&EPA
EPA/600/R 07/085 | March 2008 | www.epa.gov/ncea
United States
Environmental Protection
Agency
Climate Change Effects
on Stream and River
Biological Indicators:
A Preliminary Analysis
National Center for Environmental Assessment
Office of Research and Development, Washington, DC 20460
-------
Photo Credits - Cover Page
Top left: Emerging caddisfly larvae, Oregon State University
Top right: Drought conditions in the Little Medicine Bow River at Boles Spring, R. Swanson,
USGS
Bottom left: Dimmer Road East Fork, West Branch, J. Grabarkiewicz
Bottom right: Rainbow trout, W. Davis, http://www.epa.gov/bioindicators
-------
EPA/600/R-07/085
March 2008
Climate Change Effects on Stream and River Biological
Indicators: A Preliminary Analysis
Global Change Research Program
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
ABSTRACT
Climate change is projected to affect aquatic ecosystems through changes in water
temperature, hydrological cycles, and degree days. These effects will manifest themselves
through changes in community composition, phenology, number of reproductive cycles,
evolutionary adaptations, and genetic selection. These changes also serve as indicators of
climate change effects on ecosystems and could be used to document ecosystem condition. State
and tribal water quality agencies use biological indicators to assess ecosystem condition as
required by the Clean Water Act. These assessments rely on comparisons of reference and non-
reference sites. Climate change, however, will affect organisms at both types of sites, unlike
traditional stressors. Therefore, understanding how biological indicators respond to the effects of
climate change, what novel indicators may be available to detect effects, how well current
sampling schemes may detect climate-driven changes, and how likely it is that current sampling
schemes will continue to detect impairment, are important issues in need of discussion. This
report is meant to initiate this discussion by providing information on the potential effects of
climate change on biological indicators, outlining initial strategies to modify assessment
activities to account for climate change effects, and highlighting possible next steps.
Preferred citation:
U.S. Environmental Protection Agency (U.S. EPA). (2008) Climate change effects on stream and river biological
indicators: A preliminary analysis. Global Change Research Program, National Center for Environmental
Assessment, Washington, DC; EPA/600/R-07/085. Available from the National Technical Information Service,
Springfield, VA, and online at http://www.epa.gov/ncea.
11
-------
CONTENTS
Page
LIST OF TABLES v
LIST OF FIGURES vi
PREFACE viii
AUTHORS, CONTRIBUTORS, AND REVIEWERS ix
EXECUTIVE SUMMARY x
1. INTRODUCTION 1-1
1.1. BIOINDICATORS, BIOCRITERIA, AND THE CLEAN WATER ACT 1-3
1.2. CLIMATE CHANGE EFFECTS ON AQUATIC ORGANISMS AND
ECOSYSTEMS 1-4
1.2.1. Changes in Ranges, Distributions of Species, and Community
Composition 1-4
1.2.2. Changes in Phenology 1-7
1.2.3. Evolutionary Effects 1-7
1.2.4. Ecosystem Effects 1-8
1.3. ORGANISMAL/BIOLOGICAL INDICATORS OF CLIMATE CHANGE 1-9
2. STATE BIO ASSESSMENT PROGRAMS—RIVERS AND STREAMS 2-1
2.1. BIO ASSESSMENTS OF RIVERS AND STREAMS 2-1
2.2. BIOINDICATORS USED IN STATE PROGRAMS—RIVERS AND STREAMS
2-2
3. SENSITIVITY TO CLIMATE CHANGE OF BIOLOGICAL INDICATORS USED IN
STATE BIOCRITERIA PROGRAMS 3-1
4. CASE STUDY 1—ASSESSING TRENDS: THE POWER OF BIOLOGICAL
ASSESSMENTS TO DETECT CLIMATE CHANGE 4-1
4.1. OBJECTIVES 4-1
4.2. ANALYSIS APPROACH 4-1
4.2.1. Ability to Detect Change-Power Analysis 4-2
4.2.2. The Maryland Biological Stream Survey Data Set 4-3
4.2.3. Information on TaxaLoss Rates 4-4
4.2.4. Prediction of Expected Taxa Losses with Projected Temperature Increases
4-6
4.3. KEY FINDINGS 4-7
4.3.1. How Long Must Monitoring be Conducted to have a Fixed Probability of
Detecting a Change in the Mean Native Taxon Richness of the Reference
Site Population? 4-7
4.3.2. How Long Must Monitoring be Conducted to have a Fixed Probability of
Detecting a Change in the Mean Native Taxon Richness for a Particular
Site? 4-10
4.4. KEY CONCLUSIONS 4-14
5. CASE STUDY 2—ACCOUNTING FOR TRENDS: BIOLOGICAL ASSESSMENT IN
THE PRESENCE OF CLIMATE CHANGE 5-1
5.1. OBJECTIVES 5-1
5.2. ANALYSIS APPROACH 5-1
iii
-------
5.2.1. Datasets Evaluated 5-2
5.2.2. Metrics 5-2
5.2.3. Regional Data Partitions 5-3
5.2.4. Climate Data 5-3
5.2.5. Hydrologic Attributes 5-6
5.2.6. Specific Analyses 5-6
5.3. KEY FINDINGS 5-8
5.3.1. Observed Stressor-Responses 5-8
5.3.2. Estimates of Climate Change Effects 5-9
5.4. DISCUSSION AND KEY CONCLUSIONS 5-21
5.4.1. Reference Conditions 5-22
5.4.2. Importance of Monitoring 5-22
5.4.3. Analytical Methods 5-23
5.4.4. Stressor Identification 5-23
5.4.5. Biocriteria 5-24
5.4.6. Universal Scale to Measure Biological Condition 5-24
6. RECOMMENDATIONS FOR U.S. EPA TO IMPLEMENT A FOUNDATION FOR
STATE/TRIBAL BIOASSESSMENT/BIOCRITERIA PROGRAMS TO CONSIDER
CLIMATE CHANGE 6-1
6.1. RECOMMENDATIONS FOR U.S. EPA 6-1
6.2. RECOMMENDATIONS FOR STATES AND TRIBES 6-3
7. POTENTIAL NEXT STEPS 7-1
8. CONCLUSIONS 8-1
REFERENCES 1
APPENDIX A: REGIONAL PATTERNS OF CLIMATE CHANGE PROJECTSION
AND CONSEQUENCES FOR RIVERS AND STREAMS A-l
APPENDIX B: DESCRIPTIVE AND SUPPLEMENTAL RESULTS DETAILS FOR
CASE STUDY 2: BIOLOGICAL ASSESSMENT IN THE PRESENCE
OF CLIMATE CHANGE B-l
IV
-------
LIST OF TABLES
Page
2-1. Table of benthic macroinvertebrate metrics taken from the Rapid Bioassessment
Protocols 2-4
2-2. Fish metrics used in various bioassessment programs 2-5
3-1. Summary of expectations for responses of common categories of stream and
river biological indicators to climate change influences on water temperature
and hydrologic regime 3-3
3-2. Novel indicators that may be sensitive to climate change 3-6
4-1. Average annual temperature increases expected by region of the U. S 4-6
4-2. The time (years) to achieve a fixed probability of detecting a statistically
significant effect of temperature increases on macroinvertebrate and fish taxa
loss across different regions under maximum and minimum temperature
projections 4-12
5-1. Discrimination efficiencies of IBIs and EPT taxa under 3 climatic conditions 5-17
-------
LIST OF FIGURES
Page
1-1. Conceptual diagram of how climate change affects aquatic ecosystems and the
possible ecosystem responses that can be measured using biological indicators 1-5
4-1. Effects of confidence level (a [a] and P [b]) on time to detect a climate effect on
macroinvertebrate taxa loss due to climatic warming at high taxa loss rates in the
Northeast/Mid-Atl antic U.S 4-8
4-2. Effects of confidence level (a [a] and P [b]) on time to detect a climate effect on
macroinvertebrate taxa loss due to climatic warming at low taxa loss rates in the
Northeast/Mid-Atl antic U.S 4-9
4-3. Effects of confidence level (a [a] and P [b]) on time to detect a climate effect on
fish taxa loss due to climatic warming in the Northeast/Mid-Atlantic U.S 4-11
4-4. Effects of sample size on time to detect a climate effect on macroinvertebrate
taxa loss due to climatic warming at high taxa loss rates in the Northeast/
Mid-Atlantic U.S 4-13
4-5. Effects of sample size on time to detect a climate effect on macroinvertebrate
taxa loss due to climatic warming at low macroinvertebrate taxa loss rates in
the Northeast/Mid-Atl antic U.S 4-15
4-6. Effects of sample size on time to detect a climate effect on fish taxa loss due
to climatic warming in the Northeast/Mid-Atl antic U.S 4-16
5-1. Maryland MBSS sampling stations showing regional divisions 5-4
5-2. Monthly Palmer HDI for the 30-year period 1970-1999 5-5
5-3. Fish richness vs. temperature in Highland reference streams 5-10
5-4. (a) Macroinvertebrate richness vs. temperature; (b) EPT richness vs.
