting water
bodies for criteria assessment and listing decisions  (U.S.  EPA  2003a, 2005b).
Published EPA guidance states that waters can be partitioned "to represent homo-
geneity in physical, biological or chemical conditions." The EPA recommends that
jurisdictions use similar principles in deciding to subdivide the larger Chesapeake
Bay assessment units. A state's decision to sub-segment an existing segment should
be based on: 1) clear physical, biological, or chemical differences that can be docu-
mented; 2) homogeneity of water quality in the water body under consideration; and
3) confirmed future availability of monitoring data in the new sub-segment to provide
the capability to assess conditions and allow a determination regarding its 303(d) list
status. In all cases, there should be a priori knowledge of the conditions that support
a decision to subdivide,  and preferably specific data that demonstrate how conditions
Figure 111-1. Segment CB7PH covering the southeastern
portion of the Chesapeake Bay in Virginia.
Source: U.S. EPA 2004b.
              chapter iii  •  Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
28
                    differ in the area under scrutiny. Documentation of this information should be made
                    available for review as part of the 303 (d) listing cycle for which a new subdivided
                    segment is initially assessed. Jurisdictions need to  ensure that any sub-segmentation
                    is fully consistent with their state's water quality standard regulations.
                    The EPA  discourages states from subdividing segments simply to remove smaller
                    areas from an impaired waters list. Given the tidal exchange that occurs among all
                    segments, conditions in one segment can potentially affect adjacent segments. A sub-
                    segment that is prematurely removed from the impaired waters list might require
                    placement back on the impaired waters list in the next listing cycle due to adverse
                    conditions in the original segment.

                    Maryland and Virginia have already adopted specific sub-segments into their state's
                    water quality standards regulations in several tidal tributaries and embayments. The
                    2004 addendum to the 2003 Chesapeake Bay use attainability and designated-use
                    document contains detailed documentation supporting these state-defined, adopted
                    sub-segmentations (U.S. EPA 2004c).


                                                USE  IN                      BAY


                    To assess Chesapeake Bay water quality criteria attainment, the data used must prove
                    adequate.  Consistent with the 2003 EPA Chesapeake Bay criteria assessment guid-
                    ance, the  data should be of known quality and adequate  quantity,  as well as
                    representative of the tidal water designated use habitat under assessment (U.S. EPA
                    2003a). Documented QA/QC programs should ensure data quality; such documen-
                    tation should be publicly available for evaluation. A sufficient amount of data should
                    exist to provide  a defensible  degree of accuracy and precision given the expected
                    level of variability in the assessed tidal water body. The data should also be repre-
                    sentative of the spatial assessment unit as a whole  so the resulting assessment is not
                    biased toward any one portion. While the EPA provides no minimum requirements
                    for each of these data characteristics, they  should be maximized to the extent
                    possible to ensure that criteria assessments are scientifically defensible.

                    Opinions range broadly on the quantity of data required for criteria assessment. On
                    one extreme, some believe that sufficient data should be collected to capture all the
                    temporal and  spatial variability to  ensure that the criteria and designated uses are
                    attained in space and time. On the other extreme, some suggest that the state agency
                    manager should determine if a designated use is being attained based on available
                    information—even if it is anecdotal.

                    For the Chesapeake Bay and its tidal tributaries,  the EPA recommends basing all
                    water quality criteria assessments on monitoring data. These data should be collected
                    over a three-year period immediately prior to the year of the listing cycle, unless
                    non-attainment is definitively established in less time (as described in Chapter 7).
                    Furthermore, the monitoring  program for  data collection should optimize quality,
                    quantity, and representativeness as described above.

  chapter iii  «  Application  of  Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
                                                                                                29
The Chesapeake Bay Program partners continue to fund and conduct an extensive
baywide, coordinated water  quality monitoring  program,  much of which supports
water quality criteria assessment. Water quality monitoring takes place at more than
150 sites throughout the mainstem Chesapeake Bay and its tidal tributary waters
(Figure III-2). Samples are collected at each of the fixed stations on a monthly or semi-
monthly basis with data gathered since the mid-1980s (Chesapeake Bay Program
1989). The fixed-station network provides consistent data over  the entire mainstem
Bay, major tidal tributaries, and larger embayments. The data are useful in assessing
the published Bay water quality criteria in the open-water, deep-water, deep-channel,
and migratory and spawning designated uses.

Use of the fixed-station network is limited for criteria assessments in the shallow-
water designated use habitats because the data scale is not appropriate. This network
also  proves limited in many smaller tidal tribu-
taries and embayments, which have no or very few
stations. To address these limitations, the Chesa-
peake Bay  Program   partners  developed  a
Shallow-water Monitoring   Program  to provide
data collected intensively in  space and time in the
Bay's shallow-water habitats. Chapter 7 describes
this program and the details of data application for
criteria assessment.

The  2003  EPA  Chesapeake Bay water quality
criteria  document describes the extent of  data
collection needed to assess the state's Chesapeake
Bay  water quality standards (U.S. EPA 2003a).
Three levels of effort are described for each crite-
rion: marginal, adequate, and recommended (see
pages 178-196 in U.S. EPA 2003a). The "mar-
ginal" level of monitoring is the minimum data
collection needed to support criteria assessment.
At this level, data may not be of the right type or
in sufficient quantity to assess all of the applicable
criteria  components.  In general, this level of
monitoring  assumes that only the fixed-station
data are available  for  criteria assessment.  The
"adequate" level of monitoring assumes that the
fixed-station  monitoring   program   will  be
combined with limited intensive data collection
(e.g., temporally  continuous  monitoring  for
dissolved oxygen) to ensure that data are collected
to support the assessment  of all  the  applicable
criteria components (e.g., 30-day, 7-day, and 1-
day  means,  instantaneous   minimum)  in  some
spatial assessment units. The  "recommended"
level of monitoring assumes  that the fixed-station
                     Figure 111-2. Locations of the sites that make up the
                     fixed station network of the Chesapeake Bay Water
                     Quality Monitoring Program.
                     Source: Chesapeake Bay Program 1989.
              chapter
Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
30
                    monitoring program will be combined with intensive data collection in all spatial
                    assessment units. Funding is not currently available to support monitoring at the
                    "recommended" level. The fixed-station monitoring is expected to continue into the
                    future, so data should be available at the "marginal" level for all spatial assessment
                    units. With the implementation of the Chesapeake Bay shallow-water monitoring
                    program in 2001, combined with a growing network of high-frequency observing
                    system deployed in the Bay tidal waters, monitoring will reach the "adequate" level
                    across all spatial assessment units with time.
                    To enhance the monitoring information from the coordinated Chesapeake Bay water
                    quality and shallow-water monitoring  programs,  jurisdictions are encouraged to
                    include data from  other sources as appropriate. Consistent with the  2003 EPA-
                    published Chesapeake Bay water quality criteria assessment guidance, the states and
                    the District are encouraged to compile data from sources such as state and federal
                    monitoring agencies, local governments, universities, environmental organizations,
                    and citizen monitoring groups (U.S. EPA 2003a). Such data could prove significant
                    in enhancing  the spatial coverage of the existing Chesapeake Bay water quality
                    monitoring program. The jurisdictions must ensure, however, that the data are appro-
                    priate for use in the Chesapeake Bay criteria attainment assessment methodology.
                    Data need to be of documented quality and adequate quantity as indicated above.
                    The jurisdictions also must ensure that the data are collected at an appropriate scale
                    and are representative of a given area or volume of a specific spatial assessment unit.
                    The Chesapeake Bay Program spatial interpolator uses data collected at all locations
                    and defines how much of that area or volume can be characterized by data from a
                    particular location (see Chapter 2 and Appendix D for details). Thus, a small tidal
                    embayment may be characterized by data from a single site.  If that site is not located
                    properly (e.g., in a small creek, off a pier in shallow water,  off a beach), the assess-
                    ment of the entire embayment may rest on potentially biased information. Similarly,
                    if data are collected intermittently at some sites, the spatial  assessment unit may be
                    characterized inconsistently at times.
                    To use data collected through non-Chesapeake Bay  Program monitoring programs in
                    Chesapeake Bay water quality criteria assessments, they must be merged with the Chesa-
                    peake Bay Program monitoring program data appropriately. The assumption is that these
                    water quality data were collected on different time (more infrequent) and space (well
                    away from the mid-channel river,  mainstem) scales than the Chesapeake  Bay Water
                    Quality Monitoring Program data. Therefore, these other data will be assigned a cruise
                    designation based on the monthly collection time so that they can be interpolated along
                    with the Chesapeake Bay Water Quality Monitoring Program data to generate the cumu-
                    lative frequency distribution (see Chapter 2 and Appendix B for details).
                    The states are encouraged to seek data from sources beyond the Chesapeake Bay
                    Water  Quality Monitoring Program, but should use such data with care to avoid
                    biasing the assessment results for any particular portion of the tidal waters. Ideally,
                    the states would work with the collecting agencies and institutions in advance to

  chapter iii  »  Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
                                                                                                31
ensure that the data are collected appropriately for use in interpolation and the
overall CFD-based criteria assessment methodology.

In addition to data collected by government and non-profit agencies, the states are
also encouraged to work with agencies, organizations, or other entities subject to
regulation, but with an interest in contributing data for use in the criteria attainment
assessment process. Such agencies may be able to provide  additional monitoring
resources and significant amounts of supplementary data. Provided that an adequate
QA/QC program is in place to ensure that the data are accurate, representative, and
of known quality, these  regulated agencies or entities may significantly benefit the
criteria assessment process.

The Hampton Roads Sanitation District in Virginia is one such example. The District
has worked with the Virginia Department of Environmental Quality and the Virginia
Institute of Marine Science to establish its own shallow-water monitoring program.
The Virginia Department of Environmental Quality can use the data generated by the
program to assess the state's  dissolved oxygen, water clarity,  and chlorophyll a
criteria in the lower tidal James River. Other similar organizations of regulated stake-
holders may also wish to provide similar data.
                 mr-
The  criteria assessment methodology  developed for the  Chesapeake  Bay water
quality criteria standards will require continued refinement into the future. The tech-
nical  details  of the methodology continue to be refined through research and
experience with application. This document describes many new refinements that
will assist the jurisdictions with their criteria assessment process and listing deci-
sions. More refinements are expected over the coming years. Furthermore, better
understanding is developing with time as more data are collected. New  monitoring
programs (e.g., shallow-water  monitoring)  are offering  new  insight  into the
processes that affect water quality conditions in the Chesapeake Bay. This enhanced
understanding will help fine-tune the requirements necessary for protection of the
Bay ecosystem. Given that continued refinements of the criteria assessment method-
ology are expected, it is recommended that the jurisdictions plan continued updates
to their Chesapeake Bay water quality  standards regulations through their existing
triennial review process. The EPA commits to providing the information needed for
updating the  states' water quality standards through  publication of recommended
refinements to the criteria assessment procedures (such as in this addendum). The
publication of any future addendums to  the 2003 Chesapeake Bay criteria document
will come in advance  of the jurisdictions' triennial  reviews for use in justifying
needed changes to the state's water quality standards regulations.

One example of the expected refinements to the criteria assessment  methodology is
the development of a statistical basis for decision-making using the CFD  (see
Chapter 2 and Appendices A and C for further details). Since the Chesapeake Bay
criteria assessment methodology was first published in 2003, interest has grown in

             chapter iii  «  Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
32
                     developing an accounting  of error in the assessment process. Research has been
                     underway over the past years to develop such a methodology. The technical details
                     are challenging, however, and research has not yet led to a solution.  Progress has
                     occurred over the last year; a statistical framework could possibly be developed for
                     adoption into the state's water quality standards in upcoming 303(d) listing cycles.
                     Other refinements may be developed for monitoring programs and the interpolation
                     procedures. The EPA encourages states to prepare for adopting such refinements to
                     their criteria assessment procedures into future regulations.

                     Reference curves provide a second example of expected refinements. As more data
                     are collected, the capability  for better defining the amount and pattern of criteria
                     exceedance that the system can withstand will continually improve.  While major
                     changes to the reference curves are not expected, updating the reference curves with
                     additional data will improve the states' ability to assess Chesapeake tidal waters
                     accurately. With the prior  agreement of the watershed jurisdictions, the EPA will
                     update the reference curves with new  data and publish the revised curves in future
                     criteria document addenda. The jurisdictions will then need to adopt the new refer-
                     ence curves into their  water quality  standards regulations through  their regular
                     triennial review processes.
                     Chesapeake Bay Program. 1989. Chesapeake Bay Monitoring Program Atlas-Volume 1:
                     Water Quality and Other Physio chemical Monitoring Programs. CBP/TRS 34/89. U.S. Envi-
                     ronmental Protection Agency Chesapeake Bay Program Office, Annapolis, MD.
                     U.S. Environmental Protection Agency (U.S. EPA). 2003a. Ambient Water Quality Criteria for
                     Dissolved Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal Trib-
                     utaries. EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.
                     U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
                     Bay Designated Uses and Attainability. EPA 903-R-03-004. Region III Chesapeake Bay
                     Program Office, Annapolis, MD.
                     U.S. Environmental Protection Agency. 2004a. Ambient Water Quality Criteria for Dissolved
                     Oxygen, Water  Clarity and Chlorophyll a Chesapeake Bay  and its Tidal Tributaries-2004
                     Addendum. EPA 903-R-04-005. Region III Chesapeake Bay Program Office, Annapolis, MD.
                     U.S.  Environmental Protection  Agency.  2004b.  Chesapeake Bay  Program  Analytical
                     Segmentation Scheme: Revisions, Decisions and Rationales 1983-2003. EPA 903-R-04-008.
                     CBP/TRS 268/04. Region III Chesapeake Bay Program Office, Annapolis, MD.
                     U.S. Environmental Protection Agency. 2004c. Technical Support Document for Chesapeake
                     Bay Designated Uses and Attainability'-2004 Addendum. EPA 903-R-04-006.  Region III
                     Chesapeake Bay Program Office Annapolis, Maryland.
                     U.S.  Environmental Protection  Agency.  2005a.  Chesapeake Bay  Program  Analytical
                     Segmentation Scheme: Revisions, Decisions and Rationales 1983-2003. 2005 Addendum.
                     EPA  903-R-05-004. CBP/TRS  278-06.  Region III Chesapeake Bay Program  Office,
                     Annapolis, MD.
                     U.S. Environmental Protection Agency. 2005b. Guidance for 2006 Assessment, Listing and
                     Reporting Requirements Pursuant to  Sections  303(d) and  314 of the Clean  Water Act.
                     Watershed  Branch,  Assessment  and Watershed Protection  Division. Office of Wetlands,
                     Oceans and Watersheds, Office of Water, U.S. EPA, Washington, D.C.


  chapter iii  * Application of Chesapeake Bay Water Quality Criteria Assessment Procedures

-------
                                                                                 33
                       chapter
 Refinements  to  the Chesapeake
   Bay Dissolved Oxygen  Criteria
         Assessment  Procedures
                        BACKGROUND

In 2003, the EPA published detailed  criteria for dissolved  oxygen tailored to
different habitats within the Chesapeake Bay and its tidal tributaries (U.S. EPA
2003a) (Table IV-1). Oxygen is critical to most forms of life in the Bay; it must be
available in adequate concentrations to support overall ecosystem health. Minimum
concentrations of oxygen must be  present to support the wide range of species
requiring protection as well as their various life stages.
Dissolved oxygen criteria were established for Chesapeake Bay that varied in space
and time to provide levels of protection for different key species and communities.
The criteria were also designed around several lengths of time to reflect the varying
oxygen tolerances for different life  stages (e.g., larval, juvenile, adult) and effects
(e.g., mortality, growth, behavior). Thus, the dissolved oxygen  criteria include
multiple components. Each component includes  a target of dissolved oxygen
concentration, the duration of time  over which the concentration is averaged, the
space (designated-use area) where the criterion applies, and a time (season, month)
when the criterion applies.

The dissolved oxygen criteria include 30-day, 7-day, and 1-day means along with an
instantaneous minimum. Each of these criteria components applies to a specific
season, such as the migratory spawning nursery period or the summer months (June
through September) or all-year round. Each also relates to one of four tidal-water
designated uses, according to the species and biological communities to be protected
(U.S. EPA 2003a, 2003c). The EPA published, and the states adopted into their water
quality standards regulations, dissolved oxygen criteria protective of migratory and
spawning, open-water, deep-water, and deep-channel designated-use habitats (U.S.
EPA 2003a) (Table IV-1).
     chapter iv • Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
34
        fa:'; e IV
Chesapeake Bay dissolved oxygen criteria.
Designated
Use
Migratory
fish
spawning
and
nursery use
Shallow-
water Bay
grass use
Open-water
fish and
shellfish use
Deep-water
seasonal fish
and shellfish
use
Deep-
channel
seasonal
refuge use
Criteria
Concentration/Duration
7 -day mean > 6 mg- liter"1
(tidal habitats with 0-0.5 ppt
salinity)
Instantaneous minimum > 5
mg-liter"1
Protection Provided
Survival/growth of larval/juvenile tidal-
fresh resident fish; protective of
threatened/endangered species
Survival and growth of larval/juvenile
migratory fish; protective of
threatened/endangered species
Open-water fish and shellfish designated-use criteria apply
Open-water fish and shellfish designated-use criteria apply
30-day mean > 5.5 mg-liter"
(tidal habitats with 0-0.5 ppt
salinity)
30 -day mean > 5 mg-liter"
(tidal habitats with > 0.5 ppt
salinity)
7 -day mean > 4 mg-liter"
Instantaneous minimum > 3.2
mg-liter"
30 -day mean > 3 mg-liter"
1-day mean > 2.3 mg-liter"1
Instantaneous minimum > 1 .7
mg-liter"1
Growth of tidal-fresh juvenile and adult
fish; protective of threatened/
endangered species
Growth of larval, juvenile and adult fish
and shellfish; protective of threatened/
endangered species
Survival of open-water fish larvae
Survival of threatened/endangered
sturgeon species1
Survival and recruitment of bay
anchovy eggs and larvae
Survival of open-water juvenile and
adult fish
Survival of bay anchovy eggs and larvae
Open-water fish and shellfish designated-use criteria apply
Instantaneous minimum > 1
mg-liter"1
Survival of bottom-dwelling worms and
clams
Open-water fish and shellfish designated-use criteria apply
Temporal
Application
February 1 -
May 31
June 1 -
January 3 1
Year-round
Year-round
June 1 -
September 30
October 1 -
May 31
June 1 -
September 30
October 1 -
May 31
       1 At temperatures considered stressful to shortnose sturgeon (> 29EC), dissolved oxygen concentrations above an
        instantaneous minimum of 4.3 mg-liter1 will protect survival of this listed sturgeon species.
  chapter iv

-------
                                                                                                 35
Assessing  dissolved  oxygen  criteria  attainment  is challenging because of the
complexity of both the criteria and the Bay itself. To fully assess all the criteria
components, data need to be collected at a spatial intensity that adequately represents
the four designated-use habitats of Chesapeake Bay tidal waters at different times of
the year (U.S. EPA 2003c, 2004b). Similarly, data must be collected during all the
applicable seasons and at frequencies sufficient to address the various criteria dura-
tion  components.  The  different  dissolved oxygen criteria apply to different
designated-use areas and multiple criteria apply to the same designated-use area. The
dissolved  oxygen  criteria components also apply over different time  periods to
protect species during critical life stages or during particularly stressful times of the
year. To fully assess each dissolved oxygen component in each  designated-use
habitat over  the appropriate time periods  will require an extensive  monitoring
program and a detailed assessment methodology. The  Chesapeake Bay Program
currently conducts extensive water quality monitoring throughout the Bay  tidal
waters and the EPA published a detailed  dissolved oxygen criteria  assessment
methodology with the new water quality criteria (Chesapeake Bay  Program 1989;
U.S. EPA 2003a, 2004a). The existing Bay water quality monitoring was not suffi-
cient  to  cover all the  criteria  components,  however, and  some  details in the
assessment methodology remain unresolved.
For the 2006 303(d) listing cycle, the states'
listing decisions  were  based primarily  on
previous listings. Tidal waters that had been
listed as impaired in 2004 were not removed
from part 5 of their listing unless all the appli-
cable  criteria  components  were shown in
attainment (see Chapter 8 for further details).
The Chesapeake Bay Program partners had the
capacity (data,  assessment methodology) to
assess only the  30-day  mean dissolved oxygen
criteria and, in some cases, the instantaneous
minimum  dissolved  oxygen criteria. The
remaining dissolved oxygen criteria  were not
assessed because  the  existing water  quality
monitoring programs and the published assess-
ment  methodologies were inadequate for full
assessment.  In  most spatial assessment units,
the 30-day mean criterion was not attained and
those assessment units  would have been listed
whether or not the other applicable dissolved
oxygen criteria were also assessed (Figure IV-
1). In many smaller tidal tributaries,  however,
the 30-day  mean  criterion was attained and
those spatial assessment units were listed either
as "impaired" (part 5) due to previous listing or
   DO Impairment
   30 Day Mean
   All Designated Uses
   • Afiairwwitof Criteria
   -  msufftteni Data !o Assess

   ~ N/A
Figure IV-1. Listing status of the Chesapeake Bay open-water
designated use based on dissolved oxygen standards.
      chapter iv  •  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
36
                    as having "insufficient data to assess" (part 3). As nutrient loads are reduced and Bay
                    water quality improves, assessing the complete array of applicable dissolved oxygen
                    criteria to remove spatial assessment units from the "impaired" list will become
                    more critical.

                    Since Chesapeake Bay dissolved oxygen criteria were published in 2003, the capa-
                    bility of fully assessing all the dissolved oxygen criteria for all four designated uses
                    over all applicable time periods has progressed, but some limitations remain. The
                    refined and expanded dissolved  oxygen criteria  assessment  methodologies  docu-
                    mented in this chapter replace the methodologies previously published by U.S. EPA
                    (2003a,  2004a). Work should continue in  refining these methodologies to reduce
                    uncertainty further and to increase confidence in the resulting assessments. Devel-
                    oping, validating, and publishing EPA-recommended methodologies for assessing
                    the full array of Chesapeake Bay dissolved oxygen criteria duration components will
                    also  prove critical.
                                                            FOR
                                   OF
                    To assess dissolved oxygen criteria attainment, the time span over which the criteria
                    apply must be clearly defined. In some cases, the temporal period is defined implic-
                    itly as part of the criteria.  For example, the dissolved oxygen criteria protective of
                    the migratory fish spawning and nursery habitat designated use apply only to that
                    time of year when spawning fish (and the resultant eggs and early juveniles) require
                    higher dissolved oxygen levels  compared to the rest of the year. In this example,
                    dissolved oxygen criteria  attainment should be assessed over the entire spawning
                    season (February  1 through May 31) (U.S. EPA 2003a). Similarly, dissolved oxygen
                    criteria in the deep-water  and deep-channel designated uses apply only during the
                    summer months — June 1 through September 30 — when the Bay stratifies and
                    naturally reinforces the potential for lower dissolved oxygen  concentrations  in
                    deeper waters. Therefore, assessment of dissolved oxygen criteria attainment in the
                    deep-water and deep-channel designated uses should also be performed over the
                    entire 4-month summer season (U.S. EPA 2003a). In  all these cases, data are
                    collected over the entire criteria season in each of the three years of the assessment
                    period and these data are used to develop the cumulative frequency diagram (CFD)
                    for assessing dissolved oxygen criteria attainment (see Chapter 2 and Appendix B for
                    additional details).

                    Open water is the only tidal water designated use in which the dissolved oxygen
                    criteria apply year-round (U.S. EPA 2003a). In general, the Bay is most vulnerable to
                    low dissolved oxygen during the summer when temperatures are high, oxygen solu-
                    bility is low, and biological consumption of oxygen rises to its greatest level. Periods
                    of low dissolved oxygen can also occur during the rest of the year, sometimes caused
                    by high loading with subsequent slow consumption of organic material. The open-
  chapter iv  *  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                                37
water dissolved oxygen criteria are designed to provide protection of open-water
habitat fish and shellfish communities at all times of the year. In spite of the year-
round application of these criteria, natural processes complicate the use of a single,
year-round assessment. Cooler temperatures affect the solubility of oxygen and allow
higher concentrations compared to  similar organic loading conditions in  warmer
months. Consequently, dissolved oxygen concentrations have a large natural vari-
ability range. Detecting human effects in the presence of that greater variability often
proves difficult. For this reason, as part of the dissolved oxygen criteria development
process, the EPA originally intended that the year-round open-water dissolved oxygen
criteria (see Table 111-10, page 66 in U.S. EPA 2003a) be assessed in each season (see
pages 150-151 in U.S. EPA 2003a).  During the 2006 303(d) listing cycle, confusion
arose as to the appropriate time period for open-water dissolved oxygen assessment.
The criteria were clearly defined over the full annual cycle,  but the stated intent was
to assess them on a seasonal basis. Furthermore, the  2003 EPA Chesapeake Bay
criteria document itself did not provide consistent guidance; it referred to assessment
on both an annual basis and a seasonal one (U.S. EPA 2003a).

Based on a re-evaluation of the underlying scientific basis for Chesapeake Bay
dissolved oxygen criteria, the EPA recommends that jurisdictions assess attainment
of the open-water  dissolved oxygen criteria  separately over two time periods:
summer (June 1 through September 30) and non-summer (January 1 through May 31
and October 1 through  December 31). The open-water dissolved oxygen criteria
were largely derived to protect open-water species during the summer when elevated
temperatures, higher  salinities, and naturally low dissolved oxygen levels  occur
(U.S. EPA 2003a). Given that summer is a critical period for many species, it should
be assessed separately. The potential for dissolved oxygen impairments are lower in
the non-summer period due to greater natural dissolved oxygen solubility and lower
biological oxygen consumption—both due to lower water column temperatures.
Nevertheless, low dissolved oxygen levels sometimes occur during other times of the
year making a separate dissolved oxygen criteria assessment necessary for the non-
summer period. The  separate criteria assessments for summer  and non-summer
seasons will support year-round protective dissolved oxygen concentrations in the
open-water designated-use habitats.
The open-water designated-use boundary is explicitly defined as including "tidally
influenced waters extending horizontally from the shoreline to the adjacent shore-
line" (see page 71 in U.S. EPA 2003c). Further, on page 68, the U.S. EPA (2003c)
states that:

    The shallow-water bay grass designated use is intended specifically to delin-
    eate the habitats where  the water clarity criteria would apply.   The
      chapter iv  *  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
38
                        open-water fish and shellfish designated use and the accompanying dis-
                        solved oxygen criteria will fully  protect  the  biological  communities
                        inhabiting shallow-water habitats. The open-water designated use extends
                        into the intertidal zone and protects shallow-water organisms beyond under-
                        water bay grasses.

                    Unless a state has specifically delineated a sub-segment within a segment, attainment
                    of the open-water designated use will be based on dissolved oxygen criteria attain-
                    ment for the entire volume of the open-water designated use within the segment.
                    Neither the need nor the requirement exists for a separate assessment of dissolved
                    oxygen criteria attainment strictly within shallow waters (0-2 meters in depth). The
                    importance of acquiring better temporal and spatial coverage of dissolved oxygen
                    conditions in these shallow-depth habitats is not diminished however, since condi-
                    tions in these areas vary greatly from the open water of the mid channels where the
                    fixed stations are located. Shallow-water monitoring will provide the data needed to
                    characterize dissolved oxygen conditions in shallow-water habitats more  fully (see
                    Chapter 7 for further details).
                                                     OF


                    Historically, the Chesapeake Bay Water Quality Monitoring  Program  consisted
                    primarily  of fixed-station monitoring conducted on  a monthly or twice-monthly
                    basis (Chesapeake Bay Program 1989). This sampling design was primarily intended
                    to assess  long-term trends in  water  quality and the status of  living resources,
                    capturing  variability over decadal, annual, and seasonal time scales. The fixed-
                    station monitoring was adapted to assess the 30-day mean dissolved oxygen criteria
                    to measure dissolved oxygen throughout the Bay and its tidal tributaries and embay-
                    ments. This system ensures at least one set of measurements for each month.
                    The individual monthly estimates are considered accurate, although imprecise, since
                    the sample sizes are small (n = 1 or 2). This imprecision is likely to be mitigated by
                    the many estimates of monthly means (e.g., multiple months over the 3-year assess-
                    ment period), which are combined into each single assessment of criteria attainment
                    (see Chapter 2 and Appendix B  for additional details). The monthly and  twice-
                    monthly fixed-station data are not adequate to assess attainment  of the 7-day and
                    1-day mean dissolved oxygen criteria directly because the sampling frequency rests
                    outside the defined time intervals and is unable to capture the short-term variability.

                    For the 2006 303(d) listing cycle, only three of the dissolved oxygen criteria compo-
                    nents were assessed. The 30-day mean open-water criterion was determined in all of
                    the assessment units of Chesapeake Bay using the fixed-station data and the CFD
                    assessment methodology. In spatial assessment units where deep-water and/or deep-
                    channel designated uses  exist, the  30-day mean deep-water criterion  and the
  chapter iv  «  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                             39
instantaneous minimum deep-channel  criterion were  also determined using the
fixed-station data.

The rationale behind the assessment of the instantaneous minimum deep-channel
criterion was based on the long-term fixed station data record in the deep-channel
locations which shows that dissolved oxygen does not vary strongly through time in
the deep channel during the summer months because of the physical isolation from
the atmosphere  and the photic zone. Dissolved oxygen concentrations remain rela-
tively constant;  therefore, a 30-day mean should  be similar  to any instantaneous
measure (see section below).

No assessments were made of the 7-day and 1-day mean dissolved oxygen criteria
because the  data were considered inadequate  (as described above). In most cases,
this situation did not affect listing decisions because  many spatial assessment units
did not attain the 30-day mean criterion (see Figure IV-1) and all criteria components
need to be attained to justify removal from the impaired list (part 5). The 30-day
mean criterion was attained in some cases. These spatial assessment units, if not
previously listed on part 5, were placed in part 3 of the states' lists for waters with
insufficient data (see Chapter 8 for further details). As water quality conditions
improve in Chesapeake Bay, a method to assess higher frequency dissolved oxygen
criteria will be needed so that spatial assessment units in attainment with all appli-
cable dissolved oxygen criteria components  can be removed from the  state's
impaired waters list (see Appendix E).

Until the EPA publishes methodologies for assessing the 7-day mean, 1-day mean
and instantaneous minimum open-water and deep-water  dissolved oxygen criteria
components, the agency recommends that the states rely strictly on the assessment
of the 30-day mean open-water and deep-water dissolved oxygen criteria for listing
decisions. For those open-water and deep-water designated-use  segments in which
the 30-day mean criteria are not in attainment, the jurisdictions should list the desig-
nated-use-segment on part 5 as impaired in the absence of data and/or methodologies
for assessing the remaining criteria components. For those designated-use segments
in which the 30-day mean criteria are in attainment, the jurisdictions should generate
additional data and apply the  criteria assessment procedures to assess attainment of
the 7-day mean, 1-day mean,  and instantaneous minimum criteria components.
         DISSOLVED OXYGEN  REFERENCE  CURVES

SUMMER OPEN-WATER AND DEEP-WATER DISSOLVED OXYGEN
CRITERIA REFERENCE CURVES
Reference curves for both the 30-day mean open-water (June 1-September 30 only)
and 30-day mean deep-water dissolved oxygen criteria were based on criteria levels
that would not impair biological communities (U.S. EPA 2003a). Reference areas for
      chapter iv •  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
40
                    derivation of the 2003 published deep-water reference curves were identified using
                    a measure of benthic community health—the Chesapeake  Bay benthic index of
                    biological integrity or benthic-IBI (Weisberg et al. 1997). Sessile benthic communi-
                    ties are good indicators  of the water quality of the overlying waters. Although
                    relatively tolerant of lower oxygen concentrations, a dissolved oxygen concentration
                    of 2 mg-liter1 is the  threshold below which benthic infaunal communities become
                    severely stressed (numerous references cited in Chapter 3 of U.S. EPA 2003a). A
                    healthy benthic community, therefore, could indicate  allowable time  and  space
                    exceedances of the dissolved oxygen criteria that will not impair the biological
                    community.

                    Benthic infaunal community samples are collected as part of the long-term Chesa-
                    peake Bay Benthic Monitoring Program at fixed and random locations  during the
                    summer, usually in August to September. If the benthic-IBI of that sample is "good,"
                    (in this case 3 or more on a scale of 1 to 5), dissolved oxygen conditions were likely
                    adequate for the previous one to two months (Dauer et al. 2005).
                    In order to ensure greater consistency in deriving the  open-water and deep-water
                    reference  curves, factor in the  state-adopted designated-use boundaries and take
                    advantage of a full two decades on monitoring data, both reference  curves were
                    updated.  To develop updated open-water and  deep-water  reference curves,  the
                    monthly fixed and random station locations for the benthic-IBI data from 1985 to
                    2005 were matched  with the monthly open-water and  deep-water designated-use
                    boundaries for the same time period. This updated approach differs from the original
                    method published by the EPA (2003a), which used a single designated-use boundary
                    coverage for the entire data record. An additional difference is that previously  this
                    method was used to define only the deep-water reference curve. The open-water
                    reference curve was based on an analysis in which "good" water quality conditions
                    were defined for reference segments by year (see Appendix H in U.S. EPA 2003a).
                    Reference  locations  were identified by  sorting the resulting data set by year,
                    segment, and designated use. If a designated use in a given segment in a given year
                    had only "good" benthic-IBI scores (>3), then the dissolved oxygen data for  that
                    segment, designated use, and summer period (June-September) can be used to
                    compute a reference  curve. Appendix F lists these use-segment-year combinations.
                    Separate CFDs were  generated for open-water and deep-water designated-use habi-
                    tats from the entire data  set of  summer dissolved oxygen data from all reference
                    locations over the 1985-2005 data record. Figures IV-2 and IV-3 respectively illus-
                    trate the resultant June-September open-water and  deep-water dissolved oxygen
                    criteria reference curves. Appendix G documents the equations for the reference
                    curves.
  chapter iv  »  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                                      41
Percent of Time ^


0%
Open Water Monthly Dissolved Oxygen Biological Reference Curve









'••
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume
       IV-2, Chesapeake Bay open-water 30-day mean dissolved oxygen criterion
biological reference curve applicable only during the June 1 through September 30
assessment period.
100% -
90%
80%
70%
d>
Efin% -
P
"O cno/
C
0>
o
Q.
ono/i.
m%
no/, _
Deep Water Monthly Dissolved Oxygen Biological Reference Curve










0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume
       IV-3, Chesapeake Bay deep-water 30-day mean dissolved oxygen criterion
biological reference curve.

-------
42

                    The default reference curve, illustrated in Figure II-4 in Chapter 2, should be used in
                    the assessment of the 30-day mean, open-water dissolved oxygen criteria during the
                    non-summer months (January 1 through May 31 and October 1 through December
                    31). The necessary biological indices and data were not available to support deriva-
                    tion of a biologically based reference curve for open-water habitats during the
                    non-summer months.



0)
F
H
o
c
01
o
Q_







100%'
90%-
ono/ 1

50%.




40%',

in%-
20%-

10%-
0

Deep Channel Dissolved Oxygen Biological Reference Curve














' 1
i««Hafi
v^
i**ia^f«''-**MiiS.s,MB,
v^"iffl'™'™«.,.«,,jav,w&_mi ^
















% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume
                           IV-4, Chesapeake Bay deep-channel dissolved oxygen criterion biological reference
                    curve.
                                     ur DJ


                    The April 2003 Chesapeake Bay water quality criteria document provides conflicting
                    guidance in the use of reference curves for assessing attainment of the four instanta-
                    neous minimum dissolved oxygen criteria. Pages  170 to  173 in U.S. EPA 2003a
                    display and discuss reference  curves for migratory spawning and  nursery,  open-
                    water, deep-water, and deep-channel criteria attainment assessment. All four sets of
                    designated-use specific criteria include a use-specific instantaneous minimum crite-
                    rion. With the exception of the deep-channel criteria (page  173 in U.S. EPA 2003a),
                    none  of these sections specifically describe whether a reference curve  should be
  chapter iv «  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                                 43
applied in assessing attainment of the respective instantaneous minimum criteria.
The  reader is left with the sense that the published reference  curves  should be
applied to all the dissolved oxygen criteria, regardless of the stated duration.
All four instantaneous minimum criteria for protection of the four designated uses—
migratory spawning and nursery, open-water, deep-water, and deep-channel—protect
against mortality from very short-term exposure to low dissolved oxygen  concentra-
tions (U.S. EPA 2003a). The other dissolved oxygen criteria with specific averaging
periods  (30-day, 7-day,  and 1-day means) protect against impairments—including
growth, respiration, and behavioral/avoidance—for which the  impairments  will not
impact the designated use. The 2003 EPA criteria guidance stated that there were no
"biologically  acceptable exceedances of the applicable criteria" for the instantaneous
minimum criteria, given that the impairment is  death (page 151  in U.S. EPA 2003a).
While updating the methodology for deriving the open-water and deep-water desig-
nated-use dissolved oxygen criteria reference curves for the  30-day mean criteria
(described above), there were times and locations in the Chesapeake Bay for which
healthy benthic infaunal  communities still existed despite exceedance of the 1
mg-liter1 instantaneous minimum criterion. The EPA recommends, therefore,  that
attainment  assessment  of the instantaneous minimum  deep-channel  dissolved
oxygen criteria be conducted with the CFD methodology using the deep-channel
biological reference curve (Figure IV-4; Appendices F and G).
      USE  OF                                  AS

Several Chesapeake Bay scientists have called for future publication of dissolved
oxygen criteria based on percent saturation (not concentration) and for state adop-
tion of such percent-saturation-based criteria into the states' water quality standards
regulations. They  cite fisheries physiology  research  showing  that the pressure
gradient between the surrounding water and the blood running through the fishes'
gills that truly determines whether sufficient oxygen exists in the water to support
aquatic life. For example, Dutil and Chabot (2001) reported:
    Fishes have developed several mechanisms to secure more oxygen from their
    environment in critical situations such as low oxygen availability (Hoar and
    Randall 1984). When the partial pressure of oxygen  in the environment
    drops below some critical limit, however,  the pressure gradient between
    blood and water may not allow the fish  to deliver as  much oxygen to its
    tissues as needed to meet metabolic requirements associated with ingestion,
    digestion, growth and activity. Thus critical thresholds may vary through
    time in demersal fish  species and are best described in terms of partial pres-
    sure of oxygen or percent saturation.
      chapter iv  *  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
44
                     These scientists also note that the amount of dissolved oxygen dissolved declines as
                     temperature and salinity increase. For example, fully saturated freshwater at 20°C
                     holds 9.28 mg-liter1 of oxygen, but fully saturated seawater at the same temperature
                     only contains 7.58 mg-liter1 of oxygen. Seawater at 1°C can hold  11.38 mg-liter1 of
                     oxygen; at 30°C it can hold only 6.37 mg-liter"1  oxygen. As  for the aquatic organ-
                     isms, research indicates that percent saturation drives the oxygen diffusion supplying
                     their respiratory demands.

                     Concentration-based, not percent-saturation-based, criteria were published given the
                     lack of reporting dissolved oxygen concentrations in terms of percent saturation in the
                     extensive effects database used to derive the Chesapeake Bay dissolved oxygen criteria
                     (U.S. EPA 2000). In addition, the lack of salinity and temperature values for each data
                     point in the laboratory-based low dissolved oxygen effects database prevented calcula-
                     tion of the concentration-based effects data into percent saturation numbers.
                     Following publication of the Ambient Aquatic Life Water Quality Criteria for Dissolved
                     Oxygen (Saltwater): Cape Cod to Cape Hatteras, EPA scientists evaluated the implica-
                     tions  of recommending dissolved  oxygen  criteria  as percent  saturation versus
                     concentration (U.S. EPA 2000). In an addendum to the 2000 Virginian Province salt-
                     water dissolved oxygen criteria document, the U.S. EPA (2003b) reported:
                        A standard based  on percent saturation has a wide range of differences in
                        partial pressure (2.14-4.01 Torr), which decreases with increasing temper-
                        ature.  The opposite is  more desirable, however, since respiratory demand
                        increases with temperature. Thus standards based on percent saturation are
                        likely to overprotect during winter and potentially underprotect in summer,
                        when organisms need the most oxygen. A standard based on  concentration
                        provides a more uniform difference in partial pressure over the temperatures
                        used (2.45-2.72 Torr).  Even though the range of difference is smaller, it still
                        increases with  temperature. Thus a standard based on absolute concentra-
                        tion is more likely to  create stable physiological  conditions for animals
                        throughout the year.
                     Scientists from the EPA have generated a version of the EPA Virginian Province salt-
                     water dissolved oxygen criteria as  percent saturation for the State  of Maine (G.
                     Thursby, personal communication). At this time, however, the  EPA does not have the
                     scientific basis to recommend a set of Chesapeake Bay dissolved oxygen criteria in
                     terms of percent saturation.
  chapter iv »  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                                     45
Chesapeake Bay Program. 1989. Chesapeake Bay Monitoring Program Atlas - Volumel:
Water Quality and Other Physio chemical Monitoring Programs. CBP/TRS 34/89. U.S. Envi-
ronmental Protection Agency, Chesapeake Bay Program Office, Annapolis, MD.

Dauer, D.M., M.F. Lane, and RJ. Llanso. 2005. Addendum to the Report: Development of
Diagnostic Approaches to Determine Sources of Anthropogenic Stress Affecting  Benthic
Community Condition in the Chesapeake Bay. Prepared for U.S. Environmental Protection
Agency, Chesapeake  Bay Program Office, by Department  of Biological  Sciences,  Old
Dominion University, Norfolk, VA,  and Versar, Inc., Columbia, MD.

Dutil, J. D. and D. Chabot. 2001. Impact of hypoxia on Atlantic Cod in the Northern Gulf of
St. Lawrence, p. 51-60 In R.V. Thurston (Ed.), Fish  Physiology, Toxicology, and Water
Quality. Proceedings of the Sixth International Symposium, La Paz, Mexico, January 22-26,
2001. U.S. Environmental Protection Agency Office of Research and Development,  Ecosys-
tems Research Division, Athens, GA. EPA/600/R-02/097. 372 pp.

Hoar, W.S. and DJ. Randall. 1984.  Fish Physiology. Volume X. Gills. Part A. Anatomy,  Gas
Transfer, and Acid-Base Regulation. Academic Press, 456 pp.

U.S. Environmental Protection Agency (U.S.  EPA). 2000.  Ambient Aquatic Life Water
Quality Criteria for Dissolved Oxygen (Saltwater): Cape Cod to Cape Hatteras. EPA 822-
R-00-012. U.S. Environmental Protection Agency, Office of Water, Washington, D.C.

U.S. Environmental Protection Agency. 2003a. Ambient Water  Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a for the Chesapeake  Bay and its Tidal Tributaries.
EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.

U.S. Environmental Protection Agency. 2003b. National Saltwater Criteria for Dissolved
Oxygen: Potential Addendum to Virginian Province Saltwater Criteria for Warmer and
Colder Waters. AED-03-113. U.S. Environmental Protection Agency, Office of Research and
Development, National Health and Environmental Effects Laboratory, Atlantic Ecology Divi-
sion, Narragansett, RI.

U.S. Environmental Protection Agency. 2003c. Technical Support Document for Chesapeake
Bay Designated  Uses and Attainability. EPA 903-R-03-004. Region III Chesapeake  Bay
Program Office, Annapolis, MD.

U.S. Environmental Protection Agency. 2004a. Ambient Water  Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a Chesapeake Bay  and its Tidal Tributaries - 2004
Addendum. EPA 903-R-04-005. Region III Chesapeake Bay Program Office, Annapolis, MD.

U.S. Environmental Protection Agency. 2004b. Technical Support Document for Chesapeake
Bay Designated Uses and Attainability - 2004 Addendum. EPA 903-R-04-006. Region III
Chesapeake Bay Program Office, Annapolis, MD.
       chapter iv  •  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
46
                     Weisberg, S.B., J.A. Ranasinghe, D.M. Dauer, L.C. Schaffner, RJ. Diaz, and J.B. Frithsen.
                     1997. An estuarine benthic index of biotic integrity (B-IBI) for Chesapeake Bay. Estuaries
                     20: 149-158.
  chapter iv  •  Refinements to the Chesapeake Bay Dissolved Oxygen Criteria Assessment Procedures

-------
                                                                                   47
                        chapter

              Refinements to  the

  Shallow-Water  Designated-Use

         Assessment  Procedures



                        BACKGROUND

Submerged aquatic vegetation (SAV) is a critically important component of the
Chesapeake Bay ecosystem. These underwater plants provide habitat used by many
fish and shellfish species and provide food for migratory waterfowl,  while also
improving water quality by generating oxygen, stabilizing sediment, and taking up
nutrients. Historically, the Chesapeake Bay was once known for its extensive SAV
beds. During the 1960s, however, much of the SAV disappeared. Poor water clarity,
caused by excessive algal growth and high levels of suspended sediment (Dennison
et al. 1993), was the primary factor in the decline of these beds. Both of these water
quality impairments result from human activities in the Chesapeake watershed that
cause excessive nutrient and sediment loadings to the Bay.

In 2003, after consultation with the watershed jurisdictions, the EPA published water
clarity criteria, SAV restoration goals, and shallow-water Bay grass designated-use
delineations for the Chesapeake Bay as well as its tidal tributaries and embayments
(U.S. EPA 2003a, 2003b). When applied as state water quality standards regulations,
these standards define the water clarity needed in delineated shallow-water habitats
to support SAV restoration to agreed-upon acreages.

The water clarity criteria and SAV restoration goals were designed to define attain-
ment of the shallow-water Bay grass designated use in three ways (U.S. EPA 2003a,
2004a). First, once the targeted acreage of SAV in a given segment is reached, that
segment is considered in attainment of the shallow-water Bay  grass designated use.
Measurement of SAV goal restoration attainment is based on  annual aerial surveys
in which the beds are photographed  and mapped, acreages quantified, and the single
best year of acreage determined. Second, if sufficient shallow-water area with the
water clarity necessary to achieve restoration of the targeted SAV acres exists, then
the segment is considered in attainment. These "water clarity acres"  are measured by
routinely mapping water clarity using data from the Chesapeake Bay Shallow-water
Monitoring Program (see Chapter 7 for details). Third, if the water-clarity criteria
were attained throughout the shallow-water designated use reaching to  a specific

              chapter v •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
48
                    depth contour (segment-specific water clarity criteria application depth) based on the
                    cumulative frequency diagram (CFD) assessment methodology, then the segment is
                    also considered in attainment of this designated use. Like the water clarity acres
                    approach, the  CFD-based assessment would be performed using data from  the
                    shallow-water monitoring program (see Chapter 7 for details).

                    For the 2006 Impaired Water 303(d) listing cycle, insufficient data existed to assess
                    water clarity  criteria attainment  in nearly all of the Chesapeake Bay segments'
                    shallow-water bay grass designated-use habitats. The SAV acreages have been quan-
                    tified for many years  (annually since 1984), however, and this  data collection is
                    expected to continue. Thus, the 2006 assessments used SAV acreages over the three-
                    year assessment period from 2001 through 2004. If the single best  year of SAV
                    coverage from that period exceeded the established, state-adopted SAV restoration
                    goal, then the segment's shallow-water designated use was deemed in attainment. If
                    the SAV restoration goal was not attained, then the segment's shallow-water desig-
                    nated use was listed either as impaired (category 5) or as insufficient data (category
                    3) since shallow-water monitoring data were unavailable for the segment.

                    The procedures  for assessing attainment  of the Chesapeake  Bay shallow-water
                    designated  use using the  water clarity criteria and SAV restoration acreages, first
                    published by EPA in  2003, were broadly defined and had not been extensively
                    applied in the Chesapeake  Bay prior to the 2006 303(d) listing cycle (U.S. EPA
                    2003a, 2003b). The jurisdictions  and the EPA identified and resolved many issues
                    during the first baywide application. This chapter provides detailed and refined guid-
                    ance on the assessment of the water clarity criteria and the SAV restoration goals.
                    Ultimately, the chapter evaluates attainment of the shallow-water bay grass desig-
                    nated use. This guidance replaces the applicable criteria assessment methodologies
                    previously published by the U.S. EPA (2003a, 2003b, 2004a, 2004c).
                    The shallow-water bay grass designated use is considered in attainment if sufficient
                    acres  of SAV are observed within the segment or enough acres of shallow-water
                    habitat meet the applicable water clarity criteria to support restoration of the desired
                    SAV  acreage for that segment (U.S.  EPA  2003a, 2003b).  Assessment of either
                    measure, or a combination of both, serves as the basis for determining attainment or
                    impairment of the shallow-water bay grass designated use.
                    Given SAV is the ultimate biological measure of attainment of the designated use, in
                    the absence of sufficient shallow-water monitoring data necessary to determine the
                    available water clarity acres or assess water clarity criteria attainment using the
                    CFD-based criteria assessment procedure, the EPA recommends the States assess
                    shallow-water bay grass designated use attainment/impairment based on the acres of
                    aerial mapped SAV.
  chapter v  * Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                                49
If a shallow-water bay grass designated  use  segment meets  its SAV restoration
acreage, that designated use-segment is considered in attainment of the designated
use and should be listed on part 2.

If a shallow-water bay grass designated use segment does not meet its SAV restora-
tion acreage  and  sufficient  shallow-water  monitoring data  is  available,  the
jurisdiction can then assess attainment of the designated use using water clarity acres
or water clarity criteria as described below.  If the water clarity acres/water clarity
criteria are met/attained based on  assessment  of spatially intensive shallow-water
monitoring data, then that designated use-segment is considered in attainment of the
shallow-water bay  grasses designated use and should be listed on part 2.

If the water clarity acres/water clarity criteria are not met/attained based on assess-
ment of shallow-water monitoring data, or if  there is insufficient data to make a
determination using water clarity acres, then that designated use-segment is consid-
ered not in attainment of the shallow-water bay  grasses designated use and should be
listed on part 5.

For those  segments that contain the shallow-water bay grass  designated use, attain-
ment of this use  should be assessed with the following procedure:
    If the segment's single best year  SAV acreage (described below)  drawn
    from the most recent three-year period of available data is equal to or greater
    then the state adopted SAV restoration acreage  for that segment, then that
    segment is considered to  be in attainment of its  shallow-water bay grass
    designated use. If the segment's single best year  SAV acreage is less than
    the state adopted SAV restoration acreage for that  segment, the state should
    then proceed to assess water clarity acres if sufficient shallow-water data is
    available,  otherwise, the segment is not in attainment.
    If the segment's water clarity acres  (defined below) calculated from the
    most recent  three-year period of available  shallow-water monitoring water
    clarity data is equal to or greater than state  adopted water clarity restoration
    acreage for that segment, then that segment is considered to be in attainment
    of its shallow-water bay grass designated use. If the segment's water clarity
    acres are less than the state adopted water clarity restoration acreage for that
    segment, then that segment is considered not  to be in attainment of its
    shallow-water designated use unless SAV acreage  data indicate attainment.
A jurisdiction may also choose to apply the  CFD-based  assessment of water
clarity  criteria,  described in more detail below, in  place of water clarity acres, to
assess attainment of the segment's shallow-water bay grass designated use.

Given that SAV  is the ultimate biological measure of attainment of the designated
use, the EPA recommends a specific sequence of criteria assessment: assessment of
SAV acres meeting the segment-specific restoration  acres first, followed by assess-
ment of water clarity acres or water clarity criteria attainment. In the absence of
sufficient  shallow-water monitoring data  to determine the available water clarity
acres or assess water clarity criteria attainment using the CFD-based procedure, the
                 chapter v  •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
50
                    EPA recommends  that the states assess  shallow-water bay  grass designated-use
                    attainment based on the acres of mapped SAV (see Chapter 8).

                    A5SI-;SSM!>J.  "  -.'..- >; .. >' M THE SIM :•  -  - K "' '., -r'.K •' '•!  >;W

                    Bay wide and segment-specific SAV restoration goals were defined for the Chesa-
                    peake Bay by evaluating the historical (1930s-1970s) and more recent (1980s-2000)
                    SAV distributions (U.S. EPA 2003b).  Historical aerial photographs, available for
                    selected years in the 1930s, 1950s, and 1960s, were converted to digital maps. Then
                    acreages of SAV for all photographed shallow-water areas in Chesapeake Bay, its
                    tidal tributaries and embayments were quantified. To set restoration goals for the
                    Bay, the single best year  of SAV coverage was defined as the restoration goal for
                    each segment. The combined individual restoration acreages yielded a baywide goal
                    of 185,000 acres. (See pages 105-122 in U.S. EPA 2003b for detailed documenta-
                    tion on the entire goal-setting process.)
                    This baywide restoration  goal was established "to  reflect the  historical  abundance,
                    measured as acreage and density from the  1930's to present" as committed to in the
                    Chesapeake  2000 agreement (Chesapeake Executive Council  2000) and essentially
                    represents the "existing use" as defined by the Clean Water Act. The single-best-year
                    approach was necessary because a common basis was needed to define the area of
                    SAV that should be present. The historical photography was not consistent through
                    time for all areas of the Bay and SAV acreages varied through time.  Since at  least
                    some coverage was available for most of the Bay, the single best year offered the best
                    option for setting goals (in selected cases with little or incomplete historical data, a
                    composite of multiple years of historical data was used to define the "single best
                    year") (U.S.  EPA 2003b).
                    Because the segment-based SAV restoration goals were established based  on the  prin-
                    ciple of a "single best year," the assessment of attaining that goal within an individual
                    Chesapeake Bay Program segment's shallow-water bay grass designated-use habitat is
                    defined in a similar manner. Attainment of the SAV restoration goal is reached when
                    the single best year  of SAV acreage during the applicable preceding three-year assess-
                    ment period equals or exceeds the established goal (defined as "SAV restoration acres"
                    in the states' water quality standards regulations) for that segment.
                    In nine segments, SAV restoration goals were not published in 2003 because no  SAV
                    was mapped in the available historical photography or through the baywide aerial
                    survey (U.S. EPA 2003b). At the same time, existing information does  not support
                    delineation of these entire segments as SAV no-grow zones following the detailed
                    decision rules documented by the  U.S. EPA (2003b). The EPA recommends the
                    jurisdictions maintain the shallow-water designated use in the nine segments that
                    didn't have an SAV restoration goal published in 2003 but were previously deter-
                    mined not to be an  SAV no-grow zone (Table V-l).
  chapter v •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                                   51
     V-1. Recommended tidal-water designated uses by Chesapeake Bay Program segment and state-adopted sub-
segment. Updated version of Table IV-3 originally published on pages 62-63 of the 2003 Technical Support Document for
Identification of Chesapeake Bay Designated Uses and Attainability (U.S. EPA 2003b). The asterisks (*) indicate that no
numerical SAV restoration acreage goal was published in 2003 for the shallow-water designated use of that segment.
See Table V-2 for the nine new segment numerical SAV restoration averages. The absence of an "X"  in the shallow-water
designated-use column indicates that segment has been entirely delineated as an SAV no-grow zone and the shallow-
water bay grass designated  use should  not apply to that segment.

                                                         Migratory
Spawning Shallow-
and Open- Deep- Deep- Water
Nursery Water Water Channel (SAV
Chesapeake Bay Program CBP Juris- (Feb. 1- (Year- (June 1- (June 1- growing
Segment Name Segment diction May 31) Round) Sept. 30) Sept. 30) season)
Northern Chesapeake Bay
Northern Chesapeake Bay
Upper Chesapeake Bay
Upper Central Chesapeake Bay
Middle Central Chesapeake Bay
Lower Central Chesapeake Bay
Lower Central Chesapeake Bay
Western Lower Chesapeake Bay
Eastern Lower Chesapeake Bay
Mouth of the Chesapeake Bay
Bush River
Gunpowder River
Gunpowder River
Middle River
Back River
Patapsco River
Magothy River
Severn River
South River
Rhode River
West River
Upper Patuxent River
Western Branch (Paruxent R.)
Middle Paruxent River
Lower Paruxent River
Lower Paruxent River
Lower Paruxent River
Lower Paruxent River
Lower Paruxent River
Lower Paruxent River
Upper Potomac River
Upper Potomac River
Upper Potomac River
CB1TF1
CB1TF2
CB2OH
CB3MH
CB4MH
CB5MH
CB5MH
CB6PH
CB7PH
CB8PH
BSHOH
GUNOH1
GUNOH2
MIDOH
BACOH
PATMH
MAGMH
SEVMH
SOUMH
RHDMH
WSTMH
PAXTF
WBRTF
PAXOH
PAXMH1
PAXMH2
PAXMH3
PAXMH4
PAXMH5
PAXMH6
POTTF
POTTF
POTTF
MD
MD
MD
MD
MD
MD
VA
VA
VA
VA
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
DC
MD
VA
X
X
X
X
X





X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X



X
X
X
X
X
X






X








X
X
X
X
X
X






X
X
X
X


























X
X
X
X
X
X
X
X
X
X
X
X
X
X
X*
X
X
X
X
X
X
X
X*
X
X
X
X
X
X
X
X
X
X
                                                                                                  continued
                 chapter v  *  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
     52
Tab
(continued)
Anacostia River
Anacostia River
Piscataway Creek
Mattawoman Creek
Middle Potomac River
Middle Potomac River
Middle Potomac River
Middle Potomac River
Lower Potomac River
Lower Potomac River
Upper Rappahannock River
Middle Rappahannock River
Lower Rappahannock River
Corrotoman River
Piankatank River
Upper Mattaponi River
Lower Mattaponi River
Upper Pamunkey River
Lower Pamunkey River
Middle York River
Lower York River
Mobjack Bay
Upper James River
Upper James River
Appomattox River
Middle James River
Chickahominy River
Lower James River
Mouth of the James River
Western Branch Elizabeth River
Southern Branch Elizabeth River
Eastern Branch Elizabeth River
Lafayette River
Mouth of the Elizabeth River
Lynnhaven River
Northeast River
C&D Canal
C&D Canal
Bohemia River
Elk River
Elk River
Sassafras River
ANATF
ANATF
PISTF
MATTF
POTOH1
POTOH2
POTOH3
POTOH
POTMH
POTMH
RPPTF
RPPOH
RPPMH
CRRMH
PIAMH
MPNTF
MPNOH
PMKTF
PMKOH
YRKMH
YRKPH
MOBPH
JMSTF1
JMSTF2
APPTF
JMSOH
CHKOH
JMSMH
JMSPH
WBEMH
SBEMH
EBEMH
LAFMH
ELIPH
LYNPH
NORTF
C&DOH
C&DOH
BOHOH
ELKOH1
ELKOH2
SASOH1
DC
MD
MD
MD
MD
MD
MD
VA
MD
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
DE
MD
MD
MD
MD
MD
X
X
X
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
X


X
X
X
X
X
X







X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X








X
X


X







X
X











X
















X
X


X




















X








X
X
X
X
X
X
X
X
X
X
X
X*
X
X
X
X
X*
X
X*
X
X
X
X
X
X
X
X
X
X





X
X
X
X
X
X
X
X
      chapter v  •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                               53
Table V-1.  (continued)
Sassafras River
Upper Chester River
Middle Chester River
Lower Chester River
Eastern Bay
Upper Choptank River
Middle Choptank River
Lower Choptank River
Mouth of the Choptank River
Little Choptank River
Honga River
Fishing Bay
Upper Nanticoke River
Upper Nanticoke River
Middle Nanticoke River
Lower Nanticoke River
Wicomico River
Manokin River
Manokin River
Big Annemessex River
Big Annemessex River
Upper Pocomoke River
Middle Pocomoke River
Middle Pocomoke River
Lower Pocomoke River
Lower Pocomoke River
Tangier Sound
Tangier Sound
Tangier Sound
SASOH2
CHSTF
CHSOH
CHSMH
EASMH
CHOTF
CHOOH
CHOMH2
CHOMH1
LCHMH
HNGMH
FSBMH
NANTF
NANTF
NANOH
NANMH
WICMH
MANMH1
MANMH2
BIGMH1
BIGMH2
POCTF
POCOH
POCOH
POCMH
POCMH
TANMH1
TANMH2
TANMH
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
DE
MD
MD
MD
MD
MD
MD
MD
MD
MD
VA
MD
VA
MD
MD
VA
X
X
X
X

X
X
X
X


X
X
X
X
X
X
X
X
X
X
X
X
X
X
X



X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X



X
X



























X
X
























X
X*
X
X
X

X
X
X
X
X
X

X*
X
X
X
X
X
X
X

X*
X*
X
X
X
X
X
Source: U.S. EPA 2003b, 2004b, 2004c, 2005
To determine attainment of the shallow-water bay grass designated use, SAV restora-
tion goals for these nine segments were established based on the total surface acre
between the shoreline and the 0.5-meter depth contour divided by  the 2.5 water
clarity acres multiplier (Table V-2). Any SAV no-grow zones within  the individual
segments were removed before conducting the above calculations.

                         ON

The EPA has determined that the shallow-water designated use is protected when
there is restoration of SAV to the targeted restoration acreages or when a sufficient
area of shallow-water habitat contains required levels of water clarity, accounting for
                chapter v »  Refinements to the Shallow-Water Pesignated-Use Assessment Procedures

-------
54
                          ¥-2,  SAV restoration goals for segments that had no published acreage goals
                               in 2003.
Chesapeake
Bay Program
Segment
ANATF (MD)
BACHOH
C&DOH (DE)
C&DOH (MD)
CHSTF
MPNOH
NANTF (DE)
PAXMH3
PAXMH6
PMKOH
POCOH (MD)
POCOH (VA)
RPPOH
WBRTF
Segment Name
Anacostia River-Maryland
Back River
C&D Canal-Delaware
C&D Canal-Maryland
Upper Chester River
Lower Mattaponi River
Upper Nanticoke River-Delaware
Lower Patuxent River Sub-Segment 3
Lower Patuxent River Sub-Segment 6
Lower Pamunkey River
Middle Pocomoke River-Maryland
Middle Pocomoke River- Virginia
Middle Rappahannock River
Western Branch Patuxent River
Shallow-Water
Habitat Area1
(Acres)
3
850
15
83
574
323
370
3
3
423
56
167
1,226
3
SAV
Restoration
Goal2
(Acres)
3
340
6
33
230
129
148
3
3
169
22
67
490
3
                     Determined as total surface area of the segment from adjacent shoreline out to the 0.5-meter depth
                      contour at mean low water minus any delineated SAV no-grow zone within the segment.
                     Calculated as the shallow-water habitat area divided by 2.5 (the water clarity acres multiplier) (see
                      U.S. EPA 2003a).
                     3No (or very limited) bathymetry data were available, therefore, no shallow-water habitat area or
                      SAV restoration goal acreage could be calculated.
                     vegetated bottom habitat. Based on the decades long record of published documen-
                     tation on SAV light requirements  (Batiuk et al. 1992, 2000; Dennison et al 1993;
                     Kemp et al. 2001; U.S. EPA 2003a, 2004a), the EPA recommends that an attainment
                     determination based on water clarity acres be based on 2.5 times each acre needed
                     to meet the restoration goal acreage.
                     A water clarity acre is defined as an acre of shallow-water bay grass designated-use
                     bottom  habitat, located anywhere between the 2-meter depth contour and the adja-
                     cent shoreline  inclusively, which has  been observed to achieve the  applicable
                     salinity-regime-specific water clarity criteria. A water clarity acre cannot be defined
                     within a delineated SAV no-grow zone (see pages 41-55 in U.S. EPA 2004c for
                     narrative descriptions and maps of those zones). For segments in which the resultant
                     water clarity acreage  exceeds  the total acres  of shallow-water habitat from the
                     shoreline out to the 2-meter depth contour, the water clarity restoration acreage will
                     be set at the total acreage out to the 2-meter depth contour.
                     Assessment of attaining a segment's water  clarity  restoration acreage should be
                     based on calculation of the arithmetic mean of the year-by-year arithmetic means of
                     a month-by-month accounting  of water clarity acres  over the  three-year SAV
                     growing season assessment period.  Calculation of water clarity  acres should be
  chapter v  *  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                               55
based on spatially intensive shallow-water monitoring turbidity data converted to Kd
(light attenuation coefficient), interpolated as described in Chapter 2  and then
compared to the corresponding Kd threshold assigned to each interpolator grid cell.
The total acreage of an interpolator grid cell is  added to the running total water
clarity acres for a segment when the interpolated Kd for that cell is less than or equal
to the Kd threshold assigned to that cell.

The Kd value based on achieving the applicable water clarity criteria at the 2-meter
depth will apply to all interpolator grid cells with centroids within the 2-meter to 1-
meter depth contours. All interpolator grid cells  with centroids that lie within the
area bounded by the shoreline and the 1-meter contour will be assigned the Kd value
for the 1-meter depth.

If the segment's single best year of water clarity acres, as calculated above, is equal
to or greater than the segment's water clarity restoration acreage, then that segment
has attained the shallow-water bay grass designated use. If the segment's single best
year of water clarity acres is less than the segment's water clarity restoration acreage,
then the segment is in non-attainment of this designated use.
The EPA recommends the states  adopt one of two approaches to calculating water
clarity acres. Both methodologies directly account for progress towards meeting the
SAV restoration goal acreage and measurement  of suitable shallow water habitat
acreage necessary to support restoration of the remaining SAV beds needed to reach
the goal acreage.

The first methodology was originally published in the 2004 Chesapeake Bay water
quality criteria addendum (U.S. EPA 2004a). This methodology assesses attainment
of the shallow-water bay grass designated use in  a segment through a combination
of mapped SAV acreage and  meeting the applicable water clarity criteria in an addi-
tional, unvegetated shallow water surface area equal to 2.5 times the remaining SAV
acreage necessary to meet the segment's restoration  goal  (SAV restoration goal
acreage minus the mapped SAV acreage). In other words, a segment's shallow-water
bay grass designated use would  be considered in attainment if there is  sufficient
acres of shallow-water habitat meeting the applicable water clarity criteria to support
restoration of the  remaining  acres of SAV, beyond the SAV beds already mapped,
necessary to reach that segment's SAV restoration goal acreage.  These  measure-
ments of SAV acreages and water clarity levels  would be drawn from three years of
data as previously described  in the Regional Criteria Guidance (U.S. EPA 2003a).
Here's a hypothetical example of this first methodology for determining attainment
of the shallow-water bay grass designated use using both mapped SAV acreage and
shallow-water habitat acreage meeting the water clarity criteria.  Segment X has an
SAV restoration goal acreage of 1,400 acres. Over the past three years, SAV beds
totaling 1,100 acres have been mapped within the segment. Therefore, the remaining
SAV  acreage  necessary  to  meet the segment's  restoration goal is 1,400 acres
(segment SAV restoration goal) minus 1,100 acres (SAV acres currently mapped) or
300 acres.   Beyond the currently vegetated shallow-water habitat, an additional
                chapter v •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
56
                    750 acres of shallow-water habitat (2.5 multiplier times 300 acres) is needed to attain
                    the water clarity criteria to determine this segment is attaining its shallow-water bay
                    grass designated use.

                    The second methodology directly  accounts for mapped acres of SAV within  the
                    calculation of water clarity acres.  As part of the month-by-month accounting of
                    water clarity acres, over the three-year SAV growing  season assessment period,
                    interpolator cells containing any mapped SAV beds are counted towards the total
                    water clarity acres.
                    Here's a hypothetical example of this second methodology for determining attain-
                    ment of the shallow-water bay grass use using both mapped SAV acreage and
                    shallow-water habitat acreage meeting the water clarity criteria.  Segment Y has an
                    SAV restoration  goal  acreage  of  1,400 acres.  Applying the  2.5  multiplier, this
                    segment also has a water clarity restoration acreage of 3,500 acres.  Over the past
                    three years, SAV beds totaling  1,100 acres have been mapped within the segment
                    each year. During each growing season's accounting of water clarity acres, these
                    1,100 acres of mapped SAV beds are directly counted towards the growing season
                    arithmetic mean water clarity acreage. Therefore, accounting directly for 1,100 acres
                    of mapped SAV beds as water clarity acres, an additional 2,400 acres (3,500 water
                    clarity restoration acres minus 1,100 acres of mapped SAV) of shallow-water habitat
                    is needed to attain the water clarity  criteria to determine this segment is attaining its
                    shallow-water bay grass designated use.

                    ASSESSMENT BASED ON  CFD-BASED WATER CLARITY
                    CRITERIA ATTAINMENT

                    A jurisdiction may choose to apply the CFD-based assessment of water clarity
                    criteria to evaluate attainment of the segment's shallow-water bay grass designated
                    use  (U.S. EPA 2003a, 2004a). To attain the designated use, the segment must meet
                    the applicable water clarity criteria throughout the applicable shallow-water habitat
                    (from the shoreline out to the  segment-specific water  clarity criteria application
                    depth contour) (see Table IV-13 on pages 115-117 in U.S. EPA 2003b) over three
                    SAV growing seasons,  factoring in allowable exceedances  using the appropriate
                    salinity-regime-based biological reference curve (see Figures V-l, V-2). Chapter 2
                    and Appendix B document the  application of the  CFD-based  criteria attainment
                    assessment in detail. Chapter 7 deals with the specific elements of the shallow-water
                    criteria attainment assessment procedures using a CFD-based evaluation.
                                SHALLOW-WATER  DESIGNATED  USES
                                      AND  SAV NO-GROW  ZONES
                    Shoreline habitats of 2 meters or less (where SAV is never expected to grow due to
                    extreme wave energy,  permanent physical alterations, natural discoloration of the
                    water,  and no  functional  shallow-water habitat  from  river  channeling) were
  chapter v •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                               57
designated as SAV no-grow zones (see pages 108-110 in U.S. EPA 2003b). In the 39
segments with SAV no-grow zones, 31 of the segments have such zones extending over
a portion of the segment (see Table V-l  on page 42 in U.S. EPA 2004c). In these
segments, an area delineated as an SAV no-grow zone should simply be left out of any
assessment of shallow-water designated-use attainment based on water-clarity acres or
on a CFD-based assessment of water clarity criteria attainment.

In the case of the eight segments where the entire shallow-water area was delineated
as an SAV no-grow zone (see pages 108-110 in U.S. EPA 2003b), the best available
information indicates the shallow-water bay grass designated use is not appropriate.
The EPA recommends that this designated use not apply to  (or that it be removed
from) any segment in which the area encompassing  the entire 2 meters or less
shallow-water habitat be delineated as an SAV no-grow zone (Table V-l).

Table V-l is an updated version of Table IV-3 originally published on pages 62-63
in the 2003 Technical Support Document for Identification of Chesapeake Bay
Designated Uses and Attainability (U.S. EPA 2003b). This revised table documents
the  above-described segments that are  entirely  SAV  no-grow  zones  (where the
shallow-water bay grass  designated use does not apply) or had no previously estab-
lished SAV restoration goal. This table includes a list of all the Chesapeake Bay
Program segments in the Chesapeake Bay, its tidal tributaries, and its embayments
(U.S.  EPA 2004b, 2005) as well as the sub-segments delineated by Maryland and
Virginia (U.S. EPA 2004c).
             R
The original 2003 Chesapeake Bay water quality criteria document included biolog-
ical reference curves to assess attainment of the water clarity criteria using the CFD
methodology (see pages 173-176 and Appendix H in U.S. EPA 2003a). Those refer-
ence curves were developed using data collected as part of the Chesapeake Bay
Water Quality Monitoring Program in which the monitoring stations are located in
open, mid-channel areas  of Chesapeake Bay, its tidal tributaries,  and its embay-
ments. Use of the fixed-station, mid-channel water quality data was necessary even
though these data are not necessarily representative of the Bay's shallow-water habi-
tats;  sufficient data more representative  of the shallow-water habitats were  not
available (see Chapter 9 in Batiuk et al.  2000).