temperature in Highland reference streams; (c) EPT vs. tempterature relation;
and (d) fish richness vs. temperature relation in reference sites in Piedmont
streams 5-11
5-5. Benthic IBI performance and climatic condition 5-13
5-6. EPT performance and climatic condition 5-14
5-7. Fish IBI performance in three climatic conditions 5-16
VI
-------
LIST OF FIGURES continued
Page
5-8. (a) Relationship of EPT richness to conductivity under drought (red), base
(blue), and wet (black) conditions and (b) conditional probability in
impairment for the same three relationships 5-18
5-9. Conductivity CDF s—Piedmont 5-19
5-10. (a) Relationship of EPT richness to impervious surface under drought (red),
base (blue), and wet (black) conditions, and (b) conditional probability of
impairment for the same three relationships 5-20
vn
-------
PREFACE
This report was prepared by Tetra Tech, Inc. and the Global Change Research Program
(GRCP) in the National Center for Environmental Assessment of the Office of Research and
Development at the U.S. Environmental Protection Agency (U.S. EPA). It is intended for
managers and scientists working on biological indicators, bioassessment, and biocriteria. The
report will provide them with (1) information on the potential effects of climate change on
indicator organisms used, (2) initial strategies for adapting their programs to accommodate these
environmental changes, and (3) highlight possible next steps. The GCRP established a
partnership with the Health and Ecological Criteria Division within the U.S. EPA's Office of
Water with State Water Quality Agencies, and with some Tribal Environmental Agencies in an
effort to develop a foundation for linking climate change to their monitoring and assessment
programs. The background information and research results in this report were presented at a
workshop with state and tribal biocriteria managers and scientists from the U.S. EPA and the
academic community. The "Introductory Workshop on climate Change Effects on Biological
Indicators" was held in Baltimore, MD in March 2007 and focused on climate change effects on
river and stream ecosystems. The Workshop provided state and tribal biocriteria managers with
information on how climate change may affect their monitoring and assessment programs for
water resource protection and restoration. The Workshop included keynote presentations on the
current state of scientific understanding of climate change effects on aquatic ecosystems,
particularly rivers and streams, climate change trends in the past, present, and future, and models
and tools that managers can use to monitor and assess climate change effects. Workshop
attendees also participated in breakout sessions in an effort to identify (1) current biological
indicators of environmental condition, (2) vulnerabilities of biocriteria programs in water quality
agencies, and (3) adaptations of program elements to recognize effects of climate change. Case
studies were presented to aid in understanding the technical ramifications of adapting existing
biocriteria programs. This report includes background information about climate change effects
on rivers and streams and the initial elements of a framework that state and tribal biocriteria
managers can use to modify their programs in response to these effects. The framework
elements described in this report and presented at the workshop are (1) an approach for
identifying biological indicators sensitive to climate change, (2) an analysis for detecting climate
change effects, and (3) methods for continuing to detect impairment under climate change.
Workshop participants made recommendations about research needs and information gaps. This
information is presented in this report and was used in part to derive the proposed next steps that
conclude this report.
Vlll
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
The Global Change Research Program, within the National Center for Environmental
Assessment, Office of Research and Development, is responsible for publishing this report. This
document was prepared by Tetra Tech, Inc. under Contract No.GS-10F-0268K, U.S. EPA Order
No. 1106. Dr. Britta Bierwagen served as the Technical Project Officer. Dr. Bierwagen
provided overall direction and technical assistance, and she contributed as an author.
AUTHORS
Center for Ecological Sciences, Tetra Tech, Inc., Owings Mills, MD
Anna Hamilton, Michael Barbour, Jeroen Gerritsen, Michael Paul
U.S. EPA
Britta G. Bierwagen
REVIEWERS
This report benefited greatly from the comments and suggestions of the following
reviewers.
U.S. EPA Reviewers
Mark Bagley, Evan Hornig*, Sue Norton, Lester Yuan
Other Reviewers
Paul Kazyak (Maryland Department of Natural Resources, Steve Ormerod (Cardiff University,
UK), Mark Southerland (Versar)
ACKNOWLEDGMENTS
The authors thank the Global Change Research Program staff in NCEA, particularly
Susan Julius, Tom Johnson, Chris Weaver, and Jordan West, for their input and advice
throughout the development of this project. We also thank staff in the Office of Science and
Technology in the U.S. EPA's Office of Water for their input and assistance. We appreciate all
of the comments from workshop participants. Their thoughts and ideas contributed greatly to
this report and the formulation of further research.
* Present affiliation with U.S. Geological Survey.
ix
-------
EXECUTIVE SUMMARY
Climate change can have a variety of effects on aquatic species. Changes in air
temperature and precipitation patterns are reflected in changes in water temperature, hydrological
cycles, and degree days. These alterations, in turn, affect aquatic ecosystems, whose responses
can be documented through changes in community composition, phenology, number of
reproductive cycles, evolutionary adaptations, and genetic selection. One method for
documenting changes in ecosystems is through indicators that are particularly sensitive to the
changes or stressor of interest—in this case climate. Some potential indicators of climate change
effects include the following: ratios of drought tolerant to intolerant mussel species; ratios of
invertebrate response guilds that indicate hydrological status; and changes in community
composition. Indicators of changes in composition include shifts from cold- or cool-water fishes
to warm-water fishes; shifts from species associated with hydrologically stable to variable
conditions; and declines in particularly sensitive species, such as salmon, brook trout, or darter
species. The goals of this report are to provide managers and scientists working on biological
indicators, bioassessment, and biocriteria with information on the potential effects of climate
change on indicator organisms used and initial strategies for adapting their programs to
accommodate these environmental changes, and to highlight possible next steps.
Biocriteria programs exist in state and tribal water quality agencies to assess the
biological status and health of ecosystems as required by the Clean Water Act. The United
States Environmental Protection Agency (U.S. EPA)'s Office of Water has developed guidance
documents for states and tribes on bioassessment methods and biocriteria establishment. River
and stream ecosystems were among the first for which methods were developed. The general
approach of bioassessment includes defining reference conditions so that impaired sites can be
defined by comparison with these natural or minimally impacted sites. Currently, there is no
mandate or guidance for biocriteria programs to include climate change effects in the design of
monitoring programs or assessment of impairment. However, as a stressor, climate change will
impact both reference and non-reference sites, unlike the more conventional anthropogenic
stressors currently considered in bioassessment.
Because climate change will affect both, reference and non-reference sites, bioassessment
programs would benefit from the collection of data on climate change effects in their systems.
These data may come from indicators that detect such effects. Biological indicators that are
currently used in bioassessment programs have been selected for their sensitivity to certain
environmental stressors. Knowledge about their sensitivities allows a general extrapolation of
their response to other environmental variables related to climate change. Therefore, most
indicators that are sensitive to the conventionally considered anthropogenic stressors are also
-------
expected to have some sensitivity to climate change: For example, changes in temperature and
precipitation patterns that affect stream flow. Biological indicators that are only (or at least
predominantly) sensitive to climate change may be possible to define, but are likely to be novel,
at least in terms of their application in state bioassessment programs. Although review of the
scientific literature may lead to the identification of novel indicators, these indicators will need to
be easy to measure and practical to implement by state programs in order to be widely adopted.
Indicators that are specifically sensitive to climate change effects are only one approach
that programs could use to detect effects. The case studies discussed in this report present
additional methods and considerations to aid in the detection of climate change effects. The first
case study discusses issues of sampling power needed to detect effects using one type of
indicator of climate change and the second case study examines how climate change may affect
the ability of current monitoring programs to detect impairment due to conventional stressors.
The first case study, Assessing Trends, focuses on sampling power. The power to detect
effects depends on the effect size. In this case, that means the species' loss rate due to increases
in water temperature. The case study explores a low and high species loss rate and low versus
high temperature change scenarios. Using data from one long-term data set and sampling
scheme, it would take 15 years to detect effects due to climate change under the high loss rate
and high temperature change scenario. The other extreme, low loss rate and low temperature
change, would take more than 100 years to detect. As expected, an increasing number of
samples will help detect effects sooner.
The second case study, Accounting for Trends, examines the ability of bioassessment
programs to continue to detect impairment due to conventional stressors in the face of climate
change. The analysis shows that climate change effects will decrease the ability of states to
discriminate between reference and impaired sites, particularly if reference sites are already
somewhat stressed. These results underscore the importance of monitoring sentinel sites, sites
that are revisited during each sampling cycle, in order to detect deterioration of condition at
reference sites due to climate change.
The results of these case studies, preliminary analyses of indicator sensitivities, and
reviews of the literature of climate change effects on aquatic ecosystems were presented at the
Workshop for state biocriteria managers of rivers and streams. Their responses to this
information led to recommendations for additional research and a variety of mechanisms for
assistance to states from U.S. EPA concerning climate change effects in these ecosystems. The
large number of recommendations suggests that it is important to continue this dialogue by
conducting further research and other activities leading to more specific recommendations and
assistance for state programs. This information could then be used to modify state programs to
account for climate change effects and to ensure that management goals continue to be reached.
XI
-------
State biocriteria managers also outlined a number of actions that they could take now in response
to the information presented. In particular, their response reflects an understanding that climate
change will affect the entire ecosystem, and, therefore, regular and repeated monitoring of
reference and sentinel sites to collect biological, hydrological, and temperature information will
be particularly valuable to detect and control for climate change effects.
The recommendations for further research also lead to potential next steps. These steps
include (1) conducting another workshop for biocriteria managers of other aquatic ecosystems;
(2) conducting more in-depth analyses of climate change effects on river and stream
bioassessment programs in different regions of the U.S.; (3) disseminating information on
regional climate change effects on biological indicators; and (4) coordinating information across
U.S. EPA and state agencies to evaluate trends in bioassessment data.
xn
-------
1. INTRODUCTION
Changes in climate affect ecosystems, and, therefore, also their management. State and
tribal water quality (WQ) programs fulfilling Clean Water Act of 1972 CWA) requirements will
need to take these climate change-induced effects into account. Currently, state and tribal
assessment programs include biological monitoring in addition to physical and chemical
parameters, because organisms, and the communities they comprise, integrate the effects of
physico-chemical changes over time and space and provide any trends detected with ecological
relevance. It is for this reason that understanding the biological and ecological responses to
climate change, as well as interactions between climate change effects and other environmental
stressors, is important for continued operation of bioassessment programs and interpretation of
bioassessment results.
Human activities and natural factors have already changed the climate; these trends are
likely to continue into the future (IPCC, 2007). There is now high confidence that anthropogenic
emissions of greenhouse gases and aerosols have resulted in warming, with evidence of globally
increasing air and ocean temperatures, melting of snow and ice, and rising sea levels (IPCC,
2007; Rahmstorf et al., 2007). Global air temperatures have increased about 0.6°C over the last
30 years and 0.8°C over the last century, and global ocean temperatures are probably as warm
now as they were during the Holocene maximum (about 5000 to 9000 years ago) (Hansen et al.,
2006). Observed increases are greater over land masses than over oceans, and they are greatest
at high latitudes in the Northern Hemisphere (Hansen et al., 2006). Extreme cold days, cold
nights, and frost have been less frequent; hot days, hot nights and heat waves have been more
frequent (IPCC, 2007). The third Intergovernmental Panel on Climate Change (IPCC) report
(2001) revealed that the diurnal temperature range was decreasing; however, evaluation of more
extensive data in the fourth assessment report shows that daytime and nighttime temperatures are
actually increasing at comparable rates (IPCC, 2007). An understanding of the potential
consequences of these climatic changes for aquatic ecosystems is an initial step that will assist
water resource managers in modifying their programs to ensure that they will continue to meet
their management goals.
The goals of this report are to provide managers and scientists working on biological
indicators, bioassessment, and biocriteria with information on the potential effects of climate
change on indicator organisms used, initial strategies for adapting their programs to
accommodate these environmental changes, and highlight possible next steps. This report
supports these goals by presenting background information about climate-change effects on
rivers and streams (see Section 1), an overview of bioassessment programs (see Section 2), and
the initial elements of a framework that state and tribal biocriteria managers can use to modify
1-1
-------
their programs in response to these effects. The framework elements described in this report are
(1) a proposed approach for identifying biological indicators sensitive to climate change (see
Section 3), and (2) two preliminary case studies evaluating selected aspects of biological
assessment programs (see Sections 4 and 5).
Biological assemblages integrate effects from all impinging sources of stress, including
"conventional" anthropogenic stressors, which are commonly the focus of state programs
assessing and regulating water quality, and any other significant source of environmental change,
including climate change. This integrative characteristic makes biological assemblages effective
monitoring tools, but it also means that the analyses must reasonably account for all major
sources of stress in order to attribute observed responses to particular sources of stress in a
reliable manner. This attribution allows for more effective regulation of the stressor and/or
management of the resource. The ongoing success of biological monitoring and assessment
programs will require an understanding of both an understanding of what climate-associated
changes are occurring in monitored aquatic communities and how monitoring programs can
account for them. Accounting for climate-change influences will support effective attainment of
management goals using monitoring program results as a foundation.
The two case studies included in this report examine how climate-change effects can be
taken into account through program design and/or analytical approaches and how climate change
may affect the ability of biological monitoring and assessment programs to meet key goals.
Climate change can be viewed as a "global stressor" that affects both reference and non-
reference locations monitored for the effects of more "conventional" stressors. The ability to
account for climate change requires (1) an understanding of how vulnerable monitoring data are
to climate-change effects, and (2) how effectively differences that are a result of climate change
can be detected within existing monitoring programs.
The first case study (see Section 4) describes several important temporal aspects of
change detection. The second case study (see Section 5) examines methods for continuing to
detect impairment under climate change. More detailed examinations of these and other
important components of biological assessment programs are essential before comprehensive
recommendations for adaptation of bioassessment programs to climate change can be made.
Preliminary recommendations, derived from these analyses and the participants in the
"Introductory Workshop on Climate Change Effects on Biological Indicators," and a summary of
proposed next steps conclude this report (see Sections 6 and 7).
1-2
-------
1.1. BIOINDICATORS, BIOCRITERIA, AND THE CLEAN WATER ACT
The CWA of 1972 (Federal Water Pollution Control Act, Public Law (PL.) 92-500)
as amended in 2002 (P.L. 107-303, November 27, 2002) has as a stated goal and policy
(Section 101(a)):
".. .to restore and maintain the chemical, physical, and biological integrity of the Nation's
waters."
The concept of biological integrity has received much attention since passage of the
CWA. It is commonly defined as "the ability to support and maintain a balanced, integrated, and
adaptive community with a biological diversity, composition, and functional organization
comparable to those of natural aquatic ecosystems in the region" (Karr et al., 1986; Karr and
Dudley, 1981; Frey, 1977). This wording highlights some key attributes of biological
communities that are fundamental to preserving "integrity" (diversity, composition, and
functional organization), and it alludes to a basic element of the approach used in bioassessment,
which is comparison to existing natural communities, or reference conditions.