Efforts  are  underway through  the Chesapeake Bay  Shallow-water  Monitoring
Program to collect water clarity data for use in generating more appropriate biolog-
ical reference curves.  These data are being collected (see Chapter 7 for additional
detail) in the same way that shallow-water designated use areas will be assessed. The
resulting biological  reference  curves will, therefore, be directly comparable  to the
CFD assessment curves (see Chapter 2 for further details). Further refinement of the
existing published water clarity criteria biological reference curves (e.g., updating
with more recent mid-channel data, developing four salinity-regime-based curves) is
                chapter v »  Refinements to the Shallow-Water Pesignated-Use Assessment Procedures

-------
58
                     not warranted at this time given ongoing collection of more appropriate  shallow-
                     water data. In the interim, the EPA recommends that states assess their water clarity
                     criteria using the CFD methodology which uses existing published biological refer-
                     ence curves to define the amount and pattern of allowable criteria exceedances.

                     Figure V-l illustrates the biological reference curve that states should apply in the
                     CFD-based water clarity criteria assessment of tidal fresh and oligohaline segments
                     with shallow-water bay grass designated uses. Figure V-2 illustrates the biological
                     reference curve that should be applied in the assessment of mesohaline and polyha-
                     line segments  with  shallow-water bay grass designated uses. Appendix H in this
                     document provides  the equations for the Chesapeake Bay  water  clarity criteria
                     biological reference curves. Preliminary results from evaluation of limited shallow-
                     water monitoring data  indicate that  biological reference  curves generated from
                     mid-channel data (Figures V-l  and V-2) and those generated from shallow-water
                     monitoring data (see Figure VII-11 in Chapter 7) are quite similar in overall shape
                     and levels of allowable exceedances.
Oligohaline and Tidal Fresh Monthly Clarity Biological Reference Curve
HAAO/
QAO/
Qflp/
7AO/
<1>
F
§
o
Ju dfi0/
Q_
OAO/

HAP/
AO/










0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume
                           ¥-1. Chesapeake Bay water clarity criterion biological reference curve for application
                     to tidal fresh and oligohaline shallow-water designated-use habitats.
  chapter v

-------
                                                                                                    59
1 nn°/ -,

QAO/
7O°/
(U
H
"5 jrOo/
c
(D
O
»r ArtO/
Q_
OAO/
m°/


0<
Polyhaline and Mesohaline Monthly Clarity Biological Reference Curve




S
L'X

''"*'\,,
*'"""""\,,,,^
"""*\-,m,f<,,,rnf
'~U**~~*%mmff
"5!»my,5H!%tes^^^^^^^^^
& 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent of Volume
      V-2. Chesapeake Bay water clarity criterion biological reference curve for application
to mesohaline and polyhaline shallow-water designated-use habitats.
                          LITERATURE  CITED

Batiuk, R.A., P. Bergstrom, M. Kemp, E. Koch, L. Murray, J.C. Stevenson, R. Bartleson, V.
Carter, N.B. Rybicki, J.M. Landwehr, C. Gallegos, L. Karrh, M. Naylor, D. Wilcox, K.A.
Moore, S. Ailstock, and M. Teichberg. 2000. Chesapeake Bay Submerged Aquatic Vegetation
Water Quality and Habitat-Based Requirements and Restoration Targets: A Second Technical
Synthesis.  CBP/TRS  245/00  EPA 903-R-00-014. U.S.  EPA Chesapeake Bay  Program,
Annapolis, MD.

Batiuk, R.A., R. Orth, K. Moore,  J.C.  Stevenson, W. Dennison, L. Staver, V. Carter, N.B.
Rybicki, R. Hickman, S. Kollar, and S. Bieber. 1992. Chesapeake Bay Submerged Aquatic
Vegetation Habitat Requirements and Restoration Targets: A Technical Synthesis. CBP/TRS
83/92. U.S. EPA Chesapeake Bay Program, Annapolis, MD.

Chesapeake Executive  Council.  2000.  Chesapeake  2000.  Chesapeake  Bay  Program,
Annapolis, MD.

Dennison, W.C., RJ. Orth, K.A. Moore, J.C. Stevenson, V. Carter, S. Kollar, P. Bergstrom,
and R.A. Batiuk. 1993. Assessing  water quality with submerged aquatic plants. Bioscience
43:86-94.

Kemp, W.M., R.A. Batiuk, R. Bartleson, P. Bergstrom, V. Carter, C.L. Gallegos, W. Hunley,
L. Karrh, E. Koch, J.M. Landwehr, K.A. Moore, L. Murray, M. Naylor, N.B. Rybicki, J.C.
                 chapter v  »  Refinements to the Shallow-Water Designated-Use Assessment: Procedures

-------
60
                     Stevenson, and DJ. Wilcox. 2004. Habitat requirements for submerged aquatic vegetation in
                     Chesapeake Bay: Water quality, light regime and physical-chemical factors. Estuaries 27:
                     363-377.

                     U.S. Environmental Protection Agency (U.S. EPA). 2003a. Ambient Water Quality Criteria
                     for Dissolved Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal
                     Tributaries. EPA 903-R-03-002.  Region III Chesapeake Bay Program Office, Annapolis,
                     MD.

                     U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
                     Bay  Designated Uses and Attainability.  EPA 903-R-03-004.  Region III Chesapeake Bay
                     Program Office, Annapolis, MD.

                     U.S. Environmental Protection Agency. 2004a. Ambient Water Quality Criteria for Dissolved
                     Oxygen, Water Clarity and Chlorophyll a Chesapeake Bay and Its Tidal Tributaries - 2004
                     Addendum. EPA 903-R-04-005. Region III Chesapeake Bay Program Office, Annapolis, MD.

                     U.S. Environmental  Protection  Agency. 2004b.  Chesapeake Bay  Program  Analytical
                     Segmentation Scheme: Revisions, Decisions and Rationales 1983-2003. EPA 903-R-04-008.
                     CBP/TRS 268/04. Region III Chesapeake Bay Program Office, Annapolis, MD.

                     U.S. Environmental Protection Agency. 2004c. Technical Support Document for Chesapeake
                     Bay Designated Uses and Attainability - 2004 Addendum. EPA 903-R-04-006.  Region III
                     Chesapeake Bay Program Office,  Annapolis, MD.

                     U.S. Environmental Protection Agency. 2005. Chesapeake Bay Program Analytical Segmen-
                     tation Scheme: Revisions, Decisions and Rationales (1983-2003) - 2005 Addendum. EPA
                     903-R-05-004. CBP/TRS 278-06. Region III Chesapeake Bay  Program Office, Annapolis,
                     MD.
  chapter v  •  Refinements to the Shallow-Water Designated-Use Assessment Procedures

-------
                                                                                        61
                         chapter
            Chlorophyll  a  Criteria
          Assessment Procedures
            STATE  WATER  QUALITY  STANDARDS
With publication of the April 2003 Ambient Water Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal Trib-
utaries, the EPA provided the states with a recommended narrative (non-numerical)
chlorophyll a criterion applicable to all of the Chesapeake Bay and its tidal tributary
waters (Table VI-1) (U.S. EPA 2003). From 2004  through early 2006, Virginia and
the District of Columbia adopted numerical chlorophyll a criteria for application in
the tidal James River (Virginia) and across all the District's jurisdictional tidal
waters. Both jurisdictions determined that algae-related designated use impairments
would likely persist  in these tidal waters even after  attainment of applicable
dissolved oxygen  and water clarity criteria. The technical information supporting
adoption of numerical  chlorophyll a criteria  by Virginia  and  the District was
published in the 2003 Chesapeake Bay water quality criteria document (U.S. EPA
2003). Maryland and Delaware  adopted narrative chlorophyll a  criteria into their
water quality standards regulations (Table VI-1).
Table VI-1. Chesapeake Bay narrative chlorophyll a criteria.
Concentrations of chlorophyll a in free-floating microscopic aquatic plants (algae) shall not
exceed levels that result in ecologically undesirable consequences—such as reduced water
clarity, low dissolved oxygen, food supply imbalances, proliferation of species deemed
potentially harmful to aquatic life or humans or aesthetically objectionable conditions—
or otherwise render tidal waters unsuitable for designated uses.
Source: U.S. EPA 2003.
                                      chapter vi  •  Chlorophyll a Criteria Assessment Procedures

-------
62
                             CHLOROPHYLL A CRITERIA ASSESSMENT
                                                PROCEDURES

                    CHLOROPHYLLS CRITERIA REFERENCE CURVE

                    To assess attainment of the State adopted numerical chlorophyll a concentration-
                    based criteria, it was necessary to establish a reference curve for use in the CFD
                    criteria attainment assessment process (U.S. EPA 2003). In the case of chlorophyll
                    a criteria where a biologically-based reference curve is not available, EPA recom-
                    mends the states use of the default reference curve described in Chapter 2 (see Figure
                    II-4 and Equation  1).

                    CHLOROPHYLLS CRITERIA ASSESSMENT

                    A criterion threshold is a concentration that should rarely be exceeded by a "popu-
                    lation"  of  concentration data  exhibiting healthy  levels. The  state-adopted
                    concentration-based chlorophyll a criteria values  are threshold concentrations that
                    should only be  exceeded  infrequently since a low number of naturally occurring
                    exceedances occur even in a healthy phytoplankton population. The assessment of
                    chlorophyll a  criteria attainment, therefore, should use the CFD-based assessment
                    method described in Chapter 2 that applies the default reference curve. These
                    Chesapeake Bay chlorophyll a criteria apply only to those seasons and salinity-based
                    habitats for which they were defined to protect  against applicable human health
                    and aquatic life  impairments. Each season—spring (March 1-May 31) and summer
                    (July  1-September 30)—should be assessed separately to evaluate chlorophyll a
                    criteria attainment.
                    Assessments of seasonal mean  chlorophyll a criteria should be based on seasonal
                    averages of interpolated data sets. To calculate the seasonal averages, each interpo-
                    lated  cruise within a season should be averaged  on a point-by-point basis  in
                    matching interpolator grid cells. Spatial violation rates should be calculated for each
                    seasonally aggregated interpolation in an assessment period.  For example, for a
                    summer open-water  seasonal chlorophyll a  criteria assessment of a three-year
                    assessment period, three seasonal  average interpolations representing  each season
                    (Year 1 Summer, Year 2 summer, Year 3 summer)  should be used.
                                            LITERATURE  CITED
                    U.S. Environmental Protection Agency. 2003. Ambient Water Quality Criteria for Dissolved
                    Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal Tributaries.
                    EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.
  chapter vi  •  Chlorophyll a Criteria Assessment Procedures

-------
                                                                                    63
                       chapter \/||
  Shallow-water  Monitoring  and
          Application  for Criteria
                     Assessment
    DESIGN AND APPROACH FOR CHESAPEAKE BAY
              SHALLOW-WATER MONITORING

In July 2001, the  Chesapeake Bay Program Monitoring and Analysis  Subcom-
mittee's Tidal Monitoring and Analysis Workgroup formed a Tidal Monitoring
Design Team that undertook the redesigning of the Chesapeake Bay Tidal Moni-
toring Network. Over the next two years, the Design Team set goals and objectives,
reviewed the existing Chesapeake Bay monitoring design, evaluated potential new
monitoring strategies, and made recommendations for implementing a network to
provide the requisite data and support to address the Chesapeake Bay Program's
programmatic goals and objectives.
The new Tidal Monitoring Network focused on meeting the water quality protection
and restoration goals and objectives of the Chesapeake 2000 agreement (Chesapeake
Executive Council  2000). The network's primary objective is to supply the water
quality monitoring information needed to assess the new water quality criteria for
dissolved oxygen, water clarity, and chlorophyll a — ultimately with the goal of
removing the Chesapeake Bay and its tidal rivers from the list of impaired waters.
Secondary network objectives are to provide information for defining the nutrient
and sediment conditions necessary for protecting living resources and vital habitats.
Water quality data would also support refinement, calibration, and validation of the
Chesapeake Bay Water Quality/Sediment Transport Model.
The design of the  new Tidal Monitoring Network emphasized monitoring of the
shallow-water designated use areas. In a 1999 study, the Maryland Department of
Natural Resources investigated the validity of using mid-channel data  to assess
nearshore areas. The 13-tributary study examined water quality at  127 nearshore
stations and compared the data to 54 adjacent mid-channel stations (Karrh 1999;
Batiuk et al. 2000). The study found wide variations between nearshore  and mid-
channel data, both within and between tributaries. Based on this finding, the
researchers concluded  that  decisions to  use  mid-channel data to characterize
nearshore conditions should be made on a site-by-site basis.  Figure VII-1 illustrates

                    chapter vii  • Shallow-water Monitoring and Application for Criteria Assessment

-------
64
                      this variability, showing situations in which a single, mid-channel data point would
                      not adequately represent suspended solids and chlorophyll a in shallow areas. The
                      Design Team concluded that monitoring of shallow,  nearshore waters  must have
                      greater spatial coverage to obtain an accurate representation of these parameters.
                                  Turbidity
                                   EH 0.0 - 5.0   C10.0 -15.0 • 20.0 - 25.0 • 30.0 - 40.0  • 50.0 - 60.0
                                   O 5.0 -10.0  IZ 15.0-20.0 • 25.0 -30.0 •40.0-50.0  • >60.0
                          B
                                  Chlorophyll
                                  Q 0.0 - 5.0
                                  Q 5.0 -10.0
10.0-15.0
15.0-20.0
I 20.0 - 25.0
I 25.0 - 30.0
| 30.0 - 40.0
I 40.0 - 50.0
50.0 - 60.0
>60.0
                      Figure VII-1. Spatial distribution of turbidity (A) and chlorophyll a (B) in the tidal James River.
                      Source: Virginia Institute of Marine Science—www2.vims.edu/vecos/
  chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                               65
To capture the temporal variability of dissolved oxygen, the new Tidal Monitoring
Network incorporated high-frequency monitoring stations in surface and nearshore
locations. Since then, the dissolved oxygen criteria assessment procedure has been
modified to  project the results of open-water dissolved oxygen assessments onto
adjacent shallow-water, designated-use areas,  instead of conducting a  separate
shallow-water assessment (see Chapter 4 for details). The design for collecting high-
frequency dissolved oxygen data will  likely be modified to represent dissolved
oxygen concentrations in open-water, designated-use habitats more accurately.

                   ••
The intensive shallow-water monitoring program design is based on two innovative
technologies that were extensively tested in Maryland's Magothy and Severn rivers
as well as Tangier Sound from 1999 to  2002. The Dataflow water quality mapping
component collects  high-resolution surface data from both open tidal-tributary and
shallow waters.  The shallow-water  buoy system  collects  high-frequency  (near-
continuous)  temporal data at specific locations, resulting in a data set that better
represents dissolved oxygen, chlorophyll a, and water clarity in time and  space in
smaller tidal tributaries, small embayments, and shallow-water habitats. In 2003, the
Maryland Department of Natural Resources, the University of Maryland's Chesa-
peake Biological Laboratory, the Virginia Department of Environmental Quality, and
the Virginia Institute of Marine Sciences initiated the new Chesapeake Bay Shallow-
water Monitoring Program. The two states and their partners closely coordinate
development of the monitoring  schedules, equipment, methodologies,  and quality
assurance procedures to ensure baywide compatibility and comparability.
The Shallow-water  Monitoring Program is based on two components  that collect
spatially and temporally intensive data. Known as "Dataflow," the spatially intensive
component includes a sensor array and a GPS system that provide data continuously
along a boat track in both shallow- and open-water designated-use areas. These data
can be used to develop detailed maps of water quality conditions. The temporally
intensive component is known as "continuous monitoring" and includes a sensor array
at fixed locations that provides data continuously through time. These  data reflect
episodic changes in water quality or signify extremes in water quality conditions.
The existing shallow-water monitoring design is based upon a three-year assessment
period. Data are collected from all  segments within a tidal tributary or embayment
during the same three  years. Both Dataflow  sampling and continuous buoys are
deployed for the same time period. The three-year assessment provides adequate
time to account for variation in both weather and hydrologic conditions (see page
151 in U.S.EPA 2003a). Assessments using fewer than three years of shallow-water
monitoring data are discussed in the section Schedule for Assessment of Shallow-
water Designated Use Habitats below.
To adequately assess water quality criteria in shallow-water habitats and tidal tribu-
tary  open-water designated-use  habitats,  the EPA recommends  that  the  states
                      chapter vii *  Shallow-water Monitoring and Application for Criteria Assessment

-------
66
                    conduct Dataflow monitoring from April through October in tidal fresh, oligohaline,
                    and  mesohaline  segments and  from  March  through November in polyhaline
                    segments. These assessment periods for the water clarity criteria were based on the
                    growing seasons for the salinity-based SAV plant communities (U.S. EPA 2003a).



                    Continuous monitoring data are collected to assess the variability of water quality
                    parameters throughout the day. Temporally intensive data help explain the relation-
                    ships and timing among algal blooms, low dissolved oxygen, and nutrient additions.
                    Although previous convention suggested that shallow-water habitats did not experi-
                    ence significant low  dissolved  oxygen levels, continuous  monitoring  data are
                    proving otherwise. The lowest dissolved oxygen levels often occur between 4:00 and
                    6:00 a.m. when, historically, little information has been collected.

                    The continuous monitoring program component employs automated YSI 6600 EDS
                    water quality data sondes. Maryland and Virginia have agreed to use similar instru-
                    ments,  when possible, to ensure  consistent methodology and comparability across
                    Chesapeake Bay segments. The YSI 6600 sonde directly measures dissolved oxygen,
                    fluorescence  (an  indication of chlorophyll a),  turbidity  (an  indication  of water
                    clarity), temperature,  salinity, and pH.  The  Maryland  Department  of Natural
                    Resources Chesapeake  Bay Shallow-water Monitoring Program Quality Assurance
                    Project Plan (see page 32 in Maryland Department of Natural Resources 2006) docu-
                    ments  the YSI instrument parameters, range, resolution, units,  and  accuracy.
                    Fluorescence is correlated to chlorophyll  a, the measurement used for  assessing
                    attainment of the chlorophyll a criteria. Turbidity is correlated to Kj (light attenuation
                    coefficient), the measurement used to assess attainment of the water clarity criteria.

                    The  initial design recommended two  shallow-water buoy deployments in  each
                    segment, but often, resources limit the number of buoys to one per site. The buoys
                    are programmed to take measurements every 15 minutes for the six parameters listed
                    above. They are deployed off piers or pylons, either 1-meter below the surface or at
                    a fixed depth of 0.3 meters above the bottom, generally in waters of 2-meters or less
                    in depth (Figure VII-2).

                    Instruments  are exchanged every one to two weeks, depending on biofouling and
                    following  strict calibration protocols (Virginia Institute of Marine Science 2005).
                    Field crews collect samples to calibrate fluorescence and turbidity instrument read-
                    ings, respectively, with  chlorophyll a and light attenuation. The  monitors are
                    positioned at representative sites both up- and down-river.

                    Both Maryland and Virginia have rigorous  shallow-water monitoring quality assur-
                    ance/quality control (QA/QC)  programs. The  QA/QC protocols remain consistent
                    between states; the Chesapeake  Bay Program Quality Assurance  Officer and the
                    Chesapeake  Bay Program's Analytical Methods and Quality Assurance Workgroup
                    have reviewed these protocols.
  chapter vii  »  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                               67
                                                                                       I)
Figure VII-2. Example of a continuous monitoring site and the generated 2004 dissolved oxygen data record
at Fenwick Point in the Potomac River, Maryland.
Source: Maryland Department of Natural Resources — www.eyesonthebay.net
Overlap periods occur at each continuous monitoring site by using multiple sondes
during routine biweekly maintenance runs to determine instrument drift. Instruments
are pre- and post-calibrated and must meet rigorous QA/QC protocols. Two instru-
ments are dedicated to each site. When one instrument is removed from the site for
maintenance, it is measured against the newly calibrated instrument. At the same time,
a field crew member takes a full suite of calibration samples for laboratory analysis.
Satellite and cellular telemetry are implemented at a subset of continuous monitoring
sites where resources permit. Data from these sites are assessed on a daily basis.
Maryland shallow-water continuous monitoring data are available in near- or real-
time  on the Department of Natural Resources  "Eyes  on  the  Bay"  website
(www.eyesonthebay.net) (Figure VII-3). Virginia shallow-water continuous  moni-
toring data  are available  on the Virginia  Institute of Marine Sciences website
(www2.vims.edu/vecos/).  The  Chesapeake  Bay  Program  website's data  hub
(www.chesapeakebay.net/data)  offers  access to the complete quality  assured
Shallow-water Monitoring Program datasets for Maryland and Virginia.

WATER QUALITY MAPPING  COMPONENT

The main purpose for collecting high-resolution water quality data is to provide reli-
able water quality criteria assessments. However, Dataflow monitoring also provides
insight into spatial complexities and localized phenomena and information for water
quality modeling in shallow waters (STAC 2005). The data are useful  in producing
maps of the extent  and patchiness  of algal  blooms,  seasonal and inter-annual
progressions, and localized water quality impairments.

The Dataflow system is a small, fast-moving vessel that pumps surface water contin-
uously from 0.5 meters below the water surface through a chamber surrounding the
                      chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
68
                                          Click Stations for Data
Click Legend Symbols to
 Toggle Stations On OH
     Fixed Monthly
   Stations - Current &
    Historical Data
     Real-Time
  Continuous Monitors
O     Near-Time
  Continuous Monitors
                     Figure VII-3. The 2005 fixed dataflow and continuous monitoring station map from the
                     Maryland Department of Natural Resources' "Eyes on the Bay" website.
                     Source: Maryland Department of Natural Resources — www.eyesonthebay.net.
                     probes of aYSI 6600 sonde (http://mddnr.chesapeakebay.net/sinVdataflow_instrumen-
                     tation.cfm). The system uses the same YSI 6600 sonde as the continuous monitoring
                     buoys and measures the same suite of six parameters—dissolved oxygen, fluorescence,
                     turbidity, temperature, salinity, and pH. A Global Positioning System (GPS) unit is inte-
                     grated into the computer system to measure the spatial position of each recorded
                     measurement. Data are collected every four seconds as the boat follows a cruise track
                     that traverses between shallow and open waters. These data are then interpolated to
                     provide  a high-resolution map of surface water quality conditions (see Chapter 2 for
                     further details). Each segment is mapped monthly from April through October or March
                     through November. The vessel stops at different locations throughout a segment for
                     discrete measurements of photosynthetically active radiation (PAR), Secchi depth, and
                     dissolved oxygen along with collection of water samples  for laboratory analysis of
                     chlorophyll a (for use as calibration data). These "calibration" sites often overlap with
                     existing open-water fixed-station sites and continuous monitoring sites; they represent
                     the dynamic range of water quality in that segment.
                       SCHEDULE  FOR ASSESSMENT  OF  SHALLOW-WATER
                                        DESIGNATED-USE  HABITATS

                     The current level of shallow-water monitoring  is insufficient to conduct detailed
                     water quality criteria assessments in all Chesapeake Bay shallow-water habitats by
                     the Chesapeake 2000 agreement deadline of 2010 (Chesapeake Executive Council
                     2000). Three possible actions might remedy this problem. The first is extending the
                     deadline beyond 2010 for assessment of all Bay  shallow-water habitats. The second
  chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                              69
is identifying additional resources to expand the monitoring needed to meet the 2010
deadline. The  third option is  assessing segments  for fewer than  three years if
noncompliance of the segment is established. All three options are addressed below.
Accurately assessing how many segments can be assessed by each action remains
impossible however, since determining the availability of additional resources or
establishing how many segments might need fewer than three years of monitoring if
noncompliance is established cannot be predicted.

                           -V VI

The Chesapeake Bay Program partners have  not  approved extending the shallow-
water clarity criteria assessment timeframe beyond 2010. The current deadline will
not be met due to a lack of adequate resources to implement the shallow-water moni-
toring program design agreed upon by  the Chesapeake Bay Program and the
participating states and thoroughly reviewed by the Chesapeake Bay Program Scien-
tific Technical Advisory Committee (STAC 2005). Significant  progress has been
made to accelerate the assessment schedule. Although intensive shallow-water moni-
toring water clarity monitoring data will not be available in all segments, attainment
of the shallow-water bay grass designated use  for those segments that contain an
SAV restoration acreage would be assessed by comparing each segment's single-best
SAV acreage from the most recent three-year  period with the jurisdiction's adopted
segment-specific SAV restoration acreage (see Chapters 5 and 8 for further details).
In this way, each shallow-water designated-use segment could have some assessment
completed each year.

                •U

Maryland, Virginia, and the EPA are actively seeking additional resources to expand
shallow-water monitoring in order to accelerate the schedule for completing baywide
assessments. In 2003,  when Maryland and  Virginia implemented  shallow-water
monitoring in 11 Maryland segments and seven Virginia segments, it was estimated
that it would take until 2018 to assess all 78 Chesapeake Bay Program segments over
a three-year period on a  rotating  basis. Since  the Shallow-water  Monitoring
Program's initial implementation, both Maryland and Virginia have developed part-
nerships with  county governments (e.g.,  Anne Arundel and Harford counties in
Maryland),  municipal  agencies (e.g., Hampton Roads Sanitation  District in
Virginia), and federal agencies (e.g., NOAA's National Estuarine Research System)
and secured additional state funding to accelerate monitoring of all segments. Based
on these new partnerships, current expanded resources, and segment assessment over
a three-year period, it is estimated that Maryland will complete all its shallow-water
assessments by the year 2014 and Virginia will complete all its shallow-water assess-
ments by 2015. Figures VII-4  and VII-5 depict the current  tentative schedule for
shallow-water  monitoring  and assessment of  Maryland  and Virginia  segments,
respectively. New sources of funding continue to materialize and the schedules indi-
cated by Figures VII-4 and VII-5 will change in response to funding adjustments.


                      chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
 70
T
                                   Year Assessment Complete
                                            2006   |      |  2010

                                            2007   |      |  2011

                                            2008   |      |  2012

                                            2009   f~  ~l  2014
Segments


   ] BACOH

|    | BIGMH
|    | BOHOH

|    | BSHOH
|| C&DOH

|| CB1TF
|    | CB2OH

|    | CB3MH
|    | CB4MH

|    | CB5MH
|    | CHOMH1
||    | CHOMH2

|l    | CHOOH

HH CHOTF
|    | CHSMH

|    | CHSOH
|    | CHSTF
|    | EASMH

|    | ELKOH

|    | FSBMH

|    | GUNOH
|    | HNGMH

|    | LCHMH

|    | MAGMH
   ~1 MANMH
   J MATTF

|    | MIDOH
|    | NANMH

|    | NANOH
|    | NANTF

|    | NOFTTF

|    | PATMH
|    | PAXMH

|    | PAXOH
|    | PAXTF

|    | PISTF
|    | POCMH

|    | POCOH

|    | POCTF
|    | POTMH


|    | POTTF

|    |RHDMH

|    | SASOH
|    | SEVMH

|    |SOUMH

|    | TANMH
|    | WBRTF
|    | WICMH

   ~~] WSTMH
                     Figure VII-4. Schedule for shallow-water monitoring of Maryland's Chesapeake Bay segments.
                     Source: Maryland Department of Natural Resources
                     ASSESSMENTS BASED ON REDUCED MONITORING
                     The three-year assessment period was established to account for inter-annual variations
                     in weather and hydrologic conditions (U.S. EPA  2003a). If conditions are seriously
                     degraded, a state having fewer than three years of data can establish noncompliance by
                     applying the CFD-based criteria assessment methodology as follows.
                     First, at the  start of a segment's  shallow-water monitoring, assume  100 percent
                     compliance in all three years of the coming assessment period. Second, after the first
                     year of monitoring, a state should develop a CFD based on the collected  data,
                     assuming all other planned sampling dates for the next two years had 100 percent
                     compliance with the applicable criterion. Finally, if the resultant assessment  CFD
                     indicates that the segment will be in violation (compared to the applicable reference
                     CFD) no matter what happens in the following  two years, then conclude that the
                     segment is out of compliance for the full assessment period and move the Shallow-
                     water Monitoring Program to another segment.
   chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                               71
                                                        YEAR ASSESSMENT
                                                        COMPLETE

                                                           SCHEDULE
                                                           HH 2006
                                                           ^•1 2008
                                                           I    I 2009
                                                           II 2010
                                                               2012
                                                               2013
                                                               2015
Figure VII-5. Schedule for shallow-water monitoring of Virginia's Chesapeake Bay segments.
Source: Maryland Department of Natural Resources

To illustrate this approach, two hypothetical scenarios are illustrated below. In the
first example (Figure VII-6), it is assumed that monitoring was conducted for one
year and that full attainment was achieved during all scheduled sampling dates over
the next two years. The shallow-water monitoring  over the first year indicated that
on all of the dates, the applicable criterion was violated in 15 percent or more of the
segment. The CFD indicates that the segment would be in noncompliance even if all
future sampling dates had 100 percent compliance. In this case, the state could have
decided to move the monitoring effort to a new shallow-water segment even after a
single year of study.

In the second example (Figure VII-7), the same assumptions are made and moni-
toring is conducted over one year. In this case, however, criteria exceedance is much
less extensive spatially and the  CFD indicates that full compliance could be possible
if the current level of attainment is found in future monitoring. Since neither compli-
ance nor noncompliance could be established during the first year, shallow-water
monitoring would need to continue. The  same analysis could take place after the
second year of monitoring and  the decision could be revisited. It may turn out that a
full three years of monitoring are necessary to determine if the segment remained in
full compliance.

Although determining noncompliance in fewer than three years works in theory, the
yearly segment data must be analyzed in time to adequately design and implement a
sampling scheme for a new segment.  The states  must have the flexibility to deploy
                      chapter vii •  Shallow-water Monitoring and Application for Criteria Assessment

-------
72
Cruise Cruise
Number Year

1
2
3
4
5
6
7
8
9

1
1
1
1
1
1
2
2
2
10 2
11 2
12 2
13 3
14 3
15 3
16 3
17 3
18 3
0.9
8 0.8
1
o 07
£ '
I 0.6
^
& 05
™
1 0.4
1 0.3
1
0.1
0 -

I \
t
1 \
\ X
\
\
\
\ '••!
\
^T^^^
x<-\


0 0.1 0.2
Cruise
Month

May
June
July
August
September
October
May
June
July
August
September
October
May
June
July
August
September
October










^_


0.3 0.4
%0f
Space
1
0.7
0.45
0.3
0.25
0.2
0.15
0
0
0
0
0
0
0
0
0
0
0
0

% of Ref
Time Curve
0.00 0.00
0.05 0.02
0.11 0.05
0.16 0.09
0.21 0.11
0.26 0.14
0.32 0.19
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00
0.37 1.00

	 Detailed Reference Curve
• "Scaled" Reference Curve
^— ^— Assessment Curve









— - — — — _

0.5 0.6











0.7 0.8 0.9 1
Percent Spatial Standard Exceedence
                                Scenario 1: noncompliance established after one year of shallow-water
                    monitoring.

                    resources to different systems. Often, implementation of a monitoring program for a
                    segment requires the coordination of various stakeholders and potential partners, the
                    leveraging of resources, and the allocation of field crews.
                    The Chesapeake Bay Program's Scientific and Technical Advisory Committee has
                    recommended that the tributary systems be assessed in their  entirety for the full
                    three-year  period  rather than evaluating individual segments  of a tributary in
                    different years (STAC 2005). This recommendation is particularly important for the
                    larger tidal  tributaries such as the Patuxent, Potomac, Rappahannock,  York, and
                    James rivers. These systems have tidal fresh, oligohaline, mesohaline,  and poly-
                    haline segments, all of which influence each other. To understand the vast ecosystem
  ihapter vii

-------
                                                                                                73
Cruise Cruise
Number Year
1 1
2 1
3 1
4 1
5 1
6 1
7 2
8 2
9 2
10 2
11 2
12 2
13 3
14 3
15 3
16 3
17 3
18 3
0.9 -\

0.8 I
!°M
i °-6 *.
•o
| 0.5- '•(
? 0.4
! \ \
| 0.3 N. vv
§ 0.2 ^\ *S
* V^ K
n- ^^^ -• ^ . .
0.1 ^^-ii:
0 -I 	 1 	 1 	 1 	
0 0.1 0.2
Cruise
Month
May
June
July
August
September
October
May
June
July
August
September
October
May
June
July
August
September
October







:iiife*ai^
0.3 0.4
%0f
Space
1
0.4
0.2
0.12
0.09
0.05
0.03
0
0
0
0
0
0
0
0
0
0
0
0
%of
Time
0.00
0.05
0.11
0.16
0.21
0.26
0.32
0.37
0.37
0.37
0.37
0.37
0.37
0.37
0.37
0.37
0.37
0.37
0.37

	 Detailed Reference Curve
"Scaled" Reference Curve
	 Assessment Curve





	 	 	 	 >_
0.5 0.6






0.7 0.8
Ref
Curve
0.00
0.06
0.14
0.23
0.29
0.44
0.57
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00








0.9 1
Percent Spatial Standard Exceedence
      VII-7. Scenario 2: noncompliance not established after one year of shallow-water
monitoring.

complexities and interactions between adjacent segments of a single tributary, it is
imperative to assess these tidal tributaries and segments as  whole systems and not
discontinue monitoring in one segment if noncompliance occurs after only a year or
two of assessment.
The states should make the decision whether to continue shallow-water monitoring
for the full three years or to move the monitoring to another segment after a year or
two of sampling. In making such decisions, the  state should consider the need to
gather shallow-water data for the assessment of multiple criteria (dissolved oxygen,
water clarity, and chlorophyll a) as well as other uses of the data (e.g., shallow-water
water quality  model development and calibration).  The states will  also  need to
                       ihapter vii

-------
74
                    consider if it makes sense (in terms of leveraging resources, coordinating, and under-
                    standing the relationship between segments and restoration activities) to discontinue
                    a segment's monitoring after one or two years if noncompliance of the segment is
                    shown. Finally, in the case of segments crossing two or more jurisdictional bound-
                    aries, all affected states will be involved in any decision to discontinue monitoring
                    prior to the end of the full three-year assessment period.
                    Importantly, the scenario described above and illustrated in Figure VII-6 does not
                    form an assessment that is lower in quality than one based on three years data. Non-
                    compliance is clearly established; that status would not change no matter what takes
                    place in ensuing years. The same approach  may not be viable using alternative
                    assessment strategies such as the water clarity-acres approach for the clarity criteria.
                    Since the water clarity-acres assessment method relies on the mean of three years  of
                    data, non-compliance could not be established in fewer than three years. The reverse,
                    however, may be true. However, if the segment's SAV restoration acreage goal was
                    attained during  any single year, then compliance would be established and the deci-
                    sion could be made to discontinue monitoring.



                    The states' prioritization schedule for assessing shallow water monitoring segments,
                    (Figures VII-4 and VII-5) is based on several criteria—SAV coverage, maximization
                    of resources, partnerships, and management needs such as dissolved oxygen criteria.
                    Segment prioritization through SAV coverage is based on assessing segments that
                    are close to meeting the state-adopted SAV restoration  acreage goal for the indi-
                    vidual segment. All states have  agreed  to assess attainment by each  segment's
                    single-best SAV acreage for the most recent three-year period with the jurisdiction's
                    adopted  segment-specific SAV  restoration acreage  (see Chapter  5 for  further
                    details). Many  Chesapeake  Bay segments range  between 50 and  100 percent  of
                    meeting their restoration goals.

                    Appendix G lists all the Delaware, Maryland, Virginia,  and the District of Columbia
                    segments and their relative success (by percent) in reaching their respective  state-
                    adopted SAV restoration acreages. Those segments that have already met their SAV
                    restoration acreages constitute  a lower priority  for shallow-water  assessment.
                    Segments that have not achieved any acres in meeting their SAV restoration acreage
                    form a lower priority as  well. The higher the percentage attainment in  meeting a
                    segment's SAV restoration acreage, the greater the priority was given for assessing
                    the shallow waters of that segment.
                    On the states' 2006 303(d) lists, eight Maryland segments and six Virginia segments
                    have met their  adopted SAV restoration acreages. The segments that have  already
                    attained their shallow-water designated use are low priority for shallow-water assess-
                    ment. Fourteen Maryland segments and five Virginia segments range between 50 and
                    100 percent of meeting their SAV restoration acreages (Appendix I). These segments
  chapter vii  »  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                              75
were granted the highest priority for shallow-water monitoring (see Figures VII-4
and VII-5).
      DISSOLVED  OXYGEN CRITERIA ASSESSMENTS
       USING  SHALLOW-WATER  MONITORING DATA
The Chesapeake Bay  Shallow Water Monitoring Program has provided unprece-
dented volumes of spatially  and temporally intensive Chesapeake  Bay,  tidal
tributary, and embayment data to assess water quality  criteria attainment.  This
wealth of data, however, provides  new and unique analytical challenges within the
regulatory framework. In the case of dissolved oxygen criteria, these  challenges
include: temporal variation of water quality parameters, spatial interpolations, and
scaling and interpolation issues. Specific procedures for evaluation of the 7-day, 1-
day,  and instantaneous minimum open-water  and  deep-water  dissolved oxygen
criteria have not been fully developed at this time.

The assessment of the  30-day mean dissolved oxygen criteria for open-water desig-
nated-use habitats  will rely on mid-channel fixed station data combined  with
Dataflow and Dataflow calibration profile data. As noted previously, the Dataflow
vessel stops at five to eight  locations throughout a segment to  collect  calibration
measurements. Dissolved oxygen is measured from the surface to the bottom at these
sites using the same procedure as the mid-channel  data collection. The dissolved
oxygen calibration data will provide an additional day of dissolved data each month,
at five locations instead of one or two. The dissolved oxygen Dataflow and the corre-
sponding Dataflow  dissolved  oxygen calibration data will be interpolated  and
analyzed, along with fixed-station dissolved oxygen data, using the Chesapeake Bay
Program's interpolator and the CFD approach described in Chapter 2.

TEMPORAL VARIATION

Dataflow cruises collect between 3,000 and 10,000 points over  several  hours in a
segment. Data are normally collected between 7:00 a.m. and 5:00 p.m. with the boat
traversing open and shallow waters on one side of a tidal tributary  or embayment and
repeating the process on the opposite side. The measurements can be interpolated to
produce a continuous surface of data that can be evaluated for the percentage area of
a segment that fails the applicable criterion.

The diel patterns of surface dissolved oxygen are well documented in both the liter-
ature  and  continuous monitoring  data  (www.eyesonthebay.net). In summer,
dissolved oxygen normally declines to its lowest level during the early morning
hours (3:00 a.m.) when algal  and plant communities have been respiring throughout
the night; it reaches its peak in mid-afternoon (3:00 p.m.) following photosynthetic
activity. In some cases, this diel fluctuation can reach  more than 15 mg-liter1
                      chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
76
                    dissolved  oxygen. When interpolating  water quality mapping  data collected
                    throughout the day, this variability presents a potential problem that is best illus-
                    trated by a map. Figure VII-8 shows that data collected early in the morning on one
                    side of the Severn River in Maryland is substantially lower than data collected later
                    in the day on the other side. If these measures were interpolated, it would appear that
                    one side of the river is faring more poorly than the other when, in fact, the dichotomy
                    merely represents a temporal artifact.

                    To produce a more representative spatial interpolation of surface dissolved oxygen
                    data, estimating the diel dissolved oxygen trend from continuous monitoring instru-
                    ments and using  that trend estimate to adjust the Dataflow dissolved oxygen may
                    prove more feasible.  The University of Maryland investigated this procedure by
                    comparing data from a nearshore continuous meter with those from a mid-channel
                    continuous buoy. They found that the  dissolved  oxygen  in  the two locations
                    responded differently  to the local habitats and that nearshore dissolved oxygen
                    dropped at night  and the mid-channel  dissolved oxygen was  highly variable, often
                    exceeding  dissolved oxygen saturation during  the  day. Although  the  adjustment
                    procedure  improved the data set, the prediction  error was high. Further research is
                    needed to integrate the spatial and temporal monitoring data.
                             Rising DO Values
                              During the Day
                               Severn River
                                6/28/2001
                                                                           <: rA
                                                                                  s-x
                     Figure VII-8. Illustration of rising dissolved oxygen concentrations during the day
                     (June 28, 2001) in the Severn River, Maryland.
  chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                               77
                         '."   .' .)N
The frequency and spatial coverage of water quality sampling will always remain
lower in relation to the temporal and spatial scales at which estuarine phenomena
occur. To overcome this reality, researchers must use innovative sample designs and
statistical methods. Throughout the Chesapeake Bay Program's tidal data analysis
and monitoring network design meetings, many of these issues regarding the inter-
pretation of shallow water monitoring data have been raised, but all were not solved.
The major issues relating to dissolved oxygen are highlighted below.
Water quality mapping of dissolved oxygen uses measures from a half-meter below
the surface. Some consider this type of measurement a weakness  given that most
hypoxic events occur in deep-water or deep-channel habitats. The last five years of
water quality mapping, however, have revealed that hypoxic events can affect surface
and shallow waters more than initially recognized. Each mapping cruise collects
calibration  samples and water quality depth profiles at  five to  eight stations per
segment. In much  the same fashion that fixed station profiles  are interpolated in
three-dimensions  using  the Chesapeake  Bay  interpolator (see  Chapter 2  and
Appendix D), the surface mapping data could be interpolated along  with calibration
station and  mid-channel, fixed-station depth profiles to enhance volumetric estimates
of dissolved oxygen. Advancements in monitoring attainment technology that enable
deployment of automated vertical profilers and surface and bottom buoy monitors
could also support this effort. Overall, the integration of data types such as contin-
uous monitoring, mapping, remote sensing, and fixed-station profiles poses one of
the greatest challenges in criteria assessment.
Water quality mapping cruises cannot cover every shallow-water cove and creek in
a segment,  thus presenting a problem for spatial extrapolation of the data. Criteria
assessment using the CFD method requires the use of an interpolated/extrapolated
surface from the entire segment and does not allow for exclusion of unsampled areas.
Almost certainly,  many of the areas outside of the sampling boundary have far
different conditions than those measured in the shallow waters of the main segment.
These  areas represent only a small percentage of each segment, but the question
remains whether they contain more valuable habitat than the space they occupy on a
percentage  basis.
Annually, many of the larger fish kills in Chesapeake Bay occur in these small tidal
creeks  and  embayments due to anthropogenic influences or natural  conditions. Two
months after  torrential rains in  June  2006, a Maryland Department of Natural
Resources aerial photography survey of the state's Eastern Shore tributaries revealed
that most small embayments were still clouded by silt and algal blooms to a far
greater extent than adjacent open waters.  To assess conditions adequately in these
shallow-water tidal creeks and embayments, a probabilistic approach may be needed
in conjunction with current shallow-water sampling design in which representative
small tidal  creeks and embayments are sampled by the  surface mapping and the
                      chapter vii  *  Shallow-water Monitoring and Application for Criteria Assessment

-------
78
                    results become a surrogate for the percentage area that these creeks represent in a
                    segment.

                    A STAC-convened expert panel (described in detail in Chapter 2) has reviewed the
                    interpolation of spatial  data. Several standardization  decisions for interpolation
                    methodology will need  to be made  to address the panel's recommendations for
                    addressing shallow-water monitoring  data (STAC 2006).

                     The water  clarity assessment uses  data from  the  shallow-water water quality
                     mapping to obtain high-resolution data in nearshore shallow waters. This section
                     describes the data analysis protocols for application of high-resolution turbidity
                     measurements to assess attainment of state-adopted water clarity criteria in shallow-
                     water monitored tidal tributaries and embayments of the Chesapeake Bay.

                     During each day of water quality mapping with the Dataflow, the operator stops at
                     five to eight locations  (calibration stations) to measure photosynthetic active radia-
                     tion (PAR) so that the light attenuation coefficient (Kj) can be calculated  and
                     correlated with the in situ turbidity values recorded  simultaneously. The protocol
                     followed to derive this correlation is described below.

                     The Chesapeake Bay  water clarity criteria were  published as  the percent of light
                     through water (see Table IV-1  on page 96 in U.S. EPA 2003a). Through the applica-
                     tion of the equation:
                                                    PLW= 100 exp(-KdZ)               Equation 3

                     the  appropriate percent light-through-water  value and the selected water clarity
                     criteria application depth (Z)  are  inserted and the equation is solved  for Kj. The
                     methodology developed by the  Chesapeake Bay Program  for assessing criteria
                     attainment involves a sequence of steps that leads to a cumulative frequency diagram
                     (CFD)  as described in eight  steps in Table  II-l  in  Chapter 2. As part  of step 3,
                     equating the in situ collected values of turbidity to estimated IQ values becomes
                     necessary to  determine exceedance of the water clarity  criterion.  It is  critical to
                     convert in-situ turbidity to estimates of IQ prior to any data interpolation in order to
                     reduce the error potential.

                     The relationship between turbidity and Kj, therefore, needs to be quantified to deter-
                     mine  the  turbidity  threshold of  the  applicable  water clarity  criteria. This
                     determination narrows the scope considerably from the traditional calibration curve
                     in which the estimation of Kj is based on measurements for a wide range of turbidity
                     concentrations. In the current application,  it is only necessary to accurately estimate
                     Kj from in situ measurements of turbidity in the neighborhood of the exceedance of
                     the water clarity criteria.
  chapter vii  *  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                                 79
In conducting the analysis to formulate the decision rules and calibration curves that
relate in situ turbidity measurements with calibration station K
-------
80
                            5J
                            3
                        73
                        O
                        Q.   21
                            04
 01     234567
                            1 .5 root Turb
system  - Chickahominy    — James
-- Mattaponi
—  York
                            -• Piankatank
                                                                            9    10

                                                                          — Lynnhaven
                                                                          — Pumunkey
                                                                                      11
                                                                                           12
                     Figure VII-9. Simple linear regression of predicted Kd versus the 1.5 root of measured
                     turbidity using shallow water monitoring data from seven Virginia tidal tributaries
                     (2003-2005).
                     Source: Virginia Institute of Marine Science—www2.vims.edu/vecos.
                     coefficients and intercepts occur to form groupings of tributary data for calibration
                     purposes. The groupings developed to date reflect a strong geographic pattern, which
                     strengthens their validity.

                     INTERPOLATION

                     The very dense in situ measurements of turbidity from each sampling cruise track
                     (Figure VII-10) are first converted to Kj. The natural log of the converted IQ values
                     are then interpolated using a standardized ordinary kriging procedure with ARC/GIS
                     into a 25-meter square grid over the segment's entire surface area. Once interpo-
                     lated, the resultant interpolated Kd values  are transformed back. Each interpolator
                     cell within a segment's shallow-water  area is  then assessed against a specific Kd
                     value for each  applicable water clarity  criterion application depth. An interpolator
                     cell value equal to or below this IQ value is considered in attainment of the appli-
                     cable water clarity criterion.  A number above this value has failed to meet the
                     applicable water clarity criterion.
                     The entire area within the shallow-water  designated-use zone for each sampling
                     cruise is then aggregated on an interpolator cell-by-cell basis to determine the total
                     area either in attainment  or failing  to  meet the applicable water clarity criterion.
                     Water clarity attainment acres are determined for the total area within the shallow-
                     water area of each segment from the shoreline out to the 2-meter depth contour
                     excluding the delineated SAV no-grow zones (see Chapter 5 for details).
  chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                                 81
                                                          Mobjack
                                                          Bay
               York
               River
YRKPH
                                                                 '15!-
                                    76°30'
Figure VII-10. Example lower York River polyhaline segment YRKPH Dataflow sam-
pling cruise track on August 25, 2003.
Source: Virginia Institute of Marine Science—www2.vims.edu/vecos..
Water clarity criteria attainment can also be assessed in each segment's  shallow-
water designated-use habitat through application of the CFD-based methodology
described in Chapter 2 for each three-year assessment period. Exceedance is the
cumulative frequency distribution of the portion of this zone that failed the Kd-
equivalent of the application depth specific water clarity criterion determined for that
segment compared to a reference CFD curve.
Naturally, environmental conditions will result in periodic exceedances of bay grass
water clarity requirements; such exceedances are allowable for bay grass survival
(U.S. EPA 2003a). Since allowable exceedances can be specific to salinity-based bay
grass communities, biologically based reference curves are  applied using measured
water clarity exceedances established from existing bay grass beds for each salinity
region using mid-channel water quality data (see Figures VI-1 and VI-2 in Chapter
5). Figure VII-11 shows a preliminary example of a biological reference  curve of
water clarity exceedances based on shallow-water monitoring data for established
bay grass beds in the polyhaline lower York River segment (YRKPH) during the
2003 and 2004 growing seasons. This curve is plotted along with the previously
published water clarity reference curve for mesohaline/polyhaline shallow-water bay
grass designated-use habitats (U.S. EPA 2003a).
                       chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
82
100 •
90 ;
80 :
I eo.

0 20 40 60 80 100
Percent of Area Exceeding Criteria
                           VII-11.  Comparison of the published mesohaline/polyhaline water clarity criteria
                     biological reference curve based on mid-channel water clarity measurements and a
                     preliminary example of a shallow-water monitoring-based water clarity criteria biological
                     reference curve.
                     Source: : U.S. EPA 2003a; Virginia Institute of Marine Science—www2.vims.edu/vecos.
                                               A


                     Attainment of the chlorophyll a criteria in the shallow-water designated use areas
                     will be based upon the adjacent open-water designated use assessments.  As with
                     dissolved oxygen assessments, open-water chlorophyll a assessments will rely on
                     the mid-channel fixed station data combined with Dataflow and Dataflow calibration
                     profile data.  These  data will be interpolated  and analyzed,  along with the fixed-
                     station chlorophyll  a data, using the Chesapeake Bay Program's interpolator and
                     CFD approach described in Chapter 2. The following sections describe the rationale
                     for and development of protocols for using the in-situ fluorescence measurements
                     from the  Dataflow system to assess chlorophyll a criteria attainment in shallow and
                     open-water tidal tributaries and embayments of Chesapeake Bay.

                     The Dataflow system generates a data set that better represents the spatial variability
                     of chlorophyll. The Dataflow cruise track transverses both  the open and shallow
                     water designated  use areas (see Figure  VII-10), recording hundreds of fluorescent
                     measurements,  very quickly and less expensively than the collection and laboratory
                     analysis of individual samples.  However, the conversion of the fluorescence data to
                     chlorophyll a must be done carefully to ensure that  they are comparable to the
                     chlorophyll a data upon which the chlorophyll a criteria were based.
  chapter vii

-------
                                                                                                 83
The in-situ fluorescence method is more susceptible to bias and interferences than
the laboratory method.  Instrument manufacturers recognize that low temperatures
and high turbidities can affect the fluorescence response and note that different
phytoplankton species can fluoresce differently in-situ even if the actual chlorophyll
content is the same (YSI, Inc. 1999). To overcome these effects, it is a common prac-
tice to  "calibrate"  the in-situ data to the laboratory  results by collecting and
analyzing a set of chlorophyll a samples in the laboratory concurrent with in-situ
measurements, and establishing a quantitative relationship, or "calibration" between
the methods via simple linear regression. The calibration may be done for each day
of sampling but better estimates may result if greater numbers of  observations are
incorporated into  a statistical model.

7  '    •'

The usual approach for calibrating in situ fluorescence to in vitro chlorophyll is to
develop a model of the form:
              Chlorophyll = f(fluorescence, other variables).          Equation 4
Usually the function f is a linear regression model and the estimates  of the coeffi-
cients for this model are obtained using least squares. With this model, a measured
value of fluorescence may be used as an argument to obtain a predicted chlorophyll
value.  By  evaluating other water quality  variables measured by the monitoring
program, it was  determined that fluorescence, temperature,  turbidity,  pH, and
seasonal variables be used as independent variables as described above.
One problem with this  standard approach is that least squares estimation requires
that data used as independent variables be measured without error. Clearly this
assumption is not satisfied for fluorescence. An alternative approach that treats both
in vitro and in situ chlorophyll as variables with measurement error  estimates the
logarithm of their ratio with a linear regression model:

             Log (R) = LogCChlj / Chl2) = f (other variables)           Equation 5

where:
Chlj =  in vitro chlorophyll
Chl2 =  in situ chlorophyll (note: fluorometers used to collect data for this
       study convert the fluorescence signal to chlorophyll with a standard
       algorithm and this is the number recorded); and
R   = the ratio of these two chlorophyll measures.
An estimate of in vitro chlorophyll is obtained from the in situ measurement by first
estimating the logarithm of R given the independent variables, back-transforming to
obtain an estimate of the ratio, and multiplying the in situ chlorophyll by the ratio to
estimate the in vitro chlorophyll.
                       chapter vii  *  Shallow-water Monitoring arid Application for Criteria Assessment

-------
84
                                   .\
                     Continuous monitoring data for Maryland and Virginia were analyzed to determine
                     a method of post-calibrating fluorescence/chlorophyll to match extractive chloro-
                     phyll more precisely. Because the instruments are identical, it was  assumed that the
                     relationships between the Dataflow fluorescence and chlorophyll a would show
                     similar patterns.  Maryland data were available for 2003 through 2005 for approxi-
                     mately 21 tidal tributaries (not all tributaries were sampled  in  all three  years).
                     Virginia data came from the York River. Initial tests indicated that no more variation
                     occurred between Maryland and Virginia data than among the tidal tributaries in
                     Maryland. This finding simplified the post-calibration model geographically  by
                     allowing combination of data from both states.

                     A second test of the data evaluated potential differences among years. This test also
                     proved negative, which signified that all three years of data could be combined when
                     developing  the post-calibration model. Tests of season and tributary differences
                     suggested that the final model would need to account for temporal and spatial differ-
                     ences. Further  analyses indicated the need for two tributary groups and two season
                     groups, meaning that four calibration curves will be required. Significant variables
                     in the  model  also included water temperature, turbidity,  and  pH. Significance is
                     defined here as a p-value of less than 0.05.

                     Initial results indicate that four calibration curves would be needed, two for season
                     and two for tributary.  All four models contain fluorescence, water temperature,
                     turbidity, and pH.



                     Several issues  were  addressed in conducting the analysis to formulate the decision
                     rules and calibration curves. Similar to the turbidity/IQj relationship, many of the
                     issues related directly to the decision to lump or  divide the data when computing
                     calibration curves and decision rules. The argument in favor of lumping (to perform
                     the analysis on a data aggregate) reasons that better estimates result when large
                     numbers of observations are averaged. On the other hand, the in situ to in vitro rela-
                     tionship may not be consistent across all subsets of the data (i.e., between different
                     tidal tributaries and embayments). If so, dividing the data and developing algorithms
                     for each set may lead to better overall precision.



                     Because species composition can affect the relationship of  in situ to in vitro chloro-
                     phyll  measurements, this  relationship may  change  with  the  seasons.  Thus, one
                     aggregate-or-divide issue requiring resolution is the effect of seasons.
                     The in situ/in vitro difference generally follows a seasonal pattern consistent with
                     known species composition patterns for Chesapeake Bay and its tidal tributaries. In
                     situ chlorophyll measurements have a negative bias when phytoplankton populations
  chapter vii  *  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                                 85
shift toward a large component of blue-green algae. Blue-green algae increase in
abundance during mid to late summer, particularly in tidal-fresh to low-salinity habi-
tats. The calibration data from both the continuous monitors and the Dataflow water
quality mapping show that the negative bias of the in situ measure becomes greater
in summer. It was determined that two season groups would be needed.
It must be recognized that forming  two season groups implements a model  that
captures the average condition, but may not capture the condition that exists  in a
particular  tributary on a given date. The seasonal appearance of blue-green algae is
not the same across tributaries and not even the same within a tributary from year to
year.  Even if the model predictions agree well with the observed data for the  past
three years, it is quite possible that a blue-green bloom could form at some unusual
time of year in the future and lead to biased prediction. Truly reliable calibration of
in situ chlorophyll to in vitro chlorophyll requires that some information on the
concentration of blue-green cells be included in the calibration model.



Geography is another general factor that may influence the in situ to in vitro chloro-
phyll  a relationship. Again, this influence  is  likely to be a phytoplankton species
composition effect. Other factors (e.g.,  turbidity), however, may play a role.  It is
recommended that the analysis model the  geography by treating locations (fixed-
stations for  continuous  monitors  or  river  systems for  Dataflow)  as  discrete
categorical predictors. If these predictors are statistically significant, the geography
portion of the model should be simplified using surrogate variables, such as salinity
and turbidity.
Spatial  patterns emerge  with  data  set analysis.  These patterns,  when viewed
geographically, appear to follow arrangements expected based  on  phytoplankton
species composition. In the Virginia Dataflow data, the trend is longitudinal within
the estuaries. In the tidal-fresh region, the in situ and in vitro measurements appear
similar, with a negative bias of in situ relative to in vitro emerging in downstream
stations (Figure VII-12).  In the upper tidal Mattaponi River,  one region occurs in
which in situ has a positive bias relative to in vitro. This  situation may occur due to
high background fluorescence from tannins (dissolved organic carbon) in the water.
In Maryland, the negative bias (yellow squares) appears in regions where blue-green
populations have been identified;  however, the data do not show  a longitudinal
gradient similar to the Virginia data (Figure VII-13).

Diel
In continuous monitoring data, many locations exhibit distinct diel patterns in the in
situ chlorophyll. This diel pattern often shows that chlorophyll is higher at night and
lower during the day. Other research has shown that fluorometric chlorophyll read-
ings made  in  direct  sunlight will  be  biased  low because  sunlight  inhibits
phytoplankton fluorescence. This finding, coupled with the observed pattern of lower
in situ chlorophyll during the day, raised the concern that continuous monitoring of


                      chapter vii *  Shallow-water Monitoring arid Application for Criteria Assessment

-------
86
                     Figure VII-12. Locations of the Virginia Chesapeake Bay Shallow-
                     water Monitoring Program calibration stations. In each location, a
                     circle indicates that no significant difference occurs between the in
                     situ chlorophyll measures and the in vitro chlorophyll measures. A
                     square indicates that the in situ measures are less than the in vitro
                     measures. An X indicates that in-situ measures are greater than the
                     in-vitro measures.
                     Source: Virginia Institute of Marine Science—www2.vims.edu/vecos.

                     in situ chlorophyll might be biased low during the day because of this measurement
                     problem. A special study was conducted at the Jug Bay station on the tidal Patuxent
                     River collecting hourly calibration samples for 24 hours. One set of samples was
                     collected monthly from March to December in 2005. Analysis of the in situ/in vitro
                     difference shows a very slight diel pattern in these data, but this variability became
                     trivial when compared to other sources of variance.

                     Collection Agency
                     The two principal agencies  collecting  these  data—the Maryland Department of
                     Natural Resources and the Virginia Institute of Marine Sciences—have  devoted
                     considerable effort to maintaining comparable shallow-water monitoring program
                     field collection methodologies, instrumentation, and QA/QC procedures. Even so,
                     because the two  agencies work in geographically distinct regions, comparing results
                     between agencies to determine if these data can be combined to estimate calibration
                     curves should prove  useful. Initial data evaluations indicate that no more variation
                     exists between Maryland and Virginia data than among the tidal tributaries in Mary-
                     land. These evaluations suggest that any differences between Maryland and Virginia
                     data may actually result from variations among the tidal tributaries and not from
                     dissimilarities between the data-collecting agencies.

  chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                                87
                                .^r.   r-t-r*
                              •• KMJJ.VN   >* *Jr~-x  VK
                              l/A *,       .•""HP'  V	
                                        'S*''   >» "•
                                ,.\^,^     ^
 Figure VII-13. Locations of the Maryland Chesapeake Bay Shallow-water
 Monitoring Program continuous monitors. In each location, a circle indicates
 that no significant difference exists between the in situ chlorophyll measures
 and those for in situ chlorophyll. A square indicates that the in situ measures
 are less than the in vitro measures.
 Source: Department of Natural Resources--www2.eyesonthebay.net.
Background Fluorescence

In some Bay areas, the background fluorescence constitutes a significant component
of the total fluorescence signal due to freshwater input from blackwater streams.
Background fluorescence is the fluorescence measured on filtered water. This study
will identify those areas where background fluorescence requires measurement and
develop an algorithm to adjust for background fluorescence. Analysis indicates that
background fluorescence is not significant in the systems assessed to date.

Ancillary Data

While conventional wisdom  holds that in vitro methods produce more accurate
measures of chlorophyll than in situ methods, both are still subject to  error. Using
data collected independently of either type, the relative accuracy of the two method-
ologies will  be assessed. For  example, measurements taken as part of the nutrient
suite (e.g., particulate nitrogen, total nitrogen, etc.) have some predictive power for
chlorophyll. In cases where the in situ and in vitro measurements differ by more than
expected due to sampling error, these ancillary data may resolve which is more reliable.

Often a time series of both in situ and in  vitro  chlorophyll will show that the two
measurements compare quite well for much of the data record, with occasional large
discrepancies. Because these large  discrepancies are most  problematic  from  a
                      chapter vii  •  Shallow-water Monitoring and Application for Criteria Assessment

-------
                    decision-rule point of view, they warrant special consideration. If one of the methods
                    is more likely to be in error when these discrepancies occur, this finding will affect
                    use of that method in the regulatory process.

                    To address this issue, separate models of in situ chlorophyll and in vitro chlorophyll
                    need to be developed for which the independent variables are taken from the suite of
                    nutrient measurements (e.g., total nitrogen, paniculate nitrogen, etc.). A pilot project has
                    shown that these  models are fairly predictive. In a case where a large discrepancy
                    between the in situ and the in vitro measurements exists, if one is in agreement with its
                    predictive model and the other is not, then the one out of agreement is likely in error.
                    Batiuk, R.A., P. Bergstrom, M. Kemp, E. Koch, L. Murray, J.C. Stevenson, R. Bartleson, V.
                    Carter, N.B. Rybicki, J.M. Landwehr, C.  Gallegos,  L. Karrh, M. Naylor, D. Wilcox, K.A.
                    Moore, S. Ailstock, and M. Teichberg. 2000. Chesapeake Bay Submerged Aquatic Vegetation
                    Water Quality and Habitat-Based Requirements and Restoration Targets: A Second Technical
                    Synthesis.  CBP/TR 245/00  EPA  903-R-00-014. U.S. EPA Chesapeake  Bay  Program,
                    Annapolis, MD.
                    Chesapeake  Executive Council. 2000.  Chesapeake  2000. Chesapeake  Bay  Program,
                    Annapolis, MD.
                    Karrh, L. 1999. Comparison ofNearshore and Midchannel Water Quality Conditions. 200
                    pp. Chesapeake Bay Program, Annapolis, MD.
                    Maryland Department of Natural Resources. 2006. Quality Assurance Project Plan for the
                    Maryland Department of Natural Resources Chesapeake Bay Shallow Water Quality  Moni-
                    toring Program for the period of July 1,  2006 - June  30, 2007. Maryland Department of
                    Natural Resources, Annapolis, MD.
                    Scientific and Technical Advisory Committee (STAC). 2005. Final Report of the Chesapeake
                    Bay Scientific and Technical Advisory Committee Workshop: Evaluating the Design and
                    Implementation of the Chesapeake Bay Shallow Water Monitoring Program Chesapeake Bay
                    Program Scientific and Technical Advisory Committee Publication 05-003.
                    Scientific and Technical Advisory Committee (STAC). 2006. The Cumulative  Frequency
                    Diagram Method for Determining Water Quality Attainment: Report of the Chesapeake Bay
                    Program STAC Panel to Review Chesapeake Bay Analytical Tools. STAC Publication 06-003.
                    9 October 2006. Chesapeake Bay Program Scientific and Technical Advisory Committee.
                    Chesapeake Research Consortium, Edgewater, MD.
                    U.S. Environmental Protection Agency. 2003a. Ambient Water Quality Criteria for Dissolved
                    Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and its Tidal Tributaries.
                    EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.
                    Virginia Institute  of Marine Science. 2005. Quality Assurance Project Plan for Shallow
                    Water Monitoring. Virginia  Institute of Marine  Science, College of William and  Mary,
                    Gloucester Point, VA.
                    YSI, Inc. 1999. Environmental Monitoring Systems, 6-Series Operations Manual.
chapter vii  *  Shallow-water Monitoring and Application for Criteria Assessment

-------
                                                                                89
                     chapter\/l||
 Framework for Chesapeake  Bay
         Tidal Waters 303(d)  List
                Decision-Making
                        BACKGROUND

Section 303(d) of the Clean Water Act and EPA Regulation 40CFR 130.7 requires
biennial identification of water segments that are not attaining water quality stan-
dards.  These segments must have a total maximum daily load (TMDL) analysis
completed and allocations established that result in water quality standards attain-
ment. The states comply  with this requirement through a process known as the
Integrated Reporting Requirements which covers the assessment and listing require-
ments through Clean Water Act sections 305(d), 305(b), and 314 (U.S. EPA 2005b).

Given  that the 2006 integrated reporting documents would be the first prepared
under the states' newly adopted Chesapeake Bay water quality standards regulations,
a collaborative effort (among the EPA and watershed states) began in spring 2005 to
develop a decision-making framework  for that portion of the 2006 submittals
addressing the Chesapeake Bay system. The Chesapeake Bay Program partners
reached agreement on several key assessment and listing issues. This chapter docu-
ments  these agreements and presents the resultant flowchart for Chesapeake Bay
tidal-water listing decisions to guide Delaware, Maryland, Virginia, and the District
of Columbia in future 303(d) listing cycles.
              LISTING CATEGORY DECISIONS
Each state-adopted, tidal-water designated use by Chesapeake Bay Program segment
(or formally adopted state sub-segment) is considered an individual spatial assess-
ment unit for the purposes of each state's 303(d) list (U.S. EPA 2003a, 2003b, 2004a,
2004b, 2005a).
If a segment has been previously listed in category 5—recognizing the recent adop-
tion of new Chesapeake Bay water quality standards—the original listing decision
             chapter viii •  Framework for Chesapeake Bay Tidal Waters 303(d) List  Decision-Making

-------
90
                    should stand until sufficient data are available to fully assess attainment for all appli-
                    cable criteria  components  in each designated-use segment assessment unit.  With
                    sufficient data, states can justify moving an individual designated-use  segment, or
                    the segment as a whole, to another listing category. The lack of sufficient data for
                    full assessment of the applicable criteria is not justification for moving a category 5
                    (impaired) segment to category 3 (insufficient data).

                    If a segment's designated use was not previously listed in category 5, it can be listed
                    under category  3 if insufficient data exist to assess attainment of  all applicable
                    criteria components. Because an individual segment may have up to five tidal-water
                    designated  uses (see Table  V-l in Chapter 5), the states  can place individual
                    segments in multiple listing categories based on the criteria assessment results for
                    each designated use in the segment.



                    The preceding  chapters document  the different Chesapeake  Bay  water  quality
                    criteria assessments. Across all Bay  criteria,  non-attainment is defined  as any
                    percentage of non-attainment (even less than 1 percent) given that the CFD-based
                    criteria attainment  assessment  method already factors in the small  percentage of
                    circumstances (in time  and space) in which the criteria may be exceeded and still
                    fully protect the tidal-water designated use (U.S. EPA 2003a).
                    Given that multiple criteria often protect an individual designated use (e.g., separate
                    30-day mean, 7-day mean, and instantaneous minimum criteria required for protec-
                    tion of the  open-water fish and  shellfish  designated use),  full attainment of the
                    dissolved oxygen criteria must involve assessment of each applicable criterion indi-
                    vidually (U.S. EPA 2003a). In designated-use-segment assessment units for which
                    data are available to assess all applicable dissolved oxygen criteria, the states can
                    proceed with a full assessment of attainment of that segment's designated use. For
                    those units with insufficient data for one or more of these criteria, states should not
                    make any decisions on removing that designated-use segment from part 5 during that
                    listing cycle.

                    Until the EPA publishes methodologies for assessing the 7-day  and 1-day mean,
                    along with the instantaneous minimum open-water and deep-water dissolved oxygen
                    criteria components, the EPA recommends the states rely strictly on the assessment
                    of the 30-day mean open-water and deep-water dissolved oxygen criteria for listing
                    decisions. For those open- and deep-water designated-use segments for which the
                    30-day mean criteria are in non-attainment, the jurisdictions should list the segment
                    on  part 5 as impaired in the  absence of data or methodologies  for assessing the
                    remaining criteria components. For those designated-use segments in which the 30-
                    day mean open- or deep-water criteria  are in attainment, the jurisdictions should
                    generate additional data  and apply criteria  assessment procedures to determine
  chapter viii «  Framework for Chesapeake Bay Tidal Waters 3Q3(cl) List  Decision-Making

-------
                                                                                                91
attainment of the 7- and 1-day means as well as the instantaneous minimum criteria
components. If a segment was first listed in 2006 based on the 30-day mean open-
water and/or deep-water criteria and subsequent 30-day mean open-water and/or
deep-water criteria data now shows the segment to be in attainment, then the segment
may be delisted for these criteria.
The shallow-water bay grass designated use is in attainment if a sufficient number of
acres of SAV occur within the segment or if enough acres of shallow-water habitat
exist that meet the applicable water clarity criteria to support restoration of the
desired acreage of SAV for that segment (U.S. EPA 2003a, 2003b). Assessment of
either measure, or a combination of both, can serve as the basis for determining
attainment or impairment of the shallow-water bay grass designated use.
Since SAV is the ultimate biological measure of attainment of the designated use, in
the absence  of sufficient shallow-water monitoring data necessary to determine the
available water clarity acres  or assess water clarity  criteria attainment using the
CFD-based  criteria  assessment procedure, EPA recommends  the  States assess
shallow-water bay grass designated use attainment/impairment based on the acres of
mapped SAV.
If a shallow-water  bay grass designated-use segment meets  its SAV restoration
acreage,  that designated use-segment is in attainment of the designated use and
should be listed on part 2.

If such a segment does not meet  its restoration acreage,  the jurisdiction can then
assess  attainment using water clarity acres or water clarity criteria as described in
Chapter 5. If the  water clarity acres or water clarity criteria are attained based on
shallow-water monitoring data, then that segment is in attainment of the shallow-
water bay grasses designated use and should be listed on part 2.