The use of biological monitoring and assessment to establish criteria is mandated in
Section 303(c)(2)(B) and 304(a)(8) of the CWA. Biological criteria (biocriteria), derived from
biological monitoring and assessment, provide narrative and numeric targets that define the
desired condition of communities of aquatic organisms inhabiting streams and rivers where water
quality is subject to regulation. Biocriteria and biological assessment (bioassessment) thus
provide a valuable and direct regulatory mechanism for protecting biological resources at risk
from chemical, physical, and biological impacts. The CWA requires that states and tribes
designate aquatic life uses (i.e., environmental goals) for their waters that appropriately address
biological integrity and adopt biological criteria necessary to protect those uses (Barbour et al.,
2000).
Biologists and other natural resource scientists develop biocriteria by using accepted
scientific principles to characterize the regional reference conditions for the different water
bodies found within a state or tribal nation (Barbour et al., 2000). Effects of climate change are
broad and need to be considered in tracking trends in regional reference conditions that form the
basis of assessing ecological condition. Biological assessment programs are now widespread
throughout the states (U.S. EPA, 2002) and are best served by indicators that can be used for
multiple purposes. Water resource agencies in the 50 states and several tribes are in various
stages of development and implementation of bioassessment methods (Barbour et al., 2000).
Essentially, the multiple purposes of bioassessment can be reduced to two basic questions:
(1) asking whether a water body meets, or exceeds, an impairment threshold, and (2) asking
whether the biological condition of a water body is degraded or improved compared to an earlier
1-3
-------
time, an upstream or a nearby site (Barbour and Gerritsen, 2006). Climate change influences
both of these questions.
1.2. CLIMATE CHANGE EFFECTS ON AQUATIC ORGANISMS AND
ECOSYSTEMS
There is a substantial weight of evidence, summarized in several reviews and meta-
analyses, of ecological changes that are linked to existing climate change (Walther et al., 2002;
Root et al., 2003; Parmesan, 2006). Figure 1-1 is a conceptual model of how climate changes
affect aquatic ecosystems, and some possible ways in which ecological systems respond, which
can be measured using indicators. These analyses identify several categories of ecological
responses expected from climate change, including (1) changes in range and distribution of
species; (2) changes in phenology; and (3) evolutionary effects on morphology, behavior, and
genetic frequencies, due to altered selection regimes. These, in turn, are predicted to alter
community composition and interactions, as well as ecosystem processes, including production
and material cycling. A few of the more common examples are discussed here.
1.2.1. Changes in Ranges, Distributions of Species, and Community Composition
Water temperature drives many biological functions in aquatic invertebrates and fish,
including growth and metabolic rates, reproduction, feeding, and survival. Many fish and insect
species have fairly narrow temperature range tolerances, and these narrow ranges influence their
distribution. Temperature regime determines distributions of species in relation to temperature
tolerances and adaptations combined with competitive interactions, effects on food supply, and
other factors (e.g., Sweeney and Vannote, 1978; Vannote and Sweeney, 1980; Matthews, 1998).
Changing thermal regimes are expected to shift species ranges to the north (and/or to
higher elevations); species at the southern limits of their ranges will migrate or suffer local
extinctions. However, in many areas northward or upstream migrations of certain aquatic
species may be limited by barriers to dispersal such as habitat fragmentation due to dams and
reservoirs, deforestation, and water diversions (Poff et al., 2002; Moore et al., 1997; Covich et
al., 1997; Smith, 2004; Hawkins et al., 1997). There is some experimental evidence for dispersal
of Ephemeroptera-Plecoptera-Trichoptera (EPT) taxa across watersheds in Wales, such that
natural fragmentation of river basins is not directly an impediment to recovery from large-scale
disturbance (Masters et al., 2007). However, northward migrations may be limited in regions
including the southwest and southern Great Plains of the U.S. where most drainages flow east
and west (Poff et al., 2002). Species that are already restricted to headwater streams may be
displaced (Poff et al., 2002). In the U.S., from 36% (Mohseni et al., 2003) to 50% (Eaton and
Scheller, 1996) of cold-water fish habitat, and up to 15% of cool-water habitat may disappear
1-4
-------
Climatic
changes
Effects in
aquatic
ecosystems
Increased
CO2in
atmosphere
>
Increased
air
temperature
>
Altered
precipitation
regimes
Biological and
ecological
responses
Assessment
of responses
Increased water
temperature
Altered evapo-
transpiration
Altered flow |
Reduced ice
cover
Increased
snowmelt
Increased
sea levels
Increased
salinity /
altered water
chemistry
Increased
CO2in
waters
Altered
stratification
regime
Ecosystem
Altered energy flow
and cycling
Prtfii ,1 &3 &*nn
Altered demographic
rates
Community
Altered species
tolerances and
interactions
individual
Altered vital rates
Responses can be measured using indicators
Figure 1-1. Conceptual diagram of how climate change affects aquatic
ecosystems and the possible ecosystem responses that can be measured using
biological indicators.
1-5
-------
(Mohseni et al., 2003) due to the warming projected for a doubling of atmospheric CC>2
concentrations. Fish with the smallest geographic ranges are the most vulnerable. Rahel et al.
(1996) estimate habitat losses for cold-water fish species in the Platte River, Wyoming ranging
from 7-76% for temperature increases of 1-5°C. They anticipated potential population
fragmentation as cold-water species were progressively limited to colder headwater stream
reaches; fragmented populations (based on models for cut-throat trout) may be more susceptible
to extirpation if populations are isolated with limited carrying capacity or interactions with other
populations (Hilderbrand, 2003). In the Mid-Atlantic Appalachian mountains, cold-water brook
trout are near the southern limit of their range, and suitable habitat is mainly found at higher
elevations. Projected temperature increases could raise the elevation at which acceptable
temperatures occur by 700 m, effectively eliminating most brook trout habitat in this region
(Moore et al., 1997).
Daufresne et al. (2003) documented species replacements, range shifts, and variations in
community composition for both fish and macroinvertebrates in the Upper Rhone River in
France associated with increasing water temperatures from atmospheric warming. Increased fish
abundances were associated with increased temperatures and lower flows during the
reproductive period (April-June). In moorland and forest streams in Wales, directional climate
change (increasing temperatures ) decreased spring macroinvertebrate abundances over a 25 year
period, yielding an estimated average of 21% reduction in abundance per 1°C of temperature
increase, and in combination with the North American Oscillation (NAO) accounted for 70% of
interannual variation (Durance and Ormerod, 2007).
There are several other examples of community responses to climate variables in
different regions of the U.S. In the Southeast, freshwater mussels are especially vulnerable to
drought, along with the corresponding low flows and depressed dissolved oxygen (DO) levels,
leading to increased mortalities and local extinctions (Golladay et al., 2004). In the Great Plains,
where many fishes already exist at or near their thermal tolerance limits as a result of high
temperatures and low flows typical of shallow water habitats, increasing temperatures due to
climate change are expected to result in increased extinctions of endemic and local species
populations (Covich et al., 1997). Finally, in the Southwest, the stream fauna is typically highly
resilient and adapted to disturbance, but nonetheless is vulnerable to habitat losses that could
accompany increased runoff variability (Grimm et al., 1997). Biological effects may be
manifested as changes in relative abundance, species losses (local extinctions), and reduced
diversity.
In addition to temperature effects, projected changes in stream flow from climate change
may alter community structure. When considering climate change alone, the Sacramento River
could lose 10-18% (low and high climate change scenarios) of its fish species by 2080; the
1-6
-------
Colorado River 0-5% offish species; the Rio Grande River 0-5%; and the Sabine River 11-13%
(Xenopoulos et al., 2005). Xenopoulos et al. (2005) predict high risk of species extinctions for
subtropical and tropical rivers with a rich endemic fauna and note the vulnerability offish
species that require seasonal floodplain connection for life cycle completion. Using a similar
approach, Xenopoulos and Lodge (2006) estimated potential fish richness losses associated with
20-90% reductions in discharge for several rivers in two regions of the U.S., and report a range
of 2-30% offish species lost in rivers of the Lower Ohio-Upper Mississippi Basin, and 3-38%
offish species lost in the Southeastern U.S. Changes in timing of spring flows resulting from
climate change may have the greatest effects on spring spawning fishes in the Northeast and may
alter the survival of Atlantic salmon by changing migration timing and coincidence with optimal
conditions for survival (Hayhoe et al., 2007). It should be noted that climate-driven changes in
hydrology are considered potentially difficult to define because inter-annual rainfall variability is
large relative to trends predicted from climate change (see Wilby, 2006).
1.2.2. Changes in Phenology
Warmer water may increase the growth rates of aquatic invertebrates and result in earlier
maturation (Poff et al., 2002). In a mesocosm experiment using the mayfly Cloeon dipterum,
temperature increases alone had little effect on nymph abundance, and only small effects on
body length, though emergence began earlier in the year (McKee and Atkinson, 2000). McKee
and Atkinson (2000) also show that for treatments with both increased temperatures and
nutrients, both nymph abundance and size increase. For a Japanese species of mayfly (Ephoron
shigae), cumulative degree days and time of emergence are significantly correlated, explaining
80-90% of the variation in emergence date, depending on whether the analysis is done for all
individuals or separately by sex (Watanabe et al., 1999). For at least this species and most likely
for related species, increasing water temperatures associated with climate change will likely
result in earlier emergence of mayflies due to an earlier accumulation of degree days.
1.2.3. Evolutionary Effects
Evolutionary changes may play a small role in species' responses to climate change
through adaptation (Parmesan, 2006; Berteaux et al., 2004; Hogg and Williams, 1996). These
include processes that have been documented for range shifts due to hybridization and novel
adaptations, chromosomal inversions that allowed tolerance of warmer temperatures in southern
range sub-populations, and body size responses to increasing temperatures due to genetic
plasticity (Parmesan, 2006). However, capacity for evolutionary responses of species will be
limited by range of genetic diversity and generation time, with species characterized by small,
short-lived and abundant individuals more likely to respond adaptively (Bradshaw and
1-7
-------
Holzapfel, 2006; Berteaux et al., 2004; Hogg et al., 1995). Extinctions are still expected as a
likely consequence of directional climate change even with evolutionary changes: in part
because the mean phenotypes lag behind optimal phenotypes, and the rates of environmental
change can outpace estimated maximum sustainable rates of evolution (Bradshaw and Holzapfel,
2006; Berteaux et al., 2004; Burger and Lynch, 1995).
Parmesan (2006) points out that while there are local examples of adaptations to
changing environmental conditions, there is little evidence in the geologic record of the
appearance of novel genotypes in species in response to the larger climate changes associated
with glaciations and interglacial periods. It is expected that species' responses to climate change
will primarily be through range shifts and extinctions rather than through evolution.
1.2.4. Ecosystem Effects
There is evidence that projected increases in CC>2 will reduce the nutritional quality of
leaf litter to macroinvertebrate detritivores. Reduced litter quality would result in lower
assimilation and slower growth (Tuchman et al., 2002). While seemingly a secondary climate-
change effect, changes in these processes could have food web implications: altered stream
productivity that impacts fish and other consumers. In contrast to this, Bale et al. (2002) found
little evidence of the direct effects of CO2 on insect herbivores and instead discuss a range of
temperature effects (including interactions with photoperiod cues) on various life history
processes that affect ecological relationships.
It is not clear whether changes in nutrient loading due to climate change will have any
effects on streams and rivers. Effects of nutrient enrichment in streams are highly variable, due
to questions about which primary nutrient (nitrogen or phosphorus) is limiting, shading (light
availability), water clarity, flow regime, and available substrate for periphyton growth (e.g.,
Dodds and Welch, 2000). In general, nutrient enrichment leads to changes in the algal and
diatom community composition of a stream, and sometimes, in some streams, to increased
production and chlorophyll concentrations, leading to changes in primary invertebrate consumers
(e.g., Gafner and Robinson, 2007) which could cascade through the community (Power, 1990;
Rosemond et al., 1993).
Changes in the distribution and intensity of precipitation may induce related changes in
nutrient loading to streams from runoff. However, it is not clear if total nutrient loading to a
stream will change with altered precipitation. For example, increased precipitation does not
increase nitrogen available on the land surface to run off. However, changes in precipitation
patterns combined with other changes in land use, for example, may affect nutrient loadings.
1-8
-------
1.3. ORGANISMAL/BIOLOGICAL INDICATORS OF CLIMATE CHANGE
Since organisms respond to climatic variability and trends, some of these responses may
be useful as indicators of climate change (Figure 1-1). This section describes several candidate
indicators based on the current literature. Section 3 identifies novel indicators and species traits
that may be sensitive to climate change.
Golladay et al. (2004) surveyed mussel species during drought conditions, identifying
several low flow-sensitive species (e.g., Lampsilis straminea claibornensis, Villosa villosa, and
Lampsilis subangulata). Other species (Pleurobemapyriforme, Mediuniduspenicillatus) show
signs of drought intolerance due to decreased DO concentrations during low flows. Mussel
species less affected by drought-induced low flow and low oxygen levels include Elliptio
complanata/icterina, Villosa vibex, and Villosa lineosa (Golladay et al., 2004). A comparison of
drought intolerant to drought tolerant mussel species may be an indicator of hydrologic
variability or drought possibly due to climate change.
Golladay et al. (2004) suggest that wetland invertebrates could be divided into four
response guilds to indicate hydrologic status that may be adaptable to river/stream systems:
(1) overwintering residents that disperse passively, including snails, mollusks, amphipods, and
crayfish; (2) overwintering spring recruits that require water availability for reproduction,
including midges and some beetles; (3) overwintering summer recruits that only need saturated
sediment for reproduction, including dragonflies, mosquitoes, and phantom midges; and
(4) non-wintering spring migrants that generally require surface water for overwintering,
including most water bugs and some water beetles. Changes in density-weighted ratios of these
response guilds could be used as indicators of climate driven changes in hydrologic conditions
over time.
Monitoring changes in community composition, including shifts from cold- and cool-
water dominated systems to warm-water communities, may be another good indicator. It is
expected that cool-water and warm-water fishes will be able to invade freshwater habitats at
higher latitudes, while cold-water fish will disappear from low latitude limits of their distribution
where summer temperatures already reach fish maximum thermal tolerances (Carpenter et al.,
1992; Tyedmers and Ward, 2001). In east-west drainages fish may not be able to find thermal
refuge and may experience local extinctions (Carpenter et al., 1992). However, cold-water fish
that do persist at higher altitudes and latitudes may not experience as many winter stresses, and
their ranges at may expand with increased duration of optimal temperatures (Carpenter et al.,
1992;Melacketal., 1997).
Salmon species are known to prefer cold water temperatures and a number of studies
have investigated the impact of potential climate changes on these fish species. Pacific salmon
may be particularly sensitive to climatic changes because suitable habitat is projected to decrease
1-9
-------
due to altered thermal regimes (Schindler et al., 2005). Research has linked increased river
temperatures with increased mortality of sockeye salmon, particularly in species which migrate
during the summer when river temperatures are at their highest (British Columbia Ministry of
Water, Land and Air Protection, 2002). Warmer waters cause increased energy use and
bacterial/fungal infections in salmon, decreasing the likelihood that they will survive their
migration and be equipped to spawn (British Columbia Ministry of Water, Land and Air
Protection, 2002). Melack et al. (1997) suggest that higher temperatures will lead to reduced
growth and increased mortality of sockeye salmon in freshwater and marine waters. In
freshwater, Melack et al. (1997) suggest that there will be greater inputs of nutrients during the
winter season rather than in the spring as well as a longer period of thermal stratification, which
would likely lead to lower planktonic productivity and smaller juvenile sockeye salmon.