Finally, if the water clarity restoration acres or water clarity criteria are not attained
using  the same data, or if there are insufficient data to make a determination using
water clarity acres or water clarity criteria, then that segment is not in attainment of
the shallow-water bay grasses designated use and should be listed on part 5.
Any attainment/non-attainment determination of water clarity criteria based on mid-
channel-based monitoring is strictly diagnostic. These mid-channel data should not
directly form the basis for any listing decision based on attainment/non-attainment
of a segment's shallow-water bay grass designated use.
As described in Chapter 6, numerical chlorophyll a criteria attainment is assessed by
applying the appropriate numerical criteria over the applicable season for three years
using the CFD-based criteria assessment methodology.
               chapter viii  *  Framework for Chesapeake Bay Tidal Waters 303(d) List Decision-Making

-------
92
                                       OF
                    The benthic community health assessment is conducted in three phases to support
                    the states' tidal waters listing decisions (Llanso et al. 2005) (Appendices J and K).
                    Phase I evaluates the sample size from the segment during the five-year assessment
                    window.  An impairment assessment based on benthic  community health  is not
                    possible if the sample size requirement is not met. The data, however, may still prove
                    useful  as an adjunct to other aquatic life use  data. If the sample size satisfies the
                    requirements of the statistical method (n > 10), a formal assessment of status (i.e.,
                    impaired vs. supports aquatic life use) is determined using the "percent degraded
                    area" statistical methodology (Phase II).
                    Phase II assesses aquatic life use impairment based on a comparison of the Chesa-
                    peake Bay benthic index of biotic integrity or benthic-IBI scores (Weisberg et al.
                    1997).  This assessment is possible only when the  number of benthic-IBI  scores
                    within  a segment is sufficient to meet the sample size requirement of the approved
                    statistical method (n > 10). Phase II can result in one of two possible outcomes:  1)
                    the segment is  not impaired for aquatic life use due to benthic community status
                    (note that the segment may still be impaired due to failure  of the other aquatic life
                    use subcategories or criteria); or 2) the segment fails to support aquatic life use due
                    to benthic community status and is assessed as impaired (part 5).
                    Phase  III  identifies the probable causes of assessed benthic impairment  of the
                    segment using a diagnostic tool that can pinpoint potential sources of stress affecting
                    benthic community conditions in the Chesapeake Bay (Dauer et al. 2005). This
                    methodology can also identify causes of stress and quantify the magnitude of degra-
                    dation. In addition, it distinguishes stress due to contaminants from stress  due  to
                    other factors (Appendix L).



                    A Chesapeake Bay tidal-water designated-use criteria attainment assessment spread-
                    sheet has been  developed to assist the states in reporting listing decisions for each
                    designated-use  segment (Table VIII-1). The assessment reporting framework effi-
                    ciently documents relevant information as each segment goes through  the  listing
                    decision flowchart described below.
                    Table VIII-2 shows the example results of the  Chesapeake Bay benthic analysis for
                    the 2006 303(d) reporting cycle. The benthic-IBI assessments are separate from the
                    Chesapeake Bay water quality criteria attainment assessment determinations and
                    reported for the segments as stand-alone or supplemental information for the states
                    to use in their 303(d) listing cycle decisions.
  chapter viii  «  Framework for Chesapeake Bay Tidal Waters 3Q3(cl) List Decision-Making

-------
                                                                                                                                                                                           93
 E

 o
M—

 Ol
 C
 o
 Q_
 0)
 C
 01
 c
 0)
 E
 c
 ro

 TO
 u
 0)
T3
 0)
+-•
 TO
 C
 gi
't/i
 0)
T3
 TO
 TO
GO
 TO
 0)
 Q_
 TO  I
U
              (£>
              O
                 - ^
1
                                                  CN  CN
                                                  rn  ro
                   f%J fSJ

                   IT) CO

1

                                                                                                                :  Q_ Q_ -^ -^l
                                                                                                                                 : Q_  Q_  .^ J2
                                                                                                                                                -^1  D_ Q_ -^2  D_
                                                                                                                                                   Q Q     Q

                                                                                                                                                                              JL'  Ji! Q_ D_
1
                                                                                            :§•
                                                                                                   •
                                                                                          >  > O O  Q
                                                                                                      000°

-------
94
                     c
                     01
                    _Q
                    ID
                    O
                    O
                    rxl

                     t/i
                     0)
                    +J
                     TO
^H
O ^
!J1 '2
o> a
use
i. a o
a I -s
= S cs
to o T=
T3 U 2
£ o wt
0 .pj D
2-1°
s ^
rTi PQ

^ ^ -*^
^ ^ S3
S ^ * o
s-S |u
S "3 "^ -^
>_. S o ffs
5 g a S
Frt ' ' & ®
wo So
Q *
JS 11
§•£ su
| •§ .2 o
to *> ?? ^
"^ ^ o "«
Q> Q> g ^j
"» ja ca o
2 S T= H
Sf ^ §=S
Q -5s?
<1 ^
-a * TS
"!? 3 ^ ^ • ^ w
^ J .2 3 ^ o
gj g g^ rt ®
•3, « 1 •§ ii 's
|*.H
S to
B HH

^PQ
TS "S j^
•" "2 «8
'S 2 2
a wo 3
SO* ^H
hH O
"S
I
wt
t/}
"3
Kj
1
O
U
"3
1
•3
OJ
GO




O









O









o




•3
OJ
GO




O









O









o





CS'


r^

| JMSMHb



d
o








oo
oo
'







oo
oo
*/")








^2.



oo

CN


r^

| ELIPHa



O
Q
o








ON
CS







OO
oo
*/")








1^

1^



/">
^
r^i








^
O
ON






r^

| YRKPHa



O
Q
o








CS
CS







>/">
•
^








o





CS'


r^

| CHSMH



d
o
1








o









o









o







^
^

| APPTFa



•2
o
Q.
O
^H
P-l







O








o
o
*~~*








o




(N


^
^

\ CHSTF



d
o
1








o









o









o







^
^

| CHOTF



d
o
1








o









o









^.
CS


CS




^^

\ JMSTFa



d
o
1








o
CN








1^
^°.
^O








So



ON


-------
                                                                                                        95
The Chesapeake Bay Program partners reached agreement  on how to apply the
results  of the refined criteria assessment procedures most effectively in  making
Chesapeake Bay tidal-water listing decisions. The resultant decision-making frame-
work, presented here in the form of a flowchart, can guide jurisdictional decisions in
preparing  future integrated reporting  cycle submissions for the Chesapeake Bay
system (Figure VIII-1).

All of the  designated-use segment combinations for the  five possible tidal-water
designated uses—migratory fish spawning and nursery, shallow-water bay grass,
open-water fish and  shellfish,  deep-water seasonal fish,  and shellfish  and deep-
                                  Are the designated-use segment
                                    currently listed as impaired
                                       (Part 5) for the water
                                  qualityVbiological2 parameter?
                                 Yes
                                 I
           No
                           Are the data available
                            to reassess criteria
                              attainment?
          I
    Are the data available
      to assess criteria
        attainment?
              Are the available data
              sufficient to make an
              attainment decision?
                                                                       No
                                                                       I
Are the available data
sufficient to make an
attainment decision?
          Yes
     Criteria Attained?
List Part 2

List Part 5
  Dissolved oxygen, water clarity and chlorophyll a.
  2SAV acreage and benthic index of biotic integrity.
       VIII-1. The 303(d) listing decision flow chart for assessing tidal waters designated
uses in Chesapeake Bay and tidal tributaries.
                 chapter viii  «  Framework for Chesapeake Bay Tidal Waters 3(B(d) List DecisiorvMaking

-------
96
                    channel  seasonal  refuge—along  with the relevant dissolved oxygen, water
                    clarity/SAV restoration acreage, and chlorophyll a criteria are applied through this
                    listing decision flowchart. Benthic index of biotic integrity data are  also evaluated
                    for listing decisions.
                    The listing decision flowchart starts with each designated use-segment-applicable
                    criterion combination, asking whether that segment was previously listed in category
                    5 as impaired based on the specific water quality (dissolved oxygen, water clarity,
                    chlorophyll a) or biological (SAV acreage) criterion parameter. If  yes, its initial
                    listing status remains in category 5 pending new criteria attainment assessments. If
                    no, then the flowchart questions whether data now exist to assess criteria attainment.

                    SEGMENTS  PREVIOUSLY LISTED AS IMPAIRED

                    At the second level, the flowchart queries whether the available data are sufficient to
                    reassess criteria or index attainment. If yes, the  third level asks if  the applicable
                    criteria or index is attained. If all applicable criteria components and indices have
                    been attained, the designated-use segment is then listed in part 2.  If no, the desig-
                    nated-use segment remains in part 5. If insufficient data exist at the second level to
                    assess criteria attainment/index attainment, the designated-use segment previously
                    listed as impaired remains in part 5.

                    SEGMENTS  NOT  PREVIOUSLY LISTED AS IMPAIRED
                    At the second level, the flowchart queries whether the available data are sufficient
                    to reassess criteria or index attainment. If yes, the third level determines whether
                    the applicable criteria or index is  attained. If all applicable criteria components and
                    indices have been attained, the designated-use segment is  then listed in part 2. If
                    no, the designated-use segment is listed as impaired in part 5. If insufficient data
                    exist at the second level to assess criteria attainment/index attainment, the desig-
                    nated-use segment remains in part 3.

                    SHALLOW-WATER DESIGNATED-USE LISTING DECISIONS
                    If a  shallow-water designated-use segment does  not meet its SAV restoration
                    acreage, the EPA recommends that the state list this designated-use segment in cate-
                    gory 5 presuming the shallow-water monitoring data needed to assess water clarity
                    acres/criteria attainment do not exist.
                                             LITERATURE  CITED
                    Dauer, D.M., M.F. Lane, and RJ. Llanso. 2005. Addendum to the Report: Development of
                    Diagnostic Approaches to Determine Sources of Anthropogenic Stress Affecting Benthic
                    Community Condition in the Chesapeake Bay. Prepared for U.S. Environmental Protection
  chapter viii  •  Framework for Chesapeake Bay Tidal Waters 303(d) List  Decision-Making

-------
                                                                                                     97
Agency,  Chesapeake Bay  Program Office, by  Department of Biological Sciences, Old
Dominion University, Norfolk, VA, and Versar, Inc., Columbia, MD.
Llanso, R.J.,  J.H. V01stad, D.M. Dauer, and M.F. Lane. 2005. 2006 303(D) Assessment
Methods For  Chesapeake Bay Benthos. Final report submitted to Virginia Department of
Environmental Quality, September 2005. Versar Inc., Columbia, MD, and Old Dominion
University, Norfolk, VA.
U.S. Environmental Protection Agency (U.S. EPA). 2003a. Ambient Water Quality Criteria
for Dissolved Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and its Tidal
Tributaries. EPA 903-R-03-002. Region III Chesapeake  Bay Program Office, Annapolis,
MD.
U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
Bay  Designated Uses and Attainability.  EPA 903-R-03-004. Region III Chesapeake Bay
Program Office Annapolis,  MD.
U.S.  Environmental  Protection Agency.  2004a.  Chesapeake  Bay Program Analytical
Segmentation Scheme: Revisions, Decisions and Rationales 1983-2003. EPA 903-R-04-008.
CBP/TRS 268/04. Region III Chesapeake Bay Program Office, Annapolis, MD.
U.S. Environmental Protection Agency. 2004b. Technical Support Document for Chesapeake
Bay  Designated Uses and  Attainability: 2004 Addendum. EPA 903-R-04-006. Region III
Chesapeake Bay Program Office, Annapolis, MD.
U.S.  Environmental  Protection Agency.  2005a.  Chesapeake  Bay Program Analytical
Segmentation Scheme: Revisions, Decisions and Rationales 1983-2003:  2005 Addendum.
EPA  903-R-05-004.  CBP/TRS  278-06.  Region  III Chesapeake Bay  Program  Office,
Annapolis, MD.
U.S. Environmental Protection Agency. 2005b. Guidance for 2006 Assessment, Listing and
Reporting Requirements Pursuant to Sections 303(d), 305(b) and 314 of the Clean Water Act.
July 29, 2005. Office of Water, Office of Wetlands, Oceans and Watersheds, Assessment and
Watershed Protection Division. Washington, D.C.
Weisberg, S.B., J.A. Ranasinghe, D.M. Dauer, L.C. Schaffner, RJ. Diaz, and J.B. Frithsen.
1997. An estuarine benthic index of biotic integrity (B-IBI) for Chesapeake Bay. Estuaries
20: 149-158.

-------
98
Id-2
°c
CART
CBP
CDOM
CFD
cells/ml
Chla
DIN
DO
gCm-'d'1
GLM
HAB
IDW
kg m3m
km
LOAEL
m
mV
mg
inverse of the distance
squared
degrees Celsius
classification and
regression tree
Chesapeake Bay Program
colored dissolved organic
matter
cumulative frequency diagram
cells per milliliter
chlorophyll a
dissolved inorganic nitrogen
dissolved oxygen
grams of carbon per meter
squared per day
general linear model
harmful algal bloom
inverse-distance weighting
kilograms per cubic meter
per meter
kilometers
lowest observable acute
effects level
meter
cubic meters per second
milligram
mg chla m
mg liter'1
NASS
NH4
NO2
NO3
O2
PAR
PO4
ppt
PSU
QA/QC
SAV
STAC
TMDL
TSS
U.S. EPA
milligrams of chlorophyll a
per meter squared
milligrams per liter
non-algal suspended solids
ammonium
nitrite
nitrate
oxygen
photosynthetically active
radiation
dissolved inorganic
phosphorous/
orthophosphorous
parts per thousand
practical salinity unit
quality assurance/quality
control
submerged aquatic
vegetation
Science and Technical
Advisory Committee
total maximum daily load
total suspended solids
United States
Environmental
Protection Agency
//g/kg micrograms per kilogram
jug liter"1
% saturation
micrograms per liter
percent oxygen saturation

-------
                                                                     A-1
                   appendix
     The  Cumulative Frequency
          Diagram  Method for
     Determining Water  Quality
                  Attainment
        Report of the Chesapeake Bay Program STAC Panel to
         Review of Chesapeake Bay Program Analytical Tools
                   STAC Publication 06-003
                       9 October 2006
Panel Members:
David Secor, Chair (Chesapeake Biological Laboratory, University of Maryland
Center for Environmental Science)
Mary Christman (Dept. of Statistics, University of Florida)
Frank Curriero (Departments of Environmental Health Sciences and Biostatistics,
Johns Hopkins Bloomberg School of Public Health)
David Jasinski (University of Maryland Center for Environmental Science)
Elgin Perry (statistics consultant)
Steven Preston (US Geological Survey, Annapolis)
Ken Reckhow (Dept. Environmental Sciences & Policy Nicholas School of the
Environment and Earth Sciences, Duke University)
Mark Trice (Maryland Department of Natural Resources)
  appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-2
                     In accordance with the Chesapeake 2000 Agreement, the Chesapeake Bay Program
                     has  recently implemented important modifications to (1) ambient water quality
                     criteria for living resources and, (2) the procedures to determine attainment of those
                     criteria. A novel statistical tool for attainment, termed the Cumulative Frequency
                     Diagram (CFD) approach, was developed as a substantial revision of previous attain-
                     ment procedures, which  relied upon a simple statistical summary of  observed
                     samples. The approach was viewed as advantageous in its capacity to represent
                     degrees of attainment in both time and space. In particular, it was recognized that the
                     CFD could represent spatial data in a synoptic way: data that is extensively collected
                     across diverse platforms by the Chesapeake Bay Program Water Quality Monitoring
                     Program. Because the CFD approach is new to Bay Program applications, under-
                     lying statistical properties need to be fully established. Such properties are critical if
                     the  CFD  approach is to be used to rigorously define regional attainments in the
                     Chesapeake Bay.

                     In Fall 2005, the Chesapeake Bay Program  Scientific, Technical  and Advisory
                     Committee charged our working group to provide review and recommendations  on
                     the  CFD  attainment approach. As terms of reference we used guidelines of Best
                     Available Science recently  published by the American Fisheries Society and the
                     Estuarine Research Federation. Statistical issues that we reviewed included,

                       1. What are the specific analytical/statistical steps entailed in constructing CFD
                          attainment curves and how are CFDs currently implemented? (Section 2)

                       2. How rigorous is the spatial interpolation process  that feeds  into the CFD
                          approach? Would  alternative  spatial  modeling  procedures (e.g., kriging)
                          substantially improve estimation of water quality attainment? (Section 3)

                       3. What are the specific analytical/statistical steps entailed in constructing CFD
                          reference curves? (Section 4)

                       4. What are the statistical properties of CFD curves? How does sampling density,
                          levels of attainment, and spatial covariance  affect the shape of CFD curves?
                          What procedures are reliable for  estimating error bounds for CFD curves?
                          (Section 5)

                       5. From a statistical viewpoint, does the CFD approach qualify as best available
                          science? (Section 6)

                       6. What are the most important remaining  issues and what course of directed
                          research will lead to a more statistically rigorous CFD approach over the next
                          three years? (Section 7)

                     The central element of our work was a series of exercises on simulated datasets
                     undertaken by Dr. Perry to better  evaluate 1) sample densities in time and space, 2)
                     varying levels of attainment, and  3)  varying degrees  of spatial  and  temporal
  appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-3
covariance. Further, trials of spatial modeling on fixed station Chesapeake Bay water
quality data by Dr.s Christman and Curriero were conducted to begin to evaluate
spatial modeling procedures. These exercises, literature review and discussions
leading to consensus opinion are the basis of our findings. In August 2006, the
working group supplied preliminary findings and related text for use in the 2006
CBP Addendum to Ambient Water Quality Criteria that is now under review.



1. The CFD approach is feasible and efficient  in representing water quality
   attainment.

   The CFD approach can effectively represent the spatial and temporal dimensions
   of water quality  data to support inferences on whether regions within the Chesa-
   peake  Bay  attain  or exceed water quality standards.  The CFD  approach is
   innovative but could support  general application in water quality attainment
   assessments in the Chesapeake Bay and elsewhere. The CFD approach meshes
   well  within the  Chesapeake Bay  Program's monitoring  and  assessment
   approaches,  which have important conceptual underpinnings  (e.g.,  segments
   defined by designated uses).

   In accepting the CFD as the best available approach for using time-space data, the
   panel contrasted it with the previous method and those sustained by other juris-
   dictions. The previous method used by the Chesapeake Bay Program, similar to
   the approaches used in other states, was  simply  based on EPA assessment guid-
   ance in which all samples in a given spatial area were compiled and attainment
   was assumed as long as > 10% of the samples did not exceed the standard. In this
   past approach all samples were assumed to be fully representative of the specified
   space and time and were simply combined as if they were random samples from
   a uniform population. This approach was necessary at the time because the tech-
   nology was not available for a more rigorous approach. But it neglected spatial
   and temporal patterns that are known to exist in the standards measures. The CFD
   approach was designed to better characterize those spatial and temporal patterns
   and weight  samples according to the  amount of space or time that  they actually
   represent.
2. CFD curves are influenced by sampling density and spatial and temporal
   covariance. These effects merit additional research. Conditional simulation
   offers a productive means to further discover underlying statistical proper-
   ties and to construct confidence bounds on CFD curves, but further directed
   analyses are needed to test the feasibility of this modeling approach.

   The panel finds that the CFD approach  in its current form is feasible, but that
   additional research is needed to further refine and strengthen it as a statistical tool.
   The CFD builds on important statistical theory related to the cumulative distribu-
   tion function and as such, its statistical properties can be simulated and deduced.
   Through conditional simulation exercises, we have also  shown that it is feasible
   to construct confidence ellipses that support inferences related to threshold curves
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-4
                       or other tests of spatial and temporal compliance. Work remains to be done in
                       understanding fundamental properties of how the CFD represents likely covari-
                       ances of attainment in time and space and how temporal and spatial correlations
                       interact with sample size effects. Further, more work is needed in analyzing biases
                       across different types of designated use segments. The panel expects that a two-
                       three  year time frame of directed research and development will be required to
                       identify and measure these sources of bias and imprecision in support of attain-
                       ment determinations.
                     3. The success of the CFD-based assessment will be dependent upon decision
                       rules related to CFD reference curves. For valid comparisons, both reference
                       and attainment CFDs should be underlain by similar sampling densities and
                       spatial covariance structures.

                       CFD reference curves represent desired segment-designated use water quality
                       outcomes and reflect sources of acceptable natural variability. The reference and
                       attainment curves follow the same general approach in derivation: water quality
                       data collection, spatial interpolation, comparison  to biologically-based water
                       quality criteria, and combination of space-time attainment data through a CFD.
                       Therefore, the  biological reference curve allows for implementation of threshold
                       uncertainty as  long as the reference curve is sampled similarly to the attainment
                       curve. Therefore, we advise that similar sample densities are used in the deriva-
                       tion of attainment and reference curves. As this is not always feasible, analytical
                       methods are needed in the future to equally weight sampling densities between
                       attainment and reference curves.
                     4. In comparison with the current IDW spatial interpolation method, kriging
                       represents a more robust method and was needed in our investigations on
                       how spatial  covariance affects  CFD statistical inferences. Still, the  IDW
                       approach may sufficiently represent water quality data in many instances
                       and lead to accurate estimation of attainment. A suggested strategy is to use
                       a mix of IDW and kriging dependent upon situations where attainment was
                       grossly exceeded  or  clearly  met  (IDW)  versus  more-or-less "borderline"
                       cases (kriging).

                       The current modeling approach for obtaining predicted attainment values in  space
                       is Inverse Distance Weighting (IDW), a non-statistical spatial interpolator that uses
                       the observed data to calculate a weighted average as a predicted value for each loca-
                       tion on the prediction grid. IDW has several advantages. It is a spatial interpolator
                       and in general such methods have been shown to provide good prediction maps. In
                       addition, it is easy to implement and automate because it does not require any deci-
                       sion points during an interpolation session. IDW also has a major disadvantage - it
                       is not a statistical method that can account for sampling error.

                       Kriging is also a weighted average but first uses the data to estimate the weights
                       to provide statistically optimal spatial predictions. As a recognized class of statis-
                       tical methods with many years of dedicated research into model selection and

-------
                                                                                              A-5
  estimation, kriging is designed to permit inferences from sampled data in the pres-
  ence of uncertainty. Thus the quantity and distribution of the sample data are
  reflected in those inferences. Indeed, the panel's initial trials on the role of spatial
  sources of error in the CFD have depended upon the ability to propagate kriging
  interpolation uncertainty through the CFD process in generating confidence inter-
  vals of attainment.

  In comparison to IDW, kriging is more sophisticated but requires greater expertise
  in implementation. Kriging is available in commercial statistical  software and
  also in the free open source R Statistical Computing Environment, and requires
  geostatistical expertise and programming skills for those  software packages.
  Segment by segment variogram estimation  and subsequent procedures would
  require substantial expert supervision and decision-making. Thus, this approach is
  not conducive to automation. On the other hand, there may be CBP applications
  where the decision on attainment is clearly not influenced to any  substantial
  degree by the method of spatial interpolation. One suggested strategy is to use a
  mix of IDW and kriging - dependent upon situations  where attainment was
  grossly exceeded or clearly met (IDW) versus more-or-less "borderline"  cases
  (kriging).
5. More intensive  spatial  and temporal  monitoring  of water  quality  will
  improve the CFD approach but will require further investigations on the
  influence of spatial and temporal covariance structures on the shape of the
  CFD curve. This issue is relevant in bringing 3-dimensional interpolations
  and continuous monitoring streams into the CFD approach.

  In the near future, the panel sees that the CFD approach is particularly powerful
  when linked to continuous spatial data streams made available through the cruise-
  track monitoring program, and the promise of continuous temporal data through
  further deployment of remote sensing platforms in the Chesapeake Bay (Chesa-
  peake Bay Observing System: http://www.cbos.org/). These data sets will support
  greater precision and accuracy in both threshold and attainment determinations
  made through the CFD approach but will require directed investigations into how
  data  covary over different intervals of time and space. Further, there may be
  important space-time interactions that confound the CFD attainment procedure.

  Some of the assessments for the Bay such as that for dissolved oxygen require
  three dimensional interpolation, but the field of three dimensional interpolation is
  not as highly developed as that of two dimensional interpolation. Kriging can be
  advantageously applied in that it  can use information from the data to develop
  direction dependent weighted interpolations (anisotropy). Kriging can include
  covariates  like depth.  Options  for  implementing 3-D interpolation include:
  custom IDW software, custom kriging software  using GMS routines, or custom
  kriging software using the R-package.
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-6
                    The panel identified critical research tasks that need resolution in the near future.
                    The following is a list of critical aspects of that needed research. These research
                    tasks  appear roughly in order of priority. However, it must be recognized that it is
                    difficult to formulate as set of tasks that can proceed with complete independence.
                    For example, research on task 1 may show that the ability to conditionally simulate
                    the water quality surface is critical  to resolving the sample size bias issue. This
                    discovery might eliminate IDW as a choice of interpolation under task 3. The Panel
                    has made significant progress on several of these research tasks and CBP is encour-
                    aged  to  implement continued study  in a way that  maintains the  momentum
                    established by our panel.
                    1. Effects of Sampling Design on CFD Results
                       (a) Continue simulation work to evaluate CFD bias reduction via conditional
                          simulation.
                       (b) Investigate conditional simulation for interpolation methods other than
                          kriging—this may lead to more simulation work.
                       (c) Implement and apply interpolation with condition simulation on CBP data.

                    2. Statistical inference framework for the CFD
                       (a) Conduct confidence interval coverage experiments.
                       (b) Investigate confidence interval methods for non-kriging interpolation
                          methods.
                       (c) Implement and evaluate confidence interval procedures.

                    3. Choice of Interpolation Method
                       (a) Implement a file system and software utilizing kriging interpolation for CBP
                          data.
                       (b) Compare interpolations and CFDs based on kriging and inverse distance
                          weighting (IDW).
                       (c) Investigate nonparametric interpolation methods such as LOESS and spline
                          approaches.

                    4. Three-Dimensional Interpolation
                       (a) Implement 2-D kriging in layers to compare to current approach of 2-D IDW
                          in layers.
                       (b) Conduct studies of 3-D anisotrophy in CBP data.
                       (c) Investigate software for full 3-D interpolation.

                    5. High Density Temporal Data
                       (a) Develop methods to use these data to improve temporal aspect of CFD
                          implementation.
                       (b) Investigate feasibility of 4-Dimensional interpolation.
  appendix a  » The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                              A-7
In June 2000, Chesapeake Bay Program (CBP) partners adopted the Chesapeake
2000  agreement (http://www.chesapeakebay.net/agreement.htm), a strategic plan
that calls for defining the water quality conditions necessary to protect aquatic living
resources. These water quality conditions are being defined through the development
of Chesapeake Bay specific water quality criteria for dissolved oxygen, water clarity,
and chlorophyll_a to be implemented as state water quality standards by 2005. One
element of the newly defined standards is an assessment tool that addresses the
spatial and temporal variability of these  water quality measures in establishing
compliance. This tool has become known as the Cumulative Frequency Diagram
(CFD).

The (CFD) was first proposed as an assessment tool by Paul  Jacobson, of Langhei
Ecology  (www.LangheiEcology.com). At  that time  Dr. Jacobson was consulting
with the Chesapeake Bay Program as a member of the Tidal Monitoring Network
Redesign Team. Within this group, the CFD concept gained immediate recognition
and support as a novel approach that permitted independent modeling of the time and
space dimensions of the continuous domain that underlies Chesapeake Bay water
quality parameters. In addition, because  preparation  of the CFD uses spatial inter-
polation, the  approach can allow integration of data collected on different spatial
scales such as fixed station data and cruise track data.

While the benefits of the CFD approach has been recognized (U.S. EPA 2003) and
the the CBP  has begun implementation  of the approach for certain water  quality
parameters and segments of the Chesapeake Bay, investigations of the  statistical
properties revealed that the underlying shape parameters of the CFD were sensitive
not only to rates of compliance but also to sampling design elements  such as  sample
density. The  novelty of the  approach coupled with concerns about its  statistical
validity motivated the Chesapeake Bay Program to  request  that its Scientific and
Technical Advisory Committee (http://www.chesapeake.org/stac/) empanel a group
with expertise in criteria assessment, spatial data interpolation, and statistics to
assess the scientific defensibility of the  CFD. Here  we report the findings  of this
panel.

The primary goal of this panel is to provide an initial scientific review of the CFD
compliance approach. This review addresses  a wide range of issues  including: bias
and statistical rigor, uncertainty, practical implementation issues, and formulation of
reference curves. Because of the novelty of the CFD approach, the panel has endeav-
ored to research and explain the properties of the CFD and spatial modeling upon
which the CFD approach depends to provide a basis for this evaluation. These activ-
ities are beyond the scope of the typical review. However, because so little is known
about the CFD, it was necessary to expand the knowledge base.

The report is organized into 7 sections. In Section 2 of this report  we present the
CFD approach as a series of steps, each of which needs to be considered carefully in
evaluating its  statistical properties. Spatial interpolation is a critical but the most
statistically nuanced step in the CFD approach. Spatial interpolation of water quality
data in the CBP has to date received little statistical review. In Section 3 we evaluate
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-8
                     alternative geostatistical methods  as they pertain to the CFD approach. The CFD
                     approach is  an attainment procedure, which depends upon statistical comparison
                     between attainment and reference  curves. In Section 4, we present alternative types
                     of references curves and discuss statistical properties of each. In Section 5 the statis-
                     tical properties of CFD curves (applicable to both attainment and reference curves)
                     is elucidated through a series of conditional simulation trials.

                     In addition to this primary charge,  the panel is sensitive to the fact that the CFD will
                     be employed in the enforcement of water quality standards. Use as a regulatory tool
                     imposes a standard of credibility, which we review in Section 6.  We use here "best
                     available science"  and "best  science" criteria to evaluate the overall validity and
                     feasibility of the CFD approach, following guidelines established by the American
                     Fisheries Society and Estuarine Research Federation (Sullivan et al. 2006). These
                     follow other similar  criteria (e.g., The Daubert Criteria (Daubert v. Merrell Dow
                     Pharmaceuticals, Inc., 1993) and include:
                       1. A clear statement of objective
                       2. A conceptual model, which is a framework for characterizing systems, sating
                          assumptions, making predictions, and testing hypotheses.
                       3. A good experimental design  and a standardized method  for  collecting data.
                       4. Statistical rigor  and  sound logic for analysis and interpretation.
                       5. Clear documentation of methods, results, and conclusions
                       6. Peer review.

                     The panel has made progress in  better understanding statistical properties of the
                     CFD approach and overall, we recommend it as a feasible approach and  one that
                     qualifies  under most criteria  for best available  science.  Still, we believe  that our
                     efforts should only represent the beginning of a longer term effort to (1) Use simu-
                     lations and other means to support statistical comparisons of CFD curves; and (2)
                     Support the  CBP's efforts to model water quality data with sufficient rigor in both
                     spatial and temporal  dimensions.  Research and implementation recommendations
                     follow in Section 7.

                     The water quality criteria assessment methodology currently proposed by the E.P.A.
                     Chesapeake  Bay Program  (CBP)  involves the use  of a Cumulative Frequency
                     Diagram (CFD) curve. This curve is represented in a two dimensional plane of
                     percent time and percent space. This document briefly discusses the reasoning that
                     lead to the development of this assessment tool. The proposed algorithm for esti-
                     mating the CFD is given and illustrated with small data sets. Some properties and
                     unresolved issues regarding the use of the CFD are briefly discussed. In Section 5,
                     simulation studies explore in greater specificity the multiple issues related to error
                     and bias in the CFD approach.
  appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                              A-9
                        CFD

The CFD assessment methodology evolved from a need to allow for variability in
water quality parameters due to unusual events. For the water quality parameter to
be assessed, a threshold criterion is established for which it is determined that water
quality that exceeds this threshold is in a degraded state  (For simplicity, we will
speak of exceeding the threshold as representing degradation, even though for some
water quality constituents such as dissolved oxygen, it is falling below a threshold
that constitutes degradation). Because all water quality parameters are inherently
variable in  space and time,  it is unlikely that a healthy bay will  remain below the
threshold in all places  at all times. In the spatial dimension, there will be small
regions that persistently exceed the threshold due to poor flushing or other natural
conditions.  It is recognized  by CBP that these small regions of degraded condition
should not  lead to a degraded  assessment for the segment surrounding this small
region. Similar logic applies in the temporal dimension. For a short period of time,
water quality in a large proportion of a segment may exceed the threshold, but if this
condition is short lived and  the segment quickly returns to a healthy state, this does
not represent an impairment of the designated use of the segment. Recognition that
ephemeral  exceedances of the  threshold in both time and space  do not represent
persistent impairment of the segment leads to an assessment methodology that will
allow these conditions to be classed as acceptable while conditions of persistent and
wide spread impaired condition will be flagged  as unacceptable. The assessment
methodology should first ask how much of the segment (for simplicity, a spatial
assessment unit is called a segment, but more detail  is given on spatial assessment
units in Section 2) is not in compliance with the criteria (percent of space) for every
point in time. In a second step the process should ask how often (percent of time) is
a segment out of compliance by more than a fixed percent of space. The results from
these queries can  be presented in graphical form where percent  of time is plotted
against percent of space (Figure 2.1). It  is arbitrary to treat space first and time
second. A similar diagram could be obtained by first computing percent noncompli-
ance in time and then considering the cumulative distribution of percent time over
space.

If a segment is generally in compliance with the criterion, then one expects a high
frequency of dates where the percent out of compliance is low. In this case, the CFD
should descend rapidly from the upper left corner and pass not too far from the lower
left corner and then proceed to the lower right corner. The trace in Figure 2.1 shows
the typical  hyperbolic shape of the CFD. The closer the CFD passes to the origin
(lower left corner), the better the compliance of the segment being assessed. As the
CFD moves away from the origin, a higher frequency of large percents of space out
of compliance is indicated.

              an           of     CFD

The algorithm developed by CBP for estimating the CFD is most easily described as
a series of steps. These steps are given in bullet form to provide a frame work for the
overall approach. The quickly defined framework is followed by a simple example.
This in turn is followed by more detailed discussion of each step.
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-10
                            LO

                            0.9

                            0.8

                            0.7
                         "S
                         _ 0.5
                           0.3

                           0.2

                           0.1
                           0.0
                             0.0    0.1     0.2     03    0.4     0.5     0.6
                                                       Fraction of Space
                                                                         0.7
                                                                                0.8
                                                                                      0.9    LO
                            2.1. Illustration of CFD for 12 dates.


                      The steps:
                        1. Collect data from a spatial network of locations on a series of dates in a three
                           year assessment period.
                        2. For each date, interpolate the data for the entire system (e.g. mainstem bay) to
                           obtain estimates of water quality in a grid of interpolation cells.
                        3. For each interpolation cell assess whether or not the criterion is exceeded.
                        4. For each assessment unit (e.g. segment), compute the percentage of interpolator
                           cells that exceed the criterion as an estimate of the percent of area that exceeds
                           the criterion.
                        5. Rank the percent of area estimates for the set of all sample days in the assess-
                           ment period from largest to smallest and sequentially assign to these ranked
                           percents a value that estimates percent of time.
                        6. Plot the paired percent of time and percent of area data on a graph with percent
                           of area on the abscissa and percent of time on the ordinate. The resulting curve
                           is the Cumulative Frequency Diagram.
                        1. Compare the CFD from a segment being assessed to a reference CFD.  If at any
                           point the assessment CFD exceeds the reference CFD, that is, a given level of
                           spatial  noncompliance occurs more often than is allowed, then the segment is
                           listed as failing to meet it's  designated use.
   appendix a  «  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-11
For this example, assume a segment for which the interpolation grid is 4 cells by 4
cells. In reality, the number of grid cells is much larger. Also let data be collected on
5 dates. Typically data would be monthly for a total of 36 dates. Let the criterion
threshold for this fictitious water quality parameter be 3. In what follows, you will
find an illustration of the  steps  of  computing the  CFD for these  simplified
constraints. The three columns of the next page show the first three steps. Column
1 shows fictional data for five dates for five fixed locations in a 2 dimensional grid.
Column 2 shows a fictional interpolation of these data to cover the entire grid.
Column 3 shows the compliance status of each cell in the grid where 1 indicates
noncompliance and 0 indicates compliance.
  Step 1. Collect data at
  known locations.
  date 1
3


2





5


3


1
date2
1


1





3


1


1
date3
4


1





2


2


1
date4
1


4





2


4


1
dateS
1


1





2


3


1
Step 2.  Interpolate the
data to grid cells.
date 1
3
4
3
2
4
4
3
3
5
5
4
3
3
2
1
1
                                  date2
                                  date3
                                  date4
                                  dateS
1
2
1
1
2
2
3
1
3
3
2
1
1
2
1
1
4
3
2
1
3
2
2
1
2
2
1
1
2
1
1
1
1
2
3
4
2
2
3
3
3
2
2
1
4
3
1
1
1
2
1
1
2
2
1
1
3
2
1
1
3
2
1
1
Step 3.  Determine
compliance status of each
cell.

date 1
1
1
1
0
1
1
1
1
1
1
1
1
1
0
0
0
                                date2
                                date3
                                date4
                                dateS
0
0
0
0
0
0
1
0
1
1
0
0
0
0
0
0
1
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
1
1
1
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-12
         Step 4: Percent compliance by date.
sample date
date 1
date 2
date3
date 4
dateS
percent
space
75.00%
18.75%
18.75%
43.75%
12.50%
         Step 5. Rank the percent of space values
         and assign percent of time = (100*R/(M+1.0)),
         where R is rank and M is total number of dates.
sample date
date 1
date 4
date 2
date3
date5
ranked
percent
space
75.00%
43.75%
18.75%
18.75%
12.50%
cumulative
percent time
16.67
33.33
50.00
66.67
83.33
                                             Steps 6 and 7: The plot of the CFD
                                             and the comparison to the reference
                                             curve are shown in Figure 2.2. For
                                             this hypothetical case the assessment
                                             area would be judged in noncompli-
                                             ance. For a percent area of 18.75, the
                                             allowable frequency on the reference
                                             curve is about 53%. That is, 18.75%
                                             of the segment area should not be out
                                             of compliance more that 53% of the
                                             time. For date 3, the estimated
                                             frequency of 18.75% noncompliance
                                             is 66.67%. Thus the frequency of
                                             18.75% of space out of compliance is
                                             in excess of the 53% allowed. The
                                             reference curve is exceeded for dates
                                             4 and 1 as well. Note:  in this cumula-
                                             tive distribution framework, the actual
                                             date is not relevant. One should not
                                             infer that noncompliance occurred on
                                             that date if the data point associated
                                             with a date falls above the reference.
                                             Date is being used here as a label for
                                             each coordinate pair.
                £
               I
               •s
                  1.0


                  0.9


                  (LS


                  0.7
,S  0.6
                  0.5-

                  0,1

                  OJ

                  0,2

                  0.1

                  0.0-1,
                   0,0     0,1     02     0,3     0,4     0,5    0,6
                                            Fraction of Space
                                                              0,7
                                                                    0.8
                                                                          0,9
                                                                                to
              2,2, Graphical representation of CFD from the above example (' + ') with hypothetical reference
        curve ( smooth).

-------
                                                                                             A-13
              CFD

As defined above, the CFD is a data driven formulation. But the data used to formu-
late the CFD are a sample of points taken from a population. Defining the CFD
becomes complex when one considers the many different levels for which it might
be defined. At one level, the  CFD might be defined based on the true state of a
segment. Imagine that the state of a segment could be frozen for sufficient time to
permit deployment of an analog  sampler (that is one that measures water quality
continuously rather than in discrete samples) to assess the percent of area out of
compliance at that instant. Now stretch that imagination one step further to relax the
condition that the segment be frozen and allow that these analog measurements of
percent of area out of compliance be determined continuously in time. With this
information, a determination of the CFD for the true state of the segment is possible.
While  the information needed to construct the ideal CFD is not obtainable, it is
important to ask how  well the CFD based on obtainable data represents this ideal
(see also Section 5). Is a data driven CFD consistent for the ideal CFD in the statis-
tical sense? Loosely speaking, consistency implies that the data driven CFD should
get closer to the ideal CFD as more data are used. Is the data driven CFD unbiased
for the ideal CFD? Unbiasedness  implies that even with small amounts of data, the
data driven CFD on average covers the ideal CFD.

One might argue that if both the  assessment CFD and the reference CFD are data
driven, then  it is not important for the CFD to approximate the ideal. Even so, it is
important to understand the behavior of the CFD as a function of samples  size and
the relative temporal and spatial  contributions to the variance in the water quality
parameter. If the curve changes shape as a more data are used, this could  result in
unfair comparisons between assessment and reference regions. In Section 4, statis-
tical properties for both types of reference curves are evaluated further.



Two  approaches to defining the  reference curve are  being considered. One is a
biologically  based definition. The idea is to  identify appropriate reference regions
with healthy biological indicators  and compute the reference CFD for these regions.
For example, healthy  benthic IBI scores might be used as indicators of adequate
bottom dissolved oxygen. Thus after stratifying by salinity zone and perhaps other
factors, a series of dissolved oxygen reference CDF curves could be computed from
the existing  20+ year  monitoring data base.  When it is not possible to establish a
reference condition some more arbitrary device must be employed. Alternatives are
discussed in Section 4.0.

            of

Step 1 - Data Collection. One of the advantages of the CFD approach is that it will
accommodate a variety of input data and still arrive at the same assessment endpoint.
Data collection methods currently in place include: fix station data, cruise track data,
continuous monitor data, aircraft flight path data, and satellite imagery data.  Because
of the  interpolation step, all of these data can be used (and potentially combined)
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-14
                      with varying degrees of success to estimate the total spatial (to the limit of interpo-
                      lator pixel size) distribution of a water quality constituent. As noted above, one could
                      construct this process by reversing the roles of time and space. That is, first interpo-
                      late over time  and then build a cumulative distribution in space.  In theory it is an
                      abitrary choice to first standardize the data over space by interpolation and then
                      construct the cumulative distribution in time.  However, in practice, there is a greater
                      diversity of sampling designs  over space and therefore it is the sampling in the
                      spatial dimension more than the temporal that creates many types of data that must
                      be forced to a common currency.

                      Step 2 -  Interpolation. Interpolation is the step that puts data collected at various
                      spatial  intensities on a common footing.  On the one  hand,  this is advantageous
                      because data collected at many spatial intensities are  available for the  assessment
                      process. On the other hand, it can be misleading to accept interpolated surfaces from
                      different data  sources as equivalent without qualifying each interpolation with a
                      measure of the estimation error that is associated with each type of data. Clearly an
                      interpolation based on hundreds of points per segment (such as cruise track data) will
                      more accurately reflect the true noncompliance percent when compared to an inter-
                      polation based on two or three points per segment (such a fixed station data). Of the
                      various types of interpolation  algorithms available, the method proposed for this
                      assessment is kriging. Kriging  offers the best available  approach for the estimation
                      error associated with interpolation.

                      Step 3 - Pointwise Compliance. Determining the percent of compliance of each cell
                      from each interpolation would  seem to be a simple step. If the estimated value for a
                      cell exceeds the criterion then that cell is out of compliance.

                      While interpolation  allows for a standardization of many types of data, pointwise
                      compliance allows  for standardization of many  criteria.  Because  compliance is
                      determined at points in time  and space, it is possible to vary the compliance criteria
                      in time and space. If different levels of a water quality constituent are acceptable in
                      different seasons, then the criterion can vary by season. It is possible to implement
                      different criteria over space for a segment that bridges  oligohaline and mesohaline
                      salinity regimes.  It would even be possible to let the criterion be a continuous func-
                      tion of some ancillary variable  such as temperature or salinity.  All that is required is
                      that the final determination be yes or no for each interpolator cell.

                      Even the simplicity of this  concept becomes diminished when issues of interpolation
                      error are considered. Consider the assessment of two interpolator cells from an inter-
                      polation based on cruise track  data. One cell near the cruise track has an estimated
                      value is 4 and a standard error  of 0.1. A second cell far from the cruise track has an
                      estimated value of 4 and a standard error of 1.0. If the criterion were 3.0, it is fairly
                      certain that the first cell represents exceedance. It is much less certain that the second
                      cell represents exceedance. In the simple assessment of non-compliance, they count
                      the same.

                      Step 4 -  Percent Non-compliance in Space. Computing a percentage should also
                      be a simple step. The estimate is simply 100 times the number of cells out of compli-
                      ance divided by the total  number of cells. As a rule, the uncertainty of a binary
                      process can be modeled using a binomial distribution. However, the issue of uncer-

-------
                                                                                              A-15
tainty described for step 3 propagates into computing the percent of compliance for
a segment. Add to that the fact that estimated values for interpolator cells have a
complex dependence structure which rules out a simple binomial model and the rules
governing the uncertainty of this step are also complex. The number of interpolator
cells, N, is relatively constant and under an independent binomial model the variance
of the proportion of cells not in compliance, p, would be (p)(l-p)/N. Intuitively, one
expects the variance of p to decrease as the number of data points that feeds the inter-
polation increases. This expectation has been confirmed by simulation, but the
mathematical tools for modeling this propagation of error are yet to be developed.

Step 5 - Percent of Time.  While the percent of space coordinate of the CFD has
simple interpretation of the percent of the segment out of compliance on a given
date, the percent of time coordinate is not simply  the percent of time out of com-
pliance at a given point.  Instead the percent of time coordinate has an interpretation
similar to that of a cumulative distribution function. The percent of time coordinate
is the  percent of  time  that the associated spatial percent of  noncompliance is
exceeded.  For example,  if the (percent space, percent time) coordinates for  a point
on the CFD are (90,10),  one would say that the spatial percent of noncompliance is
greater than or equal to 90% about 10% of the time.

This step is very similar  to computing an empirical  distribution function which is an
estimator of a cumulative distribution function. Because of this similarity, one imme-
diately thinks of statistical inference tools associated with empirical distribution
functions,  such as  the Kolmogorov-Smirnov,  Shapiro-Wilk, Anderson-Darling,  or
Cramer-von Mises, as candidates for inference about the CFD.  These procedures
model uncertainty  as  a  function of sample size only; in this case the number of
sample dates. The fact that it does not incorporate the uncertainty  discussed the
previous steps seems unsatisfactory.

A quick review of probability plotting will reveal several methods on estimating the
percent of time coordinate in step 5. Formulae found in the literature include: (R/N),
(R - 0.5) / (N - 1). and (R - 0.375) / (N + 0.5), where R is rank and N is sample size.
These generally fall in to a family of given by (R - A)/(N - 2A + 1) for various values
of A. They are approximately equal, but the choice should be fixed for a rule.

Step 6 - Plotting the  CFD. Even the plotting of the points is subject to variation,
although these  variations  are  somewhat  minor compared to the larger issue  of
assessing the uncertainty of the assessment curve.  The simple approach used in the
figures above is to connect the points by line segments. In the statistical literature, it
is more common to use  a step function. If the graph represents an empirical distri-
bution function, each horizontal line segment is closed on the left  and open on the
right. Because the CFD is an inversion of an EDF it would be appropriate for these
line segments to be closed on the right and open on the left.

Step 7 - Comparing  the Curves.  It is at the point of comparing the assessment
curve to the reference curve that the issue of uncertainty becomes most important.
From the preceding discussion it is clear that uncertainty in the assessment curve is
an accumulation of uncertainty generated in and propagated through the preceding 6
steps. If the reference curve is biologically based, it  is derived under the same system
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-16
                     of error propagation.  Developing the  statistical algorithms to quantify this uncer-
                     tainty is challenging.

                     Even if the uncertainty can be properly quantified, the issue of who gets the benefit
                     of doubt due to this uncertainty is a difficult question to resolve. This is  a broad
                     sweeping issue regarding uncertainty in the regulatory process, not a problem
                     specific to the CFD approach. None-the-less, it must be dealt with here as well as
                     elsewhere. One option is to require that the assessment curve be significantly above
                     the reference curve to establish noncompliance. This option protects the regulated
                     party from being deemed out of compliance due to random effects, but if assessment
                     CFD curves are not accurately determined, it could lead to poor protection  of envi-
                     ronmental health  and designated uses.  A  second  option is to require that  the
                     assessment curve be significantly below the reference curve to establish compliance.
                     This results in strong protection of the  environmental resource, but could lead to the
                     regulated party implementing expensive management actions that  are not necessary.
                     Some compromise between these extremes is needed. The simplest compromise is
                     to ignore  variability and just compare  the assessment curve to the reference curve.
                     As long as unbiased estimation is implemented for both the assessment curve and the
                     reference curve, this third option will result in roughly equal numbers of false posi-
                     tive  (declaring  noncompliance when in fact  compliance exists) and false negative
                     (declaring compliance when in  fact noncompliance exists) results. This offers a
                     balanced approach, but there is no mechanism to motivate a reduction of these false
                     positive and false negative errors.
                            ,/.
                                       r
                     The Chesapeake Bay monitoring program routinely monitors 19 directly measured
                     water quality paramenters at 49 stations in the mainstem Bay and 96 stations in the
                     tidal tributaries. The Water Quality Monitoring Program began in June 1984 with
                     stations sampled once each month during the colder late fall and winter months and
                     twice each month in the warmer months. A refinement in 1995 reduced the number
                     of mainstem monitoring cruises to 14 per year. "Special" cruises may be added to
                     record unique weather events. The collecting organizations coordinate the sampling
                     times of their respective stations, so that data for each sampling event, or "cruise",
                     represents a synoptic picture of the Bay at that point in time. At each station, a hydro-
                     graphic profile is made (including water temperature, salinity, and dissolved oxygen)
                     at approximately 1 to 2  meter intervals. Water  samples for chemical analysis (e.g.,
                     nutrients and chlorophyll) are collected at the surface and bottom, and at two addi-
                     tional depths depending on the existence and location of a pycnocline (region(s) of
                     density discontinuity in the water column). Correlative data on sea state and climate
                     are also collected.

                     In addition, Chesapeake Bay Program partner organizations Maryland Department
                     of Natural Resources and the Virginia Institute of Marine Science have recently
                     begun monitoring using  a technology known as data flow. DATAFLOW is a system
                     of shipboard water quality probes that measure spatial position, water depth, water
   appendix a  « The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                              A-17
                      Baltimore wru
                   Patapsco R. «*-'•
              in
    Rappahannock  R.
                  TFX3
Figure 2.3. Map of the tidal water quality monitoring stations.
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-18
                     temperature, salinity, dissolved oxygen, turbidity (clarity of the water), and chloro-
                     phyll  (indicator of plankton concentrations) from a flow-through stream of water
                     collected near the  water body's surface. This system allows data to be collected
                     rapidly (approximately every 4 seconds) and while the boat is traveling at speeds up
                     to 20 knots.

                     In 2005, the MDDNR Water Quality Mapping Program covered 16 Chesapeake Bay,
                     Coastal Bay and Tributary systems. The St. Mary's, Patuxent, West, Rhode, South,
                     Middle, Bush,  Gunpowder, Chester, Eastern Bay,  Miles/Wye,  Little Choptank,
                     Chicamacomico and Transquaking Rivers will be mapped, as well as Fishing Bay
                     and the Maryland  Coastal Bays. In Virginia, dataflow data are available  for the
                     Piankatank, York, Pamunkey and Mataponi Rivers.

                     Beginning in 1990, Chlorophyll-a concentrations were measured over the mainstem
                     Chesapeake using aircraft remote sensing. From 1990-1995, the instrument used for
                     this study was the  Ocean Data Acquisition System (ODAS)  which had three
                     radiometers measuring water leaving radiance at 460, 490 and 520 nm. In 1996, an
                     additional instrument was added, the SeaWiFS Aircraft Simulator (SAS II). SAS II
                     has sensors at seen wavebands which improves detection of Chlorophyll in highly
                     turbid areas. Since 1990, 25-30 flights per year have been made during the most
                     productive times of year.

                     The  data  described above and  additional  information can  be obtained from:
                     www.chesapekebay.netmddnr.chesapeakebay.net/eyesonthebay/index.cfm
                     www2.vims.edu/vecos/

                                  of

                     The current Chesapeake Bay Interpolator is a cell-based interpolator. Water quality
                     predictions for each cell location are computed by averaging the nearest "n" neigh-
                     boring water quality measurements, where "n" is  normally 4, but this number is
                     adjustable. Each neighbor included in the average is weighted by  the inverse of the
                     square of Euclidean distance to the prediction cell  (IDW). Cell size in the Chesa-
                     peake Bay was  chosen to be 1km (east- west) x 1km (north-south) x 1m (vertical),
                     with columns of cells extending from surface to the bottom of the water column, thus
                     representing the 3-dimensional volume as a  group of equal  sized cells extending
                     throughout the volume.  The  tributaries  are represented by various  sized cells
                     depending on the geometry of the tributary, since the narrow upstream portions of
                     the rivers  require smaller cells to accurately model the river's  dimensions. This
                     configuration results in a total of 51,839 cells by depth for the mainstem Chesapeake
                     Bay (segments  CB1TF-CB8PH),  and a total  of 238,669  cells by depth for all 77
                     segments which comprise the mainstem Bay and tidal tributaries.

                     The Chesapeake Bay Interpolator is unique  in  the way it computes  values in 3
                     dimensions.  The interpolator code is optimized to compute  concentration values,
                     which closely reflect the physics of stratified water bodies, such as Chesapeake Bay.
                     The Bay is very shallow compared to its width or length; hence water quality varies
                     much more vertically than horizontally.  The Chesapeake Bay Interpolator uses a
                     vertical filter to select the vertical range of data that are used in each calculation. For
   appendix a *  The Cumulative hrequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                            A-19
instance, to compute a model cell value at 5m deep, monitoring data at 5m deep are
preferred.  If fewer than n (typically 4) monitoring data values are found at the
preferred depth, the depth window is widened to search up to d (normally +/-2m)
meters above and below the preferred depth, with the window being widened in 0.5m
increments until  n monitoring  values have been found for the computation. The
smallest acceptable n value is  selectable by the user. If fewer than n values are
located, a missing value (normally a -9) is calculated for that cell. A second search
radius filter is implemented to limit the horizontal distance of monitoring data from
the cell being computed. Data points outside the radius selected  by  the  user
(normally  25,000m) are excluded from calculation. This filter is included so that
only data that are near the location being interpolated are used.

In this version of the Interpolator, Segment and Region filters have  been added.
Segments are geographic limits for the interpolator model. For instance, the Main
Bay is composed of 8 segments (CB1TF, CB2OH,  ...,CB8PH). The tributaries are
composed of 77 additional segments,  using the CBP  2003 segmentation. These
segments divide the Bay into geographic areas that have somewhat homogeneous
environmental conditions. This segmentation also provides a means for reporting
results on a segment basis, which can show more localized changes compared to the
whole Bay ecosystem.

Segment and bathymetry information use by the interpolator is stored in auxiliary
files. Segment information allows the  interpolator  to report results on a segment
basis  which  can show more  localized changes  compared to  the  whole  Bay
ecosystem. These segment and bathymetry files have been created for the main bay
and all of the larger tributaries. The CBP segmentation scheme was replicated in
these files by partitioning and coding  the interpolator  cells that fall  within  each
segment.

The interpolator  also identifies the  geographic boundary that limits which moni-
toring station data are included in interpolation for a given segment through a region
file. Use of data regions ensures that the interpolator does not "reach across land" to
obtain data from an adjacent river which would give erroneous results. By using data
regions, each segment of cells can be computed from their individual subset of moni-
toring data. Each adjacent data region should overlap by some amount so that there
is a continuous gradient, and not a seam, across segment boundaries.

                           of

The Chesapeake Bay Program has initiated implementation of the CFD  as an assess-
ment tool. The Criteria Assessment Protocols (CAP) workgroup was formed in the
fall of 2005 to develop detailed procedures  for implementing criteria assessment.
This workgroup  has developed and implemented procedures that use the CFD
process and conducted a CFD evaluation of dissolved oxygen for many designated
assessment units.

The CFD methodology was first applied in the  Chesapeake Bay for the most recent
listing cycle which was completed in the Spring of 2006 and was based on data
collected over the period 2002 through 2004.  CFDs were developed and utilized
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-20
                     primarily for the  dissolved oxygen (DO) open-  and deep-water monthly mean
                     criteria because there were insufficient data collected to assess the higher-frequency
                     DO criteria components. The clarity criteria were  not assessed based on the CFD
                     because there were few systems in which there was sufficient data for an assessment.
                     Chlorophyll criteria were not available from the Chlorophyll criteria team in time to
                     implement a chlorophyll assessment.

                     In general, the CFD analysis indicated that most of the Bay waters failed one or more
                     of the open-water or deep-water DO criteria components. However, there were also
                     many tributaries in which  all of the DO criteria assessed indicated attainment. Thus
                     in  this  initial  application, the CFD method  did  appear to distinguish between
                     impaired and unimpaired systems in a manner that is consistent with the expectations
                     of the many stakeholders in the CAP workgroup.

                     In the 2006 application of  the assessment methodology, there were many details that
                     required resolution in order to fully implement the methodology. Procedures gener-
                     ally followed the theoretical description as described in Section 2.1, but some details
                     were  modified to address  unforeseen complications. The following describes some
                     of those details.

                     In  general, data were obtained from the  CBP  CIMS  data base  and parameters
                     included date, location, depth, salinity, temperature and the water quality parameter
                     being assessed. Some State data were also incorporated and those data were obtained
                     directly from the relevant State.  Once all the data were compiled, they were assigned
                     to a time period based on the sample date. Fixed-station data are normally collected
                     during a monitoring cruise that covers the entire tidal Chesapeake Bay over several
                     days. However, in order to provide a "snapshot" in water quality, the data collected
                     within a cruise are assumed to be contemporaneous in order to perform  a single
                     spatial interpolation. For any data not associated with a cruise, a cruise number is
                     assigned representing the closest cruise in time to the collection of each datum. Co-
                     located data points in the same cruise were  averaged.

                     The assessment procedure requires assessment over large areas rather than at points
                     in space. Spatial interpolation using the CBP IDW interpolator was performed for
                     each water-quality criteria parameter for each cruise. Clarity and surface chlorophyll
                     were  interpolated in the two horizontal dimensions using inverse distance squared
                     weighting. Dissolved oxygen was first linearly interpolated in the vertical dimension
                     within each column of data beginning at 0.5 meters and continuing at one meter
                     intervals, not to exceed the deepest observation in that column. Each depth was then
                     interpolated horizontally  using inverse  distance squared weighting.  Data  regions
                     were  specified for  each segment in order to  prevent the interpolation algorithm from
                     using data points in neighboring tributaries.

                     Designated uses in the Chesapeake Bay are defined vertically in order separate stable
                     water layers that have differing criteria levels for dissolved oxygen. The  surface layer
                     (open water) is that layer defined to be above the pycnocline and thus exposed to the
                     atmosphere. The middle layer (deep water) is defined to be  the layer between the
                     upper and lower pycnocline. And the lower  layer (deep channel) is defined to be the
                     layer below the pycnocline. Given that the pycnocline is dynamic and moves up and

-------
                                                                                             A-21
down with each monitoring cruise, the designated use of each grid cell must also be
defined based on the available data for each cruise.

The pycnocline is defined by the water density gradient over depth. Temperature and
salinity are used to calculate density, which in turn is used to calculate pycnocline
boundaries. Density is calculated using the method described in:  Algorithms for
Computation   of  Fundamental  Properties   of  Seawater   (Endorsed   by
UNESCO/SCOR/ ICES/IAPSO Joint Panel on Oceanographic Tables and Standards
and SCOR Working Group 51. Fofonoff, N P; Millard, R C Jr. UNESCO technical
papers in marine science. Paris , no. 44, pp. 53.  1983). For each column of temper-
ature and salinity data, the  existence of the upper and lower pycnocline boundary is
determined by looking for the shallowest robust vertical change in density of 0.1
kg/m3/m for the upper boundary and deepest change of 0.2 kg/m3/m for the lower
boundary. To be considered robust, the density gradient must not reverse direction at
the next measurement and must be accompanied by a change in salinity, not just
temperature.

The depths to the upper pycnocline boundary, where detected, and the fraction of the
water column below the lower boundary are interpolated in two dimensions. If no
lower boundary was detected the fraction was considered to be zero. The depth to the
upper pycnocline boundary tends to be stable across horizontal space and so spatial
definition of that boundary using interpolation generally worked  well. However,
interpolation of the lower  boundary is more complicated because the results can
conflict with the upper boundary definition or with the actual bathymetry of the Bay.
As a result, interpolation of the lower boundary was performed based on "fraction of
water column depth". In that way, the constraints of the upper pycnocline boundary
definition and the actual depth were imposed and errors related to boundary  conflicts
were eliminated.

Assessments were  performed based on criteria  specific averaging periods. The
instantaneous assessment for deep channel dissolved oxygen was evaluated using the
individual cruise interpolations.  All monthly assessments were based  on  monthly
averages of interpolated data sets. To calculate the monthly averages, each interpo-
lated cruise within a month was averaged on a point-by-point basis. Generally, there
were 2 cruises per month in the warmer months and 1 cruise per month in the cooler
months. Spatial violation rates are calculated for each temporally aggregated inter-
polation in an assessment period. For example, for a three-year summer open-water
dissolved oxygen assessment, the twelve monthly  average interpolations repre-
senting the four summer months over three years were used.
The CFD approach uses the proportion of space in attainment in any given month
estimated using an approach based on a statistical model. The current method uses
data collected in a specific month at a set of sampling locations within the segment
of interest to estimate the parameters of the model. The estimated model is then used
to interpolate likely values at unsampled locations, specifically at a set of prediction
locations arranged in a grid over the segment. The predictions thus obtained are used
   appendix a  «  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-22
                      to calculate the proportion of space in compliance that month. The current estima-
                      tion procedure for obtaining predicted values is Inverse Distance Weighting (IDW),
                      a non-statistical  spatial interpolator that  uses  the  observed data to calculate  a
                      weighted average as a predicted value for each location on the prediction grid. The
                      method calculates the weight associated with a given observation as the inverse of
                      the square of the distance between the prediction location and the observation.

                      The panel considered several interpolation methods in addition to  IDW. Of these,
                      kriging methods emerged as a principal alternative approach for populating the grid
                      of prediction  locations.  Non-parametric  methods  were  also considered.  These
                      include Loess regression or cubic spline methods. These approaches could be advan-
                      tageous in that they are statistical methods that provide levels of error, but panel
                      analyses and deliberations have been insufficient to provide definitive statements on
                      this class of methods. Table 3.2 which appears in Section 3.3 summarizes our deter-
                      minations.

                          '

                      Kriging is a spatial interpolation technique that arose out of the field of geostatistics,
                      a subfield of statistics that deals with the analysis of spatial data. Kriging and the
                      field of geostatistics has been employed in a wide variety of environmental applica-
                      tions and is  generally accepted  as a method for performing statistically optimal
                      spatial interpolations (Cressie 1991, Schabenberger  and Gotway  2004, Diggle  and
                      Ribeiro 2006). Applications  of kriging in water related research can be  found in
                      (Kitanidis 1997,  Wang and Liu 2005,Ouyang et al. 2006). References on kriging
                      methodology, geostatistics, and their related statistical development  can be found in
                      (Cressie  1991, Diggle et al. 1998, Schabenberger  and Gotway 2004, Diggle  and
                      Ribeiro 2006).

                      Kriging can equivalently be formulated in terms of a general linear regression model

                                      Y (s) = &, + fa Xj(s)  • • • + /3p Xp(s) + e(s)                   (1)
                      with s representing a generic spatial location vector (usually 2-D) assumed to vary
                      continuously over some domain of interest, Y(s) the outcome of interest measured at
                      s, Xj(s), . .  . ,Xp(s) potential covariates indexed by location s, and their associated
                      regression effects /?1; . . .  , /?p. Note that covariates must be known at every predic-
                      tion location. The elements of the spatial vector s can be used as covariates for
                      modeling spatial trends. On the other hand water quality measures such as salinity
                      which may have a strong association with the outcome of interest, is of limited value
                      as a covariate because it is not known at all prediction locations. The uncertainty in
                      this regression relationship is modeled with the random error term e(s) assumed to
                      have zero mean  and constant variance. Spatial data like the type sampled in the
                      Chesapeake Bay water-quality  criteria assessments often exhibit a property known
                      as (positive) spatial dependence, observations closer together are more similar than
                      those further away. This property is accounted for in model (1) by allowing e(s) to
                      have a spatial correlation  structure.

                      Some further specifics on  e(s) are warranted. Common distributional assumptions on
                          include normality or log-normality, although kriging can be performed based on
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                                A-23
other statistical distributions and data transformations (Christenson et al. 2001). The
spatial correlation in e(s) is represented by positive definite functions. These func-
tions can be assumed isotropic where correlation decay depends just on distance, or
anisotropic where correlation decay depends on distance and direction. Variograms
are another special type of mathematical function closely related to spatial correla-
tion functions that can and are more often used to represent spatial correlation. For
purposes here and in many kriging applications, variograms and spatial correlation
functions provide equivalent representations of spatial structure. For consistency in
what follows only the term variogram will be used in discussions of spatial structure.

While  there  is considerable flexibility in implementing the error structure of a
kriging model, it is possible to generalize somewhat  with respect to the error struc-
ture of Chesapeake Bay water quality data. Of the three water quality parameters
being  assessed, chlorophyll and clarity measures tend to  follow the log-normal
distribution and dissolved oxygen is  reasonably approximated by the normal distri-
bution. The  horizontal decay  rate  of  spatial  correlation does  not tend  to  be
directionally dependent. Thus if the bay is  viewed as a composite  of  horizontal
layers, isotropic variograms are appropriate for kriging each layer. In a vertical direc-
tion, water quality can change rapidly and thus spatial correlation can decay over a
short distance. A 3-D interpolation  procedure would benefit from use of  an
anisotropic variogram in order to differentiate the vertical correlation decay from the
horizontal correlation decay.

Note, in the literature model (1) is referred to as a universal kriging model. When
covariates (the X's) are not considered to influence interpolation of Y the right hand
side of model (1) contains just the constant term /?0 and e(s). The resulting model is
referred to as the  ordinary kriging model. When the spatial structure (variogram) for
model  (1) is known, statistically optimal predictions for the variable Y at unsampled
locations (outside of estimation of possible regression effects) can be derived using
standard statistical principles. The optimality criteria results in spatial predictions
that are linear in the data, statistically unbiased, and minimize mean squared predic-
tion error,  hence referred  to  as best  linear unbiased  predictions (BLUPs). The
minimized mean squared prediction error is  also taken as a measure of prediction
uncertainty.  In practice, however, spatial structure of the data is unknown, the esti-
mation of which via the variogram function is cornerstone to kriging applications.

To demonstrate let {Xsi)> • • • > Xsn)l represent a set of spatial data, for  example a
water-quality parameter such as dissolved oxygen sampled at a set of n spatial loca-
tions s1; . . .  , sn. Assume this data to  be a realization  of the ordinary kriging version
of model (1). The first step in kriging is variogram estimation. There are  several
methods available, method of moments and statistical likelihood based being two of
the more common, all of which though are based on the sample data {Xsi)> • • •  .
Xsn)l- Without going into detail, this process ends with a chosen variogram function
and its parameter estimation, describing the shape and strength (rate of decay) of
spatial correlation. There is also a determination, again  based on the sampled data,
of whether the spatial structure is isotropic or anisotropic. The estimated variogram
is then assumed known and kriged interpolations and their interpolated uncertainty
are computationally straight forward to generate at numerous locations where data
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-24
                     were not observed. Accounting for uncertainty in variogram parameter estimation
                     has commonly been explored using Bayesian methods (Diggle and Ribeiro 2006).

                     '; :•
                     The inverse distance weighting method that is currently used in the CFD approach
                     has already been described. Hence, this section provides a short review of IDW's
                     technical details and a comparison of IDW to alternative interpolation methods.
                     The IDW method is essentially a deterministic, non-statistical approach to interpo-
                     lating a two or three dimensional space. As a result it lacks  statistical rigor so that
                     estimates of the prediction errors are not calculable without additional assumptions.
                     Similar to kriging, IDW predicts a value () at an unobserved site, say at location %
                     using a weighted average of the N nearest observed neighbors (N specified by the
                     modeler):
                     where the weights, wt, are inversely related to the distance between locations s0 and
                     d(s0,Si) is the Euclidean distance between locations s0 and Sj, and the denominator of
                     the weight is to ensure that the weights sum to 1 . The IDW is an exact interpolator
                     in that the predicted values for observed locations are the observed values and the
                     maximum and minimum values  of  the interpolated  surface  can  occur only at
                     observed sites.

                     Recent research has compared IDW to other interpolation techniques, most notably
                     variations in kriging (Table 3.1). The  authors found that in some cases kriging was
                     at least as good an interpolator as IDW and in some instances better. The non-para-
                     metric techniques (splines and  similar methods) were not as precise as kriging and
                     IDW. The method used for comparison  in virtually all of the  research was some
                     variant of cross-validation, a method where some data are kept aside and not used in
                     the model estimation phase and then using the resulting model to predict values for
                     the data kept aside. The predicted and observed values are then compared and a
                     statistic is calculated that summarizes  the differences between the two sets of values
                     (observed and predicted).

                     None of these studies used datasets with highly irregular edges such as are found in
                     the Chesapeake Bay nor did they use any distance metric other than Euclidean
                     distance. Whether one method is preferable to another in these more difficult situa-
                     tions remains unexplored.
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                              A-25
     3.1.  A short list of recent articles comparing the precision of IDW to a subset of
          other possible interpolation methods.
Authors
Kravchenko (2003)
Dille, et al. (2002)
Valley, et al. (2005)
Lloyd (2005)
Reinstorf, et al.
(2005)
Zimmerman, et al.
(1999)
Methods Compared
Inverse Distance
Weighting (IDW),
Ordinary Kriging
(OK)
IDW, OK, Minimum
Surface Curvature
(MC), Multiquadric
Radial Basis Function
(MUL)
IDW, OK, Non-
parametric Detrend +
Splines
moving window
Regression (MWR),
IDW, OK, simple
kriging with locally
varying mean (SKlm),
kriging with external
drift (KED)
IDW, OK, KED +
deterministic
chemical transport
models
2 types of IDW, UK,
OK
Variables
Manipulated
spatial structure and
sample grid spacing
neighborhood size,
spatial structure,
power coefficient in
IDW, sample grid
spacing, quadrat size
spatial structure,
sample size, quadrat
size
spatial structure,
sample size
single dataset was
analyzed
spatial structure,
sampling pattern,
population variance
Conclusions
IDW better than OK
unless sample sizes
were fairly large
No interpolator
appears to be more
precise than another.
Sample grid spacing
and quadrat size were
deemed more
important.
OK tended to be more
precise but IDW was
very similar
KED and OK best
OK best
UK and OK better
than IDW
One final and important issue with IDW is that, as currently used, IDW is a deter-
ministic method which makes no assumptions as to the probability distribution of the
data being interpolated. Hence, it does not allow for estimating prediction errors, i.e.
it does not allow for the possibility of random variation at interpolation sites.  A
simple question is whether IDW can be recast in the kriging framework given the
similarity in prediction method (weighted average) and hence can a method be found
to estimate prediction errors? The short answer is no - the distance function used by
IDW, which  is an implicit assumption about the autocorrelation function in the
spatial field, does not meet the assumptions required for development of a valid vari-
ance-covariance matrix describing the spatial covariance. As a result, IDW cannot be
modified to take advantage of the statistical knowledge that has been developed for
geostatistical  analyses such as kriging. This does not imply that other approaches  to
estimating prediction error are  also not possible.

A non-parametric approach for estimating variance was proposed (Tomczak, 1998)
in which jack-knifing was used to provide error estimates. 95% confidence intervals
for the mean  were calculated and then compared to the actual observed values. Not
surprisingly, only 65% of the data were captured within their associated confidence
interval. The  method appears to have been misapplied—the jackknifing method  as
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-26
                     used estimates the standard error of the mean assuming independent observations.
                     As a result, the confidence interval is not capturing the effect of the spatial depend-
                     encies nor is it based on the fact that we are predicting a value for the unobserved
                     site rather than estimating a mean. The development described by Tomczak (1998)
                     should be explored further and other alternatives such as block  bootstrapping for
                     variance estimation as well.