However, a study in southwestern Alaska by Schindler et al. (2005) shows increased juvenile
growth rates, because the warmer water temperatures increase the length of the growing season
due to earlier ice breakup and increase zooplankton densities, prey for juvenile salmon. In
marine waters, Melack et al. (1997) note that all of the growth and gathering of excess energy
reserves is done during the time that Fraser River sockeye salmon spend in the ocean. However,
general circulation models (GCMs) forecast increases in sea surface temperatures and weaker
north-south pressure gradients over the north-east Pacific Ocean, which could weaken ocean
upwelling and reduce secondary productivity (Melack et al., 1997; IPCC, 2007). The higher
temperatures and reduced zooplankton would likely lead to smaller adult sockeye with fewer and
smaller eggs and less energy reserves (Melack et al., 1997). In addition, the Fraser River
sockeye salmon that Melack et al. (1997) focused on in their analysis already live at the southern
edge of their thermal range. Melack et al. (1997) also reviewed the potential impacts of climate
change on salmon species spawning such that increased winter flows and spring peaks may
reduce salmonid egg to fry survival. For example, higher spring peaks in flow and warmer water
temperatures may cause earlier emergence of fry and migration of pink and chum salmon fry to
estuaries at a time when their food sources have not developed adequately (Melack et al., 1997).
Similarly, low summer flow could lead to a decrease in available spawning and rearing habitat
(Melack et al., 1997). For species that spawn in the fall, including many salmonid species, an
increase in scouring resulting from higher precipitation rates in winter could result in the reduced
survival of eggs (Tyedmers and Ward, 2001).
Some research has shown that fish species living in streams and rivers in semi-arid
regions may be more susceptible to climate impacts than species living in streams and rivers in
sub-humid regions. Milewski (2001) found that species richness, number of insectivorous
cyprinid (minnow) species, and number of species intolerant of degraded water quality and
habitat were lower in the semi-arid region of their study suggesting that fish species rebound
1-10
-------
from low and high water levels more easily in sub-humid regions than in semi-arid regions. Poff
and Allan (1995) also investigated hydrologic variation in streams and the impact of hydrologic
variability on fish species. For the sites in their study, fish assemblages that were associated with
the hydrologically variable streams had the following characteristics: generalized feeding
strategies, association with silt and general substrata, slow velocity, headwater affinity, and
tolerance to silt. Fish species occurring at more than 50% of the hydrologically variable sites but
less than 50% of the stable sites included Ameiurus melas (black bullhead), Percaflavescens
(yellow perch), Notemigonus crysoleucas (golden shiner), Ameiurus natalis (yellow bullhead),
and Lepomis gibbosus (pumkinseed). Fish species occurring only at hydrologically variable sites
(often only one or two sites total) include Fundulus notatus (blackstripe topminnow),
Lepisosteus osseus (longnose gar), Lepisosteusplatostomus (shortnose gar), Amia calva
(bowfm), Anguilla rostrata (American eel), and Dorosoma cepedianum (gizzard shad) (Poff and
Allan, 1995). Fish species occurring at more than 50% of the stable sites and less than 50% of
the hydrologically variable sites include Moxostoma macrolepidotum (shorthead redhorse),
Micropterus dolomieu (smallmouth bass), Hypentelium nigricans (northern hog sucker),
Rhinichthys cataractae (longnose dace), and Notropis rubellus (rosyface shiner).
Cold-water fish species, and salmon species in particular, may be good indicators of
climate-change effects in streams and rivers. To use a salmon species or any fish species as an
indicator, one must be sure not to count or include fish that may have been stocked rather than
occur naturally in a particular stream or river. Native brook trout populations may be a useful
climate-change indicator for streams and rivers for certain regions since they often live at the
edge of their thermal tolerance; therefore a decline in brook trout numbers in a certain area may
be a sign of climate impacts. A decline in brook trout numbers would not always necessarily
indicate climate effects, however, because a decline in this species could also be due to other
stressors or even species competition. Species with widespread ranges and high thermal
tolerance such as largemouth bass, carp, channel catfish, and bluegills would generally not be
good indicators of climate impacts since they are relatively insensitive and their ranges extend
south into Mexico. Another possible effect of increased water temperatures is to reduce DO
levels in stream waters. Darter species are sensitive to benthic oxygen depletion because they
feed and reproduce in benthic habitats (U.S. EPA, 1999), making them another potential
indicator of climate change.
1-11
-------
2. STATE BIOASSESSMENT PROGRAMS—RIVERS AND STREAMS
Aquatic organisms integrate the effects of all sources of stress that impinge on them,
including "conventional" anthropogenic stressors, which are commonly the focus of state
programs assessing and regulating water quality (e.g., point and non-point sources of pollutants,
habitat alterations, landscape-level changes), and any other significant source of environmental
change, including climate change. Because organisms reflect all sources of environmental
disturbance to which they are exposed over time, assessments of biological communities can
provide information that may not be revealed by measurements of concentrations of chemical
pollutants or toxicity tests (U.S. EPA, 1999; Rosenberg and Resh, 1993; Resh and Rosenberg,
1984). Bioassessment thus provides a means of assessing not just biological condition or health
but also overall ecological integrity of stream and river systems.
Their integrative characteristic makes biological assemblages effective monitoring tools,
but it also means that all major sources of stress must be reasonably accounted in order to
reliably attribute observed responses to particular sources of stress and to effectively regulate the
stress and/or manage the resource. The ongoing success of biological monitoring and assessment
programs will require an understanding of what climate-associated changes are occurring in
monitored aquatic communities and how monitoring programs can account for them.
Accounting for climate-change influences will support effective attainment of management goals
using monitoring program results as a foundation.
2.1. BIOASSESSMENTS OF RIVERS AND STREAMS
Since the mid-1980s, the U.S. EPA has worked interactively with national, regional, and
state agency biologists and other nationally recognized experts to develop approaches and
technical guidance for implementation of biological assessment. Resulting guidance included
U.S. EPA's Rapid Bioassessment Protocols (RBPs) (U.S. EPA, 1989), which provided a
technical framework for using benthic macroinvertebrate and fish assemblage data as a direct
indicator of ecological health. These were updated with the additional consideration of
periphyton communities, in 1999 (U.S. EPA, 1999). As a complement to the bioassessment
development, procedures for developing narrative biocriteria were published in 1992 (U.S. EPA,
1992), and for developing biocriteria for streams and rivers in 1996 (U.S. EPA, 1996).
Following this initial focus on streams and rivers, bioassessment technical guidance was
developed for lakes and reservoirs (U.S. EPA, 1998), estuaries, and coastal marine waters
(U.S. EPA, 2000), and wetlands (U.S. EPA, 2002).
2-1
-------
Any well designed monitoring and assessment program (in this case, bioassessment) is
inherently anticipatory in that it will provide information for present needs and those not yet
determined (Yoder and Rankin, 1995). Programs that are adaptable to immediate and future
needs are also cost efficient (Barbour et al., 2000). Regardless of approach, all bioassessment
programs adhere to some basic technical elements: (1) selection and calibration of appropriate
biological indicators, (2) determination of reference condition or benchmarks for assessment, and
(3) use of standardized protocols that maximize the information on the indicators, optimize gear
efficiency, and minimize variability due to sampling error (Barbour et al., 2000).
Biological indicators are considered the best overall measure of ecological integrity from
multiple stressors, because of their continuous exposure to magnitude, frequency, and duration to
the synergistic effects of chemical and non-chemical stressors; therefore, these indicators need to
be calibrated on a regional basis and possess a range of sensitivity to the various stressors,
including climate change. Section 2.2 addresses the more common and relevant components of
bioindicators.
Reference conditions are established in various ways (U.S. EPA, 1996). However, the
use of actual reference sites in a regional population of minimally disturbed sites is ideal for
calibrating a quantitative means of assessing ecological condition. The influence of climate
change will affect the maintenance of stable reference conditions. A gradient of degradation of
reference sites over time is plausible, and it is an important factor for establishing a credible
bioassessment. Bioassessment programs throughout the U.S. have established viable reference
conditions for assessment. Many programs also establish sentinel sites that are assessed during
each monitoring cycle. The continued monitoring of sentinel sites within the reference
population will be important to identify where on the condition gradient a set of reference sites
may be for a state or tribal program.
Standardized protocols are a feature of all bioassessment programs. However, these may
vary among agencies, and they are not necessarily comparable between jurisdictions. As the
effects of climate change upon bioassessment programs are better described, modification of
protocols to capture more sensitive indicators or to collect specific attributes of established
indicators may be necessary.