                     33               ..in,.-ic
                     There are many variations on spatial interpolation in addition to  kriging and IDW.
                     See Cressie (1989) for a review. The committee did not have  sufficient  time to
                     compare all models, but CBP in encouraged to continue this research. One promising
                     category of models are for interpolation based on non-parametric methods that do
                     not rely on measuring and accounting for spatial autocorrelation. All of the non-para-
                     metric approaches  would be  based on the assumption that the autocorrelation
                     observed in the data is due to unobserved  explanatory variables and hence  alterna-
                     tive modeling approaches are not unreasonable. The particular set we mention are
                     the regression type analyses with the locational indices (northings, eastings) used as
                     explanatory  variables.  Examples include generalized additive models  (Hastie and
                     Tibshirani, 1990), high-order  polynomials (Kutner, Nachtsheim, Neter,  and Li,
                     2004), splines (Wahba, 1990), and locally weighted regression ("loess" or "lowess",
                     Cleveland and Devlin,  1988). In some kriging and IDW methods, large-scale trend
                     is modeled relatively smoothly using locational indices and local smaller-scale vari-
                     ation is modeled using the estimated autocorrelation in conjunction with the values
                     of the variable at nearby observed sites. The nonparametric methods replace estima-
                     tion  of the  local  variation  based on correlation  functions  with models of the
                     large-scale trend that are less smooth and more responsive to the spatial variation in
                     the observed data. A visual demonstration is given in Figure 3.1 which shows a  one-
                     dimensional dataset with Y as the variable to be predicted and X as  the location along
                     the one dimensional axis. For example, X could be distance from the mouth of a river
                     and Y could be chlorophyll a concentration.

                     One advantage of these approaches is that  each of the methods has extensive statis-
                     tical research into estimation of model parameters as well as standard errors for those
                     parameters and for  predictions at interpolation sites. Another  is  that the main
                     modeling decisions are related to bandwidth selection or degree order of polynomial
                     to fit. These  decisions can be  automated by developing rules for roughness of fit
                     based on reduction in MSE as compared to modeling a straight line (in X). Disad-
                     vantages are the same as for kriging, all model estimation is data dependent which
                     means that the spatial configuration and number of sampling sites  has a direct influ-
                     ence on the predictions and their error estimates.  In addition, a study done by Laslett
                     (1994) comparing kriging and  splines indicated that the two methods are similar in
                     predictive power but for certain sampling regimes  kriging  performs  better.  We
                     recommend more  study since the  non-parametric approaches would be easier to
                     implement than kriging.
   appendix a  * The Cumulative frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                                 A-27
figure 3.1, Bivariate fit of Y By X. Straight line is a linear large-scale trend fit (R2 = 0.19);
the moderately wavy line around the straight line is a 6th-order polynomial fit (X enters
the model as X, X2, X3, ..., and Xs; R2 = 0.25); and the jagged line is a spline fit with a
very small bandwidth (neighborhood used in local estimation at each X; R2 = 0.90).
3,3                  OF

The following describes some of the benefits and potential limitations of kriging in
regards to CBP application with some comparisons to the IDW approach towards
spatial interpolation outlined in the previous section. Nonparametric methods are not
sufficiently developed to include in this comparison. A primary benefit of the kriging
methodology compared to IDW is that it is a statistical technique. As such the field
of statistics (including kriging) is designed to make inference from sampled data in
the presence of uncertainty and the  quantity and quality of the sample  data are
reflected in those inferences. However, kriging is  a less than routine type of statis-
tical analysis and requires  a certain  level of statistical expertise to carry out the
process. The short description on variogram estimation provided above merely intro-
duces  this involved and  often  complicated step. This requirement for informed
decision making limits the degree to which kriging can be automated and still main-
tain its flexibility and optimal properties.

Further issues regarding kriging and CBP applications are listed below.
   •  Kriging is flexible in that it is based on an estimate of the  strength of spatial
     dependence in the data (variogram). Kriging  can consider direction dependent
     weighted  interpolations  (anisotropy)  and can include covariates (universal
     kriging) to potentially influence interpolations, either simple trends in easting
     and northing coordinates or water related measures such as sea surface temper-
     ature measured by satellite.
   •  A key feature of a statistical technique like kriging is that a measure of uncer-
     tainty (called  the kriged prediction variance) is  generated  along  with kriged
     interpolations. Research has  been initiated  (i.e., conditional simulation) to
   appendix a
The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-28
                          propagate this interpolation uncertainty through the CFD process for gener-
                          ating confidence intervals for estimates of attainment.
                        • Kriging can be applied in situations where the data are sparse, as in CBP fixed
                          station data, or densely sampled, as in CBP shallow water monitoring. Kriged
                          and IDW spatial interpolations may very well produce near identical results for
                          these two extreme scenarios. However it is the kriging approach that provides
                          a statistical model, the uncertainty of which is influenced by the quantity and
                          quality of data. Knowledge of interpolation uncertainty is crucial for discrimi-
                          nating the improved water quality assessment obtained from densely sampled
                          networks relative to sparsely sampled networks.

                      As alluded to earlier kriging is an advanced statistical technique and like all such
                      techniques should be carried out by well trained statistician(s) with experience in
                      spatial or geostatistical methodology and experience analyzing water quality data.
                      Assessing model fits (of the variogram and regression model) and kriging  accuracy
                      via cross validation and/or likelihood based criteria should be employed routinely.

                      To further exemplify this point consider kriging  the densely sampled shallow water
                      monitoring data which is generated by the DATAFLOW sampling. In addition to the
                      other technical complexities mentioned within, this spatial sampling  design may
                      raise other issues not immediately recognized by untrained users (Deutsch 1984).

                      For kriging in CBP applications one potential methodological drawback is  the issue
                      of non-Euclidean  distance  (Curriero  2006).  Current kriging methodology only
                      allows the use of the straight line Euclidean distance as the measure of proximity.
                      However, the irregular waterways in the Chesapeake Bay system  may very well
                      suggest other non- standard measures of distance. For example, the spatial design of
                      the fixed station data including those in the Bay  mainstem and tidal  tributaries. The
                      straight line Euclidean distance may very well intersect land particularly in regions
                      containing convoluted shorelines. There has been research initiated on this topic
                      (Curriero 2006, Jensen et al. 2006, Ver Hoef et al. 2007), however, results are not yet
                      ready for universal use.

                      Three dimensional interpolations (including depth as the third dimension) are poten-
                      tially required  for CBP  applications.  The IDW  and  kriging methodologies,
                      mathematically speaking, certainly extend to three dimensions. However the rapid
                      change of water quality over depth would lead to significant anisotropies in  the
                      application three dimensional kriging that would complicate this approach far more
                      than the  application of IDW.  On the other hand, a simplistic implementation of IDW
                      that does not recognize the rapid decay of covariance over depth would inappropri-
                      ately reach across  the  pycnocline when choosing nearest neighbors.  Clearly  the
                      special properties of water quality in a highly stratified bay require innovation for 3-
                      dimensional interpolations. Another  approach would be to apply universal kriging
                      where a  third dimension (depth) is used as a covariate. The use of depth as an inde-
                      pendent  variable is motivated by the observation that often water quality exhibits a
                      predictable trend  over depth as for example the trend  of DO decreasing with
                      increasing  depth. To include depth as a covariate, model (1) would be written as

-------
                                                                                               A-29
A third approach to interpolation in three dimensions is to implement 2-D interpola-
tion in  layers.  Note that the IDW  interpolator currently  implemented by CBP
(Section 2.2) employs a layered strategy by severely restricting (+/- 2m) the vertical
distance that may  be  searched  for nearest neighbors. A similar strategy could be
implemented using 2-D kriging  to interpolate the layers. Which of these approaches
is best suited to 3-D interpolation for the bay will depend on the data available and
assumptions related to vertical structure. Full 3-D kriging interpolation treats the 3rd
dimension as a spatial dimension in  the error term y (s). The covariate approach
requires that the change over depth be a predictable trend.  Interpolation in layers
assumes that covariance decays so rapidly over depth that it is adequate to treat the
layers  as  independent  entities. Data  sufficiency  requirements increase for all
approaches  when  considering  three dimensional  interpolations.  When data are
sparse, again a statistical based approach like kriging allows this to be reflected in
prediction uncertainty.

In many applications, attainment or lack of attainment will  be so extreme that the
assessment end point is clear even without optimizing  the  error estimation of the
CFD. In these extreme cases, IDW or  kriging simplified for automation could be
sufficient to support the attainment ruling without precise quantification of estima-
tion uncertainty. For these cases, the customized IDW algorithm that is currently
implemented by CBP provides  a tool with which to begin testing the CFD assess-
ment procedure, but kriging simplified for automation may  offer some advantages.
Kriging can be simplified for automation by fixing the variogram model to one math-
ematical form, say  exponential, for all applications. With the  variogram model fixed,
kriging becomes like IDW in assuming the same mathematical form for the spatial
dependence for all cases, but it is more flexible than IDW in that the rate of spatial
correlation decay could be  allowed  to vary among applications.  In addition, the
simplified kriging opens the door for  conditional simulation, with potential benefits
that are discussed  in Section 5. While  a  simplified kriging algorithm  offers  some
advantages, there are also some potential drawbacks. Because variogram estimation
typically entails use of an iterative procedure such  as maximum likelihood or non-
linear least squares,  there  is  the potential  that  lack  of  convergence  of  these
algorithms would be problematic for  an automated implementation of kriging.

In terms of computing, IDW is available in commercial GIS software, requiring GIS
skills for application. Kriging is available in commercial statistical software and also
in the free open source R Statistical Computing Environment (R Development Core
Team 2005, Ribeiro  and Diggle 2001) and requires programming skills for those
software packages.

In summary, kriging is more sophisticated than IDW, but requires greater expertise
during implementation to fully exploit its full benefit.  Table 3.2 provides a com-
parison of the capabilities of assessments based simply on:  1) percent of samples,
2) spatial interpolation based on IDW and 3) spatial interpolation based on kriging.
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-30
                     Table 3.2.  Comparison of the capabilities of methods available for interpreting data
                               collected for Chesapeake Bay water-quality criteria assessment.
Attributes
Provides Spatial
Prediction
Provides Prediction
Uncertainty
Uncertainty for CFD
Deal with Anisotropy
Can Include Cruise
Track/Fly over
Feasibility of 3
dimensional
interpolations
Feasibility of mainstem-
tributary interpolations
Inclusion of covariates to
improve prediction
Predictions of non-linear
functions of predicted
attainment surfaces
P(y>c)
Level of Sophistication
Automation
Sample-based
Yes
No
No
No
No
No
No
No
No
Lowest
Yes
IDW
Yes
not routine
No
Possible, but
not routine
Yes
Yes
Yes
No
No
Low
Yes
Kriging
Yes
Yes
Yes
Yes
Yes
Possible, but not
routine
Possible
Yes
Yes
Very High
Possible, but not
routine
                                       4.0

                     There are several approaches to defining reference curves that are proposed for use
                     in the CFD assessment methodology. One is a biologically based definition and other
                     approaches are based on an arbitrary allowable frequency (see Section 2).  Here we
                     review these options in greater detail.
                     The idea behind biological reference curves is to identify regions of the Bay that
                     have healthy biological indicators and are thus considered to be in attainment of their
                     designated use. CFDs would be developed for these areas in the same way that CFDs
                     would be developed elsewhere, but those curves developed for healthy areas would
                     be considered "reference" curves. For example, healthy benthic IBI scores might be
                     used as indicators of adequate bottom dissolved oxygen.

                     The success of the CFD-based assessment will be dependent upon decision rules
                     related to the biological reference curves. These curves represent desired  segment-
                     designated use water quality outcomes and reflect sources of  acceptable natural
   appendix a  « The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                               A-31
variability. The reference and attainment curves follow the same general approach in
derivation—water  quality  data  collection,  spatial interpolation,  comparison  to
biologically-based water quality criteria, and combination of space-time attainment
data through a CFD. Therefore, the biological reference curve allows for implemen-
tation of threshold uncertainty as long as the reference curve is sampled similarly to
the attainment curve. Bias and uncertainty  are driven in CFD curves by sample
densities in time and space. Therefore, we advise that similar sample densities are
used in the derivation of attainment and reference curves. As this is not always
feasible, analytical methods are  needed in the future to equally weight sampling
densities between attainment and reference curves.

4,2, CBP

In some cases, the development of biologically-based reference curve is not possible
due to lack of data describing the health of the relevant species. In such cases, a more
arbitrary approach is required since better information is not available. EPA recom-
mends the use of  a default curve  in cases where  a biologically-based one is not
available. That default curve is defined by these properties:
   1. symmetric about the 1:1 line,
   2. hyperbolic,
   3. total area = 0.1, and
   4. pass through (1,0) and (0,1)

(see EPA, 2003; page 174). The equation that describes this figure is defined by the
equation:

                              (x+b) * (y+b)  = a

Where: b = 0.0429945

        a = b2 + b

This reference curve is illustrated in Figure 4.1 by curve  1.

An alternative default reference curve might be formulated by extending the arbi-
trary allowance of 10% exceedance into the two dimensional framework of the CFD.

The criterion threshold is a value that should be rarely exceeded by a population at
healthy levels. When the population is unidimensional, say concentration in a point
source effluent, then one can obtain this upper threshold  based on the simple distri-
bution of values in a healthy population (Figure 4.2). The ninetieth percentile of this
distribution might be chosen as the criterion  threshold.  Thus in this example,  10%
noncompliance is  allowed because this level  of noncompliance is expected in a
healthy population. A standard technique for  estimating distribution percentiles is to
assume a mathematical form for the distribution, e.g., the normal distribution, and to
estimate the percentile as some number of standard deviations above the mean. The
90th percentile of the normal distribution is 1.2815 standard deviations above the
mean.
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-32
                             o.u
                                          0.2    0.3
                                                       0.4     0.5    0.8
                                                       Fraction of Space
                                                                          0,7
                                                                                0.8
                                                                                      0.9
                                                                                             1.0
                            4.1. Illustrations of three reference curves: 1) the  standard CBP reference curve
                      derived to cover 10% of the percent space by percent time plane (curve 1); 2) a reference
                      curve based on 10% exceedance frequency and a temporal-spatial variance ratio of 1.0
                      (curve 2); and 3) a reference curve based on  10% exceedance frequency and a temporal-
                      spatial variance derived from chlorophyll data (curve 3).
                      When regulating populations that are distributed in both space and time, this simple
                      concept for regulating noncompliance must be extended to account for the variability
                      in each dimension. While there is some added complexity in the mathematics, the
                      fundamental concept remains  the same: That  is, to set the criterion threshold at a
                      certain distance above the mean so that exceedance of that threshold will be rare in
                      a healthy population. In this case, the distance by which the threshold must exceed
                      the  mean  is a function of both the  spatial  and temporal variance components as
                      described below.

                      To establish these criteria thresholds for populations with two components of vari-
                      ance, assume the  simple model:
                      where:
                             IJL is the desired mean level of chlorophyll (in log space)
                             «; is a random term for variation over time with variance o2,
                             /?i(Sj) is a random term for variation over space with variance O2p
                             YJ(SJ) is a water quality constituent measured at time i and location Sj.

                      The variance of xy is O2a + o2^ = o2. The standard  dev of xis is sqrt(cr2)  = o. It is
                      common to  allow an overall 10% exceedance rate without declaring an assessment
                      unit out of compliance. We would expect 10% of the xis to fall above /u +  1.2815*cr
   appendix a

-------
                                                                                                A-33
       0.6
       0.5:
       0.4-
    I
    o
    a.
    &
       0.2:
       0.1
10
    15
Chlorophyll
                                                   20
25
      4.2. Hypothetical lognormal distribution that might be typical of Chlorophyll.
The figure illustrates the relation of the geometric mean and the criterion threshold
set at the 90th percentile.
where 1.2815 is the 90th percentile of the standard normal distribution. Thus (assum-
ing normality) a population with spatial and temporal variance characterized by O2a
and O2a that has a mean that is 1.2815*<7 below the threshold criterion should have
an exceedance rate of 10% over space and time. Note that the reference  curve is
determined by the ratio O2alo2^ and the distance in standard deviations of the mean
from the threshold. The actual values of the variance components, the mean, and the
threshold, are not important as long as the relationships hold. Thus as long as the
variance ratio is consistent, and mean to threshold distance is a fixed number of stan-
dard deviations, the same reference curve will serve for all seasons and regions.

Letting  chlorophyll observed in the decade of the 1960s serve as a reference popu-
lation, the parameters in Table 4.1 can be used to construct this reference curve based
on the variance ratio and the mean to threshold distance given in the table. The ratio
O2alo2p  is computed as the ratio of the temporal variance term and the spatial vari-
ance term. The mean to threshold distance is computed to be 1.2815cr for all regions
and seasons. Based on there parameters, a reference curve for chlorophyll can be
derived (curve 3 , Figure 4.1). For comparison a reference curve based on a variance
ratio of 1.0  (curve 2, Figure 4.1) and  the standard CBP reference curve (curve 1,
Figure 4.1) are also shown.
   appendix a

-------
A-34
             Table 4.1. Chlorophyll criteria derived by computing and upper threshold based on predicted
                      means for mid-flow1960s chlorphyll data.

Season
Spring
Summer
Spring
Summer
Spring
Summer
Salinity
Zone
OH
OH
MH
MH
PH
PH
Mean
Log(chl)
0.7684
1.1693
0.4137
0.8626
0.1386
0.218
GMmean
(chl)
5.87
14.77
2.59
7.29
1.38
1.65
Temporal
Variance
0.0233
0.0233
0.0233
0.0233
0.0233
0.0233
Spatial
Variance
0.0658
0.0658
0.0658
0.0658
0.0658
0.0658
Std
Dev
log(chl)
0.2985
0.2985
0.2985
0.2985
0.2985
0.2985
Threshold
Criterion
log(chl)
1 .2594
1 .6603
0.9047
1.3536
0.6296
0.709
Threshold
Criterion
(chl)
18.17
45.74
8.03
22.58
4.26
5.12
                    Relative to the standard reference curves, the curve based on the observed variance
                    ratio for chlorophyll is more restrictive of events where large portions of the popu-
                    lation are out of compliance. For example, the CBP standard reference (curve 1)
                    would allow 40% of area to exceed the criterion threshold up to about 6% of the
                    time. The proposed chlorophyll reference curve (curve 3) would restrict occurrences
                    of 40% of area out of compliance to about 2% of the time. Conversely, the proposed
                    curve (curve 3) allows  a higher frequency of events where a small  percentage of
                    space in out of compliance. For example, 10% of space is allowed out of compliance
                    36% of the time under the proposed curve and 27% of the time under the standard
                    curve.

                    While there is mathematical and statistical logic underpinning this proposed chloro-
                    phyll  reference curve, it is important to remember that it is based on parametric
                    models and simplifying assumptions. It is recommended that validation exercises be
                    performed  to insure  that  the general shape of CFD curves generated from data
                    collected in near reference conditions is approximated by the proposed curve.

                    4.3 ACCOMMODATING SEASONALITY IN REFERENCE CURVES

                    The degree of acceptable exceedance can vary with season. For example, benthos are
                    less tolerant of hypoxia in warmer water temperatures. In addition, the threshold
                    criterion  may never be exceeded in some seasons and frequently be exceeded in
                    others. By  combining seasons,  the acuteness of a specific seasonal exceedence is
                    diluted by data from the acceptable season(s). To some extent, seasonal differences
                    can be accommodated by changing the threshold criterion among seasons. However,
                    there may still be a need to develop separate reference curves by season.
                            5.0  REVIEW  CFD STATISTICAL  PROPERTIES
                         INCLUDING  BIAS, PRECISION, AND  INFERENCE
                    The CFD as an assessment tool is a relatively new and unstudied concept. Its close
                    relationship to the empirical distribution function does give some insight on the
                    mathematical behavior of the CFD. In this section we review some of the properties
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-35
of the CFD and discuss the complications that arise from these properties when the
CFD is used as an assessment tool. After defining the population which determines
the  CFD, we  go on to discuss the currently proposed sampling and estimation
scheme, sources of error in the estimation scheme, and problems that result from
these. The goal is to  succinctly define these problems and elucidate possible solu-
tions. This section will cover: the behavior of the CFD as a function of temporal and
spatial variance, methods for  construction CFD reference  curves, the influence of
sampling and estimation variance on the CFD shape, and feasible methods for devel-
oping statistical inference tools.

5 1           OF CFD
SM? m 1  l%i™Wli™SB \af 1 %«wi 1 imT I 1 %\a# 1 I™ 1 % 1 1 I™ W»

With any statistical  application, it is important  to  distinguish between the true
descriptive model underlying the population being sampled and the estimate of this
model derived from the data collected in a sample. As described above, the CFD has
a data driven definition where the CFD  is constructed based on a sample from a
population for some water quality parameter. This population  is  a  continuous
random process over  space and time.

In order to quantify the statistical properties of the  CFD, the CFD is defined in terms
of a population of experimental units. This  approach  is a discrete approximation of
the  continuous random process in both  time  and space. However, the estimation
scheme involves interpolation to discrete units in a spatial dimension and discrete
days in the temporal dimension. To facilitate an understanding of the relation of the
estimator to the true population, it seems reasonable to use a discrete approximation
as the model for the true population.
The population will be defined as having different sizes of experimental units in
much the way we think of a population that gives rise to a nested design or repeated
measures design. The Chesapeake Bay will be partitioned into segments. Assessment
will be done for each segment based on a three year record of the segment. Thus a
three  year period for the segment defines the entire population that will be  parti-
tioned into experimental units. The continuous time dimension is partitioned into
days to form the primary units which are the state of a segment for a day. Call this a
Segment-Day. Let there be M segment-days in the assessment period (typically 3 x
365).  The continuous spatial  dimension  is partitioned into N 3-dimensional cells
(may  range from hundreds to  thousands). The state of each cell for a day will be a
unit nested within  the segment-day. The attribute of interest will be a measure of
water quality for each cell for a day. Examples might be the mean level of Chloro-
phyll-a in the cell for one day or the minimum of dissolved oxygen in the cell during
the day. Let Y  be  a random variable  for the attribute of interest and consider the
following model

                             YJ(SJ) = fi + a{ + /?i(Sj)               Eqn 5.1.1.1
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-36
                     the vector a will be assumed to have expectation 0 and variance 2c
                     each vector ^ will be assumed to have expectation 0 and variance 2/8i.
                     i is the ordinal index for days and
                     s is a vector valued ordinal for spatial location.

                     Under this model, 2a defines the correlation over time at the segment-day level and
                     2/?i defines  correlation over space that occurs cell to cell within a day.

                     Let C ;(Sj) be a collection of threshold limits that define the acceptable criterion for
                     the measured attribute. If Y ;(Sj) exceeds C ;(Sj) in a cell, that cell is called degraded.
                     The criterion is allowed to vary in both time and space so that in theory each Y ;(Sj)
                     might be compared to a unique C j(Sj)..  It may vary over time because different
                     levels of Y may be acceptable in different seasons.  It may vary over space because
                     different levels of Y may be acceptable in different salinity regimes so that even
                     within a segment,  C may  be  a function of  salinity. As  a rule, it is anticipated that
                     C ;(Sj) will be constant for regions  of space  and time  such  as salinity zones and
                     seasons.

                     Now  convert the measured attribute Y ;(Sj) to a Boolean response as follows

                     TY ;(Sj) = I(Y ;(Sj)  > C i(Sj))     = 1  if Y ;(Sj) > C ;(Sj)                Eqn 5.1.1.2
                                                   = 0 otherwise

                     Thus  TY takes the value 1 when Y exceeds the  threshold defined by C.  Using TY,
                     we summarize the  state of a segment on one day as the fraction of that segment that
                     is out of compliance

                                                                                      Eqn 5.1.1.3
                     The CFD that we wish to estimate is one minus the cumulative distribution function
                     of the PJ'S. If P(i) represents the ordered values of the P;'s for any assessment period,
                     then let
                                           (l)p                                        Eqn 5.1.1.4

                     G defines the CFD that if it were known would be used for an exact assessment. The
                     cumulative distribution function is determined by the mean and variance of the ideal
                     population. This population is defined with  a  spatial variance component and a
                     temporal variance component.  The final CFD  shows the cumulative percent of time
                     that a certain percent of space is below the criterion threshold. If the CFD shows that
                     water quality in a segment is beyond the threshold for too much space and too much
                     time, then the segment is classified as impaired.

                     For one assessment period, G can be considered exact as defined above, but recog-
                     nize that even this is only one observation of the many possible observations of G
                     that could result from sampling different assessment periods.

                     Assume for simplicity that Y is normal. If 2a were 0 so that Y had constant expec-
                     tation over time and if 2/3 were of the form a2! then each cell on each day would
                     have constant probability of exceeding a constant value  of C given  by 1  - <&(C)
   appendix a  » The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                              A-37
where    is the normal  cumulative  density function.  In  this greatly  simplified
scenario, P; would be the outcome of N independent Bernoulli trials. The ideal CFD
would be the cumulative distribution function of M outcomes of a binomial random
variable with N trials. If we allow 21/j to have positive off diagonal elements, then the
Bernoulli trials become dependent (i.e. adjacent cells are more likely to either both
exceed or both meet the standard than distant cells). This should make the distribu-
tion of the Pj  more variable than under the independent binomial model, but the
expectation of P; would be constant over time. If we relax the assumption that ^a
is 0, then the expectation of the P; would vary over time which would increase the
variability of the P; even more.

Under the simplifying assumptions of independence, constant mean, and constant
variance, it is possible to obtain an analytical formulation for the CFD based on the
parameters of Eqn 5.1.1.1. However, when the more realistic time dependent, space
dependent model with seasonal nonstationarity is considered, an analytical formula-
tion is not tractable. The lack of an analytical formulation for this estimator under
realistic dependence assumptions, e.g. non-trivial ^a and 2^, points toward com-
puter intensive simulation techniques  to develop statistical inference procedures for
this problem. None-the-less, it is interesting to consider the behavior of the CFD
under the simplified model.

5.3 CFD                      A

In what follows, the behavior of the CFD under various parameter formulations for
Equation 5.1.1.1 are presented in graphical form. There are four parameters involved:
|u the population mean, ot the temporal  variance, os the spatial variance,  and C the
criterion threshold. In the examples that follow, three of these parameters are held
constant and the fourth is varied to illustrate the effect of the varied parameter.

In this exercise, the parameters of Equation 5.1.1.1 are simplified as follows: ^a = ot
I and ^ = os I, where I is the identity matrix. Thus in both the temporal and spatial
dimensions, independence and constant variance is assumed.
Example 1. Example 1 considers the effect of changing the population mean on the
shape of the CFD.

     5.1.  Parameter values and key for the family of curves shown in Figure 5.1.
V
5
4
3
2
1
CTt
1
1
1
1
1
CJs
1
1
1
1
1
C
5
5
5
5
5
color
Red
Orange
Brown
Green
Blue
curve
number
1
2
3
4
5
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-38
                           1.0;


                          0.9;
                          0.7;


                          0.6:
                        §• 0.4
                          0.3;


                          0.2;


                          0.1-
                                               0.3     O.I     OJ>    0.6
                                                    Proportion of Space
                                                                        0.7
                                                                              0.8
                                                                                    0.9
                            5.1. A family of curves illustrating the behavior of the CFD as the population mean
                      decreases from the criterion threshold. The parameter values for each curve and the
                      corresponding curve number are given in Table 5.1.
                     Note that when the population mean is equal to the criterion threshold, the CFD is a
                     diagonal line from upper left to lower right (Figure 5.1, curve 1). This is largely an
                     artifact of using symmetric distributions, the normal, for both the time and space
                     variance components. That is, when the population median is equal to the criterion
                     threshold, we expect an average of 50% noncompliance over time and we expect the
                     exceed 50% noncompliance 50% of the time.

                     As the overall population mean decreases from the criterion threshold, the family  of
                     curves tends to move from the diagonal line toward the lower left corner. Thus a
                     reference population, which should have a small probability of exceeding the crite-
                     rion threshold might have a shape similar to  the  green curve.  This illustrates the
                     importance of the shape of the CFD in measuring compliance. A CFD from a highly
                     compliant population will tend to hug to lower left corner similar to the blue and
                     green curves. As the population mean approaches  the criterion threshold, the CFD
                     approaches curve  1. If the population mean were to exceed the criterion threshold,
                     the CFD would tend toward the upper right corner.
   appendix a

-------
                                                                                                 A-39
Example 2. Example 2 considers the effect of changing the temporal variance on the
shape of the CFD. Note that the population mean is held constant at 3 which corre-
sponds to curve 2 of the preceding example.
      5,2,  Parameter values and key for the family of curves shown in Figure 5.2
0
3
3
3
3
3
Ot
1
2
3
4
5
os
1
1
1
1
1
c
5
5
5
5
5
color
Red
Orange
Brown
Green
Blue
curve
number
1
2
3
4
5
       1.0

       0.9

       O.S

       0.7
    a
    P  0.6
    "s
    I  0.5J

    |-  0.4'
    £
       0.3

       0.2-

       0.1

       0.0i
        0.0    0.1     0.2     0.3    0.4     OJi     0.6
                                 Proportion of Spa«»
                                                           0.8    0.9
                                                                       1.0
      5.2, A family of curves illustrating the behavior of the CFD as the temporal popula-
tion variance increases. The parameter values for each curve and the corresponding curve
number are given in Table 5.2. Note that the curve 1 here has the same parameters as
curve 2 in Figure 5.1.
As temporal variance increases, the frequency of large proportions of space going
out of compliance increases (Figure 5.2, lower right). Conversely, the frequency of
small proportions of space out of compliance  (i.e.  large proportions of space being
in compliance) decreases (Figure 5.2, upper left). That is, shifting the  daily mean
either down or up tends to shift the entire segment toward or away from compliance.
   appendix

-------
A-40
                      In preparing water clarity CFDs for reference areas defined by having successful
                      SAV beds, it is not unusual to find a curve shape similar to Figure 5.2 orange or
                      yellow curves. This pattern suggests that SAV is tolerant of ephemeral events of
                      spatially broad degraded water clarity. If water clarity is persistently degraded over
                      portions of the area, SAV may be impaired.
                      Example 3. Example 3 considers the effect of changing the spatial variance on the
                      shape of the CFD. Again the population mean is held constant at 3 which corre-
                      sponds to the curve 2 of the first example.

                      Table 5.3,  Parameter values and key for the family of curves shown in  Figure 5.3
n
3
3
3
3
3
CTt
1
2
3
4
5
ortion of Space
                                                                                 0.8    0.9
                                                                                             1.0
                            5.3, A family of curves illustrating the behavior of the CFD as the spatial popula-
                      tion variance increases. The parameter values for each curve and the corresponding curve
                      number are given in Table 5.3.
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                                  A-41
Increasing the spatial variance results in a family of curves that is complementary to
those that follow an increase in temporal variance. Increasing spatial variance results
in a higher frequency of small proportions being out of compliance. It is not so much
an all-or-nothing phenomenon.
Example 4. Example 4 considers the effect of changing both temporal and spatial
variance on the shape of the CFD.
     5,4, Parameter values and key for the family of curves shown in Figure 5.4.
u
3
3
3
3
3
CTt
1
2
o
3
4
5
CTs
1
2
o
3
4
5
c
5
5
5
5
5
color
Red
Orange
Brown
Green
Blue
curve
number
1
2
o
3
4
5
       1.0

       0.9

       0.8

       0.7

       0.6

       0-5


       OA

       0.3;

       0.2:

       0.1

       0.0-1
         0.0
               0.1
                            0.3     0.4     0.5     0.6
                                  Proportion of Spaa*
                                                      0.7
                                                             0.8
                                                                   0.9
                                                                         1.0
       5,4, A family of curves illustrating the behavior of the CFD as both temporal and
spatial  variance increases. The parameter values for each curve and the corresponding
curve number are given in Table 5.4.
   appendix a

-------
A-42
                     Increasing the  spatial and temporal variance together has the opposite effect of
                     decreasing the population mean. The CFD tends to move in a direction of noncom-
                     pliance. Thus compliance as measured by the CFD depends on the relative values of
                     the population mean, the temporal and spatial variance, and the criterion threshold.
                     Increasing the  population mean  has  the  same effect as decreasing the  criterion
                     threshold. Increasing population variance has the same effect as increasing the mean
                     or decreasing the criterion threshold. In a sense, the CFD is measuring the distance
                     between the population mean and the criterion threshold in units of variance analo-
                     gous to a simple t-test. A nuance introduced here that has no analogy in the t-test is
                     that the ratio of spatial to temporal variance controls the symmetry of the curve.

                                        I / '.»M  'M'.*«

                     In Section 5.1., it was shown that the shape of the CFD is a critical element to deter-
                     mining compliance. Thus it is important that this shape be primarily determined by
                     the state of compliance of a segment and not be influenced by factors not relating to
                     the status of compliance. Because the CFD is constructed based on data that are a
                     sample from the whole, it is clear that some uncertainty in the CFD will result. In
                     addition, the CFD is a function of the empirical distribution function (EDF) of frac-
                     tion of space in compliance. The shape of this EDF is determined by the mean and
                     variance of the sample. Thus any  factor, such as sample size, that affects the preci-
                     sion of the fraction of space estimate, will affect the shape of the CFD. In this section
                     we review the effect of noncompliance factors on the  shape of the CFD.



                     As noted, because the CFD is  a function of the EDF of estimates of "fraction of
                     space", any factor affecting the precision of the  estimate of fraction of space in
                     exceedance will affect the shape of the CFD. In particular,  the number  of samples
                     used for each p-hat (% exceedence) will affect precision. For a given segment, this
                     fraction will be estimated more accurately if twelve  samples are used to form the
                     interpolated surface rather than six.  Because of unknown spatial dependence in the
                     data, it is  difficult to analytically quantify the magnitude of this sample  size effect.
                     Therefore simulation analysis was employed to address this issue.

                     Numerous simulation tests were performed. These begin with a simulation of struc-
                     turally simple data that have no temporal or seasonal trend and progress to simulated
                     data that mimic the temporal and spatial structure of observed data. Because the
                     results from this  latter simulation are most relevant, these  are the results that are
                     presented and discussed.



                     Simulated data were created to mimic the properties of surface chlorophyll in the
                     Patuxent estuary.  Data were created to fill a 5 by  60 cell grid which approximates
                     the long and thin nature of an estuary. These data have mean zero and  a spatial
                     variance-covariance structure chosen to approximate the spatial variance-covariance
                     structure of cruise-track  chlorophyll observed in  the Patuxent  estuary. Thirty-six
   appendix a  * The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-43
grids of data were simulated to  represent 36 months in a three  year assessment
period. The temporal and spatial trends were added to the simulated data by adding
in means computed for each month and river kilometer during the period Jan 1, 1991
to Dec 31,1993. Simulated data were created using the "grf' function of the Geosta-
tistical Package "geoR" of the R-package.

After the full population of data was simulated for 3 year assessment period, a
sampling experiment was conducted to assess the effect of sample size on the shape
of the CFD. First, as  a benchmark, a CFD was computed using all of the  simulated
data. To  simulate the effect of sampling, a sample of fixed  size was  randomly
selected from each  the 36 5x60  grids  of  data. Using these  samples, kriging
(krige.conv function  of geoR)  was used to populate each monthly grid  with esti-
mates. These estimated chlorophyll surfaces were used to compute an estimate of the
CFD which was graphically compared to the benchmark (Figure 5.5). For a fixed
sample size, the process was repeated  until  it was clear whether the differences
between the benchmark CFD and the estimated CFDs were due to variance or bias.
To assess the effect of sample size, the process was repeated for several sample sizes.

The effect of sample size on the  shape of the CFD is consistent with expectations
based on the relation of the CFD to the empirical distribution function (Figure 5.5).
As sample size  decreases, the variance of the estimated values of fraction of space
increases. This increase in variance results in the estimated CFD being to  the left of
the true curve for low values of fraction of space and to the right of the true curve
for high values of fraction of space. This assessment has been repeated many times,
varying the threshold criterion, systematic vs. random sampling, the level of vari-
ability in the simulated data, and so on. This sample size effect persists for every case
where realistic estimation is employed.



As shown above (Figures 5.2-5.4) the shape of the CFD is a function of the ratio of
temporal and spatial variance. To the extent that the ratio  of these variance compo-
nents in the data represent the true state of nature, this is acceptable. However, under
a model with strong  spatial and temporal dependence, the  ratio  of these variance
components might be influenced by the scale of sampling in the spatial and temporal
dimensions. For example, samples collected far  apart in time might reflect higher
variance than samples collected close in time. If the ratio of temporal and spatial
variance is influenced by the density of sampling in each dimension, then experi-
mental design will have an effect on the asymmetry of the CFD estimate.
An investigation into the use of conditional simulation to obtain confidence bounds
for the CFD showed that not only is this a promising technique for statistical infer-
ence, but also has potential in correcting bias associated with sample size effects that
has been identified  as a central  problem in implementing the CFD  approach.
Correcting the bias of the CFD due to the sample size effect is important in obtaining
confidence bounds on the CFD that cover the true CFD for a segment. Because bias
correction is an important first step, this aspect of the conditional simulation exper-
   appendix a  -  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-44
                             CFD simulations
                              0.4       0.6



                               fraction of space
                                                                                CFD simulations
 0.4       0.6



  fraction of space
                            CFD simulations
                                                                               CFD simulations
           0.0
                    0,2
                             0.4       0.6



                              fraction of space
                                               0.8
                                                        1.0    00
                                                                       0,2
0,4       0.6



 fraction of space
                                                                                                     true



                                                                                                     n=80
                                                                                                  0,8
                                                                                                           1.0
   Figure 5.5. Illustration of the effect of sample size (n) on the shape of the CFD for sample sizes 10, 20, 40, and 80.
   appendix a  «  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-45
iments will be discussed first. Conditional simulation will then be evaluated in its
efficacy in obtaining confidence intervals.

This section first outlines the basic concept of conditional simulation and provides
an algorithm that employs conditional simulation to estimate confidence bounds for
the CFD. The results of this experiment support the potential of conditional simula-
tion for correcting the sample  size bias. A heuristic discussion of the mechanism
underlying this adjustment for sample size effect is presented with the hope of moti-
vating additional analytical investigation of this effect.

Conditional  simulation (Journel,  1974; Gotway, 1994) is a geostastical term for
simulating a population conditional on information observed in a sample. In the case
of kriging, a sample from a spatial population is used to estimate the variogram and
mean for the population. The conditional simulation procedure generates a field of
simulated values conditioned on the estimated mean and variogram from the sample.
To the extent that the estimated mean and variogram approximate the true mean and
variogram and the assumed distribution is a reasonable model for the true distribu-
tion, repeated  simulations of this virtual population will represent the variability
typical of the true population. It follows that statistics computed from the condition-
ally simulated fields will represent the expected variability of statistics from the true
distribution.  The CFD is a graphical representation of ordered statistics of percent
compliance over time and it is a reasonable to assume that  repeated conditional
simulations will lead to effective confidence bounds for the CFD.



In the computation of the CFD, conditional simulation is implemented at the inter-
polation  step for  each month. Interpolation produces  an estimate  of the  spatial
surface of the target parameter. From that estimate of the surface is obtained an esti-
mate of the percent of noncompliance. Using conditional simulation, the surface can
be reconstructed 1000 times. From the 1000 simulated surfaces are computed 1000
estimates of the proportion of noncompliance. When this is repeated for each  month
for say 36 months, the result is an array of 1000 sets of 36 values of the proportion
of noncompliance. Each of the 1000 sets of 36 can then be ranked from largest to
smallest to compute a CFD in the usual way which results in 1000 CFD estimates.
The variability among these 1000 CFDs can be used to estimate confidence intervals.

To evaluate this concept, the following simulation experiment was conducted
   1. The first step is to  simulate a population that  will  be considered the  "true"
     population for this exercise. A grid of dimensions 5x60 is populated using an
     exponential  spatial  variance  model  with variogram parameters   set to
     (0.00625026, 2.67393446). These variogram parameters  were estimated from
     Patuxent cruise track chlorophyll data. This grid is populated 36 times to repre-
     sent 36 months.  The mean and variogram are held constant  for the  36
     simulations to create a simplistic case with no seasonal or spatial trend.  Using
     this set of data, the CFD is computed in the usual way and this is considered
     the "true" CFD.
   appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-46
                       2. A sample of size 40 is selected from each of the 36 simulations at random loca-
                          tions on the grid. Ordinary kriging is used to estimate the spatial surface for
                          each simulation and from these 36 estimates of the monthly spatial surfaces, a
                          CFD is computed. This is called the 'estimated' CFD.
                       3. For each of the kriged monthly surfaces, 1000 conditional surfaces are simu-
                          lated based upon the mean and variogram estimated from the sample data. The
                          Cholesky decomposition is used to reconstruct the covariance structure indi-
                          cated by the estimated variogram. The  conditionally simulated surfaces were
                          processed  to yield 1000 estimates of the proportion of noncompliance. The
                          1000x36 noncompliance values are used to compute 1000 CFDs, which are
                          called the population of "conditionally simulated" CFDs.
                       4. Each "rank position"  of the monthly ordered proportions of noncompliance
                          has  1000 values in this simulated population. To assess variability in the simu-
                          lated population, graphs of the  miniumum,  the 2.5th percentile,  the 50th
                          percentile, the 97.5th percentile, and the maximum at each rank position are
                          plotted to illustrate a 95% confidence envelop  for the CFD (Figure 5.6).

                     To test this procedure under various conditions, this basic simulation exercise was
                     repeated varying the sample size and adding temporal and spatial trend to the simu-
                     lation of the "true" population to reflect conditions more similar to real populations.



                     The results of this simulation exercise  are presented graphically. In Figure 5.6 the
                     line 1 represents the CFD computed for the true population computed from the orig-
                     inal data. The line 2 is the estimated CFD computed from kriging estimates based on
                     samples from the true population. The line 3 lines represent the min and max of the
                     1000 conditionally simulated CFDs. The  two line 4s represent the 2.5  and 97.5
                     percentiles of the 1000 conditionally simulated CFDs, which  is the proposed 95
                     percent confidence interval. The line 5 is the  median of the 1000 CFD curves.



                     The results in Figure 5.6 are unusual in several respects. First  note that the line 2
                     shows the typical sample size bias for the CFD as described above (n=40).  Relative
                     to the true CFD (line 1) the estimated  CFD  is below line 1 for half the curve and
                     above line 1 for the remainder. The first unusual feature is that the distribution of the
                     conditionally simulated CFD curves is  not centered on estimated CFD. In fact the
                     estimated CFD is not completely within the bounds (min, max) of the conditionally
                     simulated population. A surprising feature is that the median of the simulated popu-
                     lation tracks fairly well with the true CFD (line 1). It is clear that the simulated CFD
                     population is estimating something other than what is estimated by the estimated
                     CFD (line 2). At the same time, it appears that the median of the simulated popula-
                     tion is a good estimator of the true CFD and the proposed confidence bands (line 3)
                     is reasonable confidence envelop about the true CFD.

                     What follows is a heuristic  explanation for why CFD computed from conditional
                     simulations might be  a better estimator  of the true CFD than a CFD computed from
   appendix a »  The Cumulative Frequency Diagram         for Determining Water Quality Attainment

-------
                                                                                              A-47
                               CFD simulations
CD
E

"o
o
      op
      o
      (O
      o
      CNI
      o
      o
      O
           0.0         0.2         0.4         0.6

                                 fraction of space
                                                     0.8
1.0
      5,6, Confidence bounds computed based on quantiles of fraction of space
computed on conditionally simulated surface estimates using variogram estimates
from data. The base simulation has spatial correlation and no spatial or temporal trend.
Sample size is 40.
the kriging estimator. Additional analyses test whether this property might hold in
general or is an artifact of the simple conditions (no spatial or temporal trend) under
which this experiment was performed.

In prior discussions we have noted that the CFD is the inverse of the CDF of the
population of p's where p is fraction of space out of compliance with the criterion
threshold. It is the variance of the p's that determines the steepness of the CFD: the
smaller the variance, the steeper the CFD. In real applications, estimates of the p's
have two important variance components. One variance component comes from true
variance  over time in the parameter being assessed. Another  variance component
comes from imperfect estimates due to sampling variability. In the base simulation
with no spatial or temporal trend in the data, it is this second source of variance that
controls the  shape of the CFD.
   appendix a

-------
A-48
                     Because the variance of the p's is critical to the shape of the CFD, consider the vari-
                     ance of p's computed from three sources in the experiment outlined above: 1) the true
                     data, 2) a krig estimate based on a sample from the true data, and 3) conditionally
                     simulated data based on a krig estimate of 2). To enhance our understanding of this
                     comparison, the variance  of the p's are discussed for two cases for each source. The
                     first case assumes complete independence in the base simulation and does not use
                     interpolation to estimate proportion of area out of compliance. This simplification
                     allows us to easily  infer  the behavior of the CFD using  analytical methods. The
                     second case introduces an unknown spatial dependence in the base simulation and
                     uses interpolated data to  estimate the proportion of area out of compliance. These
                     additional complexities make it difficult to implement  analytical inference but
                     conclusions may still be inferred by analogy to the simple independent case.

                     Consider the sequence of sources where the base simulations are generated under the
                     simple constraints of constant mean, constant variance and the errors for each cell of
                     the grid that are independent. For this case the exceedance probability is:
                     where :         C is the criterion threshold,
                                    xs is the data at location s,
                                    |i is the mean used in the simulation,
                                    crns the variance used in the simulation, and
                                     is the standard normal Cumulative Distribution Function.

                     The distribution of the true p's  computed from all 300 cells of the 5x60 simulation
                     grid would behave like that of a independent binomial with N=300 with a variance
                     of (p(l-p)/300). From these independent data draw a sample of size 40. Using only
                     the proportion of the  sample that is  out of compliance to estimate the p's, the distri-
                     bution of the p's would be that of a independent binomial with N = 40 and variance
                     (p(l-p)/40). Clearly the p's estimated from the sample of 40 have much larger vari-
                     ance than p's from the base simulation with 300 cells. Thus the true CFD computed
                     using data from 300  cells will be steeper than the  sample CFD  computed from 40
                     data points. This pattern is illustrated by comparing the true CFD (line 1) and the
                     estimated CFD (curve 2) in Figure 5.6. This increase in the variance of the p's due to
                     small sample size is the kernel of the  sample size problem with the CFD.  Now
                     consider the behavior of p's  computed  from conditional simulations based on the
                     sample. Compute x and s as estimates of® and (D from the sample of 40 in the usual
                     way. The conditional simulation is done by populating the 5x60 grid with data from
                     a normal distribution with mean jc; and variance s2j. The exceedance probability for
                     these simulated data for the ith month is
                     where : xss is simulated data at location s
                             jq is the estimated mean used in the conditional simulation, and
                            Sj is the estimated standard deviation used in the conditional simulation.

                     If the p' were constant over months, the variance of the p's estimated by conditional
                     simulation would be (p'(l-p')/300). The sample size component of this variance has
                     been standardized to 300 which is the same as the sample size component of the true
                     p's, but the variability of conditionally simulated p's will be greater than that of true

-------
                                                                                             A-49
p's because estimates of jc;  and s2j will vary over months. The parameter p and it's
estimate p' will be close if  x and s are close to CD and (D. In the simple case with
constant mean and independent errors, the CFD estimated by conditional simulation
will better approximate the  true CFD because both are based on binomial distribu-
tions with the same N and approximately the same p.

Now consider the same sequence of distributions where the assumption of inde-
pendence is relaxed and interpolation of the data is used to  estimate the proportion
of noncompliance. The  introduction of spatial covariance  in  the base simulation
changes distribution of the true p's to a dependent binomial. The dependent binomial
will have variance similar to an independent binomial with N < 300. Sample size that
approximates the variance of the dependent binomial is termed Nb. The variance of
the p's estimated from  spatially  dependent data is approximated  by (p(l-p)/Nb)
where Nb < 300 and thus the CFD from the independent case will be steeper than
from the dependent case. The degree to which Nb is less than N will depend on the
strength of the spatial correlation.

Next consider the effect of dependent data and interpolation on the distribution of the
p's. When we  interpolate the sample of 40 onto the grid of 300, the interpolated
surface is smooth relative to the original data (compare curves 1 and 4 in Figure 5.2).
Because of this increased  dependence in  the  krig estimates, the estimates of p
computed from the interpolated data behave more like binomial data with N=Ns (the
sample size) than like binomial data with N=Nb (the number of grid cells). Because
Ns is smaller than Nb, the variance of the population of p's computed  from interpo-
lated data will be greater. The greater variance explains why curve 1 in Figure 5.6is
much flatter than line 1.

Finally consider the effect of conditional simulation on the distribution  of the p's.
When data are conditionally simulated and the mean and variogram estimated from
the sampled data are accurate, then the character of the simulated data will be similar
to that of the true data (compare the line 1 with line 3 in Figure 5.7). Like the simple
independent case, the population  of p's computed from the conditionally simulated
data will have  a binomial variance that is similar to a binomial with sample size Nb.
The simulation experiment shows that the CFD computed from these  conditionally
simulated p's will have a shape similar to the true CFD. This effect is illustrated in
Figure 5.6 where the median of the conditionally simulated CFDs (blue line) is more
similar to the true CFD line 1 than is the CFD estimate based on kriging  (red line).
Additional analytical work is needed to formalize the heuristic concepts presented
here, but this finding indicates a productive direction in developing statistical  infer-
ence procedures in the CFD approach.



The most successful technique for computing confidence bounds for the CFD were
obtained using conditional simulation based on kriging interpolation of the sample
data. The 95% confidence bands (lines 2, Figure 5.6) are well centered over the true
CFD (line 1) for the simplistic case  where the true data have spatial dependence but
no spatial or temporal trends. When these simplistic assumptions are relaxed (Figure
5.8) and the true data are simulated to have spatial dependence and  temporal and
   appendix a  *  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-50
                                                     Center Transect
                       O)
                       o
                           o
                           o
                                                                           (1)
                                                                   ——  krig estimate  (2)
                                                                   —  Conditional Simulation (3)
                                          I
                                         10
 I
20
   I
  30

northing
 I
40
 I
50
 I
60
                      Figure 5,7. Simulated chlorophyll data, kriging estimates based on a sample of the
                      simulated data, and conditionally simulated data where the simulation is conditioned
                      on the data used obtain the kriging estimates.
                      spatial trends similar to chlorophyll data from the Patuxent estuary, the confidence
                      bands cover the true CFD in this case as well.  Experiments that varied the sample
                      size also produced confidence bands with good coverage.

                      Additional evaluation of the confidence band procedure should include a series of
                      confidence band coverage experiments to assess the true coverage rate in comparison
                      to the nominal coverage rate (e.g., 95% in this example). This series of experiments
                      should be conducted with  simulated data where the simulations are designed to
                      produce  data with properties similar to the three primary  assessment water quality
                      parameters.
   appendix a  «  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                                A-51
                                 CFD simulations
     c
    .
        CO
        o
        CD
        O
        2
        p
        O
             0.0
0.2
 I           I
0.4         0.6

fraction of space
0.8
1.0
      5,8, Confidence bounds based on quantiles of fraction of space computed on
conditionally simulated surface estimates using variogram estimates from data.
The base simulation has spatial and temporal trend  estimated from Patuxent data.
Sample size is 40.
    6.0                                                       OF
6.1.   CFD              AS

This report represents an initial expert review of the CFD compliance approach. In
addition the panel undertook simulation tests on the effects of 1) sample densities in
time and space, 2) varying levels of attainment, and 3) varying degrees of spatial and
temporal covariance. Further, trials of spatial modeling on fixed station Chesapeake
Bay water quality data were conducted to begin to evaluate spatial modeling proce-
dures. Based upon review of underlying theory, initial statistical assessments, and
implementation feasibility, the panel finds that the CFD approach currently repre-
sents best  available  science  in its application to  water quality  attainment
   appendix

-------
A-52
                     determinations in the Chesapeake Bay. Using criteria for Best Science and Best
                     Available Science developed by the American Fisheries Society and the Estuarine
                     Research Federation (Sullivan et  al.  2006), we list relevant attributes  of the CFD
                     approach (Table 6.1).

                     The CFD builds on important statistical theory related to the cumulative distribution
                     function and as such, its statistical  properties can be simulated and deduced. We have
                     also shown that it is feasible to construct confidence ellipses that support inferences
                     related to threshold curves or other tests of spatial and temporal compliance. Work
                     remains to be done in understanding fundamental properties of how the CFD repre-
                     sents  likely covariances  of  attainment in time and space and how temporal and
                     spatial correlations interact with sample size effects. Further, more work is needed
                     in analyzing biases  across regions and designated use segments. The panel expects
                     that a two-three year time  frame of directed research and development will be
                     required to identify  and measure these sources of bias and imprecision in support of
                     attainment determinations.

                     Through simulations of the CFD approach, it is feasible to analyze bias and error for
                     both temporal and spatial sources  of attainment variability. In particular, conditional
                     simulations merit additional investigation as  a  relatively unbiased approach  for
                     supporting statistical  comparisons among CFD curves. Much work remains to be
                     done  in  understanding fundamental  properties of how the CFD represents likely
                     covariances of attainment in time and space. Still, the panel finds the approach
                     feasible: one which merits additional  development, testing, and application. Indeed,
                     the CFD approach is beginning  to  attract  scientific  and management  attention
                     outside the Chesapeake Bay community.

                     As shown by  analyses in previous sections, the approach can efficiently combine
                     spatial and temporal data to support inferences on whether regions within the Chesa-
                     peake Bay attain or exceed water quality standards. On the other hand, we recognize
                     substantial bias and imprecision can occur due to small sample size, non-independ-
                     ence in temporal trends, and inadequate spatial interpolations. More work is needed
                     in analyzing these biases across regions and designated use segments.  Further,  the
                     old saw of needing more samples cannot be ignored.  In particular, the panel is opti-
                     mistic in the application of continuous spatial data streams made available through
                     the cruise-track monitoring program, and the promise of continuous temporal data
                     through  further deployment of remote sensing platforms  in the  Chesapeake Bay
                     (CBOS web site, etc). These data sets will support greater precision and accuracy in
                     both threshold and attainment determinations made through the CFD approach.

                     In classifying the CFD approach as best available science, we  seek to make several
                     important distinctions (Table 6.1). First, the CFD approach is a scientifically based
                     approach based upon its clear purpose, conceptual and design framework, empirical
                     procedures, documentation,  and intent to develop rigorous statistical and review
                     procedures (Sullivan  et al. 2006, Daubert v. Merrell Dow Pharmaceuticals, Inc.,
                     1993). That the approach permits evaluation of uncertainty also supports its classi-
                     fication as best available science  (Christman 2006). On the other hand, we do  not
                     believe that the CFD approach yet constitutes best science. Here, further analyses of
                     underlying statistical  properties of the approach (including sampling design and

-------
                                                                                              A-53
  6-1.  Evaluation of CFD approach as Best Science or Best Available Science accord-
       ing to AFS/ERF "Defining and  Implementing Best Available Science for Fisheries
       and Environmental Science, Policy, and Management" (Sullivan et al. 2006).
Attribute
Clear Objective
Conceptual
Model
Experimental
Design
Statistical Rigor
Clear
Documentation
Peer Review
Best
Science
YES
YES
NO
NO
YES
NO
Best
Available
Science
YES
YES
YES
YES
YES
YES
Current State of Development of CFD
Approach
Using biological response standards, combine
available water quality in time and space to determine
levels of attainment of Bay segments.
1 . Bay divided into functional classifications -
"Designated Uses."
2. Reference curves establish biologically
relevant threshold levels for attainment.
3. CFD combines and weights equally temporal
and spatial sources of water quality
variability.
1 . Bay segments are quasi-stratified for water
quality data collection.
2. Stratification of water quality data by
designated units does not yet occur.
3. Seasonal assessment of water quality
attainment through spatial interpolation and
the CFD approach is feasible but incompletely
developed.
1 . Procedures for quantifying uncertainty
associated with sampling design, spatial
interpolation and CFD approach are feasible
but incompletely developed.
2. Procedures for interpolating water quality data
are feasible but incompletely developed,
particularly for 3-D interpolations of
dissolved oxygen.
3. Procedures for testing inferences related to the
CFD curve are feasible but incompletely
developed.
CFD approach, water quality sampling design, and
current interpolation procedures well documented in
Chesapeake Bay Program Reports and on website.
1 . CFD approach and sampling design upon
which it is based has not been peer-reviewed
in the scientific literature.
2. This report comprises the first external review
by scientists with statistical expertise.
3. Grey literature reports produced by CBP
received expert and stakeholder input.
appendix a  »  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-54
                     interpolation elements) and vetting by outside experts is needed. Indeed, although
                     the CFD approach is beginning to get featured in scientific venues, it has not yet
                     been reviewed as part of the scientific literature. The panel sees this as an overdue
                     next step for necessary  for its acceptance, further development, evaluation, and
                     application.

                     The panel contrasted the CFD approach with existing state and jurisdictional water
                     quality criteria and attainment procedures that are based strictly upon the observed
                     sample, where site selection is not based upon probability sampling, inferences are
                     not based upon error structure,  and monitoring does not involve a scientifically
                     rational design. Indeed, standard practice for assessing compliance with water
                     quality criteria throughout the US is to sample monthly at a fixed set of stations and
                     make judgments about compliance strictly from those samples. Sampling stations
                     are typically located  for convenience (e.g., bridge overpasses), there is reluctance to
                     re-evaluate and change location (so as to maintain a time series at a fixed point), and
                     no consideration is given to representativeness of the sample for the space/time not
                     sampled. Thus the previous method used by the Chesapeake Bay Program, similar to
                     the approaches used  in other states, was  simply based on EPA assessment guidance
                     in which  all samples in a given spatial area  were  compiled and attainment was
                     assumed as long as > 10% of the samples did  not exceed the standard. In this past
                     approach all samples were assumed to be fully representative of the specified space
                     and time and were simply combined as if they were random samples from a uniform
                     population. This approach was necessary at the time because the technology was not
                     available for a more rigorous approach. But it neglected spatial and temporal patterns
                     that are known to exist in the standards measures. The CFD approach was designed
                     to better  characterize  those spatial  and  temporal  patterns  and  weight samples
                     according to the amount of space or time that they actually represent.

                     6,2 TilK CB03             ! I M,.i   P< ER

                     The panel views the CFD approach as innovative, one that has general application in
                     water quality attainment assessments, but scientific acceptance of the approach will
                     require that it is subjected to more extensive and targeted peer-review in the scien-
                     tific literature. Because the CFD is a regulatory tool, it is particularly important that
                     the approach is effectively communicated to the scientific community at large, for
                     general acceptance but more critically for the  sustained research and development
                     that the CFD, as  a nascent approach, requires. As highlighted  elsewhere, bias and
                     imprecision that  can occur  due to small sample densities,  non-independence in
                     temporal trends,  and inadequate spatial interpolations. Such  work is  novel and
                     should elicit interest  among biostatisticians as it addresses questions of both funda-
                     mental and applied consequence.

                     Although, continued working groups,  involvement through STAC  of expert biosta-
                     tisticians, and related reports such as this one will  remain important in  scientific
                     acceptance of the CFD approach, the panel recommends immediate attention in
                     subjecting the CFD to traditional peer review. One or several review papers  should
                     be submitted by CFD principals  that lay out the theory, general approach and lists
                     emergent  scientific issues to  stimulate  other  scientists  to begin  to address  such
                     issues.   Several  such papers might  be  appropriate  given  potential interest by
   appendix a  * The Cumulative hrequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                             A-55
biostatisticians and environmental and regulatory scientists. Scientific interest will
also be garnered by public and stakeholder interest. The CFD approach here presents
a challenge as it is complex in explanation. Still with careful diagrams and examples,
a brochure on the CFD approach should be extremely useful in getting uninitiated
scientists and stakeholders on the same page.

6.3. BIOLOGICAL REFERENCE CURVES

The success of the CFD-based assessment will be dependent upon decision rules
related to the biological reference curves. These curves represent desired segment-
designated use water quality  outcomes and reflect sources of acceptable natural
variability. The reference and attainment curves follow the same general approach in
derivation—water quality data collection, spatial interpolation,  comparison to
biologically-based water quality criteria, and combination of space-time attainment
data through a CFD. Therefore, the biological reference curve allows for implemen-
tation of threshold uncertainty as long as the reference curve is sampled similarly to
the attainment curve. Bias and uncertainty are driven in CFD curves by sample
densities in time and space. Therefore, we  advise that similar sample densities are
used in the derivation of attainment and reference curves. As this is not always
feasible, analytical methods are needed in the future to equally weight sampling
densities between attainment and reference  curves.

Conceptually, the CFD approach builds on the underlying view that water quality
criteria are surrogates for Designated Uses (regions that define ecosystem function).
Implicit is  a bottom up model based upon eutrophication, which is  expected to
diminish the designated use. Reference curves represent thresholds related to the
functioning of designated use regions. Therefore, choice of reference regions or
periods  and sampling design in developing  reference curve is critical to the imple-
mentation of a scientifically-rigorous CFD approach.  Choice of such regions is
beyond the scope of this review, but we emphasize several relevant statistical issues
in developing reference curves in Section 4.
  7.0  RECOMMENDATIONS  FOR FUTURE  EVALUATION
                    AND  REFINEMENT  OF  THE
              CFD ASSESSMENT METHODOLOGY

As part of its conclusions, the STAC CFD Review Panel identified critical remaining
issues that need resolution in the near future. The following is a list of critical aspects
of that needed research. These research tasks appear roughly in order of priority.
However, it must be recognized that it is difficult to formulate as set of tasks that can
proceed with complete  independence.  For example, research on task 1 may show
that the ability to conditionally  simulate the water quality surface is critical to
resolving the sample size bias issue. This discovery might eliminate IDW as a choice
of interpolation under task 3. The Panel has made significant progress on several of
these research tasks and CBP is encouraged to implement continued study in a way
that maintains the momentum established by this research group (Table 7.1.).
   appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-56
                     7.1. Research Tasks, examples of specific subtasks, and suggested time frame for
                         continued CFD research.
                 Task
Schedule
                 1. Effects of Sampling Design on CFD Results

                       (a) Continue simulation work to evaluate CFD bias reduction
                 via conditional simulation.
                       (b) Investigate conditional simulation for interpolation
                 methods other than kriging - this may lead to more simulation work.
                       (c) Implement and apply interpolation with condition
                 simulation on CBP data.
2006-2008
                 2. Statistical inference framework for the CFD

                       (a) Implement and evaluate confidence interval procedures.
                       (b) Conduct confidence interval coverage experiments.
                       (c) Investigate confidence interval methods for non-kriging
                 interpolation methods.
                	(c) Implement and evaluate confidence interval procedures.
2006-2008
                 3. Choice of Interpolation Method

                        (a) continue to investigate other more nonparametric
                 interpolation methods (e.g. loess and splines).
                       (b) implement a file system and software utilizing the "best"
                 interpolation for CBP data.
                       (b) compare interpolations and CFD's based on IDW and
                 "best" method.
2006-2008
                 4. Three-Dimensional Interpolation

                       (a) Implement 2-D kriging in layers to compare to current
                 approach of 2-D IDW in layers.
                       (b) Conduct studies of 3-D anisotrophy in CBP data.
                       (c) Investigate software for full 3-D interpolation. Examples
                 of options include:  custom IDW software, custom kriging software
                 using GMS routines, custom kriging software using the R-package, or
                 some other off the shelf product.	
2007-2009
                 5. High Density Temporal Data

                       (a) Develop methods to use these data to improve temporal
                 aspect of CFD in current implementation.
                	(b) Investigate feasibility of 4-dimensional interpolation.
2008-2010
   appendix a

-------
                                                                                           A-57
1. Effects of Sampling Design on CFD Results. The CFD is a special case of an
  unbiased estimator for a cumulative distribution function of a population. Like
  the cumulative distribution function, the CFD is a function of the mean and the
  variance of the population being assessed. And the better the mean and vari-
  ance are characterized with sample data, the more accurate the  shape  of the
  CFD will be. As the sampling density increases, the estimated CFD begins to
  approach the true CFD. However, if the sampling density is low, then sampling
  error could become important and there is potential that it could affect the
  shape of the CFD and ultimately the accuracy of the compliance assessment.
  Furthermore the potential for the sample size to affect the shape could create a
  compliance assessment bias if the reference curve and assessment curve are
  based on different sampling densities. Conditional simulation methods  devel-
  oped by STAC panel members showed promise toward resolving these  issues
  and mitigating potential biases caused by differences in sample size.
2. Statistical inference framework for the CFD. It is important in a regulatory
  process to differentiate an exceedance that is small and might have resulted
  from chance variability from those that are large and indicative of an inherent
  problem. This differentiation will require mathematical tools to quantify the
  variability in the CFD that occurs as a  result of sampling. The  STAC panel
  made progress on this issue by demonstrating a confidence interval procedure
  based  on conditional simulation  associated with kriging. It  remains  to be
  assessed whether or not  confidence intervals  produced by  this algorithm
  perform at the nominal level of coverage, fore example, does a nominally 95%
  CFD confidence interval cover the true CFD 95% of the time.
3. Choice of Interpolation Method. The STAC panel considered several inter-
  polation methods and outlined the features of each. Those features illustrate
  tradeoffs between ease  of  implementation and maximizing the information
  garnered from the data. Further work is needed to compare the features to the
  requirements  of  wide-scale implementation of assessment procedures  and
  formulate a plan  for tractable implementation that results in credible assess-
  ments. One strategy is to implement easily performed analysis (e.g. IDW) as a
  screening tool to identify cases where compliance / non-compliance is clear,
  and then implement more  labor intensive  methods (e.g. kriging) for cases
  where  compliance  is more difficult to  resolve.  One  difficulty  with imple-
  menting a full comparison of methods is that implementation of  each method
  requires considerable work in terms of setting up file systems, interfacing soft-
  ware and data, and coupling the considerable bathymetry data of the bay. Thus
  it would be prudent to narrow the choices based on theoretical considerations
  where  possible.
4. Three-Dimensional Interpolation.  Assessments  of  the  dissolved oxygen
  criteria require three-dimensional interpolation. However, the field of  three-
  dimensional interpolation is not as highly developed as that of two-dimensional
  interpolation. While the mathematics of each method extend  easily to three
  dimensions, there are relatively few examples of 3-D interpolation available in
  the literature and issues such as data density requirements for reliable results
  are not well studied. Efforts are needed to further evaluate research in  three-
  dimensional interpolation and  seek  additional  outside  scientific  input  and
 appendix a  • The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-58
                           review with the goal of implementing the best available technology for this
                           aspect of criteria assessment. One of the first efforts under this task is a study
                           of the 3-D variance stucture of the data to be interpolated. A short term option
                           is  to implement the optimal 2-D interpolator in layers as is done with the
                           current IDW interpolator.
                        5. High Density Temporal Data. As currently formulated, assessment for most
                           of the open-waters of the Bay are based on "snapshots" in time of the spatial
                           extent of criteria exceedence estimated via interpolation. Data collected for use
                           in interpolation are actually spaced over multiple days due to the large expanse
                           over which sampling must be conducted. It is clear that technology is becoming
                           available that will produce high density data in both space and time. Interpola-
                           tion should accommodate data that are collected densely in space. However, it
                           is unclear how the CFD process will accommodate data that are high density in
                           time. Further work is needed to evaluate methods to fully utilize the temporally
                           intensive data that is currently being collected.

                      The panel discussed several mechanisms for the CBP to make progress on  chal-
                      lenging tasks ahead (Table 7.1). We recommend that a review panel oversee the tasks
                      over the next 3-5 year time frame. This panel would periodically review trials and
                      other  products conducted by individual external scientists (academic scientists or
                      consultants) and  existing teams of CBP scientists  (e.g., the Criteria Assessment
                      Protocols (CAP) workgroup). Tasks 1 and 2 are most immediate and critical and we
                      recommend that these tasks by contracted out to external scientists, exploiting state-
                      of-the-art approaches and knowledge.  Task 3 could be conducted through CAP or
                      other group of CBP scientists. Task 4 and 5 are less immediate but again will require
                      substantial expertise and innovation and may be most efficiently accomplished by
                      scientific expertise outside the immediate  CBP community.
                      Christensen OF, Diggle PJ, Ribeiro PJ. 2001. Analysing positive-valued spatial data:  the
                      transformed Gaussian model. In: Monestiez P, Allard D, and Froidevaux, editors. GeoENV
                      III - Geostatistics for environmental applications. Quantitative Geology and Geostatistics.
                      Dordrecht (Netherlands): Kluwer Academic Publishers. 11:287-298.
                      Christman  MC. 2006. The  characterization and incorporation of uncertainty in fisheries
                      management, In Fisheries Ecosystem Planning for Chesapeake Bay. Bethesda (MD): Amer-
                      ican Fisheries Society. (In press).
                      Cressie N.  1989. The Many Faces of Spatial Prediction. Mathematical Geology 1:163- 176.
                      Cressie N.  1991. Statistics for Spatial Data. New York: Wiley. 928 p.
                      Curriero FC. 2006. On the Use of Non-Euclidean Distance Measures in Geostatistics. Math-
                      ematical Geology (in press).
                      Daubert v. Merrell Dow Pharmaceuticals. Inc. 1993. 509 U. S. Supreme Court. 579.
                      Deutsch CV. 1984. Kriging with Strings of Data. Mathematical Geology. 26:623-638.
                      Diggle PJ, Tawn JA, Moyeed RA. 1998. Model Based Geostatistics (with Discussion).
                      Applied Statistics 47:299-350.
                      Diggle PJ, Ribeiro PJ. 2006. Model-based Geostatistics. New York: Springer. 230 p.
   appendix a  *  The Cumulative hrequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                                     A-59
Dille JA. 2003. How good is your weed map? A comparison of spatial interpolators. Weed
Science 51:44-55.
Gotway, CA. 1994. The Use of Conditional Simulation in Nuclear Waste-Site Performance
Assessment. Technometrics 36:2:129-141.
Hastie TJ, Tibshirani RJ. 1990. Generalized Additive Models. New York: Chapman and Hall
p335.
Jensen OP, Christman MC, Miller TJ. 2006. Landscape-based geostatistics: A case study of
the distribution of blue crab  in Chesapeake Bay. Environmetrics 17:605-621.
Journel A. 1974. Geostatistics for conditional simulation of ore bodies. Economic Geology
69:673-687.
Kitanidis PK. 1997.  Introduction to Geostatistics: Applications in Hydrogeology. New York:
Cambridge University Press. 271 p.
Kravchenko AN. 2003. Influence of Spatial Structure on Accuracy of Interpolation Methods.
Soil Sci. Soc. Am. J. 67:1564-1571.
Kutner MH, Nachtsheim CJ, Neter J, Li W. 2004. Applied Linear Statistical  Models, 5th
edition. Boston: McGraw-Hill.
Laslett GM.  1994. Kriging and splines: an empirical comparison of their predictive perform-
ance in some applications. JASA 89:391-400.
Lloyd CD. 2005. Assessing the effect of integrating  elevation data into the estimation of
monthly precipitation in Great Britain. Journal of Hydrology 308:128-150.
Ouyang Y, Zhang JE,  Ou LT. 2006. Temporal and spatial distributions of sediment total
organic carbon in an estuary river. J. Environmental Quality. 35:93-100
Reinstorf F,Binder M, Schirmer M, Grimm-Strele J, Walther W. 2005. Comparative assess-
ment of regionalization methods of monitored atmospheric  deposition loads. Atmospheric
Environment 39:3661-3674.
Ribeiro PJ, Diggle PJ. 2001. geoR: A package for geostatistical analysis. R News 1:2:15-18.
Schabenberger O, Gotway CA. 2005. Statistical Methods for Spatial Data Analysis.
Boca Raton, FL:  Chapman and Hall/CRC Press. 512 p.
Spokas K, Graff C, Morcet M, Aran C. 2003. Implications of the spatial variability of land-
fill emission rates on geospatial analyses. Waste Management 23:599-607.
Sullivan PJ, Acheson J, Angermeier PL, Faast T, Flemma J, Jones CM, Knudsen EE, Minello
TJ, Secor DH, Wunderlich R, Zanatell BA. 2006. Defining and implementing best available
science for fisheries and environmental science, policy and management. Bethesda, Md:
American Fisheries  Society and Port Republic, Md:  Estuarine Research Federation. Port
Republic, Maryland, (available: www.fisheries.org/AFSniontana/AFS.ERF.BestScience.pdf)
Tomczak,  M. 1998. Spatial interpolation and its uncertainty using automated anisotropic
inverse distance weighting (IDW) - cross-validation/jackknife approach. J. Geog. Infor. and
Decision Analysis 2:18-30.
[EPA] Environmental Protection Agency (US). 2003.  Technical support documentation for
identification of Chesapeake Bay designated uses and attainability. Annapolis (MD): EPA.
177 p. EPA 903-R-03-002. (Available: http://www.chesapeakebay.net/search/pubs.htm)
Valley RD,Drake MT, Anderson CS. 2005. Evaluation of alternative interpolation techniques
for the mapping of remotely-sensed submersed vegetation abundance. Aquatic Botany 81:13-
25.
    appendix a  •  The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
A-60
                      Ver Hoef JM, Peterson, E, and Theobald, D, 2007, Spatial Statistical Models that Use Flow
                      and Stream Distance, Environmental and Ecological Statistics (in press).
                      Wahba, G. 1990. Spline Models for Observational Data. Philadelphia (PA): Society for
                      Industrial and Applied Mathematics. 169 p.
                      Wang XJ, Liu RM. 2005. Spatial analysis and eutrophication assessment for chlorophyll a in
                      Taihu Lake. Environmental Monitoring and Assessment 101:167-174.
                      Zimmerman D., Pavlik C, Ruggles A, Armstrong MR 1999. An experimental comparison of
                      ordinary and universal kriging and inverse distance weighting. Mathematical Geol. 31:375-
                      390.
   appendix a  » The Cumulative Frequency Diagram Method for Determining Water Quality Attainment

-------
                                                                                    B-1
                       appendix
 Detailed Chesapeake  Bay  Water
     Quality Criteria Assessment
                    Methodology
The methods in this appendix apply specifically to the  evaluation of dissolved
oxygen criteria. For water clarity criteria or chlorophyll a criteria evaluations, the
individual methods are very similar to those described here. See chapters 5 and 6,
respectively, for additional details. Chapter 7 also contains important information in
using shallow-water data for criteria attainment assessment of all three parameters.

Data come from the Chesapeake Bay Program's Chesapeake Information Manage-
ment System (CIMS) database or through the CIMS partners' networked databases.
The parameters extracted include date, location, depth, salinity, temperature, and the
water quality parameter under assessment. Data identified by the states, but collected
from other than the Chesapeake Bay Water Quality Monitoring Program and Chesa-
peake Bay Shallow-water Monitoring Program, are also obtained. These data must
be of known and documented quality as described in Chapter 3.

Once the data are compiled, they are assigned to a time period based on the sample
date. Fixed-station data are normally collected during a monitoring cruise that covers
the entire tidal Chesapeake Bay over several days.  To provide a "snapshot" of water
quality, however, the data collected within one cruise are  considered contempora-
neous to enable a single spatial interpolation. For information not associated with a
cruise, such as state-supplied data, a cruise number is assigned representing the
closest cruise in time to the collection of each data point. Co-located data points in
the same cruise are averaged.

The criteria assessment procedure requires evaluation over large areas rather than at
distinct points. Spatial interpolation is carried out for each water quality criteria
parameter for each cruise (see Appendix D for details on the Chesapeake Bay inter-
polator and  the  interpolation process) with water clarity  and  chlorophyll  a data
interpolated  in the two horizontal  dimensions  using inverse distance squared
weighting and natural logarithm transformation.  Dissolved oxygen data are first
             appendix b  • Detailed Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
B-2
                    linearly interpolated vertically within each column of observed data beginning at
                    0.5 meters below the water surface and continuing at one-meter intervals, without
                    exceeding the deepest observation in that water column. Data at each depth is then
                    interpolated horizontally using inverse distance squared weighting. Data regions
                    were specified for each segment to prevent the interpolation algorithm from using
                    data points in neighboring tidal tributaries (described in the section below and in
                    detail in Appendix D).

                    Some designated uses for dissolved oxygen during the summer in the Chesapeake
                    Bay and its tidal tributaries and embayments are  defined vertically to distinguish
                    stable water layers with different criteria levels (U.S. EPA 2003a, 2003b). In areas
                    and seasons for which vertical stratified criteria apply, the surface mixed layer (open
                    water) is that layer above the pycnocline and, thus, exposed to the atmosphere. The
                    transitional middle layer (deep water) is the layer between the upper and lower pycn-
                    ocline boundaries. The lower layer (deep channel) is the  water below the lower
                    pycnocline boundary. Given that the pycnocline is dynamic and moves up and down
                    with each monitoring cruise, the designated use of each interpolator grid cell must
                    also be defined based on the data for each cruise.

                    Temperature and salinity are used to calculate density; density, in turn, is used to
                    calculate pycnocline boundaries. Density is calculated using the method described in
                    Algorithms for Computation  of Fundamental Properties of Seawater1.  For each
                    column of temperature and salinity data, the upper and lower pycnocline boundaries
                    are determined by looking for the shallowest robust vertical change in density of
                    0.1 kg/m3/m for the upper boundary and the deepest change of 0.2 kg/m3/m for the
                    lower boundary. To be considered robust, the density gradient must not reverse direc-
                    tion at the subsequent measurement and must also  demonstrate a change in salinity
                    of at least 0.1 psu per meter (not merely a change in temperature). Chapter 7 in U.S.
                    EPA 2004, pages 85-87, documents the detailed method for determination of both
                    the vertical density profile and the pycnocline.

                    The depths to the upper pycnocline boundary (where detected) and the fraction of
                    the water column below the lower boundary are interpolated in two dimensions. If
                    no  lower  boundary was detected, then the fraction is  set at zero. The depth to the
                    upper  pycnocline boundary tends to remain  stable in the horizontal  dimension,
                    meaning that spatial definition of that boundary using interpolation generally works
                    well. Interpolation of the lower boundary is more  complicated because the results
                    may conflict with the upper boundary definition or with the actual bathymetry of the
                    Chesapeake Bay. Consequently, interpolation of the lower boundary is based on the
                    fraction of water column depth. In this  way, the constraints of the upper pycnocline
                    boundary definition and the actual Bay bottom depth are imposed, eliminating errors
                    related to boundary conflicts.
                     Endorsed by UNESCO/SCOR/ ICES/IAPSO Joint Panel on Oceanographic Tables and Standards and
                     SCOR Working Group 51. N.P. Fofonoff, and R.C. Millard, Jr., 1983. UNESCO Technical Papers in
                     Marine Science. Paris, France. No. 44, p. 53.
  appendix b  »  Detailed Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                                B-3
Criteria assessments are based on each component criterion's specific averaging
period. Assessments of attainment of the instantaneous minimum criteria are directly
evaluated using the individual cruise interpolations. All 30-day mean criteria assess-
ments  rely  on monthly averages of  interpolated  data  sets. To  calculate  these
averages, each interpolated cruise within a month is averaged on a point-by-point
basis in  matching interpolator grid  cells.  Generally, two cruises per month run
through the warm season with one cruise per month during the cooler period. Spatial
violation rates are calculated  for each temporally aggregated interpolation  in an
assessment period. For example, the 12 monthly average interpolations representing
the four summer months (June, July, August, September) over three years were used
for a three-year summer open-water dissolved oxygen assessment.

Cumulative frequency diagrams (CFD)  are generated from the spatial violation rates
for each assessed designated use, water quality parameter, criterion, and averaging
period using the Weibull plotting position (rank/(n+l)).

The  assessment CFD is compared to a reference CFD to determine if unallowable
exceedances of the criterion occur. The diagrams of both CFDs show three  areas:
non-exceedance (above the assessment curve), allowable exceedance (below both
curves), and unallowable exceedance (below the assessment curve and above the
reference curve). If the assessment CFD surpasses the reference CFD at any  point,
an unallowable exceedance exists.

Reference CFDs are continuous or generally have many more points than assessment
CFDs.  This situation can lead to spurious unallowable exceedances  even without
individual points in the assessment CFD topping reference CFD levels. To address
this problem, reference curves are evaluated only at the temporal axis points  in the
assessment curve (see Figure II-7 in Chapter 2). For non-continuous biological refer-
ence curves, these points are interpolated from neighboring points.

The trapezoidal rule is used to calculate the areas. This rule is a method of approxi-
mate integration, which calculates the areas of discrete trapezoids that make up the
area below a curve when summed. Since both the assessment and reference curves
are piecewise linear,  repeated application of the trapezoidal rule results in an  exact,
rather than approximate, value.

For dissolved oxygen criteria assessed without reference curves, the assessment
space is divided in two—non-exceedance and unallowable exceedance.
U.S. Environmental Protection Agency (U.S. EPA). 2003a. Ambient Water Quality Criteria for
Dissolved Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and its Tidal Trib-
utaries. EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.

U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
Bay Designated Uses and Attainability. EPA 903-R-03-004. Region III  Chesapeake Bay
Program Office Annapolis, MD.
               appendix  b  * Detailed Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
B-4
                     U.S. Environmental Protection Agency. 2004. Ambient Water Quality Criteria for Dissolved
                     Oxygen, Water Clarity and Chlorophyll a Chesapeake Bay and its Tidal Tributaries: 2004
                     Addendum. EPA 903-R-04-005. Region III Chesapeake Bay Program Office, Annapolis, MD.
  appendix b «  Detailed Chesapeake Bay Water Quality Criteria Assessment Methodology

-------
                                                                                       C-1
                        appendix
        Evaluation  of  Options  for
             Spatial  Interpolation
Interpolation constitutes a critical element of CFD-based assessment methodology.
It provides the spatial framework for data integration while allotting the appropriate
weight to all data. The spatial framework consists of a grid made up of a network of
cells that vary in size to cover the entire spatial domain. The size of the cells deter-
mines the scale of the assessment; smaller and more numerous cells in a given area
provide a more spatially detailed assessment. Estimates for all cells come from a
spatial interpolation algorithm.

To date, two spatial interpolation algorithms have been considered: inverse distance
weighting (IDW) and kriging. In IDW, estimates of water quality levels are based on
a weighted average derived from the closest measured data values. Weights depend
upon the distance between the measurement point and the cell being estimated. Thus,
measurements from the closest points are weighted most heavily and have the most
influence. The second method is kriging—a well-known statistical form of spatial
interpolation. The statistical details of kriging rest on ample research. This method,
however, has not been used for water quality criteria. Both spatial algorithm methods
can prove valuable for Chesapeake Bay water  quality criteria assessment; one or
both will likely be used in the future. Other methods (non-parametric regression
methods such as Loess regression or cubic splines) are also available and could also
be considered for future use. Further details on the IDW and kriging methods are
provided below.
  SPATIAL INTERPOLATION NEEDS SPECIFIC TO CHESAPEAKE
          BAY WATER QUALITY CRITERIA ASSESSMENT

The Chesapeake Bay water quality criteria were established using the spatial defini-
tion of designated-use areas for the tidal  waters of Chesapeake Bay (U.S. EPA
2003a, 2003b).  These spatial definitions, along with the characteristics of the Bay
itself, present several challenges for spatial interpolation. For example, the Chesa-
peake  Bay shoreline is  extremely complex  with  many small  tidal tributaries,
embayments, and inlets that occur at various scales throughout the water body. The
small inlets present a challenge for spatial  interpolation because they  require
                                    appendix c  • Evaluation of Options for Spatial Interpolation

-------
C-2
                     extrapolation from measured areas into unmeasured areas, often around numerous
                     bends and twists in a tidal river. Furthermore, they create the potential for interpo-
                     lating from one tidal tributary to another, which may be inappropriate since tidal
                     tributaries are often hydrodynamically independent. Most spatial interpolation algo-
                     rithms  operate in  two dimensions in a relatively simple spatial domain. Thus,
                     specific refinements need to be made for the algorithms used in Chesapeake Bay
                     criteria assessment.

                     The Chesapeake Bay dissolved oxygen criteria depend on designated-use areas—
                     specific volumetric areas with both vertical and horizontal dimensions (U.S. EPA
                     2003a, 2003b). Dissolved oxygen levels are naturally lower in bottom waters. There-
                     fore, the  designated-use areas were defined as vertically  stratified layers to allow
                     establishment of criteria levels that support the ecological communities residing in
                     the lower depths of the Bay. Any spatial interpolation supporting dissolved oxygen
                     criteria assessment must allow interpolation throughout the designated-use volumes
                     in three dimensions. The IDW algorithm developed and used by the Chesapeake Bay
                     Program  was designed in  this way  and has been used consistently to  provide
                     baywide  maps of  dissolved oxygen concentrations  (see Appendix D). Kriging,
                     however,  has not been used for three-dimensional interpolation in the  Chesapeake
                     Bay to date; in fact, only limited research has taken place to develop the capability
                     of three-dimensional  kriging for any purpose (STAC 2006). Thus, more research
                     may be required for the use of kriging in the assessment of dissolved oxygen criteria.

                     The complexity  of the Chesapeake Bay  shoreline presents  several obstacles  for
                     spatial interpolation in Bay tidal waters, mostly related to interpolating across land
                     area. Most spatial interpolation algorithms assume a relatively simple spatial domain
                     (e.g., rectangular)  and interpolation takes place without regard to direction. In
                     contrast, the Chesapeake Bay (for example, see Figure III-l in Chapter 3)  displays
                     tidal  flow patterns that make some locations  independent or virtually independent.
                     For  Bay water quality  criteria assessment, therefore, the influence between some
                     locations  must be limited when interpolating spatially. The current Chesapeake Bay
                     Program interpolator provides limits by using data regions in which the  data used to
                     estimate values in given locations are limited to certain areas (see Appendix D  for
                     additional details).  Similar or alternative methods may be required to apply kriging
                     broadly.

                     As described  above, the Chesapeake Bay Program collects two types of data  for
                     criteria assessment; these  two data  types  supply information at different spatial
                     scales.  The fixed-station  Chesapeake  Bay  Water Quality  Monitoring  Program
                     collects data consistently for the entire Bay as well as its tidal tributaries and embay-
                     ments. The Chesapeake Bay Shallow-water Monitoring Program offers much more
                     detailed information within Bay tidal tributaries and across all shallow-water habi-
                     tats.  Given the different spatial scales of these two monitoring programs, it is
                     unlikely that they can be used in the same interpolations. Thus, two separate inter-
                     polation  approaches—each designed for specific  types of criteria attainment
                     assessments—may prove necessary.

                     Since the Chesapeake Bay water quality criteria and the CFD-based criteria assess-
                     ment methodology were developed and published, interest has developed in creating
  appendix c  •  E<

-------
                                                                                               C-3
a statistical basis for decision-making using the CFD (see pages 164-165 in U.S.
EPA 2003a).  Such a basis would allow the incorporation of error analysis into the
criteria attainment assessment methodology.  It would also allow the differentiation
of an assessment based on a well-characterized system from one that was poorly
characterized. Estimates of interpolation error are important to develop such a statis-
tical framework. Such estimates allow decision-making to be based on the number
(density) of sampling locations and promote greater statistical certainty (i.e., greater
sampling density) in the assessment. The current Chesapeake Bay interpolation algo-
rithm does not yield spatial error  estimates (Appendix D); however, kriging is a
possible  alternative algorithm  that can provide spatial  interpolation  error (STAC
2006).

Chesapeake Bay spatial interpolation requires the potential  for automation.  For
many reasons, the Chesapeake Bay  Program must compute many interpolations
quickly. In developing the attainment figures for the 2006 listing cycle, for example,
the program performed a total of 2328 interpolations for the final criteria assessment
analysis of the 95 water quality segments). During development of the methodology,
these interpolations were carried out repeatedly. Also, water quality models are often
used to evaluate the potential benefits of management actions with the generation of
multiple  scenarios. Management action success is often defined in terms of water
quality criteria, with results evaluated similarly to the actual  measurements. Given
the large number of data sets, automating the criteria assessment methodology and
spatial interpolations would likely prove necessary. The current Chesapeake Bay
interpolator allows automation and has been used in this way (Appendix D). Kriging,
however, is a more detailed analysis that requires multiple decisions along the way,
is not conducive to automation, and may not necessarily remain consistent within
and between jurisdictions.
                          IO              (
                                       TV
As stated, the Chesapeake Bay Program redesigned the tidal monitoring program
specifically to support water quality criteria assessment. That redesign resulted in
multiple monitoring program components, all of which address one or more of the
objectives of the Chesapeake Bay Water Quality Monitoring Program. Two of the
components that serve most of the current needs of criteria assessment include the
Bay wide Fixed-station Water Quality Monitoring Program and the Shallow-water
Monitoring  Program. These two long-term  efforts  will provide  data useful at
different scales.

The fixed-station monitoring program began in the mid 1980s and was designed to
provide data for assessing long-term trends at key sites throughout the Chesapeake
Bay and its tidal tributaries (Chesapeake Bay Program 1989). The program collects
water quality samples at more than 150 sites (Figure C-l), including 49 stations in
the mainstem Chesapeake Bay and 96 stations in the  tidal tributaries. The samples
go to a network of laboratories for analysis, compiling data on 19 water quality
parameters.  Fixed-station monitoring cruises run on a monthly  basis throughout
                                       appendix c  «  Evaluation of Options for Spatial Interpolation

-------
C-4
         "
                       '*»»/*# \
                   **
                    *
                              ''
                                't
                                       >
         f
                                most of the year, but occur two times  a month
                                during the summer. At each station, samples are
                                collected at multiple  depths  depending on the
                                location of the pycnocline. In addition, techni-
                                cians  collect  water  quality  sensor  data—
                                including  water temperature,  salinity,  and
                                dissolved oxygen—along  vertical  profiles  at
                                regular intervals.

                                The fixed-station network provides data to assess
                                water quality in the mid-channel, open waters  of
                                the Bay mainstem as  well as in the major tidal
                                tributaries and embayments. The network does
                                not assess conditions in the shallows since many
                                of the stations were  purposely located in the
                                main channels and open tidal waters.

                                The Chesapeake  Bay Program recently  began
                                monitoring shallow-water habitats using a tech-
                                nology known as DataFlow (see  Chapter 7 for
                                details). This new technology uses a system  of
                                shipboard water  quality probes  that  measure
                                spatial position, water depth, water temperature,
                                salinity, dissolved oxygen, turbidity, and fluores-
                                cence from a  flow-through stream  of water
                                collected near the water surface. This system
                                allows rapid  data  collection  (approximately
                                every 4 seconds) while the boat  is traveling  at
                                speeds up to 20 knots. Due to the speed of data
                                collection,   each cruise  provides  extremely
                                detailed data sets useful for assessing highly
                                variable water quality conditions, such as those
                                expected in the Bay's shallow waters and small
                                tidal tributaries. Thus, this  monitoring  program
                                specifically assesses shallow  waters  (STAC
                                2005). The spatial density  of data collected by
                                the DataFlow system allows spatial interpolation.
The current Chesapeake Bay Program interpolation software is  not designed for data
of this density, however, so new methods of interpolation need to be developed.

Due to the cost of the Shallow-water Monitoring Program, it cannot be implemented
baywide concurrently. Rather, the program is  being put into practice on a rotating
basis, with the monitoring system deployed to selected assessment units long enough
to evaluate attainment  and then moved to another set of units (see  Chapter 7 for
further details). This set-up means that all shallow-water areas will not be assessed
simultaneously, although a full assessment will take place over time. For example,
the Maryland Department of Natural Resources' Water Quality  Mapping Program
covered 14 Chesapeake Bay and tributary systems in 2005.These systems include the
St. Mary's,  Patuxent, West, Rhode, South, Middle, Bush, Gunpowder,  Chester,
                                *<•
       -j i -' . The sites that make up the fixed-station
  network of the Chesapeake Bay Water Quality
  Monitoring Program.
  Source: Chesapeake Bay Program 1989.
  appendix c  •  Evaluation of Options for Spatial Interpolation

-------
                                                                                              C-5
Eastern Bay, Miles/Wye, Little Choptank, Chicamacomico, and Transquaking rivers.
In Virginia, Dataflow data are available for the Piankatank, York, Pamunkey, and
Mattaponi rivers. Chapter 7 discusses additional details on plans for monitoring
shallow-water systems.

Other alternative monitoring programs have been considered, but not fully imple-
mented for criteria  assessment. Beginning in 1990, chlorophyll a concentrations
have been measured over the mainstem Chesapeake using aircraft remote sensing
(Harding et al. 1992). Twenty-five to 30 flights per year took place during the most
productive time periods. In addition, satellite remote sensing data have been consid-
ered for evaluating chlorophyll a  concentrations in the Bay (Harding et al. 2004)
although no detailed evaluation of the feasibility has been completed. Water quality
sensors and data loggers mounted on buoys have  also been evaluated as the best
means to assess high-frequency dissolved oxygen criteria. This option is expensive,
however, and only  a limited (but growing) number of buoy systems  have been
deployed to date (http://www.cbos.org).
           INTERPOLATION METHODS CURRENTLY
      USED  FOR CHESAPEAKE  BAY WATER  QUALITY
                      CRITERIA ASSESSMENT

The current Chesapeake Bay Interpolator is a grid-based algorithm in which criteria
measurement data are used to estimate values for all grid cells (see Appendix D for
a detailed description). Estimates for cell locations are computed by interpolating the
nearest "n" neighboring water quality measurements for which "n" is normally 4 but
is adjustable. The interpolation uses an inverse distance weighted (IDW) algorithm
in which the estimated value of each grid cell is based on the four nearest measure-
ments. Each of the neighboring points is weighted by the inverse of the distance
squared (i.e., 1  d~2), however, so close neighbors have more influence than those
farther away.

The cell  size in the Chesapeake Bay interpolation grid is 1 km (east-west) x 1  km
(north-south) x 1 m (vertical),  with  columns of cells extending from the water
surface to the Bay bottom representing the three-dimensional volume as a group of
equal-sized  cells. Each tidal tributary  is  represented by  variously sized cells
depending on the river's geometry since the  narrow upstream portions require
smaller cells to model the dimensions accurately. Interpolator grid cells, however,
remain the same size within individual segments. This designation results in a total
of 51,839  cells by  depth  for  the  mainstem  Chesapeake  Bay  (segments
CB1TF-CB8PH), and a total of 238,669 cells by depth for all 78 segments making
up the mainstem Chesapeake Bay and its tidal tributaries and embayments.

The Chesapeake Bay interpolator is optimized to compute concentration values that
closely reflect the physics of stratified water bodies such as the Bay. Water quality
varies much more markedly vertically as opposed to horizontally. To accommodate
this attribute, each column of data is interpolated vertically to the same depths as the
centroids of the interpolator cells, (i.e. 0.5,1.5, 2.5 meters, etc). The interpolator then
interpolates only in the horizontal dimension.
                                       appendix c  •  Evaluation of Options for Spatial Interpolation

-------
C-6
                    Up to four points are used for interpolation. If fewer than four points exist, interpo-
                    lation is still carried out given at least one measured point. Without any measured
                    data, a missing value (normally a -9) is calculated for that cell. A search radius filter
                    limits the horizontal distance of monitoring data from the cell being computed. Data
                    points outside the user-selected radius (normally 25,000 m or 25 km) are excluded
                    from calculation. This filter ensures that only data near the location being interpo-
                    lated are used.

                    Segment and region filters have also been added.  Segments are aggregations of the
                    interpolator cells. For instance, eight segments make up the mainstem Chesapeake
                    Bay (CB1TF, CB2OH,.. .CB8PH). The tidal tributaries have 70 additional segments,
                    created by the Chesapeake Bay  Program's 2003  segmentation scheme (U.S. EPA
                    2004, 2005). These segments divide the Bay into geographic areas with somewhat
                    homogeneous environmental conditions. This segmentation also allows the reporting
                    of results on a segment basis, revealing more localized  changes compared to  the
                    whole Bay ecosystem.

                    The region file identifies the geographic  boundary that limits which monitoring
                    station data are included in interpolation for a given segment (see Appendix D). The
                    purpose of the data region is to select a subset of the monitoring data from the input
                    data  file and to use that subset  for computing the values for each grid cell in a
                    segment.  Use of data regions ensures that the interpolator does not "reach across
                    land" to obtain data from an adjacent tidal tributary—a process that would give erro-
                    neous results. By using data regions, each segment of grid cells can be computed
                    from its individual monitoring data subset. Each adjacent data region overlaps so that
                    a continuous gradient—not a seam—exists  across  segment boundaries. Data regions
                    for criteria assessment vary somewhat from the data regions in the standard interpo-
                    lator. These new regions were developed to exclude tributary measurements from
                    mainstem interpolations and to include additional observed data from Virginia.
                             EVALUATION  OF THE  INVERSE  DISTANCE
                                WEIGHTING  SPATIAL  INTERPOLATION
                         ALGORITHM FOR ASSESSING CHESAPEAKE  BAY
                                       WATER QUALITY  CRITERIA

                    The current  Chesapeake Bay interpolator is  based on an IDW algorithm—a non-
                    statistical spatial interpolator that uses observed data to calculate a weighted average
                    (as a predicted value) for each location on the prediction grid (Appendix D). The
                    method calculates the weight associated with a given observation as the inverse of
                    the square of the distance between the prediction location and the observation. The
                    IDW is a spatial interpolator; in general, such methods have provided good predic-
                    tion maps (STAC 2006). Additionally, implementation is  relatively simple  since
                    software exists to map IDW automatically. Further, the method does not require any
                    decisions during  an interpolation session. Commercial Geographic  Information
                    Systems (GIS) software contains IDW, requiring only GIS skills for application.

                    The IDW algorithm has several advantages  for use in Chesapeake Bay water quality
                    criteria attainment assessment  (STAC 2006). First,  since it is non-statistical,  the
  appendix c •  Evaluation of Options for Spatial Interpolation

-------
                                                                                               C-7
algorithm is not constrained by prior theoretical assumptions concerning error struc-
ture. It is, therefore, simpler mathematically and can be adapted to interpolation in
three dimensions (i.e., with depth). Second, due to its simplicity,  IDW does not
require operator decisions at interim steps. Thus, it is conducive to automation—
running large numbers of interpolation without having to make decisions as part of
the interpolation process.  The algorithm is susceptible to problems with interpo-
lating across land; however, methods exist to prevent such problems for Chesapeake
Bay application (as described in previous sections and in detail in Appendix D). It
can be applied at any scale, but is most appropriate  for large scales where three-
dimensional interpolation becomes a necessity and data collection sites may remain
too dispersed to provide good estimates of error structure no matter which algorithm
is used.

In addition to its advantages, IDW also has a major disadvantage: it is not a statis-
tical method. The method is a deterministic approach without any sampling or model
error assumed or accounted for (STAC 2006). In addition, IDW does  not account for
potential spatial autocorrelation among the observations  and, therefore, does not
fully utilize the information contained within the data. No method exists to estimate
either source of error associated with a set of predicted values when using IDW and
it cannot be used as a basis for statistical decision-making using the CFD. Dedicated
research could determine whether IDW could be made more statistically defensible.
          EVALUATION OF  KRIGING AS  A  SPATIAL
               INTERPOLATION ALGORITHM FOR
     ASSESSING CHESAPEAKE  BAY WATER QUALITY

Kriging has been considered by the Chesapeake Bay Program as a principal alterna-
tive algorithm for spatial interpolation in CFD  water  quality criteria assessment
methodology. Kriging is a spatial interpolation technique that arose from geostatis-
tics,  a subfield of statistics that  analyzes spatial  data. Kriging and the field of
geostatistics have been used in a wide variety of environmental applications and are
generally accepted methods for statistically optimal spatial interpolations (Cressie
1991, Schabenberger and Gotway 2004, Diggle and Ribeiro  2006). Kitanidis (1997),
Wang and Liu  (2005), and Ouyang et al. (2006) elaborate on  the application of
kriging in water-related research. References on kriging methodology, geostatistics,
and their related statistical development can be found in Cressie (1991), Diggle et al.
(1998), Schabenberger and Gotway (2004), and Diggle and Ribeiro (2006).

Kriging can be formulated equivalently in terms  of a general linear regression
model:

                   Y (s) = y30 + fii Xj(s) •••+/?£, Xp(s) + e(s)     Equation C-l

with s representing a generic spatial location assumed to  vary  continuously over
some domain of interest and Y (s) capturing the outcome of interest measured at s,
Xj(s), .  . .  ,Xp(s) potential covariates indexed by  location s  and their associated
regression effects /?1; . .  .  , /?p. The uncertainty  in this regression relationship is
modeled with the random error term e(s) assumed to  have  zero mean and constant
                                       appendix c  •  Evaluation of Options for Spatial Interpolation

-------
C-8
                     variance. Spatial data, similar to the type sampled in Chesapeake Bay water quality
                     criteria assessments, often exhibit a property known as (positive) spatial dependence;
                     observations closer together are more similar than those further away. This property
                     is  accounted for in the  model  by allowing g(s) to contain  a spatial  correlation
                     structure.

                     Common distributional assumptions on g(s) include normality and log normality,
                     although kriging can be based on other statistical distributions  and data transforma-
                     tions. Functions of a specific mathematical type (positive definite) represent the
                     spatial correlation in g(s) and are  assumed isotropic (correlation depends  only on
                     distance) or anisotropic (correlation depends on both distance and direction). Vari-
                     ograms constitute another special type of mathematical function—closely related to
                     spatial correlation functions—that  are more often used to represent spatial  correla-
                     tion.  In this  case,  and  in many  kriging applications, variograms  and  spatial
                     correlation functions provide equivalent representations of spatial structure. For con-
                     sistency, only the term "variogram" is used here in discussions of spatial structure.

                     In the literature, Equation  C-l is referred to as a universal kriging model. When
                     covariates (the X's) don't influence interpolation of Y, the right hand side of model
                     (Equation C-l) contains only the constant term/?0. The resulting model is called the
                     ordinary kriging model. When the spatial structure (variogram)  for the model (Equa-
                     tion C-l) is known, statistically optimal predictions  for the variable Y at unsampled
                     locations (outside of estimation of possible regression effects)  can be derived using
                     standard statistical principles. The optimality criteria result in spatial predictions that
                     are linear in the  data, statistically unbiased, and minimize mean squared prediction
                     error—known as best  linear unbiased predictions.  The  minimized mean  squared
                     prediction error is also  a measure of prediction uncertainty. In practice, however, the
                     spatial structure of the data  remains unknown. The estimation of the spatial structure
                     using the variogram function, therefore, is critical to kriging applications.

                     To demonstrate let {Xsi)> • • • Xsn)l represent a sample  set of spatial data such as
                     dissolved oxygen collected at a set of n spatial locations s1;. .  .sn. Assume this data
                     set to be a realization of the ordinary kriging version of model. The primary step in
                     kriging is  variogram estimation with several  methods  available;  the  method of
                     moments and statistical likelihood based are two of the more common. All  of these
                     methods are based on the sample data {Xsi)> • • • Xsn)l- This process ends with a
                     chosen variogram function and its parameter estimation, describing the shape and
                     strength (rate  of decay) of spatial  correlation. A determination, also based on the
                     sampled data, is made of whether the spatial structure is isotropic or anisotropic. The
                     estimated variogram is then assumed known. Kriged interpolations and their inter-
                     polated uncertainty at  numerous locations are  computationally straightforward to
                     generate.

                     The following describes some of the benefits and potential limitations of kriging for
                     the Chesapeake  Bay Program to use in criteria attainment  assessment application
                     (with some comparisons to  the IDW approach of spatial interpolation outlined in the
                     previous section). A primary benefit of kriging compared to IDW is that it is a statis-
                     tical technique. Statistics  (including kriging) can make inferences from sampled data
                     even in the presence of uncertainty; the quantity and quality of the sample  data are
  appendix c  •  E<

-------
                                                                                                 C-9
reflected in these inferences. Kriging, however, is a less-than-routine type of analysis
and requires statistical expertise to execute. The short description on variogram esti-
mation above merely introduces this involved and often complicated step.

Further issues regarding kriging and Chesapeake  Bay Program applications  are
listed below.
   •  Kriging is flexible; it is based on an estimate of the strength of spatial dependence
     in the data (variogram). Kriging can consider direction-dependent weighted inter-
     polations (anisotropy) and can include covariates (universal kriging) to influence
     interpolations—either simple trends in easting and northing coordinates or water-
     related measures such as salinity.
   •  A key feature of kriging is that a measure of uncertainty (called the kriged
     prediction variance) is generated along with kriged interpolations. Research
     has started to propagate this interpolation uncertainty through the CFD.
   •  Kriging can be applied in situations for which the data remain sparse (such as
     the Chesapeake Bay Water Quality Monitoring Program's fixed station data) or
     dense  (such as the Chesapeake  Bay  Shallow-water Monitoring Program).
     Kriged and  IDW  spatial interpolations may very well produce  near identical
     results for  these  two extreme scenarios. The kriging approach, however,
     provides a statistical model, the uncertainty of which is influenced by the quan-
     tity and quality of data. Interpolation uncertainty information is crucial for both
     sparsely and densely sampled networks.

In comparison to  IDW, kriging is more sophisticated, but requires greater expertise
in implementation. Kriging is available  in commercial statistical software  and also
in free open-source applications, such at the R Statistical Computing Environment.
Use of the technique requires geostatistical expertise programming skills for these
two software packages. Segment-by-segment variogram estimation and subsequent
procedures  would require  substantial  expert  supervision and decision-making.
Chesapeake Bay Program managers may very well view this as a limitation in using
kriging for certain Chesapeake Bay Program activities, such as criteria assessments,
applications that  need automated spatial interpolations. Furthermore, for  some
Chesapeake Bay Program applications, the decision on criteria attainment is clearly
not influenced to any  substantial degree by  the method  of spatial interpolation
because the water quality conditions remain far out of attainment. One  possible
strategy is using a mix of IDW and kriging in situations for which attainment was
grossly exceeded  or clearly met (IDW) versus borderline cases (kriging). Table C-l
provides a comparison of the capabilities of assessments based on  lumping data,
spatial interpolation based on IDW, and  spatial interpolation based on kriging.
                                        appendix c

-------
C-10
                     Table C-1.  Comparison of the capabilities of methods for interpreting data for Chesapeake
                                Bay water quality criteria assessment.
Attributes
Provides Spatial Prediction
Provides Prediction
Uncertainty
Uncertainty for CFD
Deal with Anisotropy
Can include cruise track/
fly-over data
Feasibility of 3-dimensional
interpolations
Feasibility of mainstem-
tributary interpolations
Inclusion of covariates to
improve prediction
Predictions of non-linear
functions of predicted
attainment surfaces P(y>c)
Level of sophistication
Automation
Sample-based
Yes
No
No
No
No
No
No
No
No
Lowest
Yes
IDW
Yes
No
No
Possible, but
not routine
No
Yes
Yes
No
No
Low
Yes
Kriging
Yes
Yes
Yes
Yes
Yes
Possible, but
not routine
Possible
Yes
Yes
Very High
No
                     Source: STAC 2006.
                      Chesapeake Bay Program. 1989. Chesapeake Bay Monitoring Program Atlas - Volume 1:
                      Water Quality and Other Physiochemical Monitoring Programs. CBP/TRS 34/89 U.S. Envi-
                      ronmental Protection Agency Chesapeake Bay Program Office, Annapolis, MD.
                      Cressie, N. 1991. Statistics for Spatial Data. Wiley, New York, NY, 928 pp.
                      Diggle, P.J., J.A. Tawn, and R.A. Moyeed. 1998. Model Based Geostatistics (with Discus-
                      sion). Applied Statistics 47:299-350.
                      Diggle, PJ. and P.J. Ribeiro. 2006. Model-based Geostatistics. Springer, New York, NY, 230
                      pp.
                      Harding, L.W.,  Jr., E.G. Itsweire, and W.E. Esaias. 1992. Determination of phytoplankton
                      chlorophyll  concentrations in the Chesapeake Bay with aircraft remote sensing. Remote
                      Sensing of the Environment 40: 79-100.
                      Harding, L.W., Jr., J.G. Kramer, and J. Phinney. 2004. Estuarine and Watershed Monitoring
                      Using Remote Sensing Technology  Present Status and Future Trends: A Workshop Report,
                      7-8 January 2002, Annapolis, Maryland. Scientific and Technical Advisory  Committee and
                      Maryland Sea  Grant College.  Maryland Sea  Grant  Publication UM-6-SG-TS-2004-03.
                      College Park, MD.
                      Kitanidis, PK. 1997. Introduction to Geostatistics: Applications in Hydrogeology. Cambridge
                      University Press, New York, NY, 271 pp.
                      Ouyang Y, J.E.  Zhang, and L.T. Ou. 2006. Temporal and spatial distributions of sediment
                      total organic carbon in an estuary river. Journal of Environmental Quality 35:93-100.
                      Schabenberger,  O and C.A. Gotway. 2005. Statistical Methods for Spatial Data Analysis.
                      Chapmann and Hall/CRC Press, FL, 512 pp.
   appendix c
Evaluation of Options for Spatial Interpolation

-------
                                                                                                   C-11
Scientific and Technical Advisory Committee (STAC). 2006. The Cumulative Frequency
Diagram Method for Determining Water Quality Attainment: Report of the Chesapeake Bay
Program STAC Panel to Review o  Chesapeake Bay Analytical Tools STAC Publication 06-
003. 9  October 2006. Chesapeake Bay  Program  Scientific and  Technical  Advisory
Committee. Chesapeake Research Consortium, Edgewater, MD.
U.S. Environmental Protection Agency. 2003a. Ambient Water Quality Criteria for Dissolved
Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and its Tidal Tributaries.
EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, MD.
U.S. Environmental Protection Agency. 2003b. Technical Support Document for Chesapeake
Bay Designated Uses and Attainability. EPA  903-R-03-004. Region III Chesapeake Bay
Program Office Annapolis, MD.
U.S. Environmental Protection Agency. 2004. Chesapeake Bay Program Analytical Segmen-
tation Scheme:  Revisions,  Decisions and Rationales 1983-2003.  EPA 903-R-04-008.
CBP/TRS 268/04. Region III Chesapeake Bay Program Office, Annapolis, MD.
U.S. Environmental Protection Agency. 2005. Chesapeake Bay Program Analytical Segmen-
tation Scheme: Revisions, Decisions and Rationales 1983-2003: 2005 Addendum. EPA
903-R-05-004. CBP/TRS 278-06. Region III Chesapeake  Bay Program Office, Annapolis,
MD.
Wang, XJ.  and R.M. Liu. 2005. Spatial analysis and eutrophication assessment for chloro-
phyll a in Taihu Lake. Environmental Monitoring and Assessment 101:167-174.
                                         appendix c

-------

-------
                                                                                 D-1
                      appendix

                 User Guide and
          Documentation  for the
    Chesapeake  Bay Interpolator
                       INTRODUCTION

The Chesapeake Bay  and Tidal Tributary  Interpolator computes water  quality
concentrations throughout the Chesapeake Bay and/or tributary rivers from water
quality measured at point locations. The purpose of the Interpolator is in compute
water quality concentrations at all locations in the 2-dimensional plane  (top or
bottom depth) or throughout the 3-dimensional water volume. Results of the inter-
polation  can then be compared over time to compute trends or  individual
interpolations can be overlaid with other data to visualize possible cause and effect
relationships. One example is to compare water quality with living resource (fish,
shellfish, aquatic vegetation) distributions. Results of the Chesapeake Bay Interpo-
lator have been used  since 1988 to determine trends in water quality  for the
Chesapeake Bay Program (http://www.chesapeakebay.net/).

Version 4.2 of  the VOL3D software includes new code to: 1) import data from
Microsoft ACCESS data tables; 2) draw improved graphics of tributary segments; 3)
draw colors using categories, as before, or to draw using a color ramp of 255 colors;
4) draw longitudinal sections which represent the centerline of the Bay or Tributary
River segments; 5) draw images of all Tributary Rivers in addition to  the Bay; and,
6) compute composite  images that represent the minimums or maximums over a
time series.

Another tool, DART, which must be run on  the CIMS network at the Chesapeake
Bay Program Office, creates data sets for the Interpolator for any parameter in the
historical water quality data base. DART is a very powerful tool  which can create
many data sets in a very short time. Anyone  who needs to interpolate data held by
the Bay Program, should investigate the use of DART.
                 appendix d  • User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-2
                               M.:&I

                    The Chesapeake Bay Interpolator is a cell-based interpolator. Fixed cell locations are
                    computed by interpolating the nearest n neighboring water quality measurements,
                    where n is normally 4, but this number is adjustable. Cell size in Chesapeake Bay
                    was chosen to be 1km (east-west) x 1km (north-south) x 1m (vertical), with columns
                    of cells extending from surface to the bottom of the water column, thus representing
                    the 3-dimensional volume as a group of equal sized cells extending throughout the
                    volume. The tributaries are represented by various  sized cells  depending  on the
                    geometry  of the tributary, since the narrow upstream portions of the rivers require
                    smaller cells  to accurately model the river's dimensions. This configuration results
                    in a total of 51,839 cells by depth for the Main Bay (Segments CB1TF-CB8PH), and
                    a total of 238,669 cells by depth for all 77 segments which comprise the Main Bay
                    and tributaries. Computation time on a Pentium 2 ghz PC running Windows XP is
                    approximately 15 seconds for the Bay and tributary interpolator model.

                    The Chesapeake Bay Interpolator  is unique in the way it computes values in 3-
                    dimensions. The interpolator code is optimized to compute concentration values that
                    closely reflect the physics of stratified water bodies,  such as Chesapeake Bay. The
                    Bay is very shallow compared to its width or length, hence water quality varies much
                    more vertically than horizontally. The Chesapeake Bay Interpolator  uses a vertical
                    filter  to select the vertical range  of data  that  are used in each calculation. For
                    instance, to compute a model cell value at 5m deep, monitoring data at 5m deep are
                    preferred.  If fewer than n (4) monitoring data values are found at the preferred depth,
                    the depth  window is widened to search up to d (normally +/-2m) meters above and
                    below the  preferred depth, with the window being widened in 0.5m increments  until
                    n monitoring values have been found for the computation. The smallest acceptable n
                    value is selectable by the user. If fewer than n values are located, a missing value
                    (normally a -9) is calculated for that cell.

                    A second search radius filter is implemented to limit the horizontal distance of moni-
                    toring data from the cell being computed. Data points outside the radius selected by
                    the user (normally 25,000m) are excluded from calculation. This filter is included so
                    that only data that are near the location being interpolated are  used.

                    In this version of the  Interpolator, Segment and Region filters  have been  added.
                    Segments  are geographic limits for the interpolator model. For  instance, the Main
                    Bay is composed of 8 segments (CB1TF, CB2OH, ...,CB8PH).  The tributaries are
                    composed of 69 additional segments, using the CBP  1998 segmentation scheme
                    (Figure D-l). These segments divide the Bay into geographic areas that have some-
                    what  homogeneous environmental conditions. This  segmentation also provides a
                    means for reporting results on a segment basis that can show more localized changes
                    compared to  the whole Bay ecosystem. To replicate  the segmentation scheme, the
                    segment boundaries  were used to cookie-cutter out the Interpolator  cells that fall
                    within each segment. Each set of  these cells  are then identified inside the corre-
                    sponding  *.bth file that contains  the bathymetry  definitions. To  compute the
                    interpolated values for the Main Bay,  the corresponding bathymetry file is  named
                    "cbayS.bth". This file contains the cell locations for the cells in the Main Bay Inter-
                    polator. A similar file, "bay_trib.bth" contains  the cell definitions for the Main Bay
  appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                  D-3
                                           Up RTF
                                    CBITF-.-
                                9SHOH
      D-1. Chesapeake Bay Program 1998 segmentation design.
and tributary interpolator. Other .bth files have also been created for individual river
systems. Users that need specialized processing,  such as finer resolution or addi-
tional segments in a particular area of interest, must create a new bathymetry file that
defines the bathymetry of the area of interest at the desire cell-size.

Regions  filters  (cbayS.drg, bay_trib.drg,  etc) are  files  which contain a  closed
polygon of x-y points that define an area larger than the corresponding *.bth file. The
region file identifies the geographic boundary that limits  which monitoring station
data are included in interpolation for a given segment. The purpose of the data region
is to select a subset of the monitoring data from the  input data file, and to then use
                    appendix d  «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-4
                                     CHESAPEAKE BAY PROGRAM INTERPOLATOR
                                             Data File Creation Tool
                             Figure D-2. Screen 1 includes seven navigation buttons and the Data
                             File Creation Tool for importing data from an ACCESS data base and
                             creating VOL3D data files.Once the user has selected the desired data
                             fields in the data table (Figure D-3), the Data File Creation Tool opens a
                             new screen that provides a range of options to the user for selecting
                             and subsetting data from the ACCESS data table (Figure  D-4). The Data
                             Engine allows the user to select data by parameter, by date range, to
                             set interpolation control parameters, to choose the desired bathymetry,
                             to select data by depth ranges or layers, and finally to choose how the
                             group the resulting data in one or more output files. The "Create Files"
                             button, when pushed, will generate data files  in the VOL3D ".d3d" file
                             format. These files are then ready for interpolation.
                     that subset for computing the values for each cell in a segment. Use of data regions
                     ensures that the interpolator does not "reach across land" to obtain data from an adja-
                     cent river which would give erroneous results. By using data regions, each segment
                     of cells can be computed from their individual subset of monitoring data. Each adja-
                     cent data region should overlap by some  amount so that there is  a continuous
                     gradient, and not a seam, across segment boundaries.

                     In the future, a pycnocline filter may be added to the Interpolator, so that water
                     above, within,  and below  the pycnocline are not interpolated together. Since the
                     water quality in various parts of the pycnocline can be so dramatically different, the
                     Interpolator file structure will be modified to handle this requirement.

                     INSTALLATION

                     The VolSD Interpolator code and auxiliary files have been bundled together into a
                     SETUP application and then PKZipped to reduce the overall file size. The Vol3D.zip
                     file must first be unzipped into a directory on any standard PC running the Windows
  appendix d

-------
                                                                                           D-5
95/98/XP operating system. Once unzipped, double click the SETUP.EXE file to
start the installation process. It is suggested that the application be installed in the
C:\VOL3D directory. The original zipped file can be deleted to regain disk space. A
fast Pentium machine with 256 mb ram and 1 gb disk drive will prove useful.
                THE                     BAY
               FOR         THE

Begin using the VOL3D software by double clicking the VOL3D.EXE icon on the PC.

The first screen provides 7 buttons (Geography, Parameter, Data Import, Interpolate,
Math, Graphics, and Reports) that step the user through the interpolation, graphics,
and reporting process. Also on the first screen, is a Data File Creation Tool, that can
be used to create VOL3D compatible data files from an external ACCESS data base.
The ACCESS data base needs to contain data necessary for interpolation, as identi-
fied  on Figure  D-2. Essential data fields include STATION,  LONGITUDE,
LATITUDE,  DEPTH, and  VALUE.  Other fields,  including SAMPLE_DATE,
PARAMETER name, LAYER, Qualifier (<, >), CRUISE, and CRUISEJD, provide
data that can be used to select or subset the monitoring data by cruise, layer, or date.
                CHESAPEAKE BAY PROGRAM INTERPOLATOR
                         |] Data Import | Inleipolale |  Malh  | Giaph.r,»
                        Data File Creation Tool
        Figure D-3. Example of data fields selected from the "cbpwq99" data
        table in the "data1999.mdb" ACCESS data base. The Data File
        Creation Tool allows a user to extract data from an ACCESS table into
        desired data files for VOL3D
                   appendix d  «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-6
           Database Name: |C:Wo(3D403\data1999mdb
Table Name: [Cbpwq93
inter Descriptive Title
Chesapeake Bay Wale* Quality Analysis
Choose Parameter (cHLA.Chlorophyll
DO,DissoJved Oxygen
SALINITY.Salrtty
WTEMP.Water Temperature
NH4F Ammonia
NQ2F,Nitrte
N03F,Nitrate v

Choose Start Date J 1 /s/1 999
1 /6/1 999
1/7/1999
1/11/1999
1/12/1999 v
Transform
ff Linear r Log (Base 2)

Choose Geography

-------
                                                                                                D-7

         CHESAPEAKE BAY PROGRAM INTERPOLATOR
             Paiamelet   Data Input   Interpolate
   •Choose Geographic Area for Analysis-
    •0 Bap and Tiibs
      lUillLOB'IAMMH.MUSI Pill

    • Chmtui flwm

    * Chop!** B««
    * N«Micok« HIVE.
    •it I'Dtomnke Hivei
                         • Magolhv Rivei

                         'H Patapsco Rival
Figure D-5. Geography screen. Select desired bathymetry.

          CHESAPEAKE BAY PROGRAM INTERPOLATOR
         Paiamelei I Data Inpoft I IntefpoMe
   rChoose Geographic Area tor Analysts
  Chcsapcjkc M.nnslcm
  ICB1-C88I
  Chesapeake Mains!em
  (CD1 CU8. T ANMH^MOBPH I

_ Cheslei Rhmi

0 Choptank Rivei
                          & Bappahannecfc Rl
     • N.jfilicoke liivci


     • PalUKWll Rivei

     • Hotomoke Hivt'i
                     • Pal.ipsco Rivei

                     • Back Rivei
                                                    Giaphio I  Report!
 Figure D-6. Parameter screen. Choose desired parameter for analysis.
        appendix d  «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-8
                     Click the Data Import Button to select the Data Import screen. On this screen, click
                     the "Get File Name" button to select the data file that contains the data that are to be
                     interpolated (Figure D-7). The default file extension for data files is .d3d. d3d files
                     include the X and Y coordinates (UTM Zone 18, NAD83 is recommended for inter-
                     polation. These are also required for the graphics tools.)
                               Selected .Bata.File;
                                     Look
                                 My Recent
                                 Documents
; _J VoSJD

i.^JSasBBl
J6DO97G716.D3O
J800970801.D3D
 'JB00970816.030
 'JBOO970901.030
                                  Desktop
                                f% Doeumetits
                                 My Computer
                                 My ftetwortc
                                   Places
filename:

Fifes of type:
                                                                                       Open
                                                     Data File f.d 3d)
                                                                                       Cancel
                                    D-7. Data Import screen, "Get File Name" window. Select a data
                              file (.d3d) for analysis.
                     Once the file has been selected, the other fields on this screen will populate with
                     information about the data file, including, start and end dates of the data, the number
                     of observations, the date the file was created, the parameter name and code, and title
                     (Figure D-8). Normally data do not need to be transformed, however, some data such
                     as chlorophyll or TSS should be transformed with the log-transform to normalize the
                     data. The data are transformed as they are read into the interpolator and the results
                     are back-transformed to the original units in the output file. If the parameter is to be
                     transformed  by  the natural log transform, any data values that are negative or zero
                     will be set to a value of 0.0001. If the parameter is to  be transformed by the square
                     root transform, any data values that are negative will be set to 0.0.

                     Two buttons  at the bottom of the screen can be used to convert latitude and longitude
                     coordinates to UTM coordinates, which are recommended for interpolation (Figures
                     D-9, D-10). The first converts the longitude and latitude coordinates in d3d formatted
                     files to UTM coordinates, and vice versa. This is  handy for checking data locations on
                     maps. The second converts individual longitudes  and latitudes to and from UTM coor-
                     dinates. NAD27 to NAD83 conversion is not supported in this  code. Improper use of
                     NAD27 or NAD83 can result in coordinate errors in the 100 to 300 meter range.
  appendix d

-------
                                                                                                            D-9
       CHESAPEAKE BAY PROGRAM INTERPOLATOR
     CHESAPEAKE SAY BY CRUISE BstolveiJ Oxygen • ln» IrterpoMed Data - 07JUU93M5IUU337
                          H
                                                                  Figure D-8. Data Import
                                                                  screen with fields popu-
                                                                  lated with data from the
                                                                  selected .d3d input file.
        Longitude/Latitude and UTM
    Converter for Interpolator Data Files

   Choose an input file in ,
-------
D-10
                                     lajnniON

                     Click the Interpolate Button to select the Interpolate screen (Figure D-ll). Select
                     the interpolator settings that  match your requirements. The  3D Inverse-Distance
                     Squared model is the 3-dimensional interpolator model. The 2D Inverse-Distance
                     Squared model uses the same code as the 3D interpolator model except that only one
                     layer of cells are computed—cells for each depth below the surface cell are set to
                     missing (normally  -9).  The 2D Octant Search model computes values for cells in
                     only one layer, however, the data used for computing each cell value are selected
                     from data in each surrounding octant. For instance, for a given cell, the data used for
                     calculation would include 4 data points from each surrounding octant, or a total of
                     32 data points. The model will use fewer than the total data in each octant if insuffi-
                     cient data exist. The model uses as many data as  are available for each octant, up to
                     the maximum requested number of data points. The octant search model is used to
                     reduce the bias from sampling schemes that collect continuous strings of data, such
                     as aircraft monitoring that collect many data points in well defined flight tracks. The
                     run-time for the octant  search model  is significantly longer  due to the extensive
                     sorting required to  select data from each data octant.

                     The "Trace Level"  selects the  amount of detail written to the ".LOG" file. A "Trace
                     Level" of "2"  provides general  interpolator statistics. A "Trace Level"  of "3"
                     provides information about the data values used in the computations for each region.
                     A "Trace Level" of "4" provides information about individual cell computations. A
                     "Trace Level" of "5", "6", or "7" provides increasing information about data values,
                     distances, and octants. Increasing the "Trace Level" value is useful for investigating
                     the performance of the interpolator.

                     The "Convert .EST to .TXT" button will create a .txt file that can be imported into
                     Arc/Info or Arc View. The .txt files are a full matrix of values, 57 columns wide, with
                     all missing or non-existent cell values designated as missing values (normally -9),
                     comma delimited, and column headings and text strings are enclosed in quotes. Each
                     row in the .txt file represents numbers from 1 column of water from top to bottom,
                     1 cell wide  by n cells deep. Additional columns are appended to the  .txt file for
                     bottom, minimum,  maximum, mean, and sum values.

                     The "Convert .EST to .T3D" button will create a .txt file that can be imported other
                     applications. The  .t3d files are 4  columns wide, comma delimited, contain the x
                     value, y value, negative z value, and the estimated value.  There are no column head-
                     ings. Missing values are included and are coded based on what was selected during
                     the interpolation.

                     The interpolator mode can be set to "Interactive" or "Batch". In interactive mode, the
                     chosen file is interpolated as defined in Figure D-ll. In  "Batch" mode,  a job file is
                     selected which provides the information needed to interpolate a series of files under
                     machine control (Figure D-12). The ".job" file can be built interactively by pressing
                     the "Save to Batch Job" button after selecting the run parameters for each desired file
                     (Figure D-13). The "Batch Job" can be executed by pushing the "Run Batch Job"
                     button.
   appendix, d  » User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                 D-11
         CHESAPEAKE BAY PROGRAM INTERPOLATOR
                   I Data Impoil I  Mnuti/f  I   Math  I  Giaphict 1
Figure D-11. 3D Inverse-Distance Squared Interpolate screen populated
with entries after having made choices on previous screens. The 2D
Inverse-Distance Squared Interpolate screen uses the same format as
the 3D Inverse-Distance Squared model; however, only the surface
depth value has computed values. Cell values at depths below surface
are set to missing (generally -9). The 2D Octant Search Interpolator
does not rely only on the closest data in all directions, but rather uses
data from data from surrounding octants. For example, if 4 nearest
neighbors are requested in each of 8 octants surrounding the cell being
computed—up to 32 nearest neighbor values will be used to compute
the value. If no nearest neighbor values are available, a missing value
will be computed. Other buttons are available for creating data using
specific formats for various GIS (.txt files) and graphics applications
(.t3d files).
 Choose Batch File Name
       Look in: Q Vol3D
                  JLl
   My Recent
   Dcxxarcerts
    Desktop
   My Computer
   1% Netwoik   Fte name:
     Places
             Res of type:
|bdo-9?07-9709jofa
                         Batch Job He T job)
 Figure D-12. Batch Job File Name selection window that displays after
 choosing "Batch" radio button on Interpolation screen.

-------
D-12
                                    CHESAPEAKE BAY PROGRAM INTERPOLATOR
I                                               Data Impoil 1 Mw&rt-  I   Math   I  Graphic* 1
                                              mmmmmmmmmmmmmm
                           Figure D-13. Saving or running a batch job through the Interpolate
                           screen. "Save to Batch Job" saves the values that have been entered in
                           the fields on this screen into the "Batch File ('filename'.job)". If the 'file-
                           name'.job file already exists, the new entry is appended to the existing
                           file. If it does not exist, a new file is created. The "Output File" file name
                           entry is also written to a file ('filename'.fls) for use in creating volume
                           and mass estimates by running batch jobs. The .fls file is simply a list of
                           interpolated (.est) file names that can be processed sequentially. The
                           "Test Batch Job" button executes a batch job but does not run the inter-
                           polator. This button can be used to test whether the needed files exist
                           and the batch job is sound prior to execution of the interpolator.
                     Click the Math Button if you need to conduct special operations on one or more
                     files. Four functions  are  provided:  1) Math operations which include  adding,
                     subtracting,  multiplying,  or dividing one interpolated file by  another, or by  a
                     constant; 2) Receding values to new values; 3) Conducting a change analysis over
                     time; and, 4) Calculating the minimum or maximum values from a set of files.

                     Math functionality is provided so that special parameters can be calculated. Math is
                     conducted on a cell by cell basis. For instance, to add two interpolated files, Cell  1
                     of input file A is added to Cell  1 of input file B and the sum is stored in Cell 1 of
                     output file C, and so forth. Subtracting one file from another can be used to show
                     change from one time to  another (Figure D-14). Missing values  are handled as in
                     regular math—a non-missing value becomes missing if a math operation attempts to
                     compare a real value with a missing  value. Division by zero or other  illegal math
                     operation will cause the operation to stop.

                     The "Derive New Parameter"  math operations can be performed sequentially to
                     provide additional capability. For instance, five interpolated files (.est files) could be
                     sequentially added together, then the resulting file could be divided by 5 to compute
   appendix d  *  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                              D-13
         r      _^
         [";. r';; •<- -x'

                  CHESAPEAKE BAY PROGRAM INTERPOLATOR
                D-14. Math screen with files chosen to "Derive New Parameter"
         of Dissolved Oxygen by subtracting File 2 from File 1 to create the
         output file.
the mean for the five files. Another example would be to subtract interpolated
dissolved oxygen data from an interpolated saturated  dissolved oxygen file to
compute the oxygen deficit.

The code checks whether the input files have the same number of segments. If the
input files ("Input .est File 1" and "Input .est File 2") do not have the same number
of segments, they were generated from different bathymetry files, and the cell values
in the two files can not be properly combined. An error message will be displayed if
this condition occurs.

The "Recode Parameter Value" radio button provides the means to convert calculated
values to new values (Figure D-15). The input file is not changed, but a new output
file is created with new values in each cell which classify the data into new values
or categories. For example, to compute the interaction of dissolved oxygen and water
temperature:

1) Recode the dissolved oxygen .est file so that  oxygen below 3 mg/1 is set to "1"
and oxygen above 3 is set to "0" (also set missing to -9).

2) Recode the water temperature .est file so that temperature below 25C is set to "0"
and temperature above 25C is set to "10" (also set missing to -9).

3) Derive a new parameter "WD" by adding the receded dissolved oxygen and water
temperature  .est files. The result is a wd.est file where: "0"=acceptable oxygen and
temperature; "l"=unacceptable oxygen; "10"=unacceptable temperature; and "11"=
unacceptable oxygen and unacceptable temperature (missing cells will = -9). This
file can be graphed to show  the distribution of these categories. The water column
                   appendix d * User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-14
                                      CHESAPEAKE BAY PROGRAM INTERPOLATOR
                                                I Data Import I Intrapolale I   tfM   I  Gidphici  I
                              Figure D-15. Math screen with files chosen to "Recede Parameter
                              Value" of Dissolved Oxygen by receding values from 0-3 to the new
                              value of "1", receding values of above 3 to the new value of "0" and
                              retaining missing values as -9, to create the new output file. Choosing
                              a Range File is provided to load an existing set of ranges, which can
                              then be modified on this screen for specialized analyses.
                      volume of these categories can also be computed to show critical ranges for habitat
                      analysis.

                      Missing values are handled in a special way. Since missing values have no real value,
                      they are not used in math operations. If a cell in either "Input File 1" or "Input File
                      2" are flagged as missing (normally -9), then no math is done, and the "Output File"
                      value for that cell is set to missing (The "Missing Value" is set  on the Interpolate
                      screen).



                      Interpolated values  can be  analyzed for trends. The "Change Over Time" button
                      allows the user to create a 3-dimensional (.est) file with linear percentage changes
                      over time for each cell in the bathymetry (Figures D-16 and D-17).

                      As a simple example, a station may be sampled several times over a period of time.
                      The measured values can be plotted with time on the x-axis and value on the y-axis.
                      The resulting linear regression line can be plotted through these points and the slope
                      and intercept can  be used to compute the percentage increase or decrease between
                      the beginning and end of the time series.

                      This same technique can be used with the interpolator. Each cell value from a series
                      of .est files can be used to compute a linear regression, so that each cell has its own
                      regression and resulting percentage change, either up or down, over time. By coding
   appendix d  «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                             D-15
CHESAPEAKE BAY PROGRAM INTERPOLATOR
 1 Pai.imeta I Data Import m Intwpolale I   ,t-'.t'/>   I Graphic!  1  Reports
                                                           Figure D-16. Math screen with files
                                                           chosen to "Change Over Time". The
                                                           bdo-9707-9709.fls file contains file
                                                           names of dissolved oxygen .EST files.
                                                           The corresponding Julian dates for
                                                           these .EST files are read from the do.jul
                                                           file. In this example, a new .EST file -
                                                           Change.est - is created which contains
                                                           the linear trend for each cell over the
                                                           time interval of the bdo-9707-9709.fls
                                                           file. The Change.est cell values are per-
                                                           centage change over time, categorized
                                                           by the selection criteria identified in
                                                           the pc.rng file. In  this example, "Ignore
                                                           missing values" has  not been selected.
                               CHESAPEAKE BAY
             "NJ*               Water Quality Analysis
                      Total Phosphorus Percent Change Jul 2,1984 Dec 18,1997
         PLANVIEW
-»*      i	  •«
. ,  - .     f/tvtfi  -*   ^      ~ ' -
-Wiai.t^    "*       ^
        TOPTH PROFILE
             D-17. Plot of total phosphorus as a "Change Over Time".  In this
      example, a new .EST file - tp8497.est - was created which contains the linear
      trend for each  cell over the time interval (July 1984 through December 1997)
      of the tp8497.fls file. The tp8497.est cell values are percentage change over
      time, categorized by the selection criteria identified in the pc.rng file. The
      percentage change categories for total phosphorus mass (kg) are: >10%
      increase (red);  5 to 10% increase (pink); 0 to 5% increase (yellow); 0 to  5%
      decrease (light blue);  5 to 10% decrease (dark blue); and greater than 10%
      decrease (green). In this example, "Ignore missing values" was selected so
      that a trend on any available data was calculated.
           appendix d  •  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-16
                     the result as the percent change, a .est file can be created that has a percentage
                     change value for each cell. This .est file represents a 3D file of "Change Over Time".
                     The plot of this file provides a graphical representation of the change. The categories
                     used to  display the changes graphically are defined in the pc.rng file. The default
                     pc.rng file provides categories of: >+10%, +5 to +10%, 0 to +5%, -5 to 0%,-10 to
                     -5 %, and >-10 % change. These categories should be  modified to reflect the needs
                     of the analysis.

                     Missing values in the analysis can be treated in two ways: 1) included, meaning they
                     are propagated through the analysis; or, 2) ignored. The default is to include missing
                     values. The result of including missing values is that if one value for a specific cell
                     is missing anytime  in the times  series, then that cell is set to missing. The single
                     missing value forces the whole  series of values  at that cell to be missing and no
                     percentage change is calculated.  The percentage change value is set to missing (-9
                     by default).

                     If missing values are set to be ignored, then each missing value in a time series for
                     a given cell  is ignored and the rest of the time series observations are used to
                     compute the percentage change over time. The potential problem with this approach
                     is that the trend may be skewed by the lack of having all of the desired data.

                     At least two points  are required to compute a time series change. If the  number of
                     observations for any cell is less than 2, the resulting value for the percentage change
                     is set to missing.



                     The "Min-Max" button can be selected to locate the minimum or maximum values
                     in a series of interpolated values. For instance, this function could be used to read ten
                     interpolated files, and find for Cell 1 the minimum value and write that minimum
                     value to the  output file Cell 1. This process would be repeated for each cell, so the
                     resulting output file would contain the minimum value for each cell in the  series. The
                     Maximum function could be chosen if desired to find the maximum cell values in a
                     series of files. These functions are useful for determining, for example,  the lowest
                     salinity over a 10-year period, or the highest temperature over a year period, for each
                     cell in the interpolated files. (Figures D-18 and D-19).
   appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                  D-17
      CHESAPEAKE BAY PROGRAM INTERPOLATOR
Geography • Parameter  I Data Import I Intetpolo'e 1  «*»   II  Graphki 1  Reporti
Input Ffcs | (is) »C \VonO\Mo-9707-97Mll!
  Temp File  »C \Vol3ti\IeroFJeesl
Output We (est) lc VM»\DO M«nume:>
Figure D-18. Min-Max screen to
capture the minimum oxygen
values in each cell over the July-
September timeframe in 1997.  The
bdo-9707-9709.fls file contains file
names of interpolated dissolved
oxygen  .EST  files. The Temp file is
an intermediate working file  that
can be deleted after the job is com-
pleted. In this example, a new .EST
file - do-minimum.est - is created
which contains the minimum vale
for each cell  over the time interval
of the bdo-9707-9709.fls file. In
this example, "Ignore missing
values"  has been selected.

      CHESAPEAKE BAY PROGRAM INTERPOLATOR
         Paiametef  i Data Inport I Intaipolate
                                 Selcct Mmttmim
                                Numl.ei nl LdlwJs
                                       :»§•;
                               '
Figure D-19. Math screen with files
chosen to "Change Over Time". The
do97.fls file contains file names of
dissolved oxygen .EST files. The
corresponding Julian dates for
these .EST files are read from the
do.97.jul file. In this example, a
new .EST file - dope.est - is created
which contains the linear trend for
each cell over the time interval of
the do97.fls file. The dope.est cell
values are percentage change over
time, categorized  by the selection
criteria identified
                 appendix d   «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-18
                      Click the Graphics Button to select the Graphics screen. In this version, the graph-
                      ical representation of the data is limited to a Plan view (looking down on the Bay and
                      tidal tributary rivers) and a Side view (looking at the vertical dimension of the Bay
                      and tidal tributaries from the West).

                      The Graphics screen provides a means to choose all of the variables need to create
                      the Bay/Trib graphic (Figure D-20). Most of the choices are driven by the files being
                      graphed, to help minimize typing in all of the required information. The graphic can
                      be printed or saved to a .BMP file (Figure D-21).
         Figure D-20. Graphics screen
      with default values for graphing
              the selected ".est" file.
                                      « INTERPOLATOR GRAPHICS
                                             CHESAPEAKE BAY PROGRAM INTERPOLATOR
     Figure D-21, Graphics screen with
      titles imported from the selected
      ".est" file. Click the "Interpolated
     File" button to load these titles to
    this screen. The titles can be edited
               directly on this screen.
                                             CHESAPEAKE BAY PROGRAM INTERPOLATOR
   appendix d  •  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
The Batch checkbox can be selected to process a group of files (.est file names are
read from a *.fls file) using the choices selected as shown in Figure D-22. If Batch
is chosen, the program prompts the user to choose a *.fls file. Title 3 and Graphics
File will  automatically change for each plot based on the information contained in
the .est file. Other titles and legends will display based on what is displayed when
the DRAW button is pushed. Each graphic will automatically be saved to the default
graphics file name (D-23).
                                                                                                     D-19
 « INTERPOLATOR GRAPHICS
         CHESAPEAKE BAY PROGRAM INTERPOLATOR
                                                                     Figure D-22. Graphics screen
                                                                     with legend imported from the
                                                                     selected ".rng" file. Click the
                                                                     "Categories File" button to load
                                                                     these range categories to this
                                                                     screen. These category values and
                                                                     colors can be edited directly on
                                                                     the screen.
   PLANVIEW

    Svsqitet)3K®a
    Legend:
    • 0.0-0.2
    •i 0.2-1.0
    cu 1.0-3.0
    a 3.0-5.0
    Bi >50MG/l.
    • Not Calculated
   OCPTM PROFILE
                      CHESAPEAKE BAY
                      Water Quality Analysis
                    Dissolved Ojygen - Jul7, ISWT-Jul IS, 1997
Pt«*BMPtofiW«|
                                                     .,—-j
                Figure D-23. Example graphic
                of interpolated Chesapeake Bay
                mainstem dissolved oxygen. The
                data are displayed  so that the
                worst case data (low dissolved
                oxygen, in this case), regardless
                of depth, are visible in both the
                Plan- and Side views.
                     appendix d

-------
D-20
                     Click the Reports Button to generate files for volumetric and mass analysis (Figure
                     D-24).

                     The "Layer Thickness" is set to 0-50 meters deep to include all cells in the interpo-
                     lated file. This thickness could be set, for example, to 3-6 meters to  calculate the
                     volume and mass for the water 3 to 6 meters deep (Figure D-25).
   Figure D-24. The Reports screen is
  used to compute volume and mass.
  In Interactive mode,  an interpolated
     file (.est) is processed to create a
   file of water volumes by parameter
       range category by segment. A
        file of parameter mass is also
        computed by segment. If the
    "Compute Mass by Concentration
      Range" is checked on, then the
           mass calculations are also
     separated by the same category
   ranges as the volume calculations.
                                            CHESAPEAKE BAY PROGRAM INTERPOLATOR

     Figure D-25. The Reports screen
      with "Bottom Layer Thickness"
      selected and set to the bottom
    3 meters of water column depth.
        Only the cells in the selected
    bottom layer will be processed for
      volume and mass calculations.
                                            CHESAPEAKE BAY PROGRAM INTERPOLATOR
   appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                D-21
The resulting volume (.vol) and mass (.mas) files can be used for creating numerical
or graphical reports, such as trends plots (Figure D-26).

Each successive  set of computed numbers are appended to the same specified
"Output File" (.vol and .mas).