2.2. BIOINDICATORS USED IN STATE PROGRAMS—RIVERS AND STREAMS
The choice of bioindicators has some commonality throughout the U.S. Benthic
macroinvertebrates are the most common assemblage used for bioassessment in streams and
rivers among the states and tribes (U.S. EPA, 2002). Fish assemblages are the second most
prevalent assemblage used to assess biological condition. The U.S. EPA recommends the use of
multiple assemblages in programs to increase the robustness of the overall bioassessment
2-2
-------
(U.S. EPA 1996). Periphyton or algae is of interest to many states as an added assemblage for
use in their monitoring and assessment program because of their sensitivities to stressors.
Tables 2-1 and 2-2 list the common metrics, which are measures of change in features or
attributes of the structure and/or function of the assemblage due to exposure to stressors, for both
benthic macroinvertebrates and fish. These metrics generally respond to various stressors in
different manners. The sensitivity to climate change is known, in a general sense, for some of
these attributes. Further study is needed to characterize signature responses to climate change
for specific use in bioassessment programs around the country. The aggregation of a series of
metrics into a biological index provides the primary measure of overall attainment of the desired
biological condition. However, certain bioassessment programs (e.g., Maine DEP, Oregon
DEQ) use discriminant or predictive models as primary bioindicators, which may provide a
different dimension of climate-change sensitivity.
2-3
-------
Table 2-1. Table of benthic macroinvertebrate metrics taken from the Rapid
Bioassessment Protocols (U.S. EPA, 1999)
Category
Richness measures
Composition
measures
Tolerance/
intolerance
measures
Feeding measures
Habit measures
Metric
Total number of taxa
Number of EPTtaxa
Number of
Ephemeroptera taxa
Number of Plecoptera
taxa
Number of Trichoptera
taxa
%EPT
% Ephemeroptera
Number of intolerant taxa
% Tolerant organisms
% Dominant taxon
% Filterers
% Grazers and scrapers
Number of clinger taxa
% Clingers
Definition
Measures the overall variety of the
macroinvertebrate assemblage
Number of taxa in the insect orders
Ephemeroptera (mayflies), Plecoptera
(stoneflies), and Trichoptera (caddisflies)
Number of mayfly taxa (usually genus or
species level)
Number of stonefly taxa (usually genus of
species level)
Number of caddisfly taxa (usually genus or
species level)
Percent of the composite of mayfly, stonefly,
and caddisfly larvae
Percent of mayfly nymphs
Taxon richness of those organisms considered
to be sensitive to perturbation
Percent of macrobenthos considered to be
tolerant of various types of perturbation
Measures the dominance of the single most
abundant taxon. Can be calculated as dominant
2, 3, 4, or 5 taxa.
Percent of the macrobenthos that filter FPOM
from either the water column or sediment
Percent of the macrobenthos that scrape or
graze upon periphyton
Number of taxa of insects
Percent of insects having fixed retreats or
adaptations for attachment to surfaces in
flowing water
Predicted
response to
increasing
perturbation
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Increase
Increase
Variable
Decrease
Decrease
Decrease
2-4
-------
Table 2-2. Fish metrics used in various bioassessment programs." Table
adapted from the Rapid Bioassessment Protocols (U.S EPA, 1999)
Category
Richness
measures
Metric
Total number of species
Number of native fish
species
Number of salmonid age
classes'3
Number of darter species
Number of sculpin species
Number of benthic
insectivore species
Number of darter and
sculpin species
Number of darter, sculpin,
and madtom species
Number of salmonid
juveniles (individuals)13
% round-bodied suckers
Number of benthic species
Number of sunfish species
Number of cyprinid
species
Number of water column
species
Number of sunfish and
trout species
Number of salmonid
species
Number of headwater
species
Definition
Measures the overall variety of the fish
assemblage
Those species of fish that are indigenous
Measures the life stage representation of
particular top predators in coldwater systems
Diversity of darters, which are typically in fast
flowing waters and cobble substrate
Normally coldwater bottom feeders
Those species that depend on aquatic insects for
primary food source
Combination of clean-water forms, mostly in
coldwater systems
Combination of key taxa that represent
important structure of fish assemblage in
certain systems
Density of juvenile salmon intended to evaluate
nursery function
Warm-water species of suckers representative
of good quality bottom feeders
Diversity of feeders of all benthic fauna,
including insects and non-insects
Warm-water pelagic species representative of
good water quality and habitat
Diversity of minnows that include a range of
tolerance
Indicative of good quality pools and migration
routes
Combination of species representing good
water and habitat quality
Diversity of salmon in coldwater systems able
to accommodate a variety of top carnivores
Diversity in generally depauperate systems
Predicted
response to
increasing
perturbation
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
2-5
-------
Table 2-2. continued.
Category
Richness
measures
(continued)
Tolerance/
intolerance
measures
Trophic measures
Metric
Number of sucker species
Number of sucker and
catfish species
Number of
intolerant/sensitive
species
Presence of brook trout
% stenothermal cool and
cold water species
% of salmonid individuals
as brook trout
Number of green sunfish
% common carp
% white sucker
% tolerant species
% creek chub
% dace species
% eastern mudminnow
% omnivores
% generalist feeders
% insectivorous cyprinids
% insectivores
% specialized insectivores
% juvenile trout
% insectivorous species
% top carnivores
Definition
Diversity of all suckers — round-bodied and
other
Combination of suckers and catfish in warm-
water systems to be indicative of healthy
systems
Diversity of sensitive fish species; may be
stressor dependent
Indigenous to many areas of the Midwest and
threatened by competition of other species
Narrow temperature tolerance of coldwater taxa
Compositional dominance of brook trout to
other salmonids
Tolerant of warm-water sunfish that becomes
dominant as other taxa decline
Tolerant bottom feeder
Tolerant bottom feeder
Compositional dominance of all tolerant
species
Tolerant minnow species
Tolerant minnow species
Tolerant minnow species
No particular food preference
Generalist feeders, able to deal with a variable
diet
Minnows that prefer aquatic insects as primary
diet
All fish that prefer aquatic insects
Highly specialized in food preference and
easily affected by decrease in food availability
Indicative of food source able to support
nursery function of juvenile trout
Composition of taxa with preference for aquatic
insects
Composition of taxa that prey on other fish and
non-fish higher trophic levels
Predicted
response to
increasing
perturbation
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
Increase
Increase
Increase
Increase
Increase
Increase
Increase
Increase
Increase
Decrease
Decrease
Decrease
Decrease
Decrease
Decrease
2-6
-------
Table 2-2. continued.
Category
Trophic measures
(continued)
Effort measures
Reproduction
measures
Disease measures
Metric
% pioneering species
Number of individuals (or
catch per effort)
Density of individuals
% abundance of dominant
species
Biomass (per m2)
% hybrids
% introduced species
% simple lithophills
% simple lithophills
species
% native wild individuals
% silt-intolerant spawners
% Diseased individuals
(deformities, eroded fins,
lesions, and tumors)
Definition
Those species that occur early in succession of
an ecosystem, and are usually very tolerant
Relative measure of density of fish in
ecosystem related to amount of effort to sample
the fish assemblage
Density regardless of effort
Dominance versus evenness of taxa in fish
assemblage
Relative measure of ability to sustain healthy
fish assemblage through food availability and
good habitat
Measures breakdown of distinct reproductive
guilds usually due to habitat alteration
Intentionally or non-intentionally taxa
introduced into ecosystem and competitive or
predatory upon native taxa
Composition of individual fish that spawn in
clean sand or gravel
Composition of species as lithophills
Measure of relative reproductive success for
native taxa
Need for clean substrate of larger particles than
silt; affected by sedimentation processes
Chronic exposure to stressors resulting in some
form of disease or deformation that may result
in lethal conditions
Predicted
response to
increasing
perturbation
Increase
Decrease
Variable
Increase
Variable
Increase
Increase
Decrease
Decrease
Decrease
Decrease
Increase
a Data from Karr et al. (1986), Leonard and Orth (1986), Moyle et al. (1986), Fausch and Schrader (1987), Hughes
and Gammon (1987), Ohio EPA (1987), Miller et al. (1988), Steedman (1988), Simon (1991), Lyons (1992),
Harbour et al. (1995), Simon and Lyons (1995), Hall et al. (1996), Lyons et al. (1996), Roth et al. (1997).
b Metric suggested by Moyle et al. (1986) or Hughes and Gammon (1987) as a provisional replacement metric in
small western salmonid streams.
Note: X = metrics used in region. Many of these variations are applicable elsewhere.
2-7
-------
3. SENSITIVITY TO CLIMATE CHANGE OF BIOLOGICAL INDICATORS
USED IN STATE BIOCRITERIA PROGRAMS
The sensitivity of common biological indicators to climate change is not very well
known. The review of relevant literature presented in Section 1.2, as well as in this section,
helps establish expectations for the significant modes of effect and probable categories of
responses. However, important details regarding (1) the sensitivity or robustness of specific
metrics or ecological attributes to changing climate parameters over time and among different
regions; (2) the mechanisms by which specific responses will interact with other stressors and
impact interpretation of effects and their causes; and (3) how such responses might combine to
alter biological index responses are recommended as components of needed research (see
Section 7).
To understand probable climate-change effects on stream/river biological indicators, the
linkage between climate and stream/river ecology must be defined. Anthropogenic increases in
greenhouse gases directly affect air temperature and precipitation (considered primary climate
drivers). Climate-change projections for the year 2100 include global average air temperature
increases of 1.1-2.9°C (for the lowest emissions scenario) to 2.4-6.4°C (for the highest
emissions scenario) (IPCC, 2007). Increases in precipitation are predicted for many regions,
with a higher percentage of total precipitation occurring in more frequent and intense storms.
Other predictions include more precipitation in winter and less precipitation in summer; more
winter precipitation as rain instead of snow; earlier snow-melt; earlier ice-off in rivers and lakes;
and longer periods of low flow and more frequent droughts in summer (Hayhoe et al., 2007;
IPCC, 2007; Barnett et al., 2005; Fisher et al., 1997).
Changes in these primary climate drivers will affect stream/river water and aquatic-life
resources mainly through direct and indirect alterations in hydrologic and thermal regimes.
Changes in hydrologic regime (including magnitude, timing, duration and frequency of runoff
events) will vary regionally (NAST, 2001), but they are expected to include changes in the
magnitude of flow ranging from increases of 10-40% in the northeastern U.S. to decreases in
annual flow of 10-30% in the South, Midwest, and West (Hayhoe et al., 2007; Milly et al., 2005;
Magnuson et al., 1997). Changes in patterns of flow will likely include increases in stream flow
occurring mainly in the winter and spring, lower stream flow in the summer and fall, and greater
variability and "flashiness" of stream flows (Hayhoe et al., 2007; Moore et al., 1997). These
projected alterations in stream flow dynamics are critical in structuring aquatic ecosystems
through influence on sediment supply and transport, habitat stability, channel formation and
maintenance, and water volume, which, in part, controls habitat availability and water quality
(Poff et al., 2002; Richter et al., 1996; Poff et al., 1996; Poff and Allan, 1995). Seasonal patterns
3-1
-------
of flow and other flow dynamics strongly influence the types of species that can inhabit an area,
defining the composition, structure, and functioning of aquatic assemblages (Poff et al., 2002;
Richter et al., 1996; Poff and Allan, 1995). As a result, climate-associated changes in stream-
flow magnitude are expected to modify habitat, species composition and abundance, and
ecological interactions over time.
Stream/river water temperature regimes will be altered by air temperature increases and
modified by other influences including variations in flow volume and snow melt, groundwater
influence, riparian shading, presence of deep pools, meteorology, river conditions, and
geographic setting (Cassie et al., 2006; Mohseni et al., 2003; Daufresne et al., 2003;
Hawkins et al., 1997). Thermal regime influences the distribution and abundance of aquatic
species in relation to temperature tolerances and evolutionary adaptations combined with
competitive interactions, effects on food supply, and other factors; it also drives timing of life
cycle events (phenology), biological productivity, and species interactions (e.g., Matthews, 1998;
Hawkins et al., 1997; Vannote and Sweeney, 1980; Sweeney and Vannote, 1978).
As discussed in Section 2.2, common metrics monitored as biological indicators in
existing bioassessment programs are measures of change in features or attributes of the structure
and/or function of the macroinvertebrate or fish assemblages; Tables 2-1 and 2-2 summarize
many of the widely applied categories of biological indicators. Additional research will provide
information on specific sensitivities of individual biological indicators to climate change.
However, taken by category of metric, expectations for probable responses of various biological
indicators to climate change can be summarized from literature information and projections of
future climate changes (see Sections 1.2, 1.3, and this section). Table 3-1 summarizes expected
climate change responses by category (Note: this summary of potential responses represents
examples and is not considered comprehensive. Categorization as sensitive or tolerant refers
generally to anticipated climate-change sensitivity, in particular to temperature and hydrologic
changes).
It is clear that many of the types of responses that can be expected for common categories
of biological indicators in response to climate change can be similar to changes caused by other
conventional stressors. For biological indicators that are sensitive to both conventional stressors
and climate change, the confounding interactions of climate change and other stressor effects
will influence the process of attributing cause to particular stressors. This interaction will require
the development of an approach to partition observed responses between climate change and
other stressors, so that the ability to manage resources and regulate water quality through the
process of monitoring and assessing biological indicator data remains viable.
Conceptually, this approach can include adaptations of monitoring methods in order to
account for climate change. Preliminary aspects of this component are discussed in Section 4
3-2
-------
Table 3-1. Summary of expectations for responses of common categories of
stream and river biological indicators to climate change influences on water
temperature and hydrologic regime
Category
Expected climate change effects/sensitivities
References
Macroinvertebrates
Richness and
abundance
measures
Overall richness generally expected to decline due to
temperature sensitivity and hydrologic stresses including
increased flashiness, increased instances of summer low
flows, drought, etc. However, replacements overtime
with tolerant forms may ameliorate this in some
situations. Abundance or eurytolerant species may
increase in some habitats.
Durance and Ormerod,
2007; Bradley and
Ormerod, 2001
Community
composition,
persistence
measures
Compositional changes resulting from reductions in
temperature and/or flow sensitive taxa (examples
potentially include Chloroperla, Protoneumura,
Neumura, Rhyacophila munda, Agabus spp,,
Hydrophilidae, and Drusus annulatus) and increases in
less temperature and/or flow sensitive taxa (examples
potentially include Athricops, Potamopyrgus,
Lepidostoma, Baetis niger, Tabanidae, Hydropsyche
instabilis, Helodes marginata, Caenis spp.), and/or from
shifts in range ; patterns of persistence or community
similarity that track climatic patterns; changes may also
occur in functional roles of species.
Daufresne et al., 2003;
Durance and Ormerod,
2007; Bradley and
Ormerod, 2001;
Burgmer et al., 2007;
Golladay et al., 2004;
Parmesan, 2006;
Hawkins et al., 1997
Tolerance/
intolerance
measures
Climate-change sensitivities related to temperature or
flow regime may be documented as decreases
(potentially resulting from local extinctions and/or range
shifts) in richness (number of taxa) of temperature or
flow-regime sensitive groups (see "Composition
Measure" for examples). Dominance by tolerant taxa
also may increase.
Daufresne et al., 2003;
Durance and Ormerod,
2007; Burgmer et al.,
2007; Golladay et al.,
2004; Parmesan, 2006
Feeding measures
Variable responses expected, driven by interactions
between temperature, which may increase phytoplankton
and periphyton productivity and thus increase associated
feeding type; hydrologic factors which may decrease
periphyton if habitat stability is decreased or
sedimentation is increased; CO2 concentrations, which
can directly affect leaf litter composition and
decomposition; and changes in riparian vegetation.
Gamer and Robinson,
2007; Dodds and
Welch, 2000; Tuchman
et al., 2002
Habitat measures
Number and percent composition of clingers likely to
decrease if hydrologic changes decrease habitat stability,
increase embeddedness, or