Each successive  set of computed numbers are appended to the same specified
"Output File" (.vol and .mas) (Figure D-21).
         CHESAPEAKE BAY PROGRAM INTERPOLATOR
                                                                 Figure D-26, The Reports screen
                                                                 with "Batch" mode selected. In
                                                                 Batch mode, a list of interpolated
                                                                 file names (.est) are processed
                                                                 sequentially to create a file of
                                                                 water volumes by parameter
                                                                 range category by segment. A file
                                                                 of parameter mass is also com-
                                                                 puted by segment. If the
                                                                 "Compute Mass by Concentration
                                                                 Range" is checked on, then the
                                                                 mass calculations are also sepa-
                                                                 rated by the same category
                                                                 ranges as the volume calculations.
                                                                 This example calculates volume
                                                                 and mass for the  top
         CHESAPEAKE BAY PROGRAM INTERPOLATOR
                                                                 Figure D-27. The Reports screen
                                                                 is used to compute volume and
                                                                 mass. In Interactive mode, an
                                                                 interpolated file (.est) is processed
                                                                 to create a file of water volumes
                                                                 by parameter range category by
                                                                 segment. A file of parameter mass
                                                                 is also computed by segment. If
                                                                 the "Limit Report to Selected
                                                                 Segments" is checked on, then
                                                                 the volume and  mass calculations
                                                                 are computed only for the seg-
                                                                 ments identified in the "file-
                                                                 name".1st file.
                    appendix d  «  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-22
                     Figure D-28 illustrates a time series plot of the mass of total phosphorus computed
                     by the procedure  described in.  The mass of total  phosphorus was  computed for
                     Chesapeake Bay and tidal tributary rivers using monthly mean data for each station
                     at each depth for the period of record. The "Date" and "Total" (sum of all columns
                     in mass file) columns were used from the mass file (.mas) to make this plot. The
                     linear trend line is superimposed to show the general rate of decline. This plot was
                     created by opening  the tp8497.mas file in Excel, selecting the line  chart button,
                     selecting the "Start_Date" and "Total" columns, and adjusting the legends and titles
                     as necessary to create the time series plot. The linear trend was added by selecting
                     the time series followed by "Chart:Add Linear Trendline".
                                        Mass of Total Phosphorus in Chesapeake Bay and
                                           Tidal Tributaries Computed by Interpolation
                                 6000000
                               o
                               in 1000000
                               U)
                               re
                                                               Date
                               D-28, TTime series plot of the mass of total phosphorus.
                                                                           I-.'  U U:
                     Monitoring data are required for the Interpolator to compute values. The file should
                     contain one value per depth per station for which data exist. If replicate values were
                     measured at some or all stations, they should be averaged at each station depth so
                     that only one value exists per depth per station. The overall data can represent one
                     cruise, a season of cruises, or a decade of data—there are no limitations on what the
                     data represent—that is up to the user to determine. It is best, statistically, to provide
                     as many data as possible. One method is to linearly interpolate values from surface
                     to bottom before creating the data file for the Interpolator. This will provide more
                     data for the Interpolator if it is valid to do so for the desired data.  For the 2D inter-
   appendix d

-------
                                                                                                        D-23
polation models, only one value per station should be used, since the depth value is
ignored. The file naming convention for the input file is 'filename'.d3d. The input
file has the following structure:
Line 1> contains a title that is meaningful to the user that identifies the contents of this dataset.
Line 2> contains a 2-digit parameter code, comma, and the spelled-out parameter name
Line 3> contains the start date, comma, and end date of the data
Line 4> contains the date and time the data were compiled
Line 5> contains the number of observations that follow
Lines 6+> contain the easting in UTM Zone 18 meters NAD83, comma, the northing in UTM Zone 18
meters NAD83, comma, the sample depth in meters,  comma, the measured value of the parameter,
comma, and the station ID
CHESAPEAKE BAY AND TRIES - Dissolved Oxygen - Measured Data -06JUL199315JUL1993
DO.Dissolved Oxygen
07/06/1993,07/15/1993
08/11/1997:15:11
1128
407056,4377577, 0.5,  7.7000,CB1.1
407056,4377577, 1.0,  6.8000,CB1.1
407056,4377577, 2.0,  6.2000,CB1.1
407056,4377577, 3.0,  5.7000,CB1.1

407056,4377577,4.0,  5.5000,CB1.1
407056,4377577, 5.0,  5.2000,CB1.1
411793,4365898,0.5,  5.9000,CB2.1
411793,4365898, 1.0,  5.7000,CB2.1
411793,4365898,2.0,  5.7000,CB2.1
411793,4365898,3.0,  5.7000,CB2.1

366939,4301041, 0.5,  7.5000,WT8.3
366939,4301041, 1.0,  7.3000, WT8.3
A metadata (documentation) file is created during the job. The default filename is
'filename'.met.
Check this file (using the Notepad editor) to see what calculations were performed
during the job.
Statistics Report for C:\Vol3D\BDO930701.est

Title: CHESAPEAKE BAY AND TRIES Dissolved Oxygen Measured Data
                                                                06JUL199315JUL1993
Parameter: Dissolved Oxygen  Parameter Code: DO
Data Period: 07/06/199307/15/1993
Data File Date: 08/11/1997:15:11
Observations: 1128
Maximum Number of Nearest Neighbors: 4
Minimum Number of Nearest Neighbors: 1
                      appendix d  »  User Guide arid Documentation for the Chesapeake Bay Interpolator

-------
D-24
                        Maximum Vertical Search Window: 4
                        Minimum Vertical Search Window: 0
                        Vertical Search Window Step Size: .5
                        Maximum Horizontal Search Radius: 25000
                        Missing Value: 9
                        Interpolator Model: DepthRadiusInterpolator
                        Interpolation Date: 10/6/97 10:55:55 AM
                        Bathymetry File: bay_trib.bth
                        Number of Bathymetry Regions: 68
                        Data Region File: bay_trib.reg
                        Number of Data Regions: 68
                        - Bathymetry Region ID: 1001 Region Name: CB1TF Data Points: 28 in data region 1001
                        - Cell Size  EW: 1000  NS: 1000 Vertical: 1
                        — 360 cells were interpolated in region CB1TF Subtotal: 360 total cells
                        — Region was calculated in 4 seconds.
                        - Bathymetry Region ID: 1002 Region Name: CB2OH Data Points: 48 in data region 1002
                        -  Cell Size EW: 1000 NS: 1000 Vertical: 1
                        — 1237 cells were interpolated in region CB2OH  Subtotal: 1597 total cells
                        	Region was calculated in 1 seconds.
                        Total Number of Cells Interpolated: 173805
                        Total Number of NonMissing Value Cells Interpolated: 160558
                        Total Number of Missing Value Cells Interpolated: 13247

                        Nearest Neighbors: 4  # of Cells: 101978
                        Nearest Neighbors: 3  # of Cells: 12358
                        Nearest Neighbors: 2  # of Cells: 33517
                        Nearest Neighbors: 1  # of Cells: 12705

                        173805 cells were calculated in 244 seconds.
                                            E^.fV:*r£S FILE

                        An interpolated estimates file is created during the job. The default filename is 'file-
                        name'.est. For the 3D interpolator model, this file contains the values for each cell
                        interpolated during the job, from surface to bottom for each cell location. For the 2D
                        interpolator models, this file contains the values for the top cell at each cell location.
                        While the interpolated value is  written to the surface cell  location in this file, its
                        value might represent the bottom value—i.e., the value might represent bottom layer
                        dissolved oxygen. All  cell  values below the top value will be set to missing (usually
                        -9). The file contents include:
                        Line l>Input data file name
                        Line 2>Data file description
                        Line 3> 2 digit parameter code and parameter name
                        Line 4>Start and end dates of data
                        Line 5>Date and time data file was compiled
                        Line 6>Number of data points, nearest neighbors, minimum neighbors, maximum vertical window,
                        minimum vertical window, vertical window step increase size, maximum search radius, missing value
                        Line 7>Name of interpolator used
                        Line 8>Date and time of job
                        Line 9>Bathymetry file used
   appendix d   »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                         D-25
Line 10>Data region file used
Line ll>Number of segments to interpolate
Line 12>Cell description for this segmentnumber of surface cells in segment, segment id, segment
name, cell e-w dimension in meters, cell n-s dimension in meters, cell vertical depth in meters
Line 13+>cell easting, cell northing, cells deep, interpolated values from surface to bottom.
C:\VOL3D\BDO970601.D3D
CHESAPEAKE BAY BY CRUISE - Dissolved Oxygen - Linear Interpolated Data -
  3JUN199712JUN1997
DO,Dissolved Oxygen
06/03/1997,06/12/1997
06/10/1998:8:55
1254,4,1,4,0,.5,25000,9
Interpolator Model: DepthRadiusInterpolator
6/17/98 10:24:26 AM
cbayS.bth
cbayS.reg
8
132,1001,CB1TF,1000,1000,1
403000,4384000,2,9.1,8.9
404000,4384000,5,9.1,8.9,8.8,8.8,8.8
404000,4383000,3,9.1,8.9,8.8
405000,4383000,8,10.3,9.9,9.5,8.8,8.8,8.8,8.8,8.8
405000,4382000,3,10.3,9.9,9.5
406000,4382000,1,9.9
405000,4082000,1,10.5
410000,4082000,4,10.5,10.5,10.4,10.3
404000,4081000,1,10.6
410000,4081000,3,10.5,10.5,10.4
405000,4079000,1,10.6
";(f -._E                                     I1-:  •   I •.'•'..-
Interpolated estimates files ('filename'.est) can be reformatted as 'filename'.txt files
which  can be readily imported  into  other applications, including Arc/Info and
Arc View.  The .TXT file  contains the values for each cell in the original Estimates
file, from  surface to bottom for each cell location. In addition, each line in the file is
padded with -9 values. So the  file is  a rectangular matrix  of data  with all values
having a value.  The file  is comma delimited,  and all extraneous blanks  have been
removed. The precision of the reported parameter values are assigned by the values
set in the  'parameter.sys' file. The .TXT file contents include:
Line l>Column Headings
Line 2+>cell easting, cell northing, segment name, cell e-w dimension in meters, cell n-s dimension in
meters, cell vertical depth in meters,  bathymetry depth in meters, interpolated values from surface to
bottom, additional depths padded with -9 down to layer_45, then bottom, minimum, maximum, mean,
and sum values for non-missing cells in this water column.
                      appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-26
                          UTM_X,UTM_Y,Segment,EW_dim,NS_dim,Vert_dim,n,layer_l,layer_2,layer_3,layer_4,layer_5,laye
                          r_6,layer_7,layer_8,layer_9,layer_10,layer_ll,layer_12,layer_13,layer_14,layer_15,layer_16,layer_17
                          ,layer_18,layer_19,layer_20,layer_21,layer_22,layer_23,layer_24,layer_25,layer_26,layer_27,layer_2
                          8,layer_29,layer_30,layer_31,layer_32,layer_33,layer_34,layer_35,layer_36,layer_37,layer_38,layer_
                          39,layer_40,layer_41,layer_42,layer_43,layer_44,layer_45,Bottom,Minimum,Maximum,Mean,Sum
                          403000,4384000,CBlTF,1000,1000,l,2,-9.0,-9.0,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9.0,-9.0,-9.0,-9.0,-9.0
                          404000,4384000,CBlTF,1000,1000,l,5,-9.0,-9.0,-9.0,-9.0,-9.0,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9.0,-9.0,-9.0,-9.0,-9.0

                          383000,4338000,CB3MH,1000,1000,l,6,8.7,8.2,8.0,7.4,6.2,4.7,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,4.7,4.7,8.7,7.2,43.2
                          384000,4338000,CB3MH,1000,1000,l,5,8.7,8.2,8.0,7.4,6.2,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,6.2,6.2,8.7,7.7,38.5
                          385000,4338000,CB3MH,1000,1000,l,6,8.7,8.2,7.9,7.3,6.2,4.8,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,4.8,4.8,8.7,7.2,43.1
                          386000,4338000,CB3MH,1000,1000,l,6,8.7,8.1,7.9,7.3,6.2,4.9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-
                          9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,-9,4.9,4.9,8.7,7.2,43.1
                          ' *  .» l   '   •                                       FILE

                          Interpolated estimates files ('filename' .est) can be reformatted as 'filename' .t3d files
                          which can be readily imported into other  applications, such as, NoeSys  and T3D.
                          The .T3D file contains the values for each cell in the original Estimates file, with one
                          cell value per line in the output file. The file is comma delimited, and all extraneous
                          blanks have been removed. The precision  of the reported parameter values  are
                          assigned by the values set in the 'parameter.sys' file. The .T3D file contents include:

                          Line l+>cell centroid easting, cell centroid northing, negative cell centroid depth in meters, interpo-
                          lated value for parameter

                          403000,4384000,-0.5,-9.0
                          403000,4384000,-!.5,-9.0
                          404000,4384000,-Q.5,-9.0
                          404000,4384000,-!.5,-9.0

                          391500,4304000,-9.5,0.8
                          391500,4304000,40.5,0.2
                          391500,4304000,41.5,0.1
                          391500,4304000,42.5,0.1
                                                    i

                         Each interpolator job requires a bathymetry file which defines the cell structure of
                         the desired body of water that  is  being interpolated. The  following  shows  the
                         contents of the cbayS.bth file:
                         Line l>Number of segments to interpolate
                         Line 2>Number of surface cells in segment 1, segment id, segment name, e-w cell size in meters, n-s
                         cell size in meters, cell depth in meters
   appendix d  *   User Guide and Documentation for the Chesapeake Bay interpolator

-------
                                                                                                           D-27
Line 3>Cell centroid easting in meters, cell centroid northing in meters, number of cells (>0) from
surface to bottom, cell centroid depths from surface to bottom. The Interpolator computes a value for
each cell centroid identified in the .bth file which is output to the .est file.
(Repeat 2 & 3 for each segment.)
132,1001,CB1TF,1000,1000,1
403000,4384000,2,0.5,1.5
404000,4384000,5,0.5,1.5,2.5,3.5,4.5
404000,4383000,3,0.5,1.5,2.5
405000,4383000,8,0.5,1.5,2.5,3.5,4.5,5.5,6.5,7.5
405000,4382000,3,0.5,1.5,2.5

410000,4363000,6,0.5,1.5,2.5,3.5,4.5,5.5
411000,4363000,10,0.5,1.5,2.5,3.5,4.5,5.5,6.5,7.5,8.5,9.5
412000,4363000,4,0.5,1.5,2.5,3.5
270,1002,CB 2OH, 1000,1000,1
403000,4363000,3,0.5,1.5,2.5
404000,4363000,4,0.5,1.5,2.5,3.5
405000,4363000,3,0.5,1.5,2.5
406000,4363000,3,0.5,1.5,2.5
400000,4362000,1,0.5

385000,4107000,1,0.5
381000,4106000,2,0.5,1.5
381000,4105000,2,0.5,1.5
381000,4104000,1,0.5
                     •      •  ;••

Each interpolator job requires  a data regions file  which defines  the geographic
boundary of the data for the body of water that is being interpolated. 77 data regions
have been created, one for each CBP segment. The data region is used to clip off data
that fall outside the desired geographic area that is being interpolated. A data regions
file  includes one or more data region  definitions that must match  the bathymetry
being  interpolated.  These data region  file names  are stored in a file,  regions, sys,
which is  required by the Interpolator. This file can contain 25  defined regions  files.
The order  of the entries in this  file  define the order presented  to  the user in the
GEOGRAPHY screen during the job. The structure of the regions.sys file is:

Line l+>Item identifier (sequential number of 1 to 25), comma, data region name,  comma,
corresponding bathymetry file name, comma, corresponding data region file name.
Repeat for each defined data region.
l,Bay and Tribs,bay_trib.bth,bay_trib.reg
2,Chesapeake Mainstem (CBlCB8),cbay8.bth,cbay8.reg
Each .reg file defined in the regions.sys file must have the following structure. The following shows the
contents of the cbayS.reg file:

Line l>Bathymetry file name
                      appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-28
                       Line 2>Data region file name
                       Line 3>Number of segments to interpolate
                       Line 4>First segment ID and name
                       Line 5> Data region ID
                       Line 6>Number of x-y points in this data region
                       Line 7+>Data region x-y points. First and last in each polygon must be the same to close the polygon.
                       Repeat for each data region.
                       cbayS.bth
                       cbayS.reg
                       8
                       1001,CB1TF
                       1001
                       8
                       398699,4385013
                       421073,4384159
                       413328,4355973
                       397839,4344869
                       383210,4352557
                       401281,4367077
                       398699,4385013
                       398699,4385013
                       1008,CB8PH
                       1008
                       13
                       410677,4131611
                       418631,4108780
                       422019,4095375
                       415538,4079614
                       408467,4086243
                       396978,4083150
                       384016,4087569
                       372968,4087569
                       372968,4095817
                       385194,4104508
                       374588,4117618
                       386372,4131464
                       410677,4131611
                                                 FILE

                       Each parameter is identified by a 2-digit parameter code and spelled out parameter
                       name. These codes and names are stored in the params.sys file. This file can accom-
                       modate 25 parameters. The order of the codes and names in this file determines the
                       order of the parameters in the PARAMETERS screen (Figure D-3). This file can be
                       edited as necessary by the user. The file structure is:
                       Line l>Item number (up to 25 lines), comma, spelled out parameter name, comma, 2-digit parameter
                       code, comma, number of digits precision to the right of the decimal in output file.
   appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                    D-29
1,Dissolved Oxygen,DO,l

2,Chlorophyll,CH,l

3,Salinity,SA,l

4,Water Temperature, WT,1

5,Total Nitrogen,TN,2

6, Ammonia, NH,3

7,Nitrite,N2,3

8,Nitrate,N3,3

9,Total Phosphorus,TP,2
           	::M RAX  :      :'••-,)

Several files are required to compute report files of volume and mass and to graphi-
cally portray the interpolated results. These include:

1) the "filename".est file of estimated values;

2) the cbpotiny.bmp which is a small CBP logo file;

3) the aro.bmp which is a small north arrow;

4) shore_18.bnd which is a shoreline boundary file; and,

5) the  "parameter_code".rng file. The range file  is used by the volume and mass
report procedures to subset the computed results  into categories for reporting. For
graphics, the .rng file defines how the graphics program assigns colors to each cell
value in the "filename".est file.  For drawing purposes, the first range in the .rng table
has drawing priority over the second range, which has priority over the third range,
etc, so the first range color will  paint over ranges lower in the table. This order deter-
mines which colors have priority in the final graphic. The do.rng file serves as an
example:
Line l>For dissolved oxygen values of 0.0 to but less than 0.2, color 12, pattern 0, title 0.0-0.2
Line 2>For dissolved oxygen values of 0.2 to but less than 1.0, color 13, pattern 2, title 0.2-1.0
Line 3>For dissolved oxygen values of 1.0 to but less than 3.0, color 14, pattern 12, title 1.0-3.0
Line 4>For dissolved oxygen values of 3.0 to but less than 5.0, color 11, pattern 12, title 3.0-5.0
Line 5>For dissolved oxygen values of 5.0 to but less than 25.0, color 19, pattern 21, title >5.0 MG/L
Line 6>For dissolved oxygen values of -10.0 to but less than -8.0 (-9=missing value), color 8, pattern
8, title Not Calculated

Pattern is currently ignored in this version.

To  categorize integer value  ranges, it is best to bracket the range, for instance, to
assign color 12 to the range of 2  (lower bound) to 2 (upper bound), set the lower
bound to 1.9 and the upper bound to 2.1. Set the title to "2" to convey the intent that
"2" is the range being presented.  This bracketing is required to  allow the code to
select values equal to or greater than the lower bound and less than the upper bound.
                     appendix d  *  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-30
                        Acceptable color codes are:

                        0       Black
                        1       Blue
                        2       Green
                        3       Cyan
                        4       Red
                        5       Magenta
                        6       YeUow
                        7       White
                        8       Gray
                        9       Light Blue
                        10      Light Green
                        11      Light Cyan
                        12      Light Red
                        13      Light Magenta
                        14      Light Yellow
                        15      Bright White
                        0.0,0.2,12,0,0.00.2
                        0.2,1.0,13,2,0.21.0
                        1.0,3.0,14,12,1.03.0
                        3.0,5.0,11,12,3.05.0
                        5.0,25.0,9,21,>5.0 MG/L
                        10.0,8.0,8,8,Not Calculated

                        Individual interpolator runs can be computed sequentially by saving the necessary
                        information for the run in a "Batch File" ('filename'.job). This "Batch File" is then
                        used to calculate each of the files identified in the .job file. This file can be edited
                        as necessary by the user. The file structure is:
                        Line l>bathymetry file
                        Line 2>regions file
                        Line 3>input data file
                        Line 4> output interpolated (.est) file
                        Line 5> output metadata (.met) file
                        Line 6>parameter transformation
                        Line 7>minimum number of neighbors
                        Line 8>maximum number of neighbors
                        Line 9>horizontal range (m)
                        Line 10>vertical range minimum
                        Line ll>vertical range maximum
                        Line 12>Vertical step size
                        Line 13>missing value
                        Line 14>interactive/batch flag
                        Repeat lines 1-14 for each file to be interpolated
   appendix d  •  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                  D-31
cbayS.bth
cbayS.reg
C:\Vol3D\BDO970601.D3D
C:\Vol3D\BDO970601.est
C:\Vol3D\BDO97060Lmet
None
 1
 4
 25000
 0
 4
 .5
-9
1
cbayS.bth
cbayS.reg
C:\Vol3D\BDO970701.D3D
C:\Vol3D\BDO970701.est
C:\Vol3D\BDO970701 .met
None
 1
 4
 25000
 0
 4
 .5
-9
1
Calculations on individual interpolator .est files can be computed sequentially by
reading the .est file names from a "batch file list" ('filename'.fls). This file is created
when the Batch Job File ('filename'.job) is created. This file can be edited as neces-
sary by the user. The file structure is:

Line 1 interpolator (.est) file name

Repeat for each file to be processed.
C:\Vol3D\BDO970601.est

C:\Vol3D\BDO970701.est
                 .'4, i   ^  .•  }•   ,  i!)

Calculations of "Change Over Time" require a Julian date file which contains the
dates which relate to the .est files identified in the .fls file. For this analysis, the Julian

                    appendix d  *  User Guide and Documentation for the Chesapeake Bay interpolator

-------
D-32
                     dates are the X variable of the time series and the .est files are the Y variables of the
                     time series. The Julian dates represent dates and times that are based on the decimal
                     numbering system, rather than years, months, and days (and time). The Julian (or any
                     linearly numbered scheme) date file must be created by  the user. The following
                     example was created by opening the appropriate .mas file  in Excel, converting the
                     "Start_Date" from mm/dd/yy format to decimal format, and cutting and pasting the
                     reformatted  date column into a flat file. The file structure is:
                     Line l>julian date

                     Repeat for each file to be processed.
                     30865.00

                     30895.00

                     30929.00

                     30956.00
                                               LIST FILE {.1st}
                     Reports on individual segments can be computed sequentially by reading the .est file
                     names from a "segments list file" ('filename'.1st). This file is created manually by
                     the user. The file structure is:
                     Line l>Number of segments to process
                     Line 2+>Segment name (spelling must match those in Appendix A).

                     Repeat for each segment to be processed.
                     3

                     POCTF

                     POCOH

                     POCMH
                                                      FILE
                     The volume of water that contains a specified range of concentrations of a parameter
                     can be computed and saved to a volume report file ('filename'.vol). Volume esti-
   appendix d  *  User Guide and Documentation for the Chesapeake Bay interpolator

-------
                                                                                                D-33
mates are reported in liters. The entire Bay and tributary volume based on the "Bay
and Tributary" 77 segment bathymetry is 75,199,817,500 mA3, or 75.2xlOA12 liters.
The volume of Main Bay segments CB1TF-CB8PH totals 51.839xlOA12 liters. Data
from one job to another may be appended to the same output file so that a time series
file is created that can be opened in a spreadsheet or database program for further
graphing or analysis. The first line of the  file is composed of  'column headings'
contained within quotes. This file can be edited as necessary by the user. The file
structure is:
Line l>Column headings defined by the 'Report' job that is run, including data start date, data end
date, depth of top layer analyzed, depth of bottom layer analyzed, volume for segment, volume by
concentration range for that segment,...,..., repeat for each segment,... .grand total volume
Line 2>data accumulated from the input interpolated file (.est) for each column in line 1
Repeat for each interpolated file processed.

Note: Since data that are calculated may be appended to an existing file, there is a
risk that the user may append data from different bathymetry jobs. The user must be
careful not to mix 8 segment mainstem data with 77 segment mainstem and tributary
data in this report, or else the column headings will not represent the data.
"Start              Date","End               Date","Layer              Top","Layer
Bottom" ,"CB1 TF","CB1 TF_0.0_0.2","CB1 TF_0.2_1.0","CB1 TF_1.0_3.0","CB1 TF_3.0_5.0","CB
1  TF_5.0_25.0","CB1  T  F  _  -  1   0  .  0  _  -
8.0","CB2OH","CB2OH_0.0_0.2","CB2OH_0.2_1.0","CB2OH_1.0_3.0","CB2OH_3.0_5.0","CB
2OH_5.0_25.0","CB2OH_-1  0  .  0  _  -
8.0","CB3MH","CB3MH_0.0_0.2","CB3MH_0.2_1.0","CB3MH_1.0_3.0","CB3MH_3.0_5.0","CB
3MH_5.0_25.0","CB3MH_-10.0_-
8.0","CB4MH","CB4MH_0.0_0.2","CB4MH_0.2_1.0","CB4MH_1.0_3.0","CB4MH_3.0_5.0","CB
4MH_5.0_25.0","CB4MH_-10.0_-
8.0","CB5MH","CB5MH_0.0_0.2","CB5MH_0.2_1.0","CB5MH_1.0_3.0","CB5MH_3.0_5.0","CB
5MH_5.0_25.0","CB5MH_-1  0  .  0  _  -
8.0","CB6PH","CB6PH_0.0_0.2","CB6PH_0.2_1.0","CB6PH_1.0_3.0","CB6PH_3.0_5.0","CB6
PH_5.0_25.0","CB6PH_-1   0   .  0  _  -
8.0","CB7PH","CB7PH_0.0_0.2","CB7PH_0.2_1.0","CB7PH_1.0_3.0","CB7PH_3.0_5.0","CB7
PH_5.0_25.0","CB7PH_-10.0_-
8.0","CB8PH","CB8PH_0.0_0.2","CB8PH_0.2_1.0","CB8PH_1.0_3.0","CB8PH_3.0_5.0","CB8
PH_5.0_25.0","CB8PH_-10.0_-8.0","Total"

"06/03/1997" ,"06/12/1997',0., 50., 359000000000. ,0,0,0,0, 359000000000. ,0., 12370000000
00., 0,0,0, 56000000000. ,1181000000000. ,0., 2391 000000000. ,0., 2000000000. ,366000000
000.,361000000000.,1662000000000.,0.,9237000000000.,0.,19000000000.,197400000000
0,1104000000000., 6140000000000., 0,15377000000000. ,0,0,0,851000000000.,1452600
0000000., 0., 6503000000000., 0., O.,0.,0., 6503000000000. ,0.,13488000000000. ,0.,0.,0.,0., 13
488000000000.,0.,3150000000000.,0.,0.,0.,0.,3150000000000.,0.,51742000000000.,0.,210
00000000.,2340000000000.,2372000000000.,47009000000000.,0.

"07/07/1997" ,"07/15/1997", 0., 50. ,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 2391000000000. ,13400
0000000.,51000000000.,150000000000.,168000000000.,1888000000000.,0.,92370000000
00.,1937000000000.,1253000000000.,933000000000.,620000000000.,4494000000000.,0.,
15395000000000.,1391000000000.,2236000000000.,1909000000000.,1069000000000.,87
90000000000.,0.,6503000000000.,0.,0.,524000000000.,1063000000000.,4916000000000.,
                    appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-34
                     0., 13482000000000. ,0,0, 661000000000. ,2978000000000., 9843000000000. ,0., 247400000
                     0000., 0,0,0. ,114000000000., 2360000000000. ,0., 49482000000000. ,3462000000000. ,3540
                     000000000.,4177000000000.,6012000000000.,32291000000000.,0.

                     "07/1 6/1997" ,"07/31/1997",0, 50., 360000000000. ,0,0,0,0, 360000000000. ,0., 12370000000
                     00.,0.,0.,7000000000.,24000000000.,1206000000000.,0.,2391 000000000.,0.,13300000000
                     0.,159000000000.,282000000000.,1817000000000.,0.,9237000000000.,0.,1620000000000.
                     , 1644000000000. ,942000000000. ,5031000000000. ,0.,15388000000000. ,0., 1453000000000
                     .,2710000000000.,2005000000000.,9220000000000.,0.,6503000000000.,0.,0.,5500000000
                     0., 1231000000000. ,5217000000000. ,0., 13491000000000. ,0,0,9000000000. ,17930000000
                     00.,11689000000000., 0., 3160000000000. ,0,0,0,0, 3160000000000. ,0., 51767000000000.,
                     0.,3206000000000.,4584000000000.,6277000000000.,37700000000000.,0.
                                                     :  ; "  .-j

                     The mass file report contains the mass of a parameter computed for each cell in the
                     interpolated (.est) file then summed in one of two ways. The default method (below
                     example) is to sum the mass by segment and total for all segments in the bathymetry.

                     The second method follows the format of the volume report and computes the mass
                     by concentration range for each segment.

                     The mass  that is computed and summed is saved to a mass report  file ('file-
                     name'.mas). It is assumed the input data are measured in [units]/[liter], such as mg/1
                     or ug/1 or counts/liter. In the  mass report, the resulting mass estimates are computed
                     by multiplying the [estimated concentration in the cell (often in mg/1)] * [the volume
                     of the cell in mA3 (for instance, 1000m east-west x 1000m north-south x 1m deep)]
                     * [1000 l/mA3 to convert from mA3 to liters]. Hence, if the input data were in mg/1
                     and then the concentration is estimated to be 6mg/l in a cell, the resulting mass will
                     be 6*10A9 mg for a 1km x 1km  x 1 m cell. As a second example, if the input data
                     were in mg/mA3, which is equivalent to ug/1, then the reported mass values would be
                     in micrograms to account for the volume being reported in mA3 rather than liters. If
                     the input data are counts (such as organism counts)  per liter, then the mass report
                     units would be total counts.  If the input data are counts (such as organism counts)
                     per cubic meter, then the total counts in the mass report must be divided by 1000 to
                     account for the conversion from cubic meters to liters between the input data and the
                     interpolated counts.  The mass (or counts)  for each cell is then  summed for a total
                     mass (or count) in the segment and also a grand sum of mass (or count) for the total
                     for all segments under analysis. For instance, if the input data for CHLA were meas-
                     ured as ug/1  and the  resulting mass in Segment CB2OH was  reported after
                     interpolation as  13,000,000,000,000, that represents 1.3A13 ug CHLA for Segment
                     CB2OH—i.e 1.3A13ug / 1.237M2 liters in  CB2OH=10.5 ug/1 average. As a second
                     example, if the input data were for mg biomass of organisms per cubic meter and the
                     resulting mass in Segment CB2OH was reported as 132,627,709,873,200, that repre-
                     sents 1.326A14 / 1000 mg for Segment CB2OH, since an adjustment for the input
                     data must be made for the per cubic meter to per liter basis. A quick check can be
                     made by multiplying  the average input data value by the volume of a  segment to
                     determine if the results are within reason. For instance, if there were approximately
   appendix d •  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                    D-35
150 mg biomass per cubic meter in the monitoring data for CB2OH, that would be
[150 mg/mA3] * [1,237,000,000 cubic meters in Segment CB2OH] = 1.86A11  mg
biomass in the  Segment CB2OH,  which is close to the interpolated value  of
1.326A11 mg, above.
Data from one job to another may be appended to the same output file so that a time
series file is created that can be opened in a spreadsheet or database program for
further graphing or analysis. The first line of the file is composed of 'column head-
ings' contained within quotes. This file can be edited as necessary by the user. The
file structure is:
Line l>Column headings defined by the 'Report' job that is run, including data start date, data end
date, depth of top layer analyzed, depth of bottom layer analyzed, mass of parameter by
segment,...,..., repeat for each segment,...,grand total mass
Line 2>data accumulated from the input interpolated file (.est) for each column in line 1

Repeat for each interpolated file processed.

Note: Since data that are calculated may be appended to an existing file, there is a
risk that the user may append data from different bathymetry jobs. The user must be
careful not to mix 8 segment mainstem data with 77 segment mainstem and tributary
data in this report, or else the column headings will not represent the data.
"Start               Date","End                Date","Layer                Top","Layer
Bottom","CBlTF","CB2OH","CB3MH","CB4MH","CB5MH","CB6PH","CB7PH","CB8PH","Total"
"06/03/1997","06/12/1997",0.,50.,33956000.,99418000.,165733000.,602087001.,1229455000.,56352
3000.,! 121378003.,301820000.,4117370004.
"07/07/1997","07/15/1997",0.,50.,0.,0.,149745000.,378545000.,726926999.,429706000.,871456001.,
173612000.,2729991001.
"07/16/1997","07/31/1997",0.,50.,21746000.,82626000.,150816000.,438919000.,795246002.,428770
O00.,917109001.,243880000.,3079112003.
"08/04/1997","08/14/1997",0.,50.,22096000.,73961000.,128601000.,465053000.,824646000.,456154
001.,940162001.,236345001.,3147018002.
"08/18/1997","08/28/1997",0.,50.,19932000.,82306000.,158793000.,460818001.,885485001.,470413
001.,1010405001.,231651001.,3319803005.
"09/02/1997","09/15/1997",0.,50.,19748000.,76169000.,128181000.,517683000.,1009546002.,45963
2001.,949621002.,227962000.,3388542005.
"10/06/1997","10/15/1997",0.,50.,0.,0.,173127001.,576579000.,1004260004.,482547000.,963827000.
,200433996.,3400774001.
                      ,   '  ,  ' •  ,  B
1) Double click the Vol3D.exe icon to run the Interpolator program.
2) Click the GEOGRAPHY button to display the GEOGRAPHY screen.
3) Choose "Chesapeake Mainstem (CB1-CB8)" to interpolate the Main Bay.
4) Click the PARAMETER button to display the PARAMETER screen.


                     appendix d  •  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-36
                     5) Choose "DO- Dissolved Oxygen" as the parameter.
                     6) Click the DATA IMPORT button to display the DATA IMPORT screen.
                     7) Click Get File Name button and select the "C:\Vol3D\BDO970701.D3D" data
                     file. The file name, start and end dates, number of observations, file date, parameter,
                     code, and title should appear in the Data Import screen. If not, you selected an incor-
                     rect file.
                     8) Click the INTERPOLATE button to display the INTERPOLATE screen. The
                     input file should read "C:\Vol3D\BDO970701.D3D", the output file should read
                     "C:\Vol3D\BDO970701.est", Bathy file should read "cbayS.bth", and metadata file
                     should read "C:\Vol3D\BDO970701.met".
                     9) Click the "Run Interpolation" button to create the standard *.est file or click the
                     "Also Create TXT File" to create a "filename".txt file that can be imported into
                     Arc/Info or Arc View as a table. The text Arc View file will be approximately 2.3 mb
                     in size.
                     10) If you created an interpolated .est file, you can view the results by clicking the
                     GRAPHICS button to display the GRAPHICS screen. If you created a .txt file to
                     load into an Arc View table, you can quit the Interpolator program and continue
                     working with the output file in Arc View.
                     11)   At  the   GRAPHICS   screen, the   Interpolated  file   should  read
                     "C:\Vol3D\BDO970701.est". Click  the Interpolated File button to load titles  and
                     dates for the graphic.
                     12) The Bathymetry file should read  "cbayS.bth",  the Logo File  1  should read
                     ".\cbpotiny.bmp", the Logo File 3  should  read ".\aro.bmp",  the Categories File
                     should read ".\DO.rng". Click the Categories File  button to load the categories for
                     the graphic. The Boundary File should read ".\shore_18.bnd", the output Graphics
                     File name should read "C:\Vol3D\BDO970701.bmp". The titles and legends were
                     loaded by pushing the Interpolated File  and Categories File buttons. The background
                     color of the boundary file is set by  clicking the small grey box to the right of the
                     Boundary File name. Click "Plot Points" ON if you want to display the location of
                     the monitoring stations. Click "Plot Data Regions" ON if you wish to see the Data
                     Region polygons. Choose "Minimum" if you wish to display the minimum color
                     value (where minimum is the worst case, such as dissolved  oxygen), or choose
                     "Maximum"  if you wish to display  the maximum  color value (where maximum is
                     the  worst case, such  as  temperature),  or choose  the "Top/West" edge  or
                     "Bottom/East" edge to display the desired side.
                     Titles, categories, colors, and legends can be modified  on this screen and will be
                     reflected in the resulting drawing. Clicking the "DRAW" button will draw the image
                     in a graphics window. The graphics window can be saved to a file or printed. This
                     version of the Interpolator does not  allow  graphical  editing.  The saved "file-
                     name" .bmp file can be edited in a  commercial graphics editing package, such as
                     Lview Pro or Corel Draw. The .bmp file can be converted to gif or jpeg format for
                     publication on the web.
   appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
                                                                                                         D-37
APPENDIX

Segment Name, EW-Dimension,  NS-Dimension,  Depth Dimension, Number Cells in Segment,
Segment Volume
CB1TF, 1000,1000,1,3 60,3 60000000
CB2OH, 1000,1000,1,1237,1237000000
CB3MH, 1000,1000,1,2391,2391000000
CB4MH, 1000,1000,1,9237,9237000000
CB5MH,1000,1000,1,15416,15416000000
CB6PH,1000,1000,1,6503,6503000000
CB7PH,1000,1000,1,13523,13523000000
CB8PH,1000,1000,1,3172,3172000000
NORTF,500,500,1,106,26500000
C&DOH,100,100,1,2413,24130000
ELKOH,500,500,1,405,101250000
BOHOH,250,250,1,272,17000000
SASOH,250,250,1,1347,84187500
CHSTF,50,50,1,1345,3362500
CHSOH,250,250,1,462,28875000
CHSMH,500,500,1,1821,455250000
EASMH,500,500,1,3987,996750000
CHOTF,50,50,1,6129,15322500
CHOOH,250,250,1,722,45125000
CHOMH2,500,500,1,1067,266750000
CHOMH1,1000,1000,1,945,945000000
LCHMH,500,500,1,833,208250000
HNGMH, 100,100,1,18568,185680000
FSBMH,1000,1000,1,143,143000000
NANTF,50,50,1,2646,6615000
NANOH,50,50,1,18000,45000000
NANMH,500,500,1,389,97250000
WICMH,100,100,1,5642,56420000
MANMH,500,500,1,358,89500000
BIGMH,250,250,1,698,43625000
POCTF,50,50,1,1788,4470000
POCOH,50,50,1,7200,18000000
POCMH,500,500,1,1418,354500000
T ANMH, 1000,1000,1,4019,4019000000
BSHOH,500,500,1,197,49250000
GUNOH,500,500,1,257,64250000
MIDOH,250,250,1,400,25000000
BACOH,250,250,1,358,22375000
PATMH,500,500,1,1806,451500000
MAGMH,250,250,1,1224,76500000
SEVMH,250,250,1,1815,113437500
SOUMH,250,250,1,1072,67000000
RHDMH,250,250,1,325,20312500
WSTMH,250,250,1,326,20375000
P AXTF,50,50,1,4410,11025000
PAXOH,100,100,1,2718,27180000
PAXMH,500,500,1,2244,561000000
PISTF,100,100,1,285,2850000
MATTF,250,250,1,152,9500000
POTTF,500,500,1,1939,484750000
POTOH,500,500,1,3409,852250000
POTMH,1000,1000,1,5792,5792000000
                      appendix d   »  User Guide and Documentation for the Chesapeake Bay Interpolator

-------
D-38
                       RPPTF.250,250,1,1719,107437500
                       RPPOH,100,100,1,5358,53580000
                       RPPMH,500,500,1,5929,1482250000
                       CRRMH,250,250,1,1051,65687500
                       PIAMH,250,250,1,3223,201437500
                       MPNTF,50,50,1,6135,15337500
                       MPNOH, 100,100,1,3539,35390000
                       PMKTF,50,50,1,11452,28630000
                       PMKOH,100,100,1,6668,66680000
                       YRKMH,500,500,1,1102,275500000
                       YRKPH,500,500,1,1603,400750000
                       MOBPH,500,500,1,5370,1342500000
                       APPTF,100,100,1,151,1510000
                       CHKOH,250,250,1,777,48562500
                       JMSTF,250,250,1,4579,286187500
                       JMSOH,500,500,1,1726,431500000
                       JMSMH, 1000,1000,1,977,977000000
                       JMSPH, 1000,1000,1,434,434000000
                       WBEMH,100,100,1,631,6310000
                       SBEMH, 100,100,1,2773,27730000
                       EBEMH,50,50,1,2584,6460000
                       ELIMH,100,100,1,5339,53390000
                       LAFMH,100,100,1,339,3390000
                       ELIPH,500,500,1,246,61500000
                       LYNPH, 100,100,1,1673,16730000
                       Total Volume (mA3) = 75199817500
   appendix d  »  User Guide and Documentation for the Chesapeake Bay Interpolate

-------
                                                                                    E-1
                       appendix
 Potential  Methods  for Assessing
      Shorter  Duration  Dissolved
                  Oxygen  Criteria
                    POTENTIAL METHODS

The 2003 Chesapeake Bay water-quality criteria document described three alternatives
for assessing attainment of the short duration dissolved oxygen criteria (U.S. EPA
2003). Those include: 1) logistic regression; 2) a time series statistical method and 3)
continuous dissolved oxygen data collection using meters that are deployed for an
extended period of time. Each of these  approaches has  strengths and drawbacks.
Appropriate implementation of logistic regression or time series statistical methods
may require continuous dissolved oxygen data. To develop the full capacity to assess
the shorter duration dissolved oxygen criteria—7-day mean, 1-day mean and instanta-
neous minimum, EPA recommends a phased approach in which the methods that are
easiest to implement are employed initially while continuing to work on development
and implementation of the more detailed and/or expensive methods.

LOGISTIC REGRESSION

The instantaneous minimum criteria imply the requirement that waters within the
respective designated use be at or above the defined concentration everywhere all the
time. Stated in this way, the logistic regression approach clearly has application to
the challenge of assessing attainment of instantaneous  minimum criteria. In the
context of criteria attainment, logistic equations are developed from the long term
dissolved oxygen data record, which predict the probability that the defined criteria
concentrations were met, based on observed monthly mean concentration.

The logistic regression approach  utilizes a well-established  statistical procedure
(U.S. EPA 2004) and has been employed in the past in Chesapeake Bay to estimate
instantaneous minima (Jordan et al. 1992). It is relatively simple to use and only
requires regular updating to keep the predictive models relevant to  current condi-
tions. The limitation of this approach is that it is based on an extrapolation of the
fixed-station data and is likely to have higher error than the other methods.
            appendix e •  Potential Methods for Assessing Shorter Duration Dissolved Oxygen Criteria

-------
E-2
                     The logistic regression approach could be also be adapted to assess attainment of the
                     7-day and 1-day mean criteria components as well as other duration-specific criteria,
                     where and when a body of observational data is available at frequencies relevant to
                     the time frame. High frequency 'buoy' data sited at sentinel locations, where contin-
                     uous records  extend over  days, weeks and months, would offer opportunities to
                     develop logistic models of the relationship between exceedance/attainment and the
                     temporal means. EPA recommends  that this method  be  actively developed for
                     possible employment for attainment assessments of the instantaneous minimum
                     dissolved oxygen  criteria (see next  section for details) while  additional  high
                     frequency data are  collected  and more complex, detailed methods described below
                     are being developed.



                     The time series approach utilizes a statistical procedure known as spectral analysis
                     to synthesize a complete record of dissolved oxygen concentrations at short interval
                     time steps over time. The synthetic record is developed using continuous measure-
                     ment data from nearby locations to develop a model that predicts the short-interval
                     variations in concentration. That model is combined with the long-term pattern of
                     variability derived  from  data collected routinely,  monthly to twice monthly, at the
                     fixed-stations located in the assessment unit. The synthetic dissolved oxygen record
                     can then be used in the same way that data collected using a continuous meter would
                     be used. This time series approach has only been applied in a limited way to date and
                     further development is needed in order for it to fully meet the needs of a publishable
                     Chesapeake Bay dissolved oxygen criteria assessment methodology (see pages 183-
                     185 in U.S. EPA 2003). EPA recommends that this  development work proceed
                     simultaneously with the development of the logistic regression and that the spectral
                     analysis method replace the logistic regression in the future should it prove a more
                     robust method.
                     The most rigorous approach for assessing attainment of the high frequency dissolved
                     oxygen  criteria would be  to collect  continuous measures  of dissolved oxygen
                     concentration at representative locations and depths throughout  each spatial assess-
                     ment unit. The temporal and spatial density of such data would need to be sufficient
                     to enable all of the dissolved oxygen criteria to be assessed simply by calculating
                     means at the appropriate time scales (e.g. 30-day, 7-day,  1-day) or by observing
                     violations of the instantaneous  minimum criteria  values. However, continuous
                     collection of high frequency dissolved oxygen concentration in the Bay is expensive
                     both in purchasing the equipment and maintaining it. It is also difficult or impossible
                     to find sufficiently representative locations where the equipment can be affixed to
                     buoys or fixed pilings. Finally, it is expensive  and labor-intensive to  maintain  the
                     equipment and sensor calibration once it is deployed due to the effects of weather,
                     turbulence,  biological fouling and human interferences (e.g.  accidents,  thefts).
                     Nevertheless, the collection of at least some continuous dissolved oxygen data will
                     be critical for use in the other two statistical  analysis-based assessment methods
  appendix e  •  Potential Methods for Assessing Shorter Duration Dissolved Oxygen Criteria

-------
                                                                                               E-3
described above. Therefore, EPA recommends that the States continue to seek funds
to support this type of data collection in order to directly generate the data supporting
attainment assessment of the full array of applicable dissolved oxygen criteria.

                OF                            TO
In the prior sections, it was noted that the data collection frequency of the long term,
fixed-station water quality monitoring program is inadequate to assess attainment of
short-duration criteria components. However, the greater than  20-year record of
dissolved oxygen  measurements collected relatively synoptically throughout the
mainstem Bay, tidal tributaries and embayments, and collected regularly throughout
the  annual cycle provides a very substantial data base from which to derive infer-
ences and define quantitative relationships between seasonal and  monthly mean
dissolved oxygen concentrations and the frequency of observations above and below
specified criterion concentrations. Where relationships are strong, the logistic regres-
sion  procedure  produces  models  in  the  form  of  simple  equations that
estimate/predict the likelihood that the criterion threshold concentration was attained
or violated during the period.

This method was explored originally to measure attainment of the 1992 Chesapeake
Bay dissolved oxygen restoration goal (Jordan et al. 1992) and was  adapted for
assessing  attainment  of the 2003 Chesapeake Bay dissolved oxygen instantaneous
minimum (see Chapter 5, pages 27-62, in U.S. EPA 2004). The 2003 method modi-
fications included  spatial and temporal refinements to the predictive models, with
consequent improvements to  their  goodness of fit. The early (1992) models esti-
mated exceedance based on segment-specific seasonal means and whether the means
were from depths above or below pycnocline. The 2003 method update was enriched
with an additional decade of monitoring data (1990-2000) for the regression analysis
and  provided segment-specific models for individual months and depths. Recent
progress on this  work again includes several additional years of new fixed-station
and continuous monitoring buoy  data (2001-2005) and modifications to implemen-
tation procedures that could provide results for attainment assessment through the
CFD methodology in a format consistent with other dissolved oxygen criteria.

In this latest  iteration,  logistic regression models  for the individual instantaneous
minima are developed  for each station.  The independent variables are, as before,
mean dissolved oxygen, month and water depth. The addition of a depth-squared
variable for deep stations is being tested, but not yet implemented.  The dependent
variable is an indicator that the minimum threshold (e.g., the instantaneous criterion
concentration) is violated. (Since the CFD methodology is based on percent failure,
the  dependent variable is  based on exceedance rather than attainment.) This model-
building step  currently  uses the entire  1985-2005 water quality data record at each
station. Over  time, however, if trends in ambient dissolved oxygen indicate signifi-
cant, sustained change  in a segment, then the extent of the historical record to be
included in this step should be re-examined.

The collection of station models is used to estimate  a  predicted probability of
exceedance for each station,  for each month in the 3-year,  multi-month seasonal
assessment period, at each meter of depth. Then, for each month, the predicted prob-


              appendix e  «   Potential Methods for Assessing  Shorter Duration Dissolved Oxygen Criteria

-------
E-4
                     abilities are spatially interpolated to estimate probabilities for all interpolator cells
                     that represent the bathymetry of the Bay, its tidal tributaries and embayments. The
                     interpolator cells that are contained within the designated use where the criterion
                     applies are parsed out by segment and the probabilities calculated for each cell are
                     evaluated cell-by-cell against a threshold of probability which indicates an unac-
                     ceptably high risk that the dissolved oxygen criterion was exceeded (Jordan et al
                     1992). The volume of water represented by the interpolator  cells  exceeding the
                     threshold as a percentage of the total volume in the designated use is tallied for each
                     segment, for each month in the  assessment period.

                     There are several elements of the logistic regression approach which should be eval-
                     uated as part of the attainment assessment procedure. Each of the station-specific
                     logistic models has its own goodness-of-fit measure. Each station will have a result
                     from  the predictive  model, i.e., the probability of exceeding the instantaneous
                     minimum over the assessment unit. Each segment will have an estimate of the
                     percent volume exceeding the criteria, based on spatial interpolation of the station
                     probabilities. As with other components of the dissolved oxygen criteria, these
                     results can also be assessed and visualized using the CFD methodology, although
                     this is not mandatory.

                     The  limitations of this  methodology  have been  noted earlier,  particularly the
                     temporal frequency on which the  models are based. In addition, the lack of good
                     spatial representation in the tidal tributaries and embayments is a concern. Most of
                     the fixed-stations are situated more or less longitudinally in mid-channel and there
                     is insufficient lateral coverage of the flanks, where different oxygen conditions and
                     different model relationships  may exist. Data now being collected through the
                     Chesapeake Bay Shallow Monitoring Program will  help answer where and to what
                     extent this is true.
                     Jordan, J., C. Stenger, M. Olson,  R.  Batiuk and K. Mountford. 1992.  Chesapeake Bay
                     Dissolved Oxygen Goal for Restoration of Living Resource Habitats. CBP/TRS  88/93.
                     Chesapeake Bay Program, Annapolis, Maryland.
                     U.S. Environmental Protection Agency. 2003. Ambient Water Quality Criteria for Dissolved
                     Oxygen, Water Clarity and Chlorophyll a for the Chesapeake Bay and Its Tidal Tributaries.
                     EPA 903-R-03-002. Region III Chesapeake Bay Program Office, Annapolis, Maryland.
  appendix e  »  Potential Methods for Assessing Shorter Duration Dissolved Oxygen Criteria

-------
                                                 F-1
              appendix




    Data Used in Deriving the


  Open-Water, Deep-Water and


Deep-Channel Dissolved Oxygen


    Criteria Summer Biological


Table F-1. Designated use, segment, year combinations found to be "good" using
     the Benthic-IBI summer reference curve area locator method described in
     Chapter 4.
CBP Segment
CB6PH
CB7PH
CB8PH
JMSPH
LCHMH
NANOH
RPPOH
CB6PH
CB8PH
CHOMH1
CHSMH
CHSMH
CHSMH
JMSOH
JMSPH
RPPOH
YRKMH
CB3MH
CB6PH
CB8PH
CHOMH1
CHSMH
JMSOH
JMSPH
NANMH
PMKTF
RPPMH
RPPMH
RPPOH
CB2OH
Year
1985
1985
1985
1985
1985
1985
1985
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1987
1987
1987
1987
1987
1987
1987
1987
1987
1987
1987
1987
1988
Designated Use
DW
OW
OW
OW
OW
OW
OW
DW
OW
OW
OW
DW
DC
OW
OW
OW
OW
DW
DW
OW
OW
DW
OW
OW
OW
OW
OW
DW
OW
OW
CBP Segment
CB7PH
JMSMH
JMSPH
NANMH
PAXMH
PMKTF
RPPMH
YRKMH
CB2OH
CB3MH
CB8PH
JMSPH
POTMH
CB1TF
CB7PH
CB8PH
CHOOH
CHSMH
CHSOH
JMSPH
JMSTF
PAXOH
RPPMH
CB6PH
CB7PH
CB8PH
CHOMH2
JMSMH
JMSPH
JMSTF
Year
1988
1988
1988
1988
1988
1988
1988
1988
1989
1989
1989
1989
1989
1990
1990
1990
1990
1990
1990
1990
1990
1990
1990
1991
1991
1991
1991
1991
1991
1991
Designated Use
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
OW
               appendix f • Data Used in Deriving Summer Biological Reference Curves

-------
F-2
CBP Segment
PMKTF
POTMH
RPPMH
RPPMH
CB1TF
CB2OH
CB5MH
CB6PH
CB6PH
CB8PH
CHOTF
CHSMH
CHSMH
CHSOH
ELKOH
JMSPH
JMSTF
PMKTF
POTMH
POTTF
RPPMH
SASOH
CB3MH
CB6PH
CB6PH
CB7PH
CB8PH
CHOMH2
CHSMH
CHSMH
JMSPH
JMSTF
PMKTF
CB2OH
CB5MH
CB7PH
CB8PH
CHOMH2
CHSMH
HNGMH
JMSMH
JMSPH
LCHMH
PMKTF
BSHOH
CB1TF
CB3MH
CB6PH
CB6PH
Year
1991
1991
1991
1991
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1992
1993
1993
1993
1993
1993
1993
1993
1993
1993
1993
1993
1994
1994
1994
1994
1994
1994
1994
1994
1994
1994
1994
1995
1995
1995
1995
1995
Designated Use
OW
DW
OW
DW
OW
OW
OW
OW
DW
OW
OW
OW
DC
OW
OW
OW
OW
OW
DW
OW
OW
OW
DW
OW
DW
OW
OW
OW
DW
DC
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
CBP Segment
CB8PH
JMSPH
MIDOH
NANMH
PAXTF
PMKTF
RPPMH
SASOH
SEVMH
SOUMH
TANMH
YRKPH
CB7PH
CB8PH
CHOOH
CHSMH
FSBMH
JMSPH
LCHMH
MIDOH
MPNOH
NANMH
PMKOH
RPPTF
SASOH
SEVMH
WICMH
WSTMH
BIGMH
CB3MH
CB6PH
CB8PH
CHOMH2
CHSOH
FSBMH
JMSTF
MANMH
MIDOH
MPNTF
NANMH
RHDMH
RPPTF
SOUMH
BIGMH
CB3MH
CB3MH
CB4MH
CB6PH
CB6PH
Year
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1996
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1998
1998
1998
1998
1998
1998
Designated Use
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
DW
OW
OW
DC
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
DW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
DW
OW
DW
  appendix f  •  Data Used in Deriving Summer Biological Reference Curves

-------
                                                                 F-3
CBP Segment
CB8PH
CHOMH2
CHOOH
CHSMH
CHSMH
GUNOH
JMSPH
MPNOH
MPNTF
PAXTF
POCOH
POTTF
RPPTF
WICMH
CB3MH
CB4MH
CB6PH
CB7PH
CB7PH
CB8PH
CHSMH
CHSMH
JMSPH
JMSTF
LYNPH
POCMH
RHDMH
WICMH
WSTMH
BSHOH
CB2OH
CB7PH
CB8PH
CHKOH
CHSOH
EASMH
ELKOH
HNGMH
JMSPH
JMSTF
LAFMH
MIDOH
MPNTF
NANOH
PMKOH
PMKTF
POTOH
RPPTF
SEVMH
Year
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
Designated Use
OW
ow
OW
ow
DW
OW
OW
ow
OW
OW
OW
OW
OW
OW
DW
DW
OW
OW
DW
OW
OW
DC
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
CBP Segment
YRKPH
CB2OH
CB3MH
CB6PH
CHSMH
ELKOH
FSBMH
HNGMH
MANMH
MOBPH
PMKTF
RPPTF
SASOH
WICMH
CB2OH
CB5MH
CB7PH
CHKOH
CHOMH1
CRRMH
NANOH
PAXTF
PMKTF
RPPOH
RPPTF
YRKPH
BIGMH
CB2OH
CB6PH
CB8PH
CHSOH
JMSPH
MIDOH
MPNOH
POCMH
APPTF
BOHOH
CB1TF
CB2OH
CB6PH
CB8PH
CHKOH
CHOMH1
CHOTF
CHSMH
CHSOH
CRRMH
GUNOH
MANMH
Year
2000
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2002
2002
2002
2002
2002
2002
2002
2002
2002
2002
2002
2002
2003
2003
2003
2003
2003
2003
2003
2003
2003
2004
2004
2004
2004
2004
2004
2004
2004
2004
2004
2004
2004
2004
2004
Designated Use
OW
OW
DC
DW
DW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
DW
DW
OW
OW
OW
OW
OW
OW
OW
OW
DW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
OW
appendix f

-------
F-4
                   CBP Segment   Year   Designated Use
                   MPNTF        2004        OW
                   NORTF        2004        OW
                   RPPOH        2004        OW
                   CB1TF         2005        OW
                   CB7PH         2005        DW
                   CHOMH2       2005        OW
                   FSBMH        2005        OW
                   PMKOH        2005        OW
                   SASOH        2005        OW
                   TANMH        2005        OW
  appendix f  «  Data Used in Deriving Summer Biological Reference Curves

-------
                                                                         G-1
                    appendix
  Equations for  the  Open-Water,

  Deep-Water and Deep-Channel

       Dissolved  Oxygen  Criteria

   Summer  Biological Reference

                        Curves

A biological reference curve of acceptable violation rates is generated using a cumu-
lative frequency distribution (CFD) of violation rates for "healthy" designated uses.
The violation rates are sorted in ascending order, ranked in descending order, and
graphed on a quantile plot:
  • Violation rates are plotted on the x-axis, with plotting position on the y-axis.
  • Plotting position represents the probability, i/n, of being less than or equal to a
    given violation rate, or x, and is plotted on the y-axis as a function of rank, or
    "i", and sample size, or "n".
  • The x-axis is labeled "Percentage of Volume" because the violation rate repre-
    sents the fraction of volume that is in violation.
  • The y-axis is labeled as "Percentage of Time" because "probability" represents
    the probable amount of time that a given violation rate will be observed.
  • The Chesapeake Bay Program currently uses the Wiebull plotting position to
    plot the cumulative distribution function. The Wiebull equation for calculating
    probability, y, for each violation rate with rank "i" is: y = i/(n+l); i = rank.

In order to generate a graph of the CFD:
  • X1; x2, x3,...xn= violation rates provided herein, sorted in ascending order,
    with rank (i) assigned in descending order.
    After plotting the data's violation rates and probabilities, two additional points
    should be added to the distribution in order to complete the CFD curve:
    Insert (x0, y0) = (0,1) before the first data point; and
    Insert (xn+1, yn+1) = (1,0) after the last data point.
                         appendix g  •  Equations for the Summer Biological Reference Curves

-------
G-2

rank

39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1

Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.1229698
0.1377778
0.1869919
0.192
0.1938775
0.2833333
0.3069767
0.3857374
0.5
0.6338462
0.7984496
1
1
1
Fraction
Time
1
0.975
0.95
0.925
0.9
0.875
0.85
0.825
0.8
0.775
0.75
0.725
0.7
0.675
0.65
0.625
0.6
0.575
0.55
0.525
0.5
0.475
0.45
0.425
0.4
0.375
0.35
0.325
0.3
0.275
0.25
0.225
0.2
0.175
0.15
0.125
0.1
0.075
0.05
0.025
0
  appendix g »  Equations for the Summer Biological Reference Curves

-------

                                                       G-3
rank

155
154
153
152
151
150
149
148
147
146
145
144
143
142
141
140
139
138
137
136
135
134
133
132
131
130
129
128
127
126
125
124
123
122
121
120
119
118
117
116
115
114
113
112
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
1
0.993589744
0.987179487
0.980769231
0.974358974
0.967948718
0.961538462
0.955128205
0.948717949
0.942307692
0.935897436
0.929487179
0.923076923
0.916666667
0.91025641
0.903846154
0.897435897
0.891025641
0.884615385
0.878205128
0.871794872
0.865384615
0.858974359
0.852564103
0.846153846
0.83974359
0.833333333
0.826923077
0.820512821
0.814102564
0.807692308
0.801282051
0.794871795
0.788461538
0.782051282
0.775641026
0.769230769
0.762820513
0.756410256
0.75
0.743589744
0.737179487
0.730769231
0.724358974
0.717948718
rank
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
Fraction
Volume
0
0
0
0
0
0.0011772
0.0027367
0.0053908
0.0058608
0.0071155
0.0082474
0.0086758
0.0105042
0.0119522
0.014231
0.0143416
0.015544
0.0186097
0.0186104
0.0186916
0.0229885
0.0242872
0.0290657
0.0303867
0.0341702
0.0372195
0.0394495
0.0442319
0.0468541
0.0492611
0.053407
0.0596184
0.0646766
0.0669035
0.0749625
0.0772947
0.0773381
0.0819209
0.0830704
0.0842912
0.0843786
0.0914286
0.0922064
0.096124
0.0967341
Fraction Time
0.711538462
0.705128205
0.698717949
0.692307692
0.685897436
0.679487179
0.673076923
0.666666667
0.66025641
0.653846154
0.647435897
0.641025641
0.634615385
0.628205128
0.621794872
0.615384615
0.608974359
0.602564103
0.596153846
0.58974359
0.583333333
0.576923077
0.570512821
0.564102564
0.557692308
0.551282051
0.544871795
0.538461538
0.532051282
0.525641026
0.519230769
0.512820513
0.506410256
0.5
0.493589744
0.487179487
0.480769231
0.474358974
0.467948718
0.461538462
0.455128205
0.448717949
0.442307692
0.435897436
0.429487179
appendix g

-------
G-4

rank
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
Fraction
Volume
0.0986842
0.1003289
0.1030177
0.1073883
0.1123967
0.1133005
0.1142857
0.1153846
0.1340996
0.1351351
0.1405229
0.1536643
0.1561065
0.1613475
0.1666667
0.1690574
0.177641
0.1888889
0.193999
0.2019704
0.2030651
0.2064298
0.2138837
0.2144487
0.2149758
0.2301587
0.2398477
0.2399356
0.2473721
0.2550629
0.2568941
0.2744511
0.2754491
Fraction Time
0.423076923
0.416666667
0.41025641
0.403846154
0.397435897
0.391025641
0.384615385
0.378205128
0.371794872
0.365384615
0.358974359
0.352564103
0.346153846
0.33974359
0.333333333
0.326923077
0.320512821
0.314102564
0.307692308
0.301282051
0.294871795
0.288461538
0.282051282
0.275641026
0.269230769
0.262820513
0.256410256
0.25
0.243589744
0.237179487
0.230769231
0.224358974
0.217948718
rank
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1

Fraction
Volume
0.2863962
0.2887439
0.2992831
0.304324
0.3064989
0.3065134
0.3125
0.313253
0.3192771
0.3256059
0.3313559
0.3367199
0.3522608
0.3867069
0.4039409
0.4058394
0.4066776
0.4071428
0.4091904
0.4172932
0.4230019
0.4251208
0.4340449
0.4419155
0.4548346
0.4548849
0.4679803
0.5176327
0.5266618
0.5465729
0.5878661
1
1
1
Fraction Time
0.211538462
0.205128205
0.198717949
0.192307692
0.185897436
0.179487179
0.173076923
0.166666667
0.16025641
0.153846154
0.147435897
0.141025641
0.134615385
0.128205128
0.121794872
0.115384615
0.108974359
0.102564103
0.096153846
0.08974359
0.083333333
0.076923077
0.070512821
0.064102564
0.057692308
0.051282051
0.044871795
0.038461538
0.032051282
0.025641026
0.019230769
0.012820513
0.006410256
0
  appendix g

-------

                                                       G-5
rank

868
867
866
865
864
863
862
861
860
859
858
857
856
855
854
853
852
851
850
849
848
847
846
845
844
843
842
841
840
839
838
837
836
835
834
833
832
831
830
829
828
827
826
825
824
823
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
1
0.998849252
0.997698504
0.996547756
0.995397008
0.99424626
0.993095512
0.991944764
0.990794016
0.989643268
0.98849252
0.987341772
0.986191024
0.985040276
0.983889528
0.98273878
0.981588032
0.980437284
0.979286536
0.978135788
0.97698504
0.975834292
0.974683544
0.973532796
0.972382048
0.9712313
0.970080552
0.968929804
0.967779056
0.966628308
0.96547756
0.964326812
0.963176064
0.962025316
0.960874568
0.95972382
0.958573072
0.957422325
0.956271577
0.955120829
0.953970081
0.952819333
0.951668585
0.950517837
0.949367089
0.948216341
0.947065593
rank
822
821
820
819
818
817
816
815
814
813
812
811
810
809
808
807
806
805
804
803
802
801
800
799
798
797
796
795
794
793
792
791
790
789
788
787
786
785
784
783
782
781
780
779
778
111
776
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.945914845
0.944764097
0.943613349
0.942462601
0.941311853
0.940161105
0.939010357
0.937859609
0.936708861
0.935558113
0.934407365
0.933256617
0.932105869
0.930955121
0.929804373
0.928653625
0.927502877
0.926352129
0.925201381
0.924050633
0.922899885
0.921749137
0.920598389
0.919447641
0.918296893
0.917146145
0.915995397
0.914844649
0.913693901
0.912543153
0.911392405
0.910241657
0.909090909
0.907940161
0.906789413
0.905638665
0.904487917
0.903337169
0.902186421
0.901035673
0.899884925
0.898734177
0.897583429
0.896432681
0.895281933
0.894131185
0.892980437
appendix g

-------
G-6

rank
775
774
773
772
771
770
769
768
767
766
765
764
763
762
761
760
759
758
757
756
755
754
753
752
751
750
749
748
747
746
745
744
743
742
741
740
739
738
737
736
735
734
733
732
731
730
729
728
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.891829689
0.890678941
0.889528193
0.888377445
0.887226697
0.886075949
0.884925201
0.883774453
0.882623705
0.881472957
0.880322209
0.879171461
0.878020713
0.876869965
0.875719217
0.87456847
0.873417722
0.872266974
0.871116226
0.869965478
0.86881473
0.867663982
0.866513234
0.865362486
0.864211738
0.86306099
0.861910242
0.860759494
0.859608746
0.858457998
0.85730725
0.856156502
0.855005754
0.853855006
0.852704258
0.85155351
0.850402762
0.849252014
0.848101266
0.846950518
0.84579977
0.844649022
0.843498274
0.842347526
0.841196778
0.84004603
0.838895282
0.837744534
rank
727
726
725
724
723
722
721
720
719
718
717
716
715
714
713
712
711
710
709
708
707
706
705
704
703
702
701
700
699
698
697
696
695
694
693
692
691
690
689
688
687
686
685
684
683
682
681
680
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.836593786
0.835443038
0.83429229
0.833141542
0.831990794
0.830840046
0.829689298
0.82853855
0.827387802
0.826237054
0.825086306
0.823935558
0.82278481
0.821634062
0.820483314
0.819332566
0.818181818
0.81703107
0.815880322
0.814729574
0.813578826
0.812428078
0.81127733
0.810126582
0.808975834
0.807825086
0.806674338
0.80552359
0.804372842
0.803222094
0.802071346
0.800920598
0.79976985
0.798619102
0.797468354
0.796317606
0.795166858
0.79401611
0.792865362
0.791714614
0.790563867
0.789413119
0.788262371
0.787111623
0.785960875
0.784810127
0.783659379
0.782508631
  appendix g •  Equations for the Summer Biological Reference Curves

-------
                                                            G-7

rank
679
678
677
676
675
674
673
672
671
670
669
668
667
666
665
664
663
662
661
660
659
658
657
656
655
654
653
652
651
650
649
648
647
646
645
644
643
642
641
640
639
638
637
636
635
634
633
632
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.781357883
0.780207135
0.779056387
0.777905639
0.776754891
0.775604143
0.774453395
0.773302647
0.772151899
0.771001151
0.769850403
0.768699655
0.767548907
0.766398159
0.765247411
0.764096663
0.762945915
0.761795167
0.760644419
0.759493671
0.758342923
0.757192175
0.756041427
0.754890679
0.753739931
0.752589183
0.751438435
0.750287687
0.749136939
0.747986191
0.746835443
0.745684695
0.744533947
0.743383199
0.742232451
0.741081703
0.739930955
0.738780207
0.737629459
0.736478711
0.735327963
0.734177215
0.733026467
0.731875719
0.730724971
0.729574223
0.728423475
0.727272727
rank
631
630
629
628
627
626
625
624
623
622
621
620
619
618
617
616
615
614
613
612
611
610
609
608
607
606
605
604
603
602
601
600
599
598
597
596
595
594
593
592
591
590
589
588
587
586
585
584
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.726121979
0.724971231
0.723820483
0.722669735
0.721518987
0.720368239
0.719217491
0.718066743
0.716915995
0.715765247
0.714614499
0.713463751
0.712313003
0.711162255
0.710011507
0.708860759
0.707710012
0.706559264
0.705408516
0.704257768
0.70310702
0.701956272
0.700805524
0.699654776
0.698504028
0.69735328
0.696202532
0.695051784
0.693901036
0.692750288
0.69159954
0.690448792
0.689298044
0.688147296
0.686996548
0.6858458
0.684695052
0.683544304
0.682393556
0.681242808
0.68009206
0.678941312
0.677790564
0.676639816
0.675489068
0.67433832
0.673187572
0.672036824
appendix g *  Equations for the Summer Biological Reference Curves

-------
G-8

rank
583
582
581
580
579
578
577
576
575
574
573
572
571
570
569
568
567
566
565
564
563
562
561
560
559
558
557
556
555
554
553
552
551
550
549
548
547
546
545
544
543
542
541
540
539
538
537
536
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.670886076
0.669735328
0.66858458
0.667433832
0.666283084
0.665132336
0.663981588
0.66283084
0.661680092
0.660529344
0.659378596
0.658227848
0.6570771
0.655926352
0.654775604
0.653624856
0.652474108
0.65132336
0.650172612
0.649021864
0.647871116
0.646720368
0.64556962
0.644418872
0.643268124
0.642117376
0.640966628
0.63981588
0.638665132
0.637514384
0.636363636
0.635212888
0.63406214
0.632911392
0.631760644
0.630609896
0.629459148
0.6283084
0.627157652
0.626006904
0.624856157
0.623705409
0.622554661
0.621403913
0.620253165
0.619102417
0.617951669
0.616800921
rank
535
534
533
532
531
530
529
528
527
526
525
524
523
522
521
520
519
518
517
516
515
514
513
512
511
510
509
508
507
506
505
504
503
502
501
500
499
498
497
496
495
494
493
492
491
490
489
488
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.615650173
0.614499425
0.613348677
0.612197929
0.611047181
0.609896433
0.608745685
0.607594937
0.606444189
0.605293441
0.604142693
0.602991945
0.601841197
0.600690449
0.599539701
0.598388953
0.597238205
0.596087457
0.594936709
0.593785961
0.592635213
0.591484465
0.590333717
0.589182969
0.588032221
0.586881473
0.585730725
0.584579977
0.583429229
0.582278481
0.581127733
0.579976985
0.578826237
0.577675489
0.576524741
0.575373993
0.574223245
0.573072497
0.571921749
0.570771001
0.569620253
0.568469505
0.567318757
0.566168009
0.565017261
0.563866513
0.562715765
0.561565017
  appendix g

-------
                                                            G-9

rank
487
486
485
484
483
482
481
480
479
478
477
476
475
474
473
472
471
470
469
468
467
466
465
464
463
462
461
460
459
458
457
456
455
454
453
452
451
450
449
448
447
446
445
444
443
442
441
440
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.560414269
0.559263521
0.558112773
0.556962025
0.555811277
0.554660529
0.553509781
0.552359033
0.551208285
0.550057537
0.548906789
0.547756041
0.546605293
0.545454545
0.544303797
0.543153049
0.542002301
0.540851554
0.539700806
0.538550058
0.53739931
0.536248562
0.535097814
0.533947066
0.532796318
0.53164557
0.530494822
0.529344074
0.528193326
0.527042578
0.52589183
0.524741082
0.523590334
0.522439586
0.521288838
0.52013809
0.518987342
0.517836594
0.516685846
0.515535098
0.51438435
0.513233602
0.512082854
0.510932106
0.509781358
0.50863061
0.507479862
0.506329114
rank
439
438
437
436
435
434
433
432
431
430
429
428
427
426
425
424
423
422
421
420
419
418
417
416
415
414
413
412
411
410
409
408
407
406
405
404
403
402
401
400
399
398
397
396
395
394
393
392
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fraction Time
0.505178366
0.504027618
0.50287687
0.501726122
0.500575374
0.499424626
0.498273878
0.49712313
0.495972382
0.494821634
0.493670886
0.492520138
0.49136939
0.490218642
0.489067894
0.487917146
0.486766398
0.48561565
0.484464902
0.483314154
0.482163406
0.481012658
0.47986191
0.478711162
0.477560414
0.476409666
0.475258918
0.47410817
0.472957422
0.471806674
0.470655926
0.469505178
0.46835443
0.467203682
0.466052934
0.464902186
0.463751438
0.46260069
0.461449942
0.460299194
0.459148446
0.457997699
0.456846951
0.455696203
0.454545455
0.453394707
0.452243959
0.451093211
appendix g *  Equations for the Summer Biological Reference Curves

-------
G-10

rank
391
390
389
388
387
386
385
384
383
382
381
380
379
378
377
376
375
374
373
372
371
370
369
368
367
366
365
364
363
362
361
360
359
358
357
356
355
354
353
352
351
350
349
348
347
346
345
344
Fraction
Volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.0005089
0.0005304
0.0005965
0.0006126
0.000801
0.0009697
0.0011871
0.0014519
0.0016221
0.0016335
0.0021645
0.0021645
0.0021684
0.0022422
0.0024594
0.0024882
0.0025445
0.0029858
0.0035562
0.0042328
0.0044582
0.0048011
0.0049478
0.0050706
0.0055944
0.005598
Fraction Time
0.449942463
0.448791715
0.447640967
0.446490219
0.445339471
0.444188723
0.443037975
0.441887227
0.440736479
0.439585731
0.438434983
0.437284235
0.436133487
0.434982739
0.433831991
0.432681243
0.431530495
0.430379747
0.429228999
0.428078251
0.426927503
0.425776755
0.424626007
0.423475259
0.422324511
0.421173763
0.420023015
0.418872267
0.417721519
0.416570771
0.415420023
0.414269275
0.413118527
0.411967779
0.410817031
0.409666283
0.408515535
0.407364787
0.406214039
0.405063291
0.403912543
0.402761795
0.401611047
0.400460299
0.399309551
0.398158803
0.397008055
0.395857307
rank
343
342
341
340
339
338
337
336
335
334
333
332
331
330
329
328
327
326
325
324
323
322
321
320
319
318
317
316
315
314
313
312
311
310
309
308
307
306
305
304
303
302
301
300
299
298
297
296
Fraction
Volume
0.005618
0.0056497
0.0056497
0.0056497
0.0057741
0.0060024
0.0064935
0.0064935
0.0064935
0.0064935
0.006993
0.0072816
0.0077864
0.0081425
0.0083485
0.009176
0.009772
0.0100344
0.0108696
0.011236
0.0112994
0.0113636
0.0122877
0.0129206
0.0133457
0.0135135
0.0135135
0.0138099
0.0141643
0.0145985
0.0152119
0.0153473
0.0163934
0.0169367
0.0173661
0.0178716
0.0181209
0.0190476
0.0195954
0.0201126
0.020657
0.021121
0.0213471
0.0220096
0.0222782
0.0227457
0.0228091
0.0235199
Fraction Time
0.394706559
0.393555811
0.392405063
0.391254315
0.390103567
0.388952819
0.387802071
0.386651323
0.385500575
0.384349827
0.383199079
0.382048331
0.380897583
0.379746835
0.378596087
0.377445339
0.376294591
0.375143843
0.373993096
0.372842348
0.3716916
0.370540852
0.369390104
0.368239356
0.367088608
0.36593786
0.364787112
0.363636364
0.362485616
0.361334868
0.36018412
0.359033372
0.357882624
0.356731876
0.355581128
0.35443038
0.353279632
0.352128884
0.350978136
0.349827388
0.34867664
0.347525892
0.346375144
0.345224396
0.344073648
0.3429229
0.341772152
0.340621404
   appendix g  »  Equations for the Summer Biological Reference Curves

-------
                                                      G-11

rank
295
294
293
292
291
290
289
288
287
286
285
284
283
282
281
280
279
278
277
276
275
274
273
272
271
270
269
268
267
266
265
264
263
262
261
260
259
258
257
256
255
254
253
252
251
250
249
248
Fraction
Volume
0.0238095
0.0238095
0.0238095
0.0238095
0.0238095
0.0238095
0.0242139
0.0243704
0.0248963
0.0251828
0.025533
0.0257611
0.025804
0.02595
0.0277429
0.0282486
0.0285016
0.0300546
0.0310473
0.0312355
0.0321381
0.0321381
0.0328082
0.0341186
0.0349206
0.0363636
0.0364963
0.0366795
0.0374777
0.0380952
0.0381803
0.0391608
0.0399501
0.0408163
0.0422037
0.042328
0.0423729
0.0427173
0.043956
0.0444942
0.0452794
0.0464169
0.0470397
0.0486772
0.0486772
0.0492197
0.0511788
0.0519084
Fraction Time
0.339470656
0.338319908
0.33716916
0.336018412
0.334867664
0.333716916
0.332566168
0.33141542
0.330264672
0.329113924
0.327963176
0.326812428
0.32566168
0.324510932
0.323360184
0.322209436
0.321058688
0.31990794
0.318757192
0.317606444
0.316455696
0.315304948
0.3141542
0.313003452
0.311852704
0.310701956
0.309551208
0.30840046
0.307249712
0.306098964
0.304948216
0.303797468
0.30264672
0.301495972
0.300345224
0.299194476
0.298043728
0.29689298
0.295742232
0.294591484
0.293440736
0.292289988
0.291139241
0.289988493
0.288837745
0.287686997
0.286536249
0.285385501
rank
247
246
245
244
243
242
241
240
239
238
237
236
235
234
233
232
231
230
229
228
227
226
225
224
223
222
221
220
219
218
217
216
215
214
213
212
211
210
209
208
207
206
205
204
203
202
201
200
Fraction
Volume
0.0521231
0.0529101
0.0529101
0.0530612
0.0536869
0.0536869
0.0536869
0.0541126
0.0541126
0.0541126
0.0541126
0.0541126
0.0579235
0.0586084
0.0587911
0.0596465
0.0604396
0.0614422
0.0618847
0.0620364
0.0625508
0.0653824
0.0658002
0.0684039
0.0695876
0.0708995
0.0713287
0.0714866
0.0717703
0.0739236
0.0756155
0.0757415
0.0760522
0.0779468
0.0781759
0.0784933
0.0793651
0.0813397
0.0836718
0.0852341
0.0852341
0.0852341
0.0858202
0.0859692
0.0862069
0.0867133
0.0889352
0.0924318
Fraction Time
0.284234753
0.283084005
0.281933257
0.280782509
0.279631761
0.278481013
0.277330265
0.276179517
0.275028769
0.273878021
0.272727273
0.271576525
0.270425777
0.269275029
0.268124281
0.266973533
0.265822785
0.264672037
0.263521289
0.262370541
0.261219793
0.260069045
0.258918297
0.257767549
0.256616801
0.255466053
0.254315305
0.253164557
0.252013809
0.250863061
0.249712313
0.248561565
0.247410817
0.246260069
0.245109321
0.243958573
0.242807825
0.241657077
0.240506329
0.239355581
0.238204833
0.237054085
0.235903337
0.234752589
0.233601841
0.232451093
0.231300345
0.230149597
appendix g

-------
G-12

Fraction
Volume
0.0926076
0.094402
0.0951807
0.0953661
0.0980392
0.0986222
0.0995439
0.1013514
0.1056534
0.1097062
0.1108631
0.1108631
0.1110075
0.1119293
0.1123596
0.1135513
0.113798
0.1141304
0.1160355
0.1222826
0.1235521
0.1259259
0.1260344
0.1269841
0.1270358
0.1300254
0.1310766
0.1316527
0.1342461
0.1372868
0.1389115
0.14
0.14
0.140647
0.1415645
0.1419069
0.1435523
0.1449735
0.1455978
0.1514339
0.1538896
0.1542142
0.1566434
0.1587452
0.1595922
0.1611479
0.1614429
0.162963
Fraction Time
0.228998849
0.227848101
0.226697353
0.225546605
0.224395857
0.223245109
0.222094361
0.220943613
0.219792865
0.218642117
0.217491369
0.216340621
0.215189873
0.214039125
0.212888377
0.211737629
0.210586881
0.209436133
0.208285386
0.207134638
0.20598389
0.204833142
0.203682394
0.202531646
0.201380898
0.20023015
0.199079402
0.197928654
0.196777906
0.195627158
0.19447641
0.193325662
0.192174914
0.191024166
0.189873418
0.18872267
0.187571922
0.186421174
0.185270426
0.184119678
0.18296893
0.181818182
0.180667434
0.179516686
0.178365938
0.17721519
0.176064442
0.174913694
rank
151
150
149
148
147
146
145
144
143
142
141
140
139
138
137
136
135
134
133
132
131
130
129
128
127
126
125
124
123
122
121
120
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
Fraction
Volume
0.1631854
0.1643766
0.1650579
0.1661578
0.1686848
0.1688742
0.1706892
0.1722272
0.1731602
0.1735369
0.1756757
0.1779041
0.1805116
0.1805379
0.1830601
0.190725
0.1914525
0.1941337
0.199403
0.201087
0.2013652
0.2039852
0.2189781
0.227972
0.2337085
0.2359882
0.2374406
0.2409669
0.2419833
0.2432432
0.2444856
0.2445605
0.2457132
0.2472826
0.2478753
0.2583187
0.2593284
0.2593284
0.2593284
0.2611517
0.2736842
0.2866706
0.2889755
0.2900763
0.2905174
0.3018642
0.3043062
0.3130724
Fraction Time
0.173762946
0.172612198
0.17146145
0.170310702
0.169159954
0.168009206
0.166858458
0.16570771
0.164556962
0.163406214
0.162255466
0.161104718
0.15995397
0.158803222
0.157652474
0.156501726
0.155350978
0.15420023
0.153049482
0.151898734
0.150747986
0.149597238
0.14844649
0.147295742
0.146144994
0.144994246
0.143843498
0.14269275
0.141542002
0.140391254
0.139240506
0.138089758
0.13693901
0.135788262
0.134637514
0.133486766
0.132336018
0.13118527
0.130034522
0.128883774
0.127733026
0.126582278
0.12543153
0.124280783
0.123130035
0.121979287
0.120828539
0.119677791
  appendix g

-------
                                                           G-13

rank
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
Fraction
Volume
0.319855
0.3289125
0.3301158
0.3326345
0.3344764
0.350166
0.350417
0.3507276
0.3753799
0.3762215
0.37668
0.3773917
0.3793436
0.3807623
0.3879781
0.3998882
0.4010152
0.4038889
0.4038889
0.4038889
0.4038889
0.4038889
0.4146816
0.4163347
0.4233177
0.4269663
0.429676
0.4300699
0.4304933
0.4374046
0.4423898
0.4565826
0.456869
0.4570895
0.4570895
0.4570895
0.464702
0.4692463
0.4841679
0.5085324
0.5108851
0.5263544
0.545568
0.5625
0.5701425
0.575
0.576076
0.6036122
Fraction Time
0.118527043
0.117376295
0.116225547
0.115074799
0.113924051
0.112773303
0.111622555
0.110471807
0.109321059
0.108170311
0.107019563
0.105868815
0.104718067
0.103567319
0.102416571
0.101265823
0.100115075
0.098964327
0.097813579
0.096662831
0.095512083
0.094361335
0.093210587
0.092059839
0.090909091
0.089758343
0.088607595
0.087456847
0.086306099
0.085155351
0.084004603
0.082853855
0.081703107
0.080552359
0.079401611
0.078250863
0.077100115
0.075949367
0.074798619
0.073647871
0.072497123
0.071346375
0.070195627
0.069044879
0.067894131
0.066743383
0.065592635
0.064441887
rank
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
Fraction
Volume
0.6200294
0.6237785
0.6320347
0.6437941
0.6469252
0.6567556
0.663745
0.7578948
0.7717455
0.8227364
0.8384528
0.97
0.9949544
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Fraction Time
0.063291139
0.062140391
0.060989643
0.059838895
0.058688147
0.057537399
0.056386651
0.055235903
0.054085155
0.052934407
0.051783659
0.050632911
0.049482163
0.048331415
0.047180667
0.046029919
0.044879171
0.043728423
0.042577675
0.041426928
0.04027618
0.039125432
0.037974684
0.036823936
0.035673188
0.03452244
0.033371692
0.032220944
0.031070196
0.029919448
0.0287687
0.027617952
0.026467204
0.025316456
0.024165708
0.02301496
0.021864212
0.020713464
0.019562716
0.018411968
0.01726122
0.016110472
0.014959724
0.013808976
0.012658228
0.01150748
0.010356732
0.009205984
appendix g *  Equations for the Summer Biological Reference Curves

-------
G-14

rank
7
6
5
4
3
2
1

Fraction
Volume
1
1
1
1
1
1
1
1
Fraction Time
0.008055236
0.006904488
0.00575374
0.004602992
0.003452244
0.002301496
0.001150748
0
  appendix g

-------
                                                                                    H-1
                       appendix
hi
                Equations for  the
            Water  Clarity Criteria
     Biological  Reference Curves
A biological reference curve of acceptable violation rates is generated using a cumu-
lative frequency distribution (CFD) of violation rates for "healthy" designated uses.
The violation rates are sorted in ascending order, ranked in descending order, and
graphed on a quantile plot:
  • Violation rates are plotted on the x-axis, with plotting position on the y axis.
  • Plotting position represents the probability, i/n, of being less than or equal to a
    given violation rate, or x, and is plotted on the y-axis as a function of rank, or
    "i", and sample size, or "n".
  • The x-axis is labeled "space" because the violation rate represents the fraction
    of volume that is in violation.
  • The y-axis is labeled as "time" because "probability" represents the probable
    amount of time that a given violation rate will be observed.
  • The Chesapeake Bay Program currently uses the Wiebull plotting  position to
    plot the cumulative distribution function. The Wiebull equation for  calculating
    probability, y, for each violation rate with rank "i" is: y = i/(n+l); i = rank.

In order to generate a graph of the CFD:
  • Xj ,  x2, x3,...xn=  violation rates provided herein, sorted in ascending order,
    with rank (i) assigned  in descending order.
    After plotting the data's violation rates and probabilities, two additional points
    should be added to the distribution in order to complete the CFD curve:
       Insert (x0, y0) = (0,1) before the first data point; and
       Insert (xn+1, yn+1) = (1,0) after the last data point.
                   appendix h  •  Equations for the Water Clarity Criteria Biological Reference Curves

-------
H-2

rank

406
405
404
403
402
401
400
399
398
397
396
395
394
393
392
391
390
389
388
387
386
385
384
383
382
381
380
379
378
377
376
375
374
373
372
371
370
369
368
367
366
365
364
363
362
361
360
359
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
1
0.997542998
0.995085995
0.992628993
0.99017199
0.987714988
0.985257985
0.982800983
0.98034398
0.977886978
0.975429975
0.972972973
0.970515971
0.968058968
0.965601966
0.963144963
0.960687961
0.958230958
0.955773956
0.953316953
0.950859951
0.948402948
0.945945946
0.943488943
0.941031941
0.938574939
0.936117936
0.933660934
0.931203931
0.928746929
0.926289926
0.923832924
0.921375921
0.918918919
0.916461916
0.914004914
0.911547912
0.909090909
0.906633907
0.904176904
0.901719902
0.899262899
0.896805897
0.894348894
0.891891892
0.889434889
0.886977887
0.884520885
0.882063882
rank
358
357
356
355
354
353
352
351
350
349
348
347
346
345
344
343
342
341
340
339
338
337
336
335
334
333
332
331
330
329
328
327
326
325
324
323
322
321
320
319
318
317
316
315
314
313
312
311
310
309
308
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.87960688
0.877149877
0.874692875
0.872235872
0.86977887
0.867321867
0.864864865
0.862407862
0.85995086
0.857493857
0.855036855
0.852579853
0.85012285
0.847665848
0.845208845
0.842751843
0.84029484
0.837837838
0.835380835
0.832923833
0.83046683
0.828009828
0.825552826
0.823095823
0.820638821
0.818181818
0.815724816
0.813267813
0.810810811
0.808353808
0.805896806
0.803439803
0.800982801
0.798525799
0.796068796
0.793611794
0.791154791
0.788697789
0.786240786
0.783783784
0.781326781
0.778869779
0.776412776
0.773955774
0.771498771
0.769041769
0.766584767
0.764127764
0.761670762
0.759213759
0.756756757
  appendix h  »  Equations for the Water Clarity Criteria Biological Reference Curves

-------
                                                                           H-3

rank
307
306
305
304
303
302
301
300
299
298
297
296
295
294
293
292
291
290
289
288
287
286
285
284
283
282
281
280
279
278
277
276
275
274
273
272
271
270
269
268
267
266
265
264
263
262
261
260
259
258
257
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.754299754
0.751842752
0.749385749
0.746928747
0.744471744
0.742014742
0.73955774
0.737100737
0.734643735
0.732186732
0.72972973
0.727272727
0.724815725
0.722358722
0.71990172
0.717444717
0.714987715
0.712530713
0.71007371
0.707616708
0.705159705
0.702702703
0.7002457
0.697788698
0.695331695
0.692874693
0.69041769
0.687960688
0.685503686
0.683046683
0.680589681
0.678132678
0.675675676
0.673218673
0.670761671
0.668304668
0.665847666
0.663390663
0.660933661
0.658476658
0.656019656
0.653562654
0.651105651
0.648648649
0.646191646
0.643734644
0.641277641
0.638820639
0.636363636
0.633906634
0.631449631
rank
256
255
254
253
252
251
250
249
248
247
246
245
244
243
242
241
240
239
238
237
236
235
234
233
232
231
230
229
228
227
226
225
224
223
222
221
220
219
218
217
216
215
214
213
212
211
210
209
208
207
206
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.628992629
0.626535627
0.624078624
0.621621622
0.619164619
0.616707617
0.614250614
0.611793612
0.609336609
0.606879607
0.604422604
0.601965602
0.5995086
0.597051597
0.594594595
0.592137592
0.58968059
0.587223587
0.584766585
0.582309582
0.57985258
0.577395577
0.574938575
0.572481572
0.57002457
0.567567568
0.565110565
0.562653563
0.56019656
0.557739558
0.555282555
0.552825553
0.55036855
0.547911548
0.545454545
0.542997543
0.540540541
0.538083538
0.535626536
0.533169533
0.530712531
0.528255528
0.525798526
0.523341523
0.520884521
0.518427518
0.515970516
0.513513514
0.511056511
0.508599509
0.506142506
appendix h  *  Equations for the Water Clarity Criteria Biological Reference Curves

-------
H-4

rank
205
204
203
202
201
200
199
198
197
196
195
194
193
192
191
190
189
188
187
186
185
184
183
182
181
180
179
178
177
176
175
174
173
172
171
170
169
168
167
166
165
164
163
162
161
160
159
158
157
156
155
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.503685504
0.501228501
0.498771499
0.496314496
0.493857494
0.491400491
0.488943489
0.486486486
0.484029484
0.481572482
0.479115479
0.476658477
0.474201474
0.471744472
0.469287469
0.466830467
0.464373464
0.461916462
0.459459459
0.457002457
0.454545455
0.452088452
0.44963145
0.447174447
0.444717445
0.442260442
0.43980344
0.437346437
0.434889435
0.432432432
0.42997543
0.427518428
0.425061425
0.422604423
0.42014742
0.417690418
0.415233415
0.412776413
0.41031941
0.407862408
0.405405405
0.402948403
0.4004914
0.398034398
0.395577396
0.393120393
0.390663391
0.388206388
0.385749386
0.383292383
0.380835381
rank
154
153
152
151
150
149
148
147
146
145
144
143
142
141
140
139
138
137
136
135
134
133
132
131
130
129
128
127
126
125
124
123
122
121
120
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
volume
0.0054
0.0108
0.0108
0.0108
0.0108
0.0108
0.0196
0.0215
0.0215
0.0215
0.0215
0.0215
0.0261
0.0269
0.0269
0.0278
0.0278
0.0278
0.0323
0.0323
0.0455
0.0556
0.0719
0.0784
0.1111
0.1176
0.1237
0.1307
0.1307
0.1389
0.1389
0.1389
0.1389
0.1438
0.1505
0.1667
0.1667
0.1944
0.1989
0.2151
0.2222
0.25
0.2742
0.2778
0.3116
0.3203
0.3659
0.3889
0.3889
0.4167
0.4167
time
0.378378378
0.375921376
0.373464373
0.371007371
0.368550369
0.366093366
0.363636364
0.361179361
0.358722359
0.356265356
0.353808354
0.351351351
0.348894349
0.346437346
0.343980344
0.341523342
0.339066339
0.336609337
0.334152334
0.331695332
0.329238329
0.326781327
0.324324324
0.321867322
0.319410319
0.316953317
0.314496314
0.312039312
0.30958231
0.307125307
0.304668305
0.302211302
0.2997543
0.297297297
0.294840295
0.292383292
0.28992629
0.287469287
0.285012285
0.282555283
0.28009828
0.277641278
0.275184275
0.272727273
0.27027027
0.267813268
0.265356265
0.262899263
0.26044226
0.257985258
0.255528256
  appendix h  »  Equations for the Water Clarity Criteria Biological Reference Curves

-------
                                                                           H-5

rank
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
volume
0.4444
0.4444
0.5538
0.5581
0.5752
0.5948
0.6136
0.6136
0.6202
0.6237
0.6434
0.6434
0.6434
0.6434
0.6899
0.7209
0.7209
0.7364
0.7519
0.7597
0.7597
0.7674
0.7727
0.7796
0.7974
0.8062
0.8188
0.8225
0.8225
0.8333
0.837
0.8611
0.8864
0.8864
0.8889
0.8992
0.8992
0.8992
0.9058
0.938
0.9457
0.9477
0.9545
0.9545
0.9612
0.9674
0.9722
0.9739
0.9767
0.9767
0.9773
time
0.253071253
0.250614251
0.248157248
0.245700246
0.243243243
0.240786241
0.238329238
0.235872236
0.233415233
0.230958231
0.228501229
0.226044226
0.223587224
0.221130221
0.218673219
0.216216216
0.213759214
0.211302211
0.208845209
0.206388206
0.203931204
0.201474201
0.199017199
0.196560197
0.194103194
0.191646192
0.189189189
0.186732187
0.184275184
0.181818182
0.179361179
0.176904177
0.174447174
0.171990172
0.16953317
0.167076167
0.164619165
0.162162162
0.15970516
0.157248157
0.154791155
0.152334152
0.14987715
0.147420147
0.144963145
0.142506143
0.14004914
0.137592138
0.135135135
0.132678133
0.13022113
rank
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
volume
0.9804
0.9804
0.9804
0.9935
0.9935
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
time
0.127764128
0.125307125
0.122850123
0.12039312
0.117936118
0.115479115
0.113022113
0.110565111
0.108108108
0.105651106
0.103194103
0.100737101
0.098280098
0.095823096
0.093366093
0.090909091
0.088452088
0.085995086
0.083538084
0.081081081
0.078624079
0.076167076
0.073710074
0.071253071
0.068796069
0.066339066
0.063882064
0.061425061
0.058968059
0.056511057
0.054054054
0.051597052
0.049140049
0.046683047
0.044226044
0.041769042
0.039312039
0.036855037
0.034398034
0.031941032
0.029484029
0.027027027
0.024570025
0.022113022
0.01965602
0.017199017
0.014742015
0.012285012
0.00982801
0.007371007
0.004914005
appendix h  *  Equations for the Water Clarity Criteria Biological Reference Curves

-------
H-6

rank
1
volume
1
1
time
0.002457002
0

rank

390
389
388
387
386
385
384
383
382
381
380
379
378
377
376
375
374
373
372
371
370
369
368
367
366
365
364
363
362
361
360
359
358
357
356
355
354
353
352
351
350
349
348
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
1
0.997442455
0.99488491
0.992327366
0.989769821
0.987212276
0.984654731
0.982097187
0.979539642
0.976982097
0.974424552
0.971867008
0.969309463
0.966751918
0.964194373
0.961636829
0.959079284
0.956521739
0.953964194
0.95140665
0.948849105
0.94629156
0.943734015
0.941176471
0.938618926
0.936061381
0.933503836
0.930946292
0.928388747
0.925831202
0.923273657
0.920716113
0.918158568
0.915601023
0.913043478
0.910485934
0.907928389
0.905370844
0.902813299
0.900255754
0.89769821
0.895140665
0.89258312
0.890025575
rank
347
346
345
344
343
342
341
340
339
338
337
336
335
334
333
332
331
330
329
328
327
326
325
324
323
322
321
320
319
318
317
316
315
314
313
312
311
310
309
308
307
306
305
304
303
302
301
300
299
298
297
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.887468031
0.884910486
0.882352941
0.879795396
0.877237852
0.874680307
0.872122762
0.869565217
0.867007673
0.864450128
0.861892583
0.859335038
0.856777494
0.854219949
0.851662404
0.849104859
0.846547315
0.84398977
0.841432225
0.83887468
0.836317136
0.833759591
0.831202046
0.828644501
0.826086957
0.823529412
0.820971867
0.818414322
0.815856777
0.813299233
0.810741688
0.808184143
0.805626598
0.803069054
0.800511509
0.797953964
0.795396419
0.792838875
0.79028133
0.787723785
0.78516624
0.782608696
0.780051151
0.777493606
0.774936061
0.772378517
0.769820972
0.767263427
0.764705882
0.762148338
0.759590793
  appendix h

-------
                                                                           H-7

rank
296
295
294
293
292
291
290
289
288
287
286
285
284
283
282
281
280
279
278
277
276
275
274
273
272
271
270
269
268
267
266
265
264
263
262
261
260
259
258
257
256
255
254
253
252
251
250
249
248
247
246
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
time
0.757033248
0.754475703
0.751918159
0.749360614
0.746803069
0.744245524
0.74168798
0.739130435
0.73657289
0.734015345
0.731457801
0.728900256
0.72634271 1
0.723785166
0.721227621
0.718670077
0.716112532
0.713554987
0.710997442
0.708439898
0.705882353
0.703324808
0.700767263
0.698209719
0.695652174
0.693094629
0.690537084
0.68797954
0.685421995
0.68286445
0.680306905
0.677749361
0.675191816
0.672634271
0.670076726
0.667519182
0.664961637
0.662404092
0.659846547
0.657289003
0.654731458
0.652173913
0.649616368
0.647058824
0.644501279
0.641943734
0.639386189
0.636828645
0.6342711
0.631713555
0.62915601
rank
245
244
243
242
241
240
239
238
237
236
235
234
233
232
231
230
229
228
227
226
225
224
223
222
221
220
219
218
217
216
215
214
213
212
211
210
209
208
207
206
205
204
203
202
201
200
199
198
197
196
195
volume
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.0037
0.0037
0.0037
0.0054
0.0054
0.0073
0.0094
0.011
0.0147
0.0147
0.0154
0.0154
0.0162
0.0162
0.0162
0.0162
0.0162
0.0162
0.0162
0.0204
time
0.626598465
0.624040921
0.621483376
0.618925831
0.616368286
0.613810742
0.611253197
0.608695652
0.606138107
0.603580563
0.601023018
0.598465473
0.595907928
0.593350384
0.590792839
0.588235294
0.585677749
0.583120205
0.58056266
0.578005115
0.57544757
0.572890026
0.570332481
0.567774936
0.565217391
0.562659847
0.560102302
0.557544757
0.554987212
0.552429668
0.549872123
0.547314578
0.544757033
0.542199488
0.539641944
0.537084399
0.534526854
0.531969309
0.529411765
0.52685422
0.524296675
0.52173913
0.519181586
0.516624041
0.514066496
0.511508951
0.508951407
0.506393862
0.503836317
0.501278772
0.498721228
appendix h  *  Equations for the Water Clarity Criteria Biological Reference Curves

-------
H-8

rank
194
193
192
191
190
189
188
187
186
185
184
183
182
181
180
179
178
177
176
175
174
173
172
171
170
169
168
167
166
165
164
163
162
161
160
159
158
157
156
155
154
153
152
151
150
149
148
147
146
145
144
volume
0.0204
0.022
0.022
0.0235
0.0256
0.0256
0.027
0.027
0.027
0.0282
0.0378
0.0432
0.0476
0.0476
0.0513
0.0513
0.0513
0.0612
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0623
0.0659
0.0659
0.0696
0.0703
0.0728
0.0769
0.0806
0.0811
0.0939
0.102
0.102
0.1077
0.1077
0.1081
0.1127
0.1136
0.1209
0.1221
0.1224
0.1224
0.1224
0.1224
0.1224
0.1224
0.1224
0.1224
time
0.496163683
0.493606138
0.491048593
0.488491049
0.485933504
0.483375959
0.480818414
0.47826087
0.475703325
0.47314578
0.470588235
0.468030691
0.465473146
0.462915601
0.460358056
0.457800512
0.455242967
0.452685422
0.450127877
0.447570332
0.445012788
0.442455243
0.439897698
0.437340153
0.434782609
0.432225064
0.429667519
0.427109974
0.42455243
0.421994885
0.41943734
0.416879795
0.414322251
0.411764706
0.409207161
0.406649616
0.404092072
0.401534527
0.398976982
0.396419437
0.393861893
0.391304348
0.388746803
0.386189258
0.383631714
0.381074169
0.378516624
0.375959079
0.373401535
0.37084399
0.368286445
rank
143
142
141
140
139
138
137
136
135
134
133
132
131
130
129
128
127
126
125
124
123
122
121
120
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
volume
0.1224
0.1224
0.1224
0.1224
0.1231
0.1231
0.1231
0.1231
0.1282
0.1297
0.1319
0.1351
0.1385
0.1429
0.1502
0.1538
0.1538
0.1575
0.1612
0.1612
0.1633
0.1633
0.169
0.1795
0.1837
0.1868
0.1995
0.2
0.2041
0.2088
0.2108
0.2113
0.2254
0.2308
0.2308
0.2308
0.2308
0.2378
0.2394
0.2432
0.2449
0.2449
0.2486
0.2541
0.2564
0.2582
0.2653
0.2653
0.2811
0.2857
0.2857
time
0.3657289
0.363171355
0.360613811
0.358056266
0.355498721
0.352941176
0.350383632
0.347826087
0.345268542
0.342710997
0.340153453
0.337595908
0.335038363
0.332480818
0.329923274
0.327365729
0.324808184
0.322250639
0.319693095
0.31713555
0.314578005
0.31202046
0.309462916
0.306905371
0.304347826
0.301790281
0.299232737
0.296675192
0.294117647
0.291560102
0.289002558
0.286445013
0.283887468
0.281329923
0.278772379
0.276214834
0.273657289
0.271099744
0.268542199
0.265984655
0.26342711
0.260869565
0.25831202
0.255754476
0.253196931
0.250639386
0.248081841
0.245524297
0.242966752
0.240409207
0.237851662
  appendix h  »  Equations for the Water Clarity Criteria Biological Reference Curves

-------
                                                                           H-9

rank
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
volume
0.2857
0.2973
0.3081
0.3187
0.3216
0.3239
0.3243
0.3265
0.3405
0.3451
0.3469
0.348
0.359
0.359
0.359
0.3592
0.3622
0.3692
0.3838
0.3846
0.3892
0.3919
0.392
0.4
0.4
0.4
0.4054
0.4202
0.4286
0.4324
0.439
0.4432
0.4615
0.4703
0.4769
0.4812
0.4872
0.4872
0.4872
0.5077
0.5077
0.5092
0.5094
0.5102
0.5128
0.5128
0.5231
0.5385
0.5495
0.5657
0.5692
time
0.235294118
0.232736573
0.230179028
0.227621483
0.225063939
0.222506394
0.219948849
0.217391304
0.21483376
0.212276215
0.20971867
0.207161125
0.204603581
0.202046036
0.199488491
0.196930946
0.194373402
0.191815857
0.189258312
0.186700767
0.184143223
0.181585678
0.179028133
0.176470588
0.173913043
0.171355499
0.168797954
0.166240409
0.163682864
0.16112532
0.158567775
0.15601023
0.153452685
0.150895141
0.148337596
0.145780051
0.143222506
0.140664962
0.138107417
0.135549872
0.132992327
0.130434783
0.127877238
0.125319693
0.122762148
0.120204604
0.117647059
0.115089514
0.112531969
0.109974425
0.10741688
rank
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1

volume
0.5838
0.5897
0.6
0.6054
0.6154
0.6315
0.641
0.6462
0.6667
0.6714
0.6923
0.6923
0.6923
0.7077
0.7077
0.7297
0.7347
0.7538
0.7551
0.7568
0.7949
0.7959
0.8239
0.838
0.8498
0.8571
0.8571
0.8615
0.9121
0.9385
0.9388
1
1
1
1
1
1
1
1
1
1
1
time
0.104859335
0.10230179
0.099744246
0.097186701
0.094629156
0.092071611
0.089514066
0.086956522
0.084398977
0.081841432
0.079283887
0.076726343
0.074168798
0.071611253
0.069053708
0.066496164
0.063938619
0.061381074
0.058823529
0.056265985
0.05370844
0.051150895
0.04859335
0.046035806
0.043478261
0.040920716
0.038363171
0.035805627
0.033248082
0.030690537
0.028132992
0.025575448
0.023017903
0.020460358
0.017902813
0.015345269
0.012787724
0.010230179
0.007672634
0.00511509
0.002557545
0
appendix h  *  Equations for the Water Clarity Criteria Biological Reference Curves

-------
           appendix |

  Evaluation of Maryland and
   Virginia Chesapeake Bay
 Segment SAV Acreages from
 2003 to 2005 for Prioritizing
  Shallow-water Monitoring
          by Segment
                 MARYLAND
Chesapeake Bay
Program
Segments/
Subsegments
CHSOH
BSHOH
BOHOH
CB20H
PAXTF
SASOH
C&DOH
PAXOH
GUNOH
MATTF
ELKOH
PISTF
POTTF(MD)
NORTF
SEVMH
CB1TF
MIDOH
PATMH
2003 Acres
0
390
288
212
217
371
0
106
489
612
346
212
885
46
III
7,574
391
7
2004 Acres
4
1,025
730
1,303
220
1,272
8
106
2,392
601
1,913
507
1,256
84
388
10,110
671
183
2005 Acres
228
726
918
1,071
324
1,476
9
125
1,733
770
1,964
757
2,029
78
426
9,193
454
279
2003-2005 Single
Best Year Acres
228
1,025
918
1,303
324
1,476
9
125
2,392
770
1,964
757
2,029
84
426
10,110
671
279
State-adopted
SAV Restoration
Acreage
77
350
354
705
205
1,168
7
115
2,432
792
2,034
789
2,142
89
455
12,903
879
389
Single Best Year
as % of SAV
Restoration
Acreage Status
3 Pass
3 Pass
3 Pass
2 Pass
2 Pass












Pass
Pass
Pass
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
appendix i • Evaluation of SAV Acreages from 2003 to 2005 for Prioritizing Shallow-water Monitoring

-------
                                    MARYLAND (continued)
Chesapeake Bay
Program
Segments/
Subsegments
POTOH(MD)
CB3MH
HNGMH
MAGMH
CHOMH1
POTMH(MD)
BIGMH
LCHMH
EASMH
CHSMH
TANMH(MD)
CB5MH(MD)
WSTMH
SOUMH
MANMH
FSBMH
POCMH(MD)
PAXMH
CB4MH
CHOMH2
CHOOH
NANMH
NANOH
RHDMH
WICMH
BACOH
CHOTF
CHSTF
NANTF
POCOH
POCTF
WBRTF
2003 Acres
1,384
23
2,844
169
2,972
2,430
451
784
1,639
117
4,725
700
23
14
235
15
58
37
21
0
0
0
0
0
0
0
0
0
0
0
0
0
2004 Acres
1,408
909
3,433
300
3,774
3,063
550
1,221
1,040
731
4,554
398
0
46
291
17
69
42
10
1
0
0
0
0
0
30
0
0
0
0
0
0
2005 Acres
1,888
567
4,376
308
2,293
2,893
710
260
768
462
5,801
919
0
10
410
7
69
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
2003-2005 Single
Best Year Acres
1,888
909
4,376
308
3,774
3,063
710
1,221
1,639
731
5,801
919
23
46
410
17
69
42
21
0
0
0
0
0
0
30
0
1
0
0
0
0
State-adopted
SAV Restoration
Acreage
2,802
1,370
7,761
579
8,184
7,088
2,043
4,076
6,209
2,928
24,757
8,270
238
479
4,353
197
877
1,634
2,533
1,621
72
3
12
60
3
-
-
-
-
-
-
.
Single Best Year
as % of SAV
Restoration
Acreage
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
N/A
-
-
-
-
-
-
Status
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
No SAV
No SAV
No SAV
No SAV
No SAV
No SAV
NO SAV
appendix i  •  Evaluation of SAV Acreages from 2003 to 2005 for Prioritizing Shallow-water Monitoring

-------
                                                                                        "
                                      VIRGINIA
Chesapeake Bay
Program
Segments/
Subsegments
MPNTF
PMKTF
POTOH(VA)
CHKOH
RPPTF
POTTF(VA)
CB8PH
CB7PH
JMSOH
CB6PH
MOBPH
JMSPH
POCMH(VA)
CRRMH
TANMH(VA)
CB5MH(VA)
YRKPH
LYNPH
PIAMH
RPPMH
POTMH(VA)
JMSTF
JMSMH
APPTF
YRKMH
EBEMH
ELIMH
ELIPH
LAFMH
MPNOH
PMKOH
RPPOH
SBEMH
WBEMH
2003 Acres
184
217
1,950
425
0
761
5
9,192
9
707
8,457
132
1,608
43
4,682
*
887
0
447
21
55
75
2
0
0
0
0
0
0
0
0
0
0
0
2004 Acres
179
334
2,326
432
24
1,197
6
7,157
0
488
7,549
74
1,094
224
3,990
1,833
597
9
443
33
339
12
2
0
0
0
0
0
0
0
0
0
0
0
2005 Acres
296
585
2,627
697
81
2,336
9
8,139
0
642
7,205
0
1,716
292
5,036
2,464
438
19
561
198
444
53
0
0
0
0
0
0
0
0
0
4
0
0
2003-2005 Single
Best Year Acres
296
585
2,627
697
81
2,336
9
9,192
9
707
8,457
132
1,716
292
5,036
2,464
887
19
561
198
444
75
2
0
0
0
0
0
0
0
0
4
0
0
State-adopted
SAV Restoration
Acreage
85
187
1,503
535
66
2,093
11
15,107
15
1,267
15,901
300
4,066
768
13,579
7,633
2,793
107
3,479
1,700
4,250
1,200
200
379
239
-
-
-
-
-
-
-
-
-
Single Best Year
as % of SAV
Restoration
Acreage
3
3
2








0
0
0
0
0
0
0
0
0
0
0
0
0
0

-
-
-
-
-
-
-
-
Status
Pass
Pass
Pass
Pass
Pass
Pass
Fail
Fail
Fail
Fail
Fail
Full
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
No SAV
No SAV
No SAV
No SAV
No SAV
No SAV
L No SAV
No SAV
No SAV
appendix i  •  Evaluation of SAV Acreages from 2003 to 2005 for Prioritizing Shallow-water Monitoring

-------
1-4

  Chesapeake Bay
  Program
  Segments/
  Subsegments
2003 Acres
               2004 Acres
                             2005 Acres
                State-adopted
2003-2005 Single   SAV Restoration
 Best Year Acres      Acreage
Single Best Year
 as % of SAV
  Restoration
   Acreage
                                                                                          Status
  NANTF(DE)
  POTTF(DC)
  ANATF (DC)
                                                        OF
        PASS (> 100% of Goal)

        FAIL (50%-< 100% of Goal)

        FAIL (<50% of Goal)

        No SAV Goal


  ^Partial data available that year
  appendix

-------
                                                                             J-1
                       appendix I

      Chesapeake Bay Estuarine
           Benthic Communities
        Assessment Protocol for
          Maryland  and Virginia
  305b/303d  Integrated  Reports
Maryland (Department of the Environment, Department of Natural Resources),
Virginia (Department of Environmental Quality) and U.S. EPA (Region 3 Water
Protection Division and Chesapeake Bay Program Office) reached agreement on the
protocol to assess Chesapeake Bay benthic community health. This appendix docu-
mented the assessment protocol supporting the States evaluation of Chesapeake Bay
benthic community data as part of their 305b/303d Integrated Reports. This assess-
ment protocol  builds  directly on the more detailed  assessment methods
recommended by Llanso et al. 2005 (see Appendix K).

The overall decision protocol is shown in Figure J-1. Phase I consists of the evalu-
ation of the sample size (i.e., number of B-IBI scores) available from the waterbody
segment during  the five-year assessment window. If the sample size satisfies the
requirements of  the statistical method (N > 10), a formal assessment of status (i.e.
impaired vs. supports aquatic life use) is determined utilizing the "percent degraded
area" statistical methodology (Phase II). If the sample size requirement is not met
an impairment assessment based solely on these analyses is not possible. Results for
segments with insufficient sample size should still be examined for possible use in
conjunction with other assessment data of the 305b/303d reporting process.

Phase II consists of the impairment assessment of aquatic life use attainment based
on a comparison of Benthic Index of Biotic Integrity (B-IBI) scores and can only be
performed when the number of B-IBI scores within a specified waterbody segment
is sufficient to meet the sample size requirement of the approved statistical method
(N > 10). Phase II can result in one of two possible outcomes: (1) the segment is not
impaired for Aquatic Life use  due to  benthic community status (note that the
segment may still be impaired for aquatic life use due to failure of other aquatic life
use criteria), or (2) the segment fails to support aquatic life use due to benthic
community status and is assessed as impaired. Best professional judgment can be
      appendix J  • Bay Estuarine Benthic Communities Assessment Protocol for Maryland and Virginia

-------
J-2


Phase!
Sample She
Evaluation

N<10
i\o
N>li

Yes —
Yes -.

Phase 11
Impairment Assessment

Insufficient sample size

Apply Degraded Area
Statistical method
1
Segment declared
'not impaired* for bcnthic
aquatic life communities
in 305b/303d Integrated
Report
1 No
Segment declared
'impaired' for benthic
aquatic life communities in
305b/303d Integrated
Report

Yes —
Yes -.

Ph»« in
Segment Characterization
(Identify Probable Causes)

Optional use of
B-EBE scores and diagnostic analyses
in conjunction with other available
data for305b/303d Integrated Report

Optional use of
B-IBI scores and diagnostic analyses
in conjunction with other available
data Ibr305b/303d Integrated Report

Apply diagnostic analyses for
assignment of suspected cause(s) of
degradation in 305b/303d Integrated
Report


                           J-1, Overall Chesapeake Bay benthic index of biotic integrity assessment decision
                     protocol.
                     applied to override (reverse) the outcome of the formal statistical analysis results, but
                     such reversals must be justified and documented.

                     Phase III consists of the identification of probable causes of benthic impairment of the
                     waterbody segment based upon benthic stressor diagnostic analyses. It is a two-step
                     procedure that involves (1) Site Classification, and (2) Segment Characterization.
                        1. Site classification: The first step is to assign probable cause of benthic degra-
                          dation to each individual "degraded" benthic sample. For purposed of these
                          diagnostic analyses, a sample is considered degraded if the B-IBI score is less
                          than 2.7.
                          Site Classification—Step la: The  application of a  formal statistical  linear
                          discriminant function calculates the 'inclusion probability' of each degraded
                          site belonging to a 'contaminant caused' group or an 'other causes' group,
                          based upon its B-IBI score and associated metrics. If a site is assigned to the
                          'Contaminant' Group with a probability > 0.9, this site is considered impacted
                          by contaminated sediment and no further classification is required.
                          Site Classification—Step Ib: If a site is  classified as degraded due to  'other
                          causes' (i.e., not contaminant-related), an evaluation of the relative abundance
                          (and/or biomass) of the benthos is examined.  Scores  for both abundance and
                          biomass are considered to be bipolar for the Chesapeake Bay Benthic IBI. For
                          either metric; a high score of 5, indicating desirable conditions, falls in the mid-
                          range  of the abundance/biomass  distributions,  while a  low  score  of 1,
                          indicating undesirable conditions, can result either  from insufficient  abun-
  ap pen a ix j

-------
                                                                                                J-3
     dance/biomass or excessive abundance/biomass. The scoring thresholds for
     these two metrics vary with habitat type (salinity regime and substrate type) as
     summarized in Figure J-2. In this process, a site is classified as degraded by
     "low dissolved oxygen" if the abundance (and/or biomass) metric scores a 1
     due to insufficient abundance (and/or biomass). Alternatively, if the abundance
     (and/or biomass) metric scores a 1 because of excessive  abundance (and/or
     biomass) the site is classified as degraded by "eutrophication".
  2. Segment classification: The assignment of probable causes of benthic degra-
     dation for the overall segment is accomplished using a simple 25% rule.  If the
     percent of total sites in a segment impacted by a single cause (i.e. sediment
     contaminants, low dissolved oxygen, or eutrophication) exceeds 25%, then that
     cause is assigned. If no causes exceed 25%, the cause is considered unknown.
     The cause(s) should be identified as a suspected (vs. verified) cause of benthic
     community degradation in the ADB database.

Table J-l  shows the  possible conclusions from applying the above protocol. The
States should carefully review the results from application of the protocol to ensure
all findings and conclusions  are rational and reasonable.  Best profession judgment,
common sense, and ancillary information about each segment should be utilized as
necessary and available.
Habitat
Metric
Lower Limit
(Metric
Score=1)
Upper Limit
(Metric
Score=1)

Tidal Freshwater
Abundance (# m-2)
^Biomass(gm-2)^
<800
25500

Oligohaline
Low Mesohaline
Abundance (# m-2)
Biomass (g m-2)
Abundance (# m-2)
Biomass (g m-2)
<180
<500
<1
24050
26000
230

High Mesohaline Sand
Abundance (# m-2)
Biomass (g m-2)
<1000
<1
>5000
250

High Mesohaline Mud
Abundance (# m-2)
Biomass (g m-2)
<1000
<0.5
>5000
250

Polyhaline Sand
Abundance (# m-2)
Biomass (g m-2)
<1500
<1
58000
250

Polyhaline Mud
Abundance (# m-2)
Biomass (g m-2)
<1000
<0.5
>8000
230
      J-2, Metric scoring
Source: Llanso 2002, Table 9,
for eutrophication and low dissolved oxygen causes.
pages 24-26.

-------
J-4
        ,--",  Possible conclusions from application of the assessment protocol.
 n>=10  - sufficient sample size for assessment

Scenario
1
Impairment Analysis
CL-L
(P-Po)
(Table 3 of
VERSAR
Technical
Report)
<0
Impaired:
Degraded Area
method?
(Table 3 of
VERSAR
Technical
Report)
No
Stressor Diagnostic Analyses
Samples with
contaminant
Posterior Prob.
p>= 0.90; % of
Total (Table 5 of
VERSAR Technical
Report)
review as
supplemental info
Degraded Samples with
excessive Abundance/Biomass;
% of Total w/o Cont. (Table 5 of
VERSAR Technical Report)
review as supplemental info
Degraded Samples with
Insufficient
Abundance/Biomass; % of
Total w/o Cont. (Table 5 of
VERSAR Technical Report)
review as supplemental info
   A small, non-significant fraction of IBI scores are within or below the lower range of the reference distribution so water quality conditions in this
   segment support the benthic community (no impairment).
   Where community samples are degraded, the stressor analyses may provide information that supports other assessment data.
2
>0
Yes
< 25% of Total
Samples
< 25% of Total Samples
< 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses do not suggest dominant stressors affecting community composition.  Cause of degradation is "unknown".
3
>0
Yes
> 25% of Total
Samples
< 25% of Total Samples
< 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses suggest sediment contaminants as a likely pollutant affecting benthic community structure.
4
>0
Yes
> 25% of Total
Samples
> 25% of Total Samples
< 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses suggest sediment contaminants as a likely pollutant affecting benthic community structure. Observation of high
   biomass or abundance is indicative of eutrophic conditions as an additional stressor affecting the benthic community.
5
>0
Yes
> 25% of Total
Samples
< 25% of Total Samples
> 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses suggest sediment contaminants as a likely pollutant affecting benthic community structure. Samples observed with
   low biomass or abundance are indicative of low dissolved oxygen as an additional stressor affecting the benthic community.
6
>0
Yes
< 25% of Total
Samples
> 25% of Total Samples
< 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses do not suggest sediment contaminants as a stressors affecting community composition. Samples observed with
   high biomass or abundance are indicative of eutrophic conditions (excessive nutrients) as a stressor affecting the benthic community.
7
>0
Yes
< 25% of Total
Samples
> 25% of Total Samples
> 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses do not suggest sediment contaminants as stressor affecting community composition. Samples observed with high
   biomass or abundance are indicative of eutrophic conditions within the segment while other samples observed with low biomass or abundance
   are indicative of low dissolved oxygen as another stressor within the segment.
8
>0
Yes
< 25% of Total
Samples
< 25% of Total Samples
> 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses do not suggest sediment contaminants as a stressor affecting community composition. Samples observed with
   low biomass or abundance are indicative of low dissolved oxygen as a stressor affecting the segment.
9
>0
Yes
> 25% of Total
Samples
> 25% of Total Samples
> 25% of Total Samples
   A large, significant fraction of IBI scores are within or below the lower range of the reference distribution, so water quality conditions in this
   segment do not support the benthic community (impaired condition).
   Stressor diagnostic analyses suggest sediment contaminants as a likely pollutant affecting benthic community structure. Samples observed with
   high biomass or abundance are indicative of eutrophic conditions within the segment while other samples observed with low biomass or
   abundance are indicative of low dissolved oxygen as an additional stressor within the segment.
  appendix j  •  Bay Estuarine Benthic Communities Assessment Protocol for Maryland  and Virginia

-------
                                                                                                          J-5
      J-l,  (continued)
  n<10 - small sample size, insufficient for analysis

Scenario
1
Impairment Analysis
CL-L
(P-Po)
(Table 3 of
VERSAR
Technical
Report)
n/a
Impaired:
Degraded
Area? (Table 3
of VERSAR
Technical
Report)
Unknown, Not
Assessed
Stressor Diagnostic Analyses
Samples with
contaminant
Posterior Prob.
p>= 0.90; % of
Total (Table 5 of
VERSAR
Technical
Report)
review as
supplemental info
Degraded Samples with
excessive
Abundance/Biomass; % of
Total w/o Cont. (Table 5 of
VERSAR Technical Report)
review as supplemental info
Degraded Samples with Insufficient
Abundance/Biomass; % of Total w/o
Cont. (Table 5 of VERSAR Technical
Report)
review as supplemental info
    There are too few samples to define the confidence interval of benthic sample IBIs, so in this segment -the biological community condition is
    unknown.
    Where community samples are identified as degraded, information from the stressor diagnostic analyses may provide supplemental information
    that may support other assessment data.
Llanso, RJ. 2002. Methods for Calculating the Chesapeake Bay Benthic Index of biotic
Integrity. Versar Inc.,  Columbia, Maryland http://www.baybenthos.versar.com/docs/Ches-
BayBIBI.PDF

Llanso , RJ., J.H. V01stad, D.M. Dauer, and M.F. Lane. 2005. 2006 303(D) Assessment
Methods For Chesapeake Bay Benthos. Final Report Submitted to Virginia Department of
Environmental Quality, Richmond, Virginia. September 2005.
  appendix i   *  Evaluation of SAV Acreages from 2003 to 2005 for Prioritizing Shallow-water Monitoring

-------

-------
                                                                         K-1
                    appendix
        2006  303(d) Assessment
   Methods for  Chesapeake  Bay
                      Benthos
                    Final Report Submitted to:
             Virginia Department of Environmental Quality
                      629 East Main Street
                    Richmond, Virginia 23230


                        Submitted by:
                       Roberto J. Llanso
                        Jon H. V01stad
                  Versar, Inc., Columbia, Maryland

                       Daniel M. Dauer
                        Michael F. Lane
                 Department of Biological Sciences
                    Old Dominion University
                        Norfolk, Virginia

                        September 2005
                       FOREWORD
This report, 2006 303(d) Assessment Methods for Chesapeake Bay Benthos, was
prepared by Versar at the request of the Virginia Department of Environmental
Quality, under Purchase Order # 11646 between Versar, Inc. and the Commonwealth
of Virginia. Old Dominion University contributed to the diagnostic (discriminant
tool) assessment and to project conceptualization and evaluation. The statistical
analyses for the 2006 impairment assessment were conducted by Dr. Ed Weber and
Ms. Jody Dew, of Versar. Dr. Weber also contributed to the development of the
Degraded Area method presented in this report.
                   appendix k •  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-2
                    To meet the requirements of the  Clean Water Act, the States of Maryland and
                    Virginia are using benthic biological criteria for reporting overall condition and iden-
                    tification of impaired waters in Chesapeake Bay. The Chesapeake Bay benthic index
                    of biotic integrity (B-IBI) is the basis for these biological criteria. Previous work
                    conducted by Versar and Old Dominion University had two objectives: to develop a
                    methodology for the assessment of benthic community status for 303(d) impairment
                    decisions  and to produce an assessment for each of the Chesapeake Bay segments
                    and sub-segments containing benthic community data. A statistical procedure was
                    developed that tests whether the distribution of B-IBI scores from probability-based
                    samples collected from a Bay segment is significantly different from the distribution
                    of scores  from reference sites (Llanso  et al.  2003). This procedure, a stratified
                    Wilcoxon rank sum test, was evaluated and applied to the 2004 assessment data. The
                    assessment  resulted in 26 segments considered  impaired based upon  benthic
                    community  condition.  The Wilcoxon approach, however, was sensitive to small
                    shifts in B-IBI scores relative to the reference condition, even in some cases where
                    a majority of the B-IBI scores in a segment met the restoration goals. For stratified
                    data (i.e.,  the habitat types of the B-IBI, see below)  it was not possible to estimate
                    the magnitude  of the shift, for example by using  a Hodges-Lehman confidence
                    interval. Thus, with the Wilcoxon approach we were unable to estimate the magni-
                    tude of degradation: the difference between the segment and the reference condition.
                    A small difference could be statistically significant but of little ecological relevance.
                    It was recommended that alternative methods be evaluated, especially those that take
                    into account magnitude of departure from reference conditions and whether this
                    magnitude is above specific thresholds  of protection that the States may wish to
                    implement. For the 2006 303(d) report, we developed a new method that quantifies
                    magnitude of degradation. We call this method "Degraded  Area." In the  present
                    report, we describe the Degraded Area method, apply this method and the Wilcoxon
                    approach to the 2006 assessment data, and compare the results.

                    In addition, a benthic diagnostic tool has  been developed that can be used to identify
                    potential sources of stress affecting benthic community condition in the Chesapeake
                    Bay (Dauer et al. 2002). The tool can distinguish stress  due to contaminants versus
                    stress due to other factors (e.g., low dissolved oxygen, or unknown). This screening
                    tool was used to identify which impaired segments have a high probability of sedi-
                    ment contamination. These segments could then be targeted for additional sampling
                    or evaluation. The B-IBI metric scores for abundance and biomass were also used to
                    identify (1) insufficient abundance patterns consistent with a low dissolved  oxygen
                    effect and (2) excessive abundance patterns consistent with eutrophication effects.
  appendix k  »  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                              K-3
   1. Develop a new method for the assessment of Chesapeake Bay benthic commu-
     nity status for 303(d) impairment decisions.
   2. Produce an assessment for the 2006 303(d) report using both the new method
     and the Wilcoxon approach.
   3. Apply the benthic diagnostic tool  and the insufficient/excessive abundance
     criteria to the 2006 assessment data.

Like the Wilcoxon (described in Llanso et al.  2003), the Degraded Area method
compares reference data  sets  to assessment  data sets.  The reference  data set
consisted of the calibration and validation data used to develop the Chesapeake Bay
benthic index of biotic integrity (B-IBI). The Chesapeake Bay B-IBI is described in
Weisberg et al. (1997)  and Alden et al. (2002).  The  B-IBI  consists  of benthic
community metrics and scoring thresholds (metric values) that were developed sepa-
rately for seven habitat types (Table 1). The numbers of reference samples in each
habitat used to develop the B-IBI, the Wilcoxon  approach, and the method described
in this report are listed in Table 2. The reference samples were either "good" (=unde-
graded, collected at  sites known to have good sediment and  water quality) or
"degraded" (collected at sites with low dissolved  oxygen, organic enrichment, or
high sediment contaminant concentrations and toxicity). To  develop the B-IBI, Weis-
berg et al. (1997) used averages of three replicate  samples per site for mesohaline
and polyhaline habitats, while Alden et al. (2002) used single replicate samples for
tidal fresh and oligohaline habitats. We used the same metrics values produced by
these two studies, but re-calculated B-IBI scores from these metrics to be consistent
with the  latest B-IBI methodology. The methods for the calculation of the Chesa-
peake B-IBI  are  described in the World  Wide Web at:  http://www.baybenthos.
versar.com/ referenc.htm.

The  assessment data  for the 2006 303(d) report consisted of random  samples
collected from  2000 to  2004 throughout the Chesapeake Bay. A total  of  1,430
samples (single replicates) were used, including  750 samples collected by the Mary-
land Chesapeake  Bay benthic monitoring program, 500 samples collected by the
Virginia Chesapeake Bay benthic monitoring program, 150  samples collected by the
Elizabeth River benthic biological monitoring program, and 10 samples collected for
a gear comparison study in each of Mobjack Bay,  the tidal fresh Mattaponi River,
and the Nansemond River. All assessment samples were collected with a Young grab
(440 cm2 surface area, 0.5-mm screen). For sample collection methods, see the
benthic monitoring program comprehensive reports posted at the World Wide Web
address given above.
                        appendix k  *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-4
                    Assessments were produced for each of 85 Chesapeake Bay Program segments and
                    sub-segments containing benthic data. Segments (TMWA 1999) are Chesapeake Bay
                    regions having similar salinity  and  hydrographic characteristics. In Virginia,
                    segments  were sub-divided into smaller units by the Virginia Department of Envi-
                    ronmental Quality. Sub-segments were produced for each of the mainstems of rivers
                    and bays  (e.g., James  River  mesohaline)  and for some  of the  smaller systems
                    opening into the mainstem (e.g., Pagan River). Assessment samples were assigned to
                    segments  and sub-segments using GIS software. Hydrographic data collected synop-
                    tically with the benthic data were used to assign each sample to one of seven habitat
                    classes used in the calculation of the B-IBI. These are the same habitat classes used
                    in the reference data set.

                    3.2.

                    The new  method developed for the 2006 assessment was based on the confidence
                    limit and  bootstrap simulation concepts described in Alden  et al. (2002). Specifi-
                    cally, bootstrap simulation (Efron and Tibshirani 1998) was  applied to incorporate
                    uncertainty in reference conditions. Bootstrap simulation is used to assess the accu-
                    racy of an estimate by randomly sampling  n times, with replacement,  from an
                    original data set. In our case, we wished to estimate the score corresponding  to the
                    5th percentile of the B-IBI reference distributions for the good sites (by habitat).
                    Because the reference distributions were based on small sample sizes, the percentiles
                    were not well defined and would likely vary if different sets of reference sites were
                    sampled.  Thus the need to estimate this parameter more accurately  with bootstrap
                    simulations. Bootstrap simulations make no assumptions, except that the reference
                    data are a representative sample from a "super population" of reference sites.

                    For each habitat, a threshold based on the 5th percentile B-IBI score of the reference
                    data set for the good sites (or the maximum B-IBI score observed for the degraded
                    sites, see below), was determined. This threshold was not intended to serve as a crite-
                    rion for classifying individual B-IBI scores, rather it was used to  categorize  the
                    segment as impaired or not based on the proportion of sites below the threshold (i.e.,
                    degraded  area) and the  variance associated with  this estimate. The variance  in the
                    estimates  of proportions for each segment was estimated by the simulations.

                    The B-IBI scores for the reference good and degraded sites had degrees of overlap
                    that ranged from quite high in the tidal freshwater and oligohaline habitats to moder-
                    ately low  in the mesohaline and polyhaline habitats. An assessment sample is more
                    likely  to  come from an impaired benthic community  if the B-IBI score for this
                    sample is  within the range of scores observed for sites known to be degraded. There-
                    fore, two  criteria were established for determining the threshold: its score had to be
                    within the lower bound of the good reference distribution (i.e., 5th percentile), and
                    it had to be within the upper range of observed scores for known degraded sites (i.e.,
                    the reference degraded sites). If the 5th percentile score for a simulation run was not
                    within the range of scores for the reference degraded sites, then the maximum  B-IBI
                    score for  the reference  degraded sites was selected as the threshold. Thus, in this
                    study,  sites with low B-IBI scores below thresholds were likely to be impaired and
                    unlikely to come from good reference areas.
  appendix k  *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                               K-5
In each simulation run, a subset of the reference good sites for each habitat was
selected at random, and the B-IBI threshold for this subset was determined (i.e., the
IBI score at the 5th percentile, or the maximum score for the reference degraded
samples). The scores of the assessment data for each habitat were then compared to
the threshold to estimate the proportion of sites below the threshold. By repeating
this process over and over again (5,000 runs) we were able to estimate the variance
in the proportion of sites below the threshold from the bootstrap estimates. This vari-
ance reflects variability in the thresholds as well as sampling  variability in the
assessment data.

In the final step of the method, segments were declared impaired if the proportion of
sites below the threshold (i.e., degraded area) was significantly higher than expected
under the null  hypothesis. Under the null  hypothesis,  a small number of  sites
(defined as 5 % of the sites) would be expected to have low IBI scores even if all  sites
in a segment were in good condition (i.e., no low dissolved oxygen, contaminant, or
nutrient enrichment problems). This is because of natural variability in the benthic
communities, the effects of natural stressors, and sampling and methodological error.
For a segment to be declared as impaired, the lower bound of the 95% confidence
interval of the estimate had to be higher than 5% (the expected proportion under the
null hypothesis), with a minimum sample size of 10. A 5% level was used in agree-
ment with standard statistical practice.

The steps described above are summarized below and in Appendix A:
   1. Thresholds are set for each of seven benthic habitats in Chesapeake Bay.
  2. The threshold is set as the smaller of two values: 5th percentile IBI score for
     the good reference sites or maximum observed IBI score for the degraded refer-
     ence sites.
  3. The 5th percentile score and its variance is estimated by bootstrap simulations.
  4. For each iteration of the bootstrap  simulation, a subset (of same sample size)
     of the good reference sites for each habitat is selected at random (with replace-
     ment), and the 5th percentile score determined.
  5. At each iteration, the threshold is set according to #2.
  6. At each iteration, the assessment data are compared to the reference data to
     estimate the proportion of sites (P) with scores below the threshold. This is
     done for each of one or more habitats within a segment.
  7. P is averaged over all the iterations.
  8. Under the null hypothesis, 5% of the sites (Po) would be expected to have low
     IBI scores, even if all sites in a segment were in good condition.
  9. Segments  are declared impaired if P — Po > 0 (greater than expected under the
     null hypothesis, with 95% confidence) (See Schenker and  Gentleman 2001).

          ,

A stratified Wilcoxon rank sum test was  applied as described in Llanso et al. (2003)
using Proc-StatXact 5  software (Cytel Software Corporation 2002). B-IBI scores
were grouped into three ordered condition categories (1.0-2.0, 2.1-2.9, 3.0-5.0) and
                        appendix k  »  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-6
                    the distribution of scores in each category within a segment was compared for each
                    habitat to the distribution of scores for the good reference condition. Under the null
                    hypothesis (Ho) of no impairment, the two  populations (segment  and reference)
                    were considered to have the same underlying multinomial distributions of samples
                    among the ordered categories. The assessment of impairment was based on a one-
                    sided exact test of Ho  against the alternative hypothesis that the segment had a
                    distribution shifted towards lower B-IBI scores than for the reference condition. The
                    ranking  was  done separately by  habitat,  and  then combined across habitats.
                    Segments with a minimum of 10 samples for which the test was significant at the 1%
                    alpha level and 90% power, were considered impaired under this method.

                    3.4.

                    The benthic  diagnostic tool  allows environmental managers to identify potential
                    sources of anthropogenic stress to benthic communities within Chesapeake Bay. The
                    development and application of the tool was described in detail in Dauer et al. (2002,
                    2005). The benthic diagnostic tool is based on a linear discriminant function that
                    classifies sites in Chesapeake Bay identified as having degraded benthic communi-
                    ties  into categories distinguished  by the type  of  stress experienced  by  those
                    communities. Presently, the function is capable of discriminating contaminated sites
                    from sites affected by all other potential sources of stress in any of the seven benthic
                    habitat types of Chesapeake Bay.  Sites are classified into two groups: 1) a contami-
                    nant group and 2) the other group representing  all other potential sources of stress
                    (eutrophication, low dissolved oxygen, etc.). This function is a linear combination of
                    variables that includes over 60 measures of diversity, dominance, and function of
                    benthic communities. The score for the function is used to calculate the probabilities
                    that a sample is drawn from both groups and the sample is assigned to the group to
                    which it has  the highest  probability of belonging. These probabilities  are typically
                    referred to as posterior probabilities of group membership.

                    For  this assessment, sites with B-IBI  scores < 2.7 were defined  as "degraded" for
                    benthic diagnostic tool application purposes. A score of 2.7 is used in the Chesa-
                    peake Bay benthic monitoring programs to define benthic community degradation.
                    This cutoff value may differ from the  threshold used by the Degraded Area method
                    to determine proportion of sites with degraded benthic communities, but it should be
                    very close to that threshold. Because cutoff values differ, diagnostic tool percentages
                    should only be used as a general  guide for identifying potential causes of degrada-
                    tion. For each "degraded" site, benthic metric values were submitted to the function
                    and  posterior probabilities of group membership calculated. Posterior  probabilities
                    for impaired segments were then used to identify the most likely source of stress
                    affecting benthic communities in these segments. Sites with posterior  probabilities
                    of membership in  the contaminant group that were greater than 0.50 were classified
                    as putatively  contaminated.

                                          "'HI

                    Insufficient and excessive abundance or biomass was determined from the  abun-
                    dance and  biomass  metric  scores  for all  sites not  classified  as putatively
  appendix k  «  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                            K-7
contaminated. In the B-IBI, a score of 1 is assigned to total species abundance and
total biomass if the value of these metrics for the site being evaluated is below the
5th percentile or above the  95th percentile of corresponding reference values. A
score of  1 is assigned for both insufficient and excessive abundance or biomass
because abundance and biomass of organisms respond bimodally to pollution. An
increase in abundance or  biomass is expected at polluted sites when stress from
pollution is moderate, such as at sites where there is organic enrichment of the sedi-
ment.  Excessive  abundance  and excessive biomass are  phenomena  usually
associated with eutrophic conditions. A  decrease in abundance  and biomass is
expected at  sites with high degrees  of stress from pollution; for example, sites
affected by low dissolved oxygen.  The insufficient and excessive  abundance  or
biomass criteria can then be used to determine  the likelihood of low dissolved
oxygen problems  versus  eutrophic  conditions for each of  the Chesapeake  Bay
segments evaluated.
                            4.0
Based on the bootstrap-degraded area procedure, 22 segments with sample size of at
least 10 were considered impaired (Table 3). Impaired segments were sorted according
to the lower 95% bound of the confidence  interval of the difference between the
proportion of sites in the segment below threshold (P) and the proportion of sites below
threshold under the null hypothesis (Po),  from high to low. The estimated P for the
impaired segments ranged from 28 to 76%, and the average B-IBI score was below 3.0
for most segments (Table 3). The estimates for CB4MH and CB5MH exclude the deep
trough (>12 m) of the mainstem which is not monitored because this area is subjected
to summer anoxia and has consistently be  found to be azoic.

Nineteen of the segments declared impaired in this assessment were also declared
impaired by the Wilcoxon test in the 2004 assessment. Three segments (JMSMHb,
PMKOHa, MOBPHa) were declared impaired in this assessment but not in the 2004
assessment, and seven segments (LAFMHa, POCMH, POTOH, GUNOH, TANMH,
NANMH, CB7PHa) were declared impaired in the 2004 assessment but not in the
current  assessment.  Of the new  impaired segments, the Nansemond River
(JMSMHb) and Mobjack Bay (MOBPHa) were sampled with additional effort in
2004. Previously, these two segments and  the Pamunkey River (PMKOHa) had
sample size <10. Of the segments that are no longer classified as impaired, only the
Pocomoke River mesohaline (POCMH) had sample size <10 in the current assess-
ment.

4.2.

The stratified Wilcoxon rank sum test identified 27 segments with sample size of at
least 10 as impaired (Table 3). Segments impaired by the  Wilcoxon test but not
impaired by the Degraded Area method were the lower Bay  meainstem (CB7PHa),
Tangier Sound (TANMH), the Lafayette River (LAFMHa), Severn River (SEVMH),
                       appendix k  *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-8
                    and Gunpowder River (GUNOH). Except for the Severn River, these segments were
                    also identified as impaired in the 2004 assessment.

                    4,3.
                    The diagnostic tool and the insufficient and excessive abundance/biomass criteria
                    can be used  as  ancillary information to  determine most likely source of stress
                    affecting benthic communities in segments classified as impaired. The results of this
                    part of the assessment should be used only as a screening tool to identify probable
                    causes of degradation and to prioritize segments for further study.

                    There is always a risk of misclassifying sites as affected by toxic contamination, low
                    dissolved oxygen, or nutrient enrichment, so independent measurements of sediment
                    and water quality should be made whenever possible. Table 4 presents the results of
                    the diagnostic tool and the insufficient and excessive abundance/biomass character-
                    ization for sites with contaminant group posterior probabilities >=0.50, and Table 5
                    presents the results for sites with contaminant group posterior probabilities >=0.90.
                    A general decision tree for segment assessment and characterization is provided in
                    Figure 1. Results are summarized below.



                    The percentages of degraded samples with a contaminant effect ranged from 67% in
                    the upper James River (JMSTFa)  to 78% in the  middle James River (JMSOHa) for
                    P >=0.5, with average contaminant group posterior probabilities ranging from 0.64
                    to 0.79. At P >=0.9 contaminant percentages ranged from 33-50% (Table 4). At the
                    James River mouth (JMSPHa) no samples were  classified as contaminated. In addi-
                    tion, an examination of all samples collected indicated that only one sample had
                    excessive abundance/biomass  and only one had insufficient abundance/biomass. In
                    the Nansemond River  (JMSMHb), 90% of the degraded samples were classified as
                    contaminated with an average contaminant group  posterior probability  of 0.87.
                    Eighty percent of degraded samples had contaminant group posterior probabilities of
                    at least 0.90. Only three samples were collected in the Chuckatuck River/Pagan
                    River segment (JMSMHc), and three in the Warwick River (JMSMHd). Although the
                    low number of samples makes reliable assessments difficult, degraded samples were
                    collected in both segments and each was classified as contaminated with high poste-
                    rior probabilities of contaminant  group membership. Although only three samples
                    were collected in Willoughby Bay (JMSPHd), each sample was classified as contam-
                    inated. Contaminated  samples in this segment had an average contaminant group
                    posterior probability of 0.84. Additional samples are required in these  segments to
                    determine the extent of benthic degradation and  potential sources of stress.

                    In summary, results indicate that contaminants may account for a large portion of the
                    degradation in the James River,  except for the James River mouth. The primary
                    source of degradation in the Nansemond River appears to be anthropogenic contam-
                    ination.  Sampling  was not sufficient for a reliable  assessment in the Chucktuck/
                    Pagan River and Warwick River segments.
  appendix k «  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                            K-9
Percentages of degraded samples with a contaminant effect ranged from 50% in the
lower Elizabeth River mainstem (ELIPHa) to nearly 91% in the Eastern Branch
(EBEMHa). At least 80% of degraded samples were classified as contaminated in
both the Southern Branch (SBEMHa) and the Lafayette River (LAFMHa) and 68%
were classified as contaminated in the upper Elizabeth River mainstem (ELIMHa).
Of the remaining degraded samples without a contaminant effect, excessive abun-
dance/biomass was  found in 9.1%, 12.5%,  and 5.3%  in  the Western Branch
(WBEMHa),  Southern Branch (SBEMHa) and upper Elizabeth River mainstem
(ELIMHa), respectively,  indicating  the potential of stress due to eutrophication.
Only one sample had excessive abundance in the lower Elizabeth River mainstem
(ELIPHa). Insufficient abundance/biomass was found in 12.5%, 5.9%, and 15.8% of
the  degraded samples  without a contaminant  effect in the Southern Branch
(SBEMHa),  the Lafayette  River (LAFMHa)  and the  upper Elizabeth River
(ELIMHa), respectively, indicating low dissolved oxygen as an additional source of
stress to benthic communities in these segments.

In summary, the predominant source of stress to benthic communities within the
Elizabeth  River  is anthropogenic contamination. Both  eutrophication  and  low
dissolved oxygen appear to be additional  sources  of stress within the  Southern
Branch (SBEMHa) and upper Elizabeth River mainstem (ELIMHa).



None of the upper Pamunkey River (PMKTF) samples had B-IBI scores <2.7, so
none were assessed by the diagnostic tool. Over 57% of the lower Pamunkey River
(PMKOH) degraded samples were classified as contaminated by the tool, but the
average  contaminant group posterior probability was low at 0.62. One additional
sample in this last segment was not classified as contaminated and had insufficient
abundance/biomass.  Few samples  were degraded in  the upper Mattaponi River
(MPNTFa), and 67% of these were classified as contaminated. However, the average
contaminant group posterior probability was low at 0.65 and no samples collected
had a probability of contaminant group membership >=0.90. No samples were clas-
sified as having excessive or insufficient abundance/biomass within this segment. In
the lower Mattaponi River (MPNOHa) 80% of the degraded samples were classified
as contaminated. The average contaminant group posterior  probability in  this
segment was high at 0.87 and group membership probabilities for all samples clas-
sified as contaminated were >=0.90. No uncontaminated degraded samples had
excessive or insufficient abundance/biomass. In the middle York River (YRKMHa)
64% of the degraded samples were classified as contaminated. An additional 9.1%
of degraded samples had excessive abundance/biomass and were not classified as
contaminated by the tool, while  12.1% of the uncontaminated degraded samples had
insufficient abundance/biomass. In the lower York River (YRKPHa) only 46% of the
degraded samples were classified as contaminated. An additional 9.1% and 27.3% of
uncontaminated degraded samples were found with excessive  abundance/biomass
and insufficient abundance/biomass, respectively, in this segment. In Mobjack Bay
(MOBPHa), 50% of the degraded samples were classified as contaminated, all with
contaminant group posterior probabilities >=0.90. An additional 12.5% and 25% of
                       appendix k  » 2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-10
                    uncontaminated degraded samples were found with excessive abundance/biomass
                    and insufficient abundance/biomass, respectively. Insufficient sample size in Severn
                    Creek (MOBPHe), Ware River (MOBPHf), and East River (MOBPHh), precluded
                    reliable assessments of degradation within these segments.

                    In summary,  contaminants  are likely to be substantial contributors to benthic
                    community degradation in the York River, particularly in the lower Mattaponi River
                    (MPNOHa) and the middle York River (YRKMHa). Contamination sources of stress
                    are unlikely in both the lower York River (YRKPHa) and Mobjack Bay (MOBPHa),
                    but both eutrophication and low dissolved oxygen may affect benthic communities
                    in these segments, as well as in the lower York River (YRKMHa).



                    All of the degraded samples in the upper Rappahannock River (RPPTFa) were clas-
                    sified as  contaminated. Only five  samples  were collected in the middle
                    Rappahannock River (RPPOH), making assessments of benthic community degra-
                    dation unreliable. In the lower Rappahannock  River  (RPPMHa), 67% of the
                    degraded samples were classified as contaminated, with an  average contaminant
                    group posterior probability of 0.67. The remaining degraded samples that were not
                    classified into the contaminant group had insufficient abundance/biomass.  Only
                    eight samples were collected in the  Corrotoman River. One of these samples was
                    classified as contaminated and another as uncontaminated with insufficient abun-
                    dance/biomass.

                    In summary, degradation in the upper Rappahannock River (RPPTFa) appears to be
                    the result of anthropogenic contamination while degradation in the lower Rappa-
                    hannock River may be  the result of a combination of contamination and low
                    dissolved oxygen effects. The small number of samples collected makes assessments
                    of overall  benthic community  condition in the middle Rappahannock  River
                    (RPPOHa) and Corrotoman River (CRRMHa) difficult but, the degradation observed
                    appears to be from a variety of sources in both segments.



                    Fifty percent of the degraded samples in the upper Potomac  River (POTTF)  were
                    classified as contaminated by the diagnostic tool.  None of the uncontaminated
                    degraded samples had excessive or insufficient  abundance/biomass. In the middle
                    Potomac River (POTOH), 80% of the degraded samples were classified as contami-
                    nated.  Of  the   uncontaminated  degraded  samples,  20%  had  excessive
                    abundance/biomass and  none had insufficient  abundance/biomass. In the  lower
                    Potomac River (POTMH), 31% of the degraded samples were classified as contam-
                    inated. Of the remaining degraded samples classified as uncontaminated, 65% had
                    insufficient abundance/biomass while only 2.6% had excessive abundance/biomass.
                    In summary, benthic community degradation in much of the upper Potomac River
                    (POTTF) appears to be the result of anthroprogenic contamination. In the middle
                    Potomac River (POTOH), the primary source of stress appears to be contamination;
  appendix k  * 2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                           K-11
however, eutrophication is likely to also affect benthic communities in this segment,
as indicated by the samples with excessive abundance/biomass.

The predominant source of stress in the lower Potomac River (POTMH) appears to
be from low dissolved oxygen, as indicated by the high percentage of samples clas-
sified as uncontaminated and having insufficient abundance/ biomass.



An  inadequate number of samples were  collected in the upper Patuxent River
(PAXTF) and middle Patuxent River (PAXOH) for assessing benthic community
degradation using the benthic diagnostic tool. In the upper Patuxent River (PAXTF),
two   samples  were  classified  as  contaminated  and one  had  excessive
abundance/biomass without likelihood of  contamination.  In the middle Patuxent
River (PAXOH), three samples were classified as contaminated and none had exces-
sive or insufficient abundance/biomass. In the lower Patuxent River (PAXMH), 46%
of the degraded samples were classified as contaminated, with an average posterior
probability of contaminant group membership of 0.51. Of the remaining uncontam-
inated samples, 50% had insufficient abundance/biomass while only 1.5% had
excessive abundance/biomass.

In summary, accurate assessment of benthic community degradation in the upper
Patuxent River (PAXTF) and middle Patuxent River (PAXOH) requires additional
sampling; however, available data suggest contaminants may be a source of stress in
these segments. Degradation in the lower Patuxent River (PAXMH) is likely to be
the result of a combination of contamination and low dissolved oxygen stress.



Over 38% of the degraded samples in the lower Chester River (CHSMH) were clas-
sified as contaminated. Of  the  remaining  uncontaminated  samples, 11% had
excessive abundance/biomass and 33% had insufficient abundance/biomass. Benthic
community degradation in this segment would appear to be the result of contamina-
tion, eutrophication, and low dissolved oxygen effects. All other segments in the
Chester River had low sample size.



Accurate assessment of benthic degradation the upper Choptank River (CHOTF),
middle Choptank River (CHOOH) and  Choptank River mouth  (CHOMH1) will
require additional sampling. In the lower Choptank River (CHOMH2), 67% of the
degraded samples were classified  as contaminated, with group membership proba-
bilities >0.90.  Of  the  remaining uncontaminated  degraded samples, 22% had
excessive abundance/biomass while 11% had insufficient abundance/biomass. Cont-
amination appears to account for most of the benthic community degradation in the
lower Choptank River (CHOMH2), but eutrophication and low dissolved oxygen are
also likely to play a role.
                       appendix k  »  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-12
                    Pocomoke River segments had low sample size;  however, most of the degraded
                    samples in the lower Pocomoke were classified as contaminated.



                    Again, Pocomoke Sound  had low sample size; however, 75% of the degraded
                    samples were classified as  contaminated by the benthic diagnostic tool. Twenty-five
                    percent of the uncontaminated samples had insufficient abundance/biomass. Results
                    suggest that benthic community degradation in this segment stems from a combina-
                    tion of contaminants and low dissolved oxygen.



                    Of the Maryland small Eastern Tributaries, only the Manokin River (MANMH) had
                    adequate sample size.  Seventy-five percent of the degraded samples were classified
                    as contaminated. Of the remaining uncontaminated and degraded samples, 25% had
                    insufficient abundance/biomass.



                    In the Gunpowder River (GUNOH), only 17% of the samples were classified as
                    contaminated.   Of   the  uncontaminated  samples,   50%  had  insufficient
                    abundance/biomass and another 17% had  excessive abundance/biomass. The
                    predominant source of stress to benthic communities in this segment appears to be
                    low dissolved oxygen. In the  Magothy River (MAGMH), 38% of the degraded
                    samples were  classified  as contaminated.  Excessive  abundance/ biomass was
                    observed in 13% and insufficient abundance/biomass in 50% of the uncontaminated
                    degraded samples. Results suggest a mixed source of stress. In the Patapsco River
                    (PATMH), 58% of the  degraded samples were classified as contaminated. The
                    remaining degraded  samples had  insufficient abundance/biomass,  suggesting
                    contaminants and low dissolved oxygen as sources of stress. In the Severn River
                    (SEVMH), 60% of the degraded samples were classified as contaminated. An addi-
                    tional 20%  and 40%  of the uncontaminated degraded samples had excessive and
                    insufficient abundance/biomass, respectively. Results suggest a variety of sources of
                    stress for this segment.



                    Sixty-seven percent of the upper Chesapeake Bay  (CB1TF) degraded samples had
                    possible contaminant  effects, and 17% of the remaining degraded samples had
                    excessive abundance/biomass. Segment CB2OH, on the other hand, had no degraded
                    samples. In  Segment  CB3MH, 55%  of the degraded samples were classified as
                    contaminated while 32% of the remaining  degraded samples had insufficient abun-
                    dance/biomass. In Segment CB4MH, 35% of the degraded samples were classified
                    as contaminated, 25%  of the uncontaminated degraded samples had  excessive
  appendix k  » 2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                               K-13
abundance/biomass, and 35% had insufficient abundance/biomass. Few samples in
Tangier Sound were degraded. In Segment CB5MH, 18% of degraded samples were
classified as contaminated and 82% of the uncontaminated degraded had insufficient
abundance/biomass, indicating a low dissolved oxygen effect. In the lower main-
stem, Segment CB6PH had 67% of the degraded samples classified as contaminated
and 33% of the uncontaminated degraded samples classified with insufficient abun-
dance/biomass. Segment CB7PHa had 63% of the degraded samples classified as
contaminated, but none had contaminant group posterior probabilities above  0.90
and the average probability for the segment was 0.58. Of the degraded samples not
classified as contaminated in this last segment, 13% had  excessive abundance/
biomass and 25% had insufficient abundance/biomass. Finally, none of the samples
near the Bay mouth in Segment CB8PHa were classified as contaminated.

In summary, contaminants were likely  sources of stress to benthic communities in
CB1TF and CB3MH,  while a variety  of  stresses were likely in CB4MH.  Low
dissolved oxygen was the predominant source of stress  in CB5MH, contaminants
and low dissolved oxygen in CB6PHa and CB7PHa, and low dissolved oxygen alone
in CB8PHa.
Alden, R.A. Ill, D.M. Dauer, J.A. Ranasinghe, L.C. Scott, and RJ. Llanso. 2002. Statistical
verification of the Chesapeake Bay benthic index of biotic integrity. Environmetrics 13:473-
498.
Cytel Software Corporation. 2002. Pro-StatXact for SAS users. Statistical Software for Exact
Non-Parametric Inference.
Dauer, D.M., M.F. Lane, and RJ. Llanso. 2002. Development of Diagnostic Approaches to
Determine Sources of Anthropogenic Stress Affecting Benthic Community Condition in the
Chesapeake Bay. Prepared for U.S. Environmental Protection Agency, Chesapeake Bay
Program Office, by Department of Biological Sciences, Old Dominion University,  Norfolk,
VA.
Dauer, D.M., M.F. Lane, and RJ. Llanso. 2005. Addendum to the Report: Development of
Diagnostic Approaches to Determine Sources of Anthropogenic Stress Affecting Benthic
Community Condition in the Chesapeake Bay. Prepared for U.S. Environmental Protection
Agency, Chesapeake Bay Program  Office, by Department of Biological Sciences, Old
Dominion University, Norfolk, VA., and Versar, Inc., Columbia, MD.
Efron, B. and R. Tibshirani. 1998. An Introduction to the Bootstrap. Chapman & Hall/CRC.
Llanso, RJ., J.H. V01stad, and D.M. Dauer. 2003. Decision Process for Identification of
Estuarine Benthic Impairments. Prepared for Maryland Department of Natural resources,
Tidewater Ecosystem Assessments, Annapolis, MD., by Versar, Inc., Columbia, MD., and
Department of Biological Sciences, Old Dominion University, Norfolk, VA.
Schenker, N. and J.F Gentleman. 2001 . On judging the significance of differences by exam-
ining the overlap between confidence intervals. The American Statistician 55:182-186.
TMWA (Tidal Monitoring and Analysis Workgroup). 1999.  Chesapeake Bay Program,
Analytical Segmentation Scheme for  the 1997 Re-evaluation and Beyond. Prepared for the
U.S. Environmental Protection Agency, Chesapeake Bay Program Office, by the Tidal Moni-
                        appendix k *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-14
                      toring and Analysis Workgroup of the Chesapeake Bay Program Monitoring and Assessment
                      Subcommittee, Annapolis, MD.
                      Weisberg, S.B., J.A. Ranasinghe, D.M. Dauer, L.C., Schaffner, RJ. Diaz, and J.B. Frithsen.
                      1997. An estuarine benthic index of biotic integrity (B-IBI) for Chesapeake Bay. Estuaries
                      20:149-158.
   appendix k  *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                              K-15
1,  Habitat classification for the Chesapeake Bay B-1 B1.
Habitat Class
1 . Tidal freshwater
2. Oligohaline
3. Low mesohaline
4-1 . High mesohaline sand
4-2. High mesohaline mud
5-1. Polyhaline sand
5-2. Polyhaline mud
Bottom Salinity (psu)
0-0.5
>0.5-5
>5-12
>12-18
>12-18
>18
>18
Silt-clay (<62 ji) content by
Weight (%)
N/A
N/A
N/A
0-40
>40
0-40
>40
2,  Number of samples by habitat in the original index development data files used by Weisberg et al. (1997) and
   Alden et al. (2002). Calibration (Cal) and validation (Val) samples combined. Habitat Class designations as in
   Table 1.


Cal + Val
Reference Degraded
Reference Good
Habitat Class
1
136
75
2
92
32
3
49
20
4-1
5
14
4-2
81
39
5-1
7
39
5-2
136
24
                    appendix k
2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-16
            _
         -C  O H=

         -
O

O
       T3
       o
         13

         C .ti
         -l°l
         01 ^ O  %

         t S>- P
                c
                01

         o ""»  I,
      OI '4^ i; ^  0)
Dl 0)


O 01
         ^  0)


         ' =3 ~ .
           in
           01 '
               : _ O
               • CQ u
         'C +J u  O

         I/! — "" **~

         •^ D CQ ,-t^

         2 0) CQ .§

         u cu -^  dj
             —  u
             a  c
             •*  0)

           i;-£^

           59 <" *E

             Dl u

           -C Si  o
           4-. 10 Q.

           3 .E oi-

           £ So

           £-^
                c
                ro
         O   E a> D- -^

     :3 DiE <"•§
     - Q) 0{  , 10 ^

     i •£ -7-, 'aT i-  c

     > ^^ " 8  ?

     ; £•- oJ ">  E-o
     ' ^^ TO i— —  Ql (U
         .-t  g 2. .!= - CQ c  g

         1/1  > ro^;c^ — —
       g
       g
               .

           . c o
           - 0) CL
         N  is^

         2|^r

         Q- OJ ^^ M-
            Eu aj ^,
            r- U T3
         TO  aJ c c
         ^  i_ (U D
         ^ ^ aj o


         I'.Is
         O •*-> ni C
           ^ ^  ^
           i c- *<-
           ^ S  °
           c E •£

           ell
           DI ""  jn
           01 r-  2;
           w .—  Q-
           u
               rs
           s_ J  c a.
           dj _Q  ro c

           5 ££ I
           O 3    i/i

           — c  fe =

           £ ^T  m ro
           . _ r^  TO ^

           5 rxi  Dl 10















•o
1
'53
p,
a


£

X
3
a
o
X
o
g





«
3
«
i
1
s
&•
Dl
O











o

£7
^® fS
^ A
O
A
S
z !
Is
IS
a
o
X
s

>


•a
S3 «
o
9
"3

o.



a
1
PH
£
&
-.
O

,—
PH
PH

°

O
PH
PH

&


PH

"5. «

a -3
x
Segment

C*1

"n






~


X
03





X
03

O

O
0



0
p


tN
OO
0



o
0



l>
0
•n
0
o
\0
l>
o


^

POTMH

. in
^ tN
m
•— i tN





tN tN


X X
03 03





X X
03 03
>H >H

O O

O O
0 0



0 0
p o


O tN
OO OO
0 0




0 0


Ti tN

0 0
•n m
0 0
o o
o r-
r-- \o
o o

O 00
•^t tN

SBEMHa
CB4MH

O Ti

^
m m


*f 00


tN tN


X X
03 03





X X
03 03

O O

O O
0 0



0 0
o p


OO OI
0 0




0 0



•*t •>*•
0 0
•n m
0 0
o o
Ch tN
*3- in
o o


,__, •=*•

PAXMH
PATMH



m
<3- *±


in o
•— ' 01

tN tN


X X
03 03
i* >"





X X
03 03
>H >H

0 0

o o
0 0



0 0
o p


00 Tt
0 0




0 0


oo a\

0 0
in in
0 0
o o
01 •>*•
in •=*•
o o

OI •— i
OI VO

CHSMH
CB3MH



0,
*f tN




Ti tN
tN tN


X X
03 03





X X
03 03

O O

O O
0 0



0 0
p p


"3- tN
\O 00
o o




0 0


OI tN
•st W^
0 0
in in
p p
o o
00 1>
•sj- Ti
0 0


OI •— i

ELIMHa
EBEMHa



O 00
•n Tt-



•*t 01

tN tN


X X
03 03
i* >"





X X
03 03
>H >H

0 0
o o
o o
0 0



o o
o o


in •— i
0 0




0 0


tN 00
OI OI
0 0
in in
0 0
0 0
f~- OI
01 ^t
o o

00 TJ-


RPPMHa
YRKMHa


Tt in
vo
•rf in



tN -—I

tN tN


X X
03 03





X X
03 03

O tN
O O
o o
0 0



o o
o o


-H 0
0 0




0 0


tN VO

0 0
in in
0 0
0 0
t~- •— '
01 -^
o o

VO tN
•rj- tN

JMSMHa
CHOMH2

00 00
"*• Tt-
•— i tN
VO ^O


O 00
tN -—I
l> O
tN 01


X X
03 03





X X
03 03
>H >H

0 0
o o
o o
0 0



o o
o o


OI VO
0 0




0 0



tN OI
0 0
in in
0 0
o o
tN 00
OI OI
O O


•^t tN

CB5MH
YRKPHa

oo 





tN 01


X X
03 03





X X
03 03
>H >H

g 0
o o
o o
0 0



o o
o o


in \sO
0 0




o o
0 0



Tf OI
0 0
in in
0 0
0 0
\0 \0
•* 01
o o

•—i O
•— ' tN

PMKOHa
MOBPHa
   appendix k

-------
                                                                                                                                            K-17
1  2
S  ffl
       O  -^   •—i   CO
                       (N  (N   (N
                                                                                                               H  >H   >H  fc
                     O   O    O   O
                                           ooooooooooo
       •—'OfiOOOOOl
       O  O   O  O   O   O  O
                                          O(No
                                                                                             ooof^coocoooo

                                                                                             ,_;,_;,_;o'o',_;o',_;,_;,_;
   o
       oooooooooooooooooooo
       QQ^pppppO'—i   '—''—!  i—i   '—!  i—^CNCNCNCN

       ooo^^^^^^^^^^^^^^1^^
       o  o   o  o
                                                                                          o   o   o  o   o  o   o
       pppppppppppppppppppppppppp
       oooooooooooooooooooooooooo
       oooooooooooooooooooooooooo
 01
 D
 C
'+-•
 C
 o
 u
                         cot^m-— i
           K   ^3
       K  ^
Se
           ffl   S
                5  ffl
CB
                                                                                                          tS   K   K  K   S  K   K
                                                                                                          £   S   S  o   I  S   I
                                                                                                          53   u   o  K   3  ^   2

                                                                                                          §   £   2  CQ   2^  M   U
                      appendix k

-------
K-18












•o
2
'53
&




1
B!
a
o
X
S
a



Vl
13
W)
3
1
si

o









o

*I
*3
S
z 7
A
§3
12
a
o
X
S
^
p

•o
•o «
53 «
M^
0
u
1
B



u
&
£
'o
PH
-.
U
^
PH
fe
°
O

PH
O
PH

PH

±
g"|
x
Segment

o

S
-
«

^
(N


'





,


t>
O
O
o



o
o
~

1



•

00

o
o
o
f-1
(N
O

00

CRRMHa

•n^r-c^c^c^c^oooog'r)0o'nog0

P2S9SSSgSSg°SSPPSoo
— — — — — — — — — °
^^^-^^^^,^-^^^-^^^^^^,^0

ri(^oon(^t^^^>n^O(N(^C^
mO>C*^OOO>OOOOOOOOOOOf^t~^C*^OO>OO
000000^-^^^0^0000-^00

	



	

0^).00^VO(N(NVOVOrfrsu-i00^oomu-ivo

ddddddddddocpdddddod
ooooooooooooooooooo
ooooooooooooooooooo
^^mo^t-t-^^^mo^t^^ooo^
^^t^^tc^^^cN«>«>n>n>«^-^-^-^-^-^-m

J— < , , iTJHHrt HHHHHH^ ^ T3
§§l|||§llilll§ilill
rTi HI ^ HI 1 1 o HI PI o ^ nn o rTi nn ^ TT i . HI ^

m

t~
-
«

oo
(N


'





,


r-
oo
(N
p
o



,__,
o>
o

1



•

<3\

o
o
o
^
C^l
o

m

JMSMHd

1^

G
«
«

rf
r-r>


'





,


o
o



o
00
o

1



•

00

o
o
o
m
m
o

f-1

JMSMHc

o
o

o
o
m
m

^
m


'





,


o
o



(N
^
o

1



•

u-1
o
°
o
o
o
o
o

m

MIDOH

r-

G
(N
(N

(N
f-1


'





,


(N
o



0>
o>
o

1



•

u-1

°
o
o
o
o
o

m

SASOH

o

o
-


rf
fl


'





,


(N
o



«n
00
o

1



'

«n

o
o
o
o

o

(N

RPPMHm
               T3
                0)
                C
                O
                u
  appendix k

-------
                                                                             K-19
          oooooooooo^^^oo
          (NOOOOOOO^O^^^^O
                         t^t^OOOOt^"^
       i A
       •a «
       2 2
     $
        I
              o  ^  *r*
              ^t-  i
            ^H  T3    «  O       HH M
            KHhHtijHHhH   rT  tti rn rn
            S£HS£^H^O£
            H^hwwSofcffi"
            ^^<^r^ffiffieL,Og
            &>nS>^>^oo<;mS
                      appendix k

-------
K-20
                              01
                              >  01
                    -* c
                    s?
                         Q.J'
                         T  l-\
                              ~


                       23    II-
                      -
                                .2
                    °|^|-^||
                         o  en CD — —
                         ~  01 c  °--o
                       _  ' "D 'F  E <"
                    ±;'o to H- c  CD HE
                    co o  o  J3

-^o f
^ 0 ^
Q S

o
^o "^
^ a
0 (D
Q


2
"o
<« - ^
° -2 o
0s H ^
^*
.
^1|

0 (D -<
Q S
T3
-^ Q
|1
0 (D
Q
2
"o
H
t<-i —
o J2
^0 0
o- H
"S

o "^
VO g)
Q

^
"o
H


§ o a
O PH DH

(D (D
r^ f "^
| °|
9

m
1
5b
(D
t/1
0

d



i




,







0

p
d


i



,




0


o
0
,— J

,








.




, 	 1

p
fi

APPTFa
0 0
o o
d d



1 o
q
d


1 0
o
0





0 0

o 8
d d


1 o
p
d


1 t>
vo
vo


O •— '


o t>
o "n
d co
fN
1 t>
VO
vo
vo





1 o

^
o

in Tf

t> tN
t*i t*i

CHKOHa
JMSTFa
in

vo
vo
vo




vo

p
d


Q
d


| 	 !
1 	 1
1—1


^


fN
CO
,_;
^
CO
t>
t>
t>





fN

CO
t>
o

fN
fN
0>
fN

JMSOHa
^ 0
i-; p
tN d



o o
o o
•* d


0 0
o o
o o
^ Tf




0 •*

t> o
"1 p
fN d


o o
o o
•* d


0 0
o o
•^t o
fN t>


VO t>


t*i in
•— i fN
cK vd

0 0
o o
fN O
t^ o>


^


fN O
fN O>
•* VO
VO CO
o o

vo vo

t- Tf
fN fN

JMSMHa
JMSMHb
o o
o o
d d



o o
p p
d d


0 0
o o
o o
2




-H o

§ o
d d


o o
p p
d d


0 0
o o
0 0


^ ^


f-1 f-1
t*"l t*"l
iri c^i
c^i m
0 0
O O

^ ^





in t>
in ^t
co in
0 0

fi t*i

i— i CO
C^i tN

JMSMHc
JMSMHd
o o
o o
d d



1 o
p
d


1 0
o
0





0 0

§ o
d d


1 o
p
d


1 0
o
0


o o


0 0
o o
d d
0
1 0
o

^





• 00
CO
CO
0

O c*"i

"*• ^
^ , 	 ;

JMSPHa
JMSPHd
0 0
o t>
0 r^



O CO
O CO
d "n


t^ in
fN VO
t> t>
tN •— i




^ ^

tN O
in d


o> o
o o
cK d


**• -H
vo t>
vo vo


t> ,— 1


^ ^
i— i in
tN "n
rf "n
t^ Tf
t> fN
tN CO
t> CO


^


t*"l fl
CO O>
fi t^
t> CO
o o

o> t^
•— i fN
^ ^
tN fN

WBEMHa
LAFMHa
^
vo
d



Q

p
CO
vo
0
o
Q
CO


tn


vo
CO

o

t>

p
fN

SBEMHa
0
o
d



o
o
d




o>





^

o
p
d


o
o
d


^
vo
fl
vo


t>


t>

vd
vo
| 	 !
O>
O



^


^
o
CO
o

in

fN
fN

EBEMHa
i— i CO

co in



o> t^

in vd
i i i i

*± m
co m
VO fi
^ ^




t> fN

O CO
t> CO
fN m'


vo t>
fN VO
in vd


tN t>
fi VO
vo vo
fN •— i


in •— i


rf "n

in t>
m -H
fN O
^" O
CO O
vo "n


^


CO O>
in -et
t> CO
vo •*
o o

t> t>

in co
fN fN

ELIMHa
ELIPHa
   appendix k

-------
                                                                                                                                        K-21
T3
 0)
 C
 O
 u
                  11
                  JD g
         m
                           Q  £
                        §  I  S
a
a
                                                •—'   O  
                                                l^,—I,—IOO-—'OOOO
                                     en  en  en
                                     en  -—I  en  en
                                                •—i   o  in  •—i
                                                                      O  •—i   •—i
                                            •—i  Tf  •—i  O>
                                 encNen(N
-------
K-22
                      T3
                      0)
                      C
                      O
                      u
CJ

3


1
o
o

"3
                            O
                            o
                           .
                           I
                           I
•a
es
                                  (D O
                                  o o
                                  Q fi
                                  13
                                   W
                                  i
      PH »
      |£
                                       Q S
                                       ~a °
                                       £°°
                                       Q S
            o 2
            ^j2
          111
          O PH PH
1-sl
I   I
  m
  5
                                 o •—i  •—i
                                                   O -—(  ^H
                                           fe  K
                                           I  °
                                           5  5
TF
                          § 3
                               9 y
                8||«
  appendix k

-------
                                                                                        K-23
T3
0)
C
O
u




1
rn S
"5 O
'5 ffl
|> M-|
(/; fl}
fl} O
!"!
rt S
™ M
"o ^
"^ *-
£3 H

fl} O
1







$
.ts o
^ m

U
-a s>
-§ •*•
S §
txi-Z
IB ^
Q IB
W


«
|g

3 ^

O -

s
«> 'i
o> jjj
1- o
| PH
GO














"^
to "^ ^

^ H _o
^

,
fl} "S
^^8
*£!
T3
fl}
^M "^

^ S1
*
*rt
o
H
^J
fl
Q g ,Q
vO O W
^ H o

. fl} "S
^|3
Q "^
T3
Q— -rt
O ^
^ OD
Q
2
"o
H

VO £
ff- H


=« -§
o a
,~ hn
0 
o
o
o




o
p
o

o
o
o


o



o
o
o


o
o
o

o
o
o


o


00
oo
•— '

o
p
o
o




en

oo
o


^_l
•— '



en
RPPTFa
o
o
o




o
o
o

o
o
o


o



o
o
o


o
o
o

o
o
o


^


o
o
o


o
o
o





,__,

^
o


m




en
RPPOHa
en
en
vo




V~i
vo
en

en
'xf
, 	 I


•n
en


o
o
o


o
o
o

en
en
vo
•— '

00


^
vo
en
en


en

vo

en


o
CN
VO
o


00
o^


vo

RPPMHa
o o
o o
o o




o o
o o
o o

o o
o o
o o
o

•— ' O



o o
o o
o o


o o
o o
o o

o o
o o
o o
o

O •— '


o o
o o
o o
o «n
^
o o
o o
o o
o o




t> -H

^ s
o o


•— i 
vo
vo
vo

rs



o
o
o


o
o
o

o
o
o


o


o
oi


en
en
en
en




en
o^
vo
o


00




(N
CRRMHa
o
o
o




o
o
o

o
o
o


o



o
o
o


o
o
o

o
o
o
o

^


o
o
o
o

o
o
o
o




,__,

0?
o


,__,




^
MATTF
o o
o o
o o




o o
o o
o o

o o
o o
o o


O (N



0 VO
p t>
o -^t


o o
o o
o o
(N

o o
o o
o o
V~i "3"

 «n

•— ' •— '

O O
p p
o o
«n oo




\C -H

o oo
vo vo
o o


(N •— '
•— ' (N



en en
POTTF
POTOH
«n
O"i
•^j-




T-j-
O"i
s

,__,
en
00
00

oo
vo


o
CN
(N


O

(N

0>
'xf
VO


-


^
en
vo
(N


•— ;

en

(N


^
en
vo
en
o


,__,
ON



^
POTMH
o
o
o




o
p
o

o
o
o


o



^
vo
vo


o
o
CN

o
o
»n
(N

^


en
en
en
en

o
p
o





,__,
vo
oo
en
o


vo



vo
(N
PAXTF
o vo
o en
o o
en



o o
o o
o o

O 00
o oo
>/~i o
(N 00

^

o oo
o o


0 t>
p ^
o •-<

o 
(N

^ ^


VO 00
oo vo
(N l>
^ 
O V~>

iS ^t

en


t> 0
t-^ 00
vo o
vo «n
0 0


t> 
-------
K-24
                        T3
                        0)
                        C
                        O
                        u
                                     11
                                     CD $>
                                     Q
                              •s
                              +3

                              I
                              •S
                              a
                              I
                              H
                              es
                              w
                              w
?
                                     J I
                                           "a o
                                           8 "
                                           E-i o
      O
                                           o -<
                                           Q »
                                             8
                                           la
                                           H o
                                           S O
                                           g>-°
                                           Q £
                                             ^
                                             ^
  o
    m
    5
                                                       m  o ^ o  ^
                                                       o  ^ ^
                                               ^  o  ^ m  ^ ^
                   oo  oo en en
                                               a  a  a
                                               2oo
                                               2  §  ^
                                               ra  S  w
                 a a  a a a
                 2 g  2 Q 2
                 Z 5  Z
MA
NA
C
   appendix k

-------
                                                                                              K-25
T3
0)
C
O
u
       CJ

      1
      •g
      H
       CJ
       &



CC
1
-•s .2
j> M-|
X CD
CD O
"75 fl
II
1X1 ^
IB ^

&b -^
(D y
B
KH





s
-B 1
.tS 0
* m

f J
X §
"^ rC
-§ <
«d 
o

^
^
(N
a
BACO
,_,

^


en
en
en
en


en
en
en

•—(



o
p
o


o
p
o


en

en
en


I~H

r-j
CJ
CJ

t>


vo


vq
CJ
K
O
o
o
o
CN


O
o
o


o
o
o

en



^
vq
vo


[-^
vq
vo


^
vq
vo



i~H

vq
vo


t>
vo
vo


^H



0>
vo
o

^
^
CJ
K
o
o
en

en


o
o
o


o
p


vo



CO
CO


o


en

en



vo
o
o

o
en


M


(
t~^
CO
en


t>

^

(N

vo
vo
vo

CJ



o
p
o


o
p
o


en
en
en
en


I~H

CO


t>
vo
vo
vo


vq


o
p
o
CJ


o
o
o
vo


en

CO
o
en


o
p
o
vo

en




fx|
2;


t>

CO

(N



O
p
o


o
p
o


^.j.
•— ;



«n



^
en
t>
o

CO


-------
K-26
                          T3

                           0)
                           C

                           O

                           u
                                  I
                                  So
                                  CJ
                                 C/3
                                 'u
                                  o
                                 •a
m


I

8,
                                 o

                                 o




1
r* &
^ O
'S ffl

X 03
03 C3
t/3 S
"o <

S. §
(tl
a
KH







1
.tS 0
S m
|1
||
C/3 -3
"^ J^
-S <
a g
bj, ^
03 ^
Q aS
o
X
w


+-
§
n ^

I?
"S ^
o -


^ o
03 g
1|
5^









"^
^o *rt °

^ H ^
^

,

I"3 3
^

v§

t~

^d
>J3

OJ



o
p
o


o
p
o


01
01
01
01




'~H


o
p
•n


01
r*"i



^H


O
o

^j-

CJ

EASMH
o
o
•n




O
p
o
o

o
o
o
o

^H



o
p
o


o
p
o


o
o
o





o


o
p
o


o
o
o


o


2
o

^-j-

«

FSBMH
o 01
O O1
O O1




0 O
o \o
§ vS

0 O
o ^o
o vd
O '^O

i-H CS



0 t>
o ^
o ^J


O 01
p 01
O 01


o 01
O 01
o 01





O -— (


o o
p p
o o


o o
o o
o o


o o


vo l>

*3~


o
p
in
 '

^


t> '

\£)
'^O




Tf O


•n o
o o
i — i o


t> '
^q



Tl- 0


i>
in
o

o> o
-H rt
•— i 00

CBITF
CB2OH
0>
O1
\sO




'^O
cs
S

0>

^
t>

O1
CN


Tl-

, — J


01
cs
01


oo
^o
o>





01


^
00
[-^!
cs

Tl-
co
Tf
m

o
^

00
Tl-
•n
o

,__<
\o
S

CB3MH
o
o
m
— < in




•n 01

 0
 -—<
m o
o o

01 in
t -H
S S

CB7PHa
CBSPHa
   appendix  k

-------
                                                                                                                                 K-27
            0)
            > 01
 oirSoTC. ?>•£
m £ U g +; Q- .&.
  .

 TO  01
_C     CD o CD to n


•1-  = E8>E =
 >  0)  O CD <" S —
_  ;-  u 10 -o .2 
    S^^  U  -So.
 o

 c
         CD ^ CD Ol
               .2
 j_ -O  O Q)
 o  a;  >, Q.

*-  c^-o
       o 01 CD -^ '-
      -            -
.•t: o  i/i H- c CD ^
 c o  o o ^5 10 '(/,


 E"occno,
    S  ol fe
 CD  l/l  l/l



•J3  (-  O
            '         -°
QtCQ-DicDO.b


m
                           rt
1

W

•a
 03
 ^H
 U
 >

rt



I
H m

S 8
"ft 9
a -a
a g
"o t^p'nco
ooocoinoooooinc^icjoc^in

1 'O^OOOO 'O^COOO(Nt~^
O^OOOO OOCOOOc^^)
d^coddd do^m'ind^d^d
"T 04 i— i CN i— i

• 'or-0000 'or-mo^^c^
O^OOOO OCJ^OOCOC'^
°g§§§° °stsCTNmm
^



000>J=0^--OOOmmO-l^(N


OOOOCMOOOOOVOOCTNO^OO
OOOOTiOOOOOC^t^ooO — oo
OOOO^OOOOOOl/"lc*^*3'Oooi/~i



1 '000000 'OCTNOOOOCTNt^
OOOOOO OOOOinOt^^O
OOCJOOO OCTN^t^O^^


' 'G-oooo '8S"oSSG
S-?i°oo °SSSSSS

00««vot^««00t^«vot^in«
^ cmpmmqmoo-.'TpCTN>c
OOr^^H^?ic^c^Oc^c^^w"iS^-^

1 •OmOOOO'mTt^OIN'TO
oc^oooo C^^O^T'/^OOOOO
§o c^i c^i ^d t^ — ^d o
^0 m>cl^>coomin


« tN


1 'oc^cso1/^^ 'ooc^c^^o^roo^

^OO^T^OOO"/^ C^C^t^^^f^OO
OOOOOO OOOOOOO


^'/"l^C'J^O^OC^C^OC'^^C^C^i/^C^C^
«(NT« « «(NT«m«

Ot^C^^t^^^oo^t^TTTTOCS'/^oo


aaacacs^o-aa-iscaacscsaa
feffi^aaaffiffiffiffiaffiaaaa
                             appendix  k

-------
K-28
                        T3
                        0)
                        C
                        O
                        u
rt
•w
                              m




1
_ a
-£ o
'I? rn
!> ty
M O
*& B
11
"o ^

S S
W) -^
Q S

S







a>
J= I
.tS .0
3l
d, C
S -^
M §
•^ "5
•rt
rt O
ob -^
a> M
Q o
X
W


-g

-S ^

1 ?

o ..^
^ S
•^ PH
's .s

J2 -2
||
GO













"tf
^o 13 o

^ H ^
S

.

^ 1 0

Q ~g

*^
Q— -rt
O 5
^ oc
Q
2
^s
H
=4H —
O ^3
VO 0
ff- H

T3
. p
° s
^ o1
Q

^te
1
H


0 0 2
QJ PH PH

a> a>
-^ Cj, *Q.

1 ° 1
a
ffl
(D
1
o
o
o




o
o
o


o
o
o




(N



O
p
o


o
o
o



o
p
o
^-

«


\0
en
\^j
en

o
p
o
CO



•^t


^
CO

CO
o



"
-

MPNOHa
o
o
o




o
o
o


o
o
o




o



o
p
o


o
o
o



o
o
o


o


o
p
o


o
o
o



o


I_H
o

^o
o


en

en

MPNTFa
o>
o
o>




o>
CJ
T^
•— (

0>
CN

i-H



>— <



O
p
o


o
o
o



^o
CO
CJ
^

m


t^
C-l
[-^
CJ

\0
CO
CJ



en


o
o

^o
o



"
S

PMKOHa
o
o
o




1




,






o



o
p
o


1




1




o


o
p
o


,





o


,








S

PMKTFa
CO O
en o
o> o




CO O
•— i O
CO O
•— (

o o
en o
o o
en



o o
^


CO O
en p
o> o
o

CO O
•— i O
CO O
--H O



 -H


o o
p p
T} O


CO O
•*t o
CO O



\^o o


en r-
en en

^ *—!
o o



^0
T} t>

YRKMHa
YRKMHb
-3- o
en o
o o
•— < O
^


l> o
 o
o o
CTs O



o> o
p p
o> o


•— < O


-3- o
en p
o o


t~^ o
CN O
l> O



en o


\o ^o

CN -^t

o o





MOBPHa
MOBPHe
o
o
o
o
^


o
o
o
o

o
o
o
o
"~H


>— 1



o
p
o


o
o
o



o
o
o


o


o
p
o


o
o
o



o


^

o

o




en

MOBPHf
o o
o o
o o
o
^


O '
o
o
o

O '
o
o
o
"~H


•— < O



o o
o p
o o


O '
o
o



O '
o
o


o o


0 0
o p
o o


O '
p
o



o o


CO '


••/"I
o




t> r-;

MOBPHg
MOBPHh
   appendix k

-------
                                                                                           K-29
0)
D
C

'+->
C
o
u
            'g m
              >•§
             (D ^

             Q IB


               w
                  "S £
                   a::
                   o w
                   H o
                  "i S
                  "!1
                 1||
                 O PH PH
m

5
                              •—<  O CN
                                             CO  O '—>
                       O -—(  CO O  -—( O  -—(
                         •—I  .—I O  -—( O  -—(
                         -H  O t>  -H
                       rt rt
                       PH a
                      appendix k

-------
K-30
                    T3
                    0)
                     C

                     O
                     U
                          I
                          o
                          "B
                          CJ
                          .fe
                          ^

                          I
                          6
                          O
                          1
                          H
                          a
                          es
                          w
                          -


                          es
                               a -2


                               u "S
                               03 O


                               fl
                               11
                               3 §
                               £ in
o. a




11

-------
                                                                                                K-31
T3
0)
C

O
u
      .
      I

      I
      •g
      +3

      I
es
w

•a



X
X
(H 3
•S o
'>? CO
u "33
1-3

"2^
H
(D y
Q fe
a
KH







»
.s a
*S ,0

o "53
"^H fl
1|
5 §
"^ r^
03 -^
« 03
00 -S
^ X
Q 03
X
W


•g
a °

'a 3
a 0
"S
o ^
O o
-£H CL-
5 0
CO 'C
oi> jj
& o
3 ^
c/3












"o
^



"o
^











o
xO




'o
^

























o
o

;_<
-2
|







•a
"o
H


Ti
1
00
03
Q

o








+-
0


•rt
B
1
00
03
Q


"o
•sO





t*H
o

0^


^
o
^x
°





"x
O
CM


t!_,
°
m

m

Segment


0
O
0
£


"S
0
O
o
Is

(D
1
Q

^te
•a
"o
H

oo
o

>n

0>


K
1
o
p
o



1



,



o



o
p
o



1





,




o


o
p
o


1





o



1


,__<

p


K
BOHO
o
p
o



o
o
o


o
o
o


o



o
p




o
o
o
o
"-^


o
o
o
o


CN


o
p
o


o
o
o



o



^
o

00


^o



o
p
>/~i
CJ


^H



1
o

en
rt
(
en

K
MANM
o
p
o



o
o
o


o
o
o


o



o
o
o



o
p
o



o
p
o
o


'~H


o
o



p
en

K
NANM
o
p
o



o
p
o


o
p
o


o



o
o
o



o
p
o



o
p
o



o


o
o
o


o
p
o



o



1
o

en

CN
en

K
O
t/i
o
o
o



o
p
o


o
p


,_H



o
o
o



o
p
o



o
p
o



CJ


en
en
en
en

o
p
>/~i
t^


en



en
oo
o

o>

oo


a
o
                       appendix k

-------
K-32
                                     CJ

                                     1
                                     •a
                                     H
                                     CJ

                                     &
                              C
                              o
                              u
                                              &i
                                             •a s
                                              a .a
                                             2 ,a
                                             a i
                                             Q
                                                  "83°

                                                  ^t2^
                                                  =" T3  O
                                                   ° 2 O
                                                  =* -a  o
                                                   ° 2 O

                                                  ^ S>-°
                                                     Q  £
                                                  •5 ts •§

                                                  6 (2 <§
                                                     na
S
m
                                                          O  ^H


                                                          O  ^H



                                                          O  ^H
                                                                 o  ^H cn  C^  oo
                                                                       m  m  o  ^H
                                                                    ^H o  oo  t^
                                                                              m  oo  ^H
                                                                           C^  t~^  ^H  c~J

                                                                           CN  CN  tN  tN
      a  a
      o  o
      Q  a
      <;  x
      CQ  CQ
                                                                 a  a  a  a  a  a  a
                                                                 2  §  §  I
                                                                    o
Q  >  5
a  w  o
S  «>  x
   appendix  k

-------
                                                                                             K-33
0)
D
C
'+->
C
O
u




1
rn S
"5 O
'I? rn
j> y-j
03 O
C/3 3
"03 <3
•^ *-
s. §
i>-s
Q £
>9







X
a
.a a
.tS 0
£ ffl
-1 8
f 1
5/3 3
"^ Js
•8 <

§)•£
$1
o
X
W


+-
IS

§ II

o ^

j* in
'•S o
X 'C

Q, X
3 o
9
t/i













"JH
to "^ °
"Q O
^ H .0
^

,
03 "S
^ T3 O
° 2u
^ l-a
Q £

o
'o "«

*!
%
13
o
H
^_J
fl
Q g ,Q
vo 0 ^

Is

.
03 "S
^ T3 O
0 rt Q
^

(N

0


^
VO


FSBMH
0
o
o
CN



O
o
0
o

o
o
o
0

(



o
q
o




o
q
o


0
o
0



o


o
q
o



o
o



o



vo
vo

 O

vd o




01 '
01
01
01


01 '
01
01
01


•— i O


o o
0 0
0 0



q
o



o o



,__< 1

vo

0


VO 01
in *f
(N 01

LCHMH
MIDOH
0
o
in
CN



O
q
o
m

o
o
o
in

(



O
q
•n




O
q
O
m


0
O
0
in


^


O
q
o



o
o



o



t>
vo
(N
m
0


^
rs
01

NORTF
(N O
^" O
0 0




o o
in O
(N 0
VO

§o
o
in O
t>

vo o



O O1
0 m
0 0




O O1
O O1
O O1
01


0 O
o vo
o vo
vo


o -3-


CO 01
0 m
(N 0


O O1
in 01
oi 01
'—> 01






^ -H

o> vo
(N in
0 0


CO O>
(N --H
01 01

TANMH
CBITF
o r-
O CO
O l>
CN



1 T-j-
co
in

, ^
i-H
^
o

O 01
(N


o -3-
o vo
o --5




1 01
CN
O1


i co
vo
o>



O 01


O in
q t>
O ^


1 01
q

CN


O O>



, t^
CO

m
0


0 -H
CO !>;
01 (N

CB20H
CB3MH
vo o>
CO O
(N -3-
^" 01



o -3-
O (N
O CO
VO CO

o o
q q
o o


(N l>



01 t>

1—5 oi
(N



O CO
O CO
O in


§co
CO
in in
01


t> -H


^ m
K 5



q t^
O --H



(N (N



01 m
0 (N

01 (N
0 0


CO Tf
(N ^
^ t-
(N (N

CB4MH
CB5MH
(
i-H
^H
i-H



t^

s

^
vo
vd
vo

(N



vo
•n
in




01
01
s


01
01
01
01


^


vo
•n


01

01
01






CO
t>
o

0


CO
01


CB6PHa
in 01
o> 01
01 01
i-H i-H



o o
o o
in O
l> O

o o
q q
•n O
l> O

VO 
vo o>

in O
0 0


01 in
01 "^t
01 01

CB7PHa
CBSPHa
                       appendix k

-------
K-34
                           T3
                           C
                           TO
                           +J
                           C
                           0)
                           c
                           o
                           'l/l
                           'u
                           0)
                           T3
















hH
hH
hH
ft
64
PH






a ^
O »

Characteriza
Probable Cai
8 '-B
S a
»* ^
^^



U
^
'S ®

al application
:ool for stress
racterization
a .. 54
•2-5 "S
o. g
0 a
54
•d













u
^
'S ®

al application
;ool for stress
racterization
|.a|
al "
0 a
64
•d












•s
'35*
V
c2 s
*• s!
iagnostic tool
of probable c
egradation
•° a -d
>, .2
11
•< SB

a
2










T T
J GC GC










hH
hH
ft
64
PH


a
x
v
35
%
•<
•*-
a
1
L.
=.
hH







1
V
=.
g
54
-f-
a
-
'3
1
a
hH







64
•^
•d
11
e s
"E.
^



^



•d
^ "^ .
54 W
•3 -a
— Q,
U ^_
g O
M P




O



•d
v
64 '
It
1 •§
I








T T


hH
ft
64
PH
1 §
« "54
a. 5
1 *



0
^H
V
z

o
Z

o
^H
Al
Z



   appendix k

-------
                                                                          K-35
                APPENDIX
 Benthic Index of Biotic Integrity

    (B-IBI) for 2006 303(d)  List


       Alternative Assessment

               Methodology

        Roberto Llanso, Jon Velsted, Ed Weber
                   Versar, Inc.
                   Daniel Dauer
               Old Dominion University
                    (co-Pis)
                  August 23, 2005
                  Summary

 The impairment assessment for each segment is based on the
 proportion of samples with "low" B-IBI scores (i.e., below a
 threshold)

 Two steps, estimate:

1.   Proportion of sites in a segment with scores below a threshold (P)

2.   Difference between P and the expected proportion under the null
    hypothesis (P0), i.e., if the segment were in good condition (no low
    DO, contaminant, or nutrient enrichment problems), we would still
    expect a small proportion of sites to have low" scores (e.g., because
    of natural variability); this proportion under the null hypothesis is
    defined as 5%.
          appendix k  *  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
K-36
                                                           Summary (cont.)
                                          Thresholds are set for each of seven benthic habitats in Chesapeake
                                          Bay: tidal fresh, oligohaline, low mesohaline, high mesohaline sand, high
                                          mesohaiine mud, polyhaline sand, polyhaline mud.
                                          The threshold is set as the smaller of two values:
                                         1,    5""1 percentile IBI score for the good reference distribution (i.e., sites
                                              with low scores are unlikely to come from good reference conditions)
                                         2,    Maximum observed IBI score for the degraded reference distribution
                                              (i.e, sites with low scores are likely to come from degraded
                                              conditions)
                                          See example next slide for two hypothetical habitats:  1) Habitat A, the
                                          distributions of scores for the good and the degraded reference sites do
                                          not overlap, 2) Habitat B, the distributions overlap.
                                     Habitat A
                                                       IBI scores
                                                                         2,0   2.7
                                     Habitat B
                                                              IBI scores
                                                                            2.2  3.0

-------
                                                                                                      K-37
                      Summary (cont.)
   Reference distributions are sometimes based on a small number of
   samples; therefore the 5th percentile score is not well defined

   The 5th percentile score and its variance was estimated by bootstrap
   simulations

   For each iteration of the bootstrap simulation, a subset of the good
   reference data for each habitat was selected at random, and the 511
   percentile score determined

   Over all the iterations, the 5* percentile score varied, and at each iteration
   the threshold was established according to the rule described earlier

   See next slide for the two habitat examples
                 181 scores
H.ihtll I

-------
K-38
                                                         Summary (cont.)
                                      For each iteration of the bootstrap simulation, the assessment data are
                                      compared to the reference data to estimate proportion of sites with scores
                                      below the threshold

                                      This is done for each of one or more habitats within a segment (i.e.,sonne
                                      segments have sites in more than one habitat)

                                      See next slides for the two examples
                                    Hnb tot
                                                     IBI scores
                                                                       2.0
-2.7-
 5%
                                                                                 3,3
                                                                             \ \
                                      Habitat A
                                      Habitat B

-------
                                                                                        K-39
  Hah tit B
    Habital A
    Habitat B
                   c-
                   8
                                     r\
                Summary (cont.)
Example of calculations for a hypothetical segment with two habitats:
Iteration
1
2
3


n

n
10
10
1D


10

Habitat A
threshold P<*reshold
20 0.4C
2.0 a AC
2.0 0.4C


20 0.4C

n
-0
as
iO


^0

Habitat B
threshold P 
-------
K-40
                                                     Summary (cont.)
                                    Under the null hypothesis, 5% of the sites (P0) would be expected to have
                                    low IBI scores, even if all sites in a segment were in good condition {i.e. no
                                    low DO, contaminant, nutrient enrichment problems)

                                                                     _-—JT 5% of sites
                                    Segments declared impaired if P greater than expected under the null
                                    hypothesis
                                               P- P. > 0 (with 95% confidence)
                                                     Summary (cont.)
                                    Variance components in P added
                                     1  Variance in P due to estimating thresholds - from bootstrap
                                     «  Sampling variation within segment - binomial
                                    Confidence interval of P - P0 =
                                        P - Pa ± 1.96(SEP + SEP )  = P - P., + 1,96*SQRT(Varp 4 VarP )
Var^ = Variance from bootstrap =
from segment = (pq/N-1)
                                                                       5000 -1
                                                                               plus variance
   appendix k  »  2006 303(d) Assessment Methods for Chesapeake Bay Benthos

-------
                                                                                              K-41
Advantages of new method over Wilcoxon's

 Wilcoxon

   •  evaluates differences in distributions based on ranks, cannot quantify
     magnitude of shift

   •  sensitive to small shifts in distribution of B-IBI scores

 New method

   •  estimates proportion of area below thresholds and magnitude of
     departure from reference conditions

   •  tests if this magnitude is above specific thresholds of protection

   •  incorporates uncertainty in reference conditions as well as sampling
     variability in the assessment data

   •  does not require purchase of special statistical analysis package
     (Wilcoxon does)

   *  Both methods are suitable for data segregated into multiple habitats for
     which reference distributions are not homogeneous
                 appendix k  «  2006 303(d) Assessment Methods tor Chesapeake Bay Benthos

-------
                                                              L-1
              appendix |

Addendum  to the Report
 Development of Diagnostic Approaches to
Determine Sources of Anthropogenic Stress
 Affecting Benthic Community Conditions
          in the Chesapeake Bay
                 Prepared for:

       U.S. EPA Chesapeake Bay Program Office
           410 Severn Avenue, Suite 109
            Annapolis, Maryland 21403
                 Prepared by:
               Daniel M. Dauer1
               Michael F. Lane1
               Roberto J. Llanso2
          Department of Biological Sciences
             Old Dominion University
             Norfolk, VA 23529-0456
                 2Versar, Inc.
               9200 Rumsey Road
            Columbia, Maryland 21045
                  June 2005
                                    appendix I • Addendum to the Report

-------
L-2
                                             1.  INTRODUCTION

                    Dauer et al. (2002)  submitted a report to the US EPA Chesapeake Bay Program
                    Office on the development of diagnostic approaches to determine sources of anthro-
                    pogenic stress affecting benthic community condition in the Chesapeake Bay. The
                    objective of the study was to develop analytical tools capable of classifying regions
                    in Chesapeake Bay identified as having degraded benthic communities into cate-
                    gories distinguished by the type of stress experienced by those communities. The
                    tool  was successful  at identifying regions with high probabilities of sediment
                    contamination.  However, prior to implementation, it was recommended that the
                    operational effectiveness of the diagnostic tool be further tested using additional
                    validation data sets.

                    In this Addendum the results of two additional tasks are presented.  First, the linear
                    discriminant function was independently derived to verify the accuracy of the devel-
                    opment of the function.  Second, two additional putative validation data sets were
                    used to assess the validity of the linear discriminant function.
                                2.  LINEAR  DISCRIMINANT  FUNCTION

                    In this task it was discovered that four samples from the original calibration data set
                    were not included in the derivation of the final linear discriminant function originally
                    reported in Dauer et al. 2002.  The final validation of the linear discriminant func-
                    tion with these additional four samples was identical to that reported in Table 21 for
                    the Baywide scenario, i.e. using the All Province sediment contaminant classifica-
                    tion, namely,  with  an overall percent correct classification  of 75.14%.  The new
                    coefficients for this function are given in Table 1 of this Addendum (revised Table
                    24 of Dauer etal. 2002).
                              3. ADDITIONAL  VALIDATION  DATA SETS

                    Two putative data sets were used for further validation of the Contaminant Discrim-
                    inant Tool (CDT) as presented in Dauer et al. 2002.

                    ELIZABETH RIVER WATERSHED

                    The first putative data set consisted of 125 random samples collected in 1999 from
                    the Elizabeth River watershed (Dauer and Llanso 2003). An additional 100 random
                    samples collected 25 per year from 2000-2003 were also used (Dauer 2001, 2002,
                    2003, 2004).  All samples were analyzed using the CDT function and placed into
                    categories based upon the posterior probability of inclusion into  the Contaminant
                    Group. Due to the high levels of contaminants recorded historically in the Elizabeth
                    River watershed (Hall et al., 1992, 1997, 2002; Padma et al.  1998; Conrad et al.
                    2004), the a priori expectation was that a high  percentage of samples declared
                    degraded by the Benthic Index of Biotic Integrity would be placed into the Contam-
                    inant Group. The results from the Elizabeth River watershed are compared to results
  appendix I  • Addendum to the Report

-------
                                                                                                 L-3
from the Virginia Mainstem that is characterized as having low levels of contami-
nants and accordingly classified as of no environmental concern (USEPA 1999).

Our a priori expectation was that all branches of the Elizabeth River would show a
higher percent area placed into the Contaminant Group compared to the Virginia
Mainstem. For the Virginia Mainstem the number of sites placed into the Contami-
nant Group represented 11% of the entire stratum.  Consistent with our a priori
expectation, all strata in the Elizabeth River had higher proportions placed into the
Contaminant Group, ranging from 40-92% (Table 2; Figure 1). These results indi-
cate strong support for the CDT.

                                 FOR                 BAY

The second putative data set consisted of random samples collected as part of the
Maryland and Virginia Benthic Monitoring Program from 1996-2002.  All samples
were analyzed using the CDT  function and placed into categories based upon the
posterior probability of inclusion into the sediment Contaminant Group.   The  a
priori expectation was that more samples collected near highly urbanized or indus-
trialized watersheds would be placed into the Contaminant Group.  Results are more
difficult to interpret but the pattern of location of samples placed into the Contami-
nant Group is non-random (Table 3; Figure 2), and can be interpreted to be consistent
with known patterns of sediment contaminant distributions for the entire Chesapeake
Bay (e.g.  see USEPA 1999).  GIS maps show patterns of location that agree well
with a priori expectations within highly contaminated regions of the Bay such  as
Baltimore Harbor (Figure 3) and the Elizabeth River (Figure 4).  The maps were
made with data placed  on a 100 m grid and interpolated using a two-dimensional
surface fitting algorithm.
Conrad, C.F. and CJ. Chisholm-Brause. 2004. Spatial survey of trace metal contaminants
in the sediments of the Elizabeth River, Virginia. Marine Pollution Bulletin 49:319324.

Dauer, D.M. 2001. Benthic Biological Monitoring Program of the Elizabeth River Water-
shed (2000). Final Report to the Virginia Department of Environmental Quality, Chesapeake
Bay Program, 35 pp. plus Appendix.

Dauer, D.M. 2002. Benthic Biological Monitoring Program of the Elizabeth River Water-
shed (2001) with a study of Paradise Creek.  Final Report to the Virginia Department of
Environmental Quality, Chesapeake Bay Program, 45 pp.

Dauer, D.M. 2003.  Benthic Biological Monitoring Program of the Elizabeth River Water-
shed (2002). Final Report to the Virginia Department of Environmental Quality, Chesapeake
Bay Program, 56 pp.

Dauer, D.M. 2004.  Benthic Biological Monitoring Program of the Elizabeth River Water-
shed (2003). Final Report to the Virginia Department of Environmental Quality, Chesapeake
Bay Program, 88 pp.

Dauer, D.M., M.F Fane and RJ. Flanso.  2002. Development of Diagnostic Approaches to
Determine Sources of Anthropogenic Stress Affecting Benthic Community Condition in the
                                                           appendix I  »  Addendum to the Report

-------
L-4
                     Chesapeake Bay.  Final Report to the U.S. Environmental Protection Agency, Chesapeake
                     Bay Program Office, Annapolis, Maryland, 64 pp.

                     Dauer, D.M. and RJ. Llanso. 2003. Spatial scales and probability based sampling in  deter-
                     mining levels of benthic community degradation in the Chesapeake  Bay.  Environmental
                     Monitoring and Assessment 81:175-186.

                     Hall, L.W. Jr. and R.W. Alden, III. 1997. A review of concurrent ambient water column and
                     sediment toxicity  testing in the Chesapeake Bay  watershed: 1990-1994.  Environmental
                     Toxicology and Chemistry  16:16061617.

                     Hall, L.W. Jr., R.D. Anderson and R.W. Alden, III. 2002. A ten-year summary of concurrent
                     ambient water  column and sediment  toxicity tests  in the Chesapeake  Bay watershed:
                     19901999. Environmental Monitoring and Assessment 76:311352.

                     Hall, L.W.  Jr., M.C. Ziegenfuss and S.A. Fischer.  1992.  Ambient toxicity testing  in the
                     Chesapeake Bay watershed using freshwater and estuarine water column tests.  Environ-
                     mental Toxicology and Chemistry 11:14091425.

                     Padma, T.V., R.C. Hale, andM.H. Roberts. 1998. Toxicity of water-soluble fractions derived
                     from whole creosote and creosote-contaminated sediments.  Environmental Toxicology and
                     Chemistry 17:16061610.

                     USEPA.  1999.  Targeting Toxics: A Characterization Report. A Tool for Directing Manage-
                     ment  and  Monitoring  Actions in the Chesapeake Bay's  Tidal Rivers, 1999.   U.S.
                     Environmental Protection Agency, Chesapeake Bay Program Office, Annapolis, Maryland,
                     49pp.
  appendix  I  »  Addendum to the Report

-------
                                                                                                 L-5
1.  Revised Table 24 of Dauer et al. (2002).  Coefficients and cutoff values for the Baywide linear discriminant
   function for classifying severely degraded and degraded sites into the Contaminant and Other stress
   groups using "uncorrected" data.
Variable
Isopoda abundance
Isopoda diversity
Isopoda proportional abundance
Amphipoda abundance
Amphipoda richness
Amphipoda proportional abun.
Haustoriidae abundance
Haustoriidae diversity
Haustoriidae proportional abun.
Ampeliscidae abundance
Ampeliscidae richness
Ampeliscidae proportional abun.
Corophiidae abundance
Corophiidae richness
Corophiidae proportional abun.
Mollusca abundance
Mollusca richness
Mollusca proportional abundance
Bivalvia abundance
Bivalvia richness
Bivalvia proportional abundance
Gastropoda abundance
Gastropoda richness
Gastropoda proportional abun.
Polychaeta abundance
Polychaeta richness
Polychaeta proportional abun.
Spionidae abundance
Spionidae richness
Spionidae proportional abundance
Capitellidae abundance
Capitellidae richness
Capitellidae proportional abun.
Coefficient
2.01518
-3.07226
9.45420
0.38084
-0.32010
-4.25029
-3.85522
-1.39235
34.61687
-1.57316
-1.79716
25.88958
37.26499
-18.36548
-2329.15377
2.52241
0.74909
-1.43165
-4.43466
1.28499
-0.27727
-1.23734
-0.15477
-3.82240
0.05506
0.46294
-5.08183
-0.02286
-1.89087
4.02486
0.48588
2.55550
-1.67289
Variable
Nereidae abundance
Nereidae richness
Nereidae proportional abundance
Oligochaeta abundance
Oligochaeta richness
Oligochaeta proportional abundance
Tubificidae abundance
Tubificidae richness
Tubificidae proportional abundance
Deep deposit feeder abundance
Deep deposit feeder richness
Deep deposit feeder proportional abun.
Suspension feeder abundance
Suspension feeder richness
Suspension feeder proportional abun.
Interface feeder abundance
Interface feeder richness
Coefficient
-0.28511
-0.53535
12.23099
0.43911
1.37409
-5.05367
0.33669
0.96057
-2.27273
-1.07320
-2.43057
12.57963
1.05255
-1.25065
2.17966
0.84134
-0.47052
Interface feeder proportional abundance 4.50630
Carnivore-Omnivore abundance
Carnivore-Omnivore richness
-0.05179
-0.00602
Carnivore-Omnivore proportional abun. 3.13784
Total Abundance
Total biomass
Biomass to abundance ratio
Infaunal species richness
Infaunal Shannon Wiener diversity
Infaunal species evenness
Epifauna to Infaunal abundance ratio
Epifauna species richness
Epifaunal Shannon Wiener diversity
Epifaunal species evenness


0.18311
4.75310
-123.97124
-0.04107
1.22042
-2.50732
4.41998
-0.96505
-1.11725
5.85736


                                                                            Cutoff Value = 2.56645
                                                         appendix

-------
L-6
                            2, Percent of the Elizabeth River 1999 strata placed into the sediment
                               contaminant effect group using the contaminant discriminant function of
                               Dauer et al. 2002 (posterior probability > 0.5).  Scuffletown, Gilligan, Jones,
                               and Paradise creeks are subsystems of the Southern Branch. Paradise Creek
                               sampled in 2000.  The Elizabeth River strata are compared to the Virginia
                               Mainstem Stratum.
                                 Stratum
                                          Percentage of Stratum
                                          in Contaminant Group
                                 Mainstem of the Elizabeth River
                                 Lafayette River
                                 Eastern Branch
                                 Western Branch
                                 Southern Branch
                                    Scuffletown Creek
                                    Gilligan/Jones Creek
                                    Paradise Creek (2000)
                                 Entire Elizabeth River watershed*
                                 Virginia Mainstem
                                 * Area weighted value
                                                   40
                                                   60
                                                   64
                                                   72
                                                   64
                                                   60
                                                   68
                                                   92
                                                   54
                                                   11
100
 90
 80
 70
 60
 50
 40
 30
 20
 10
   0
                                                          >0.5
                                                                                            D
                                                                   w
                             1. Percentage of stratum with a B-IBI value < 2.7 and placed into the
                       Contaminant Group with a posterior probability > 0.5.

-------
                                                                                                      L-7
     3,  Percent of the stratum placed into the sediment contaminant effect
         group using the contaminant discriminant function of Dauer et al. 2002
         (posterior probability > 0.5).  Data from 1996-2002.  Elizabeth River data
         includes the intensive 1999 event and 25 random samples of the watershed
         from 2000-2002.
Percentage of stratum
Stratum N in Contaminant Group
Lower (VA) Mainstem 175
Upper Bay Mainstem 175
MD Eastern Tributaries 175
Patuxent River 175
MD Middle Mainstem 175
MD Western Tributaries 175
Potomac River 175
James River 175
Rappahannock River 175
York River 175
Elizabeth River 275
10.9
17.7
16.6
20.0
17.1
24.6
31.4
30.9
37.1
38.3
52.4
% of stratum
50 |
40
30
20 | 	 1
:m Q D D
SI fe « s £ B.I 1
si Ii si £ il |s
§1 3i §^ | gl *f=
a E
— |
. 	 .
— 1 1 — 1


ffi & | ,_ S S
1 ™ I 1
       2, Percentage of stratum with a B-IBI value < 2.7 and placed into the
Contaminant Group with a posterior probability > 0.5.

-------
L-8
                      Figure 3. Diagnostic discriminant tool results and an interpolation fitting algorithm were
                      used to classify Baltimore Harbor benthic communities into categories distinguished by
                      the type of stress experienced by those communities.  Red shading indicates degraded
                      benthic communities stressed by toxic contamination (posterior probability in
                      Contaminant Group > 0.5), with higher  color intensity indicating higher probabilities of
                      contaminant effects (>0.5 to <0.7;  >=0.7 to <0.9;  > = 0.9).  Salmon shading indicates
                      degraded benthic communities stressed  by other sources, most likely low dissolved oxy-
                      gen (posterior probability in  Contaminant Group <=0.5).  Green indicates good benthic
                      community condition.  Middle Branch (mb), Curtis Creek (cc).  Stony Creek (sc), and Bear
                      Creek (be) show contamination as likely source of stress.  The  deep basin north of Curtis
                      Bay and the deep channel southwest of  Sparrows Point (sp) shows other stress (low DO)
                      as probable cause of degradation.
  appendix I  •  Addendum to the Report

-------
                                                                                                      L-9
Figure 4. Diagnostic discriminant tool results and an interpolation fitting algorithm used
here to classify lower James River benthic communities into categories distinguished by
the type of stress experienced by those communities.  Red shading indicates degraded
benthic communities stressed by toxic contamination  (posterior probability in
Contaminant Group > 0.5), with higher color intensity indicating higher probabilities of
contaminant effects (>0.5 to <0.7; >=0.7 to <0.9; > = 0.9). Salmon shading indicates
degraded benthic communities stressed by other sources (posterior probability in
Contaminant Group <=0.5).  Green indicates good benthic community condition. The
Elizabeth River (er), Craney Island (ci), Willoughby Bay (wb), Nansemond River (nr), and
Pagan  River (pr) show contamination as likely source of stress.
                                                               appendix I  •  Addendum to the Report

-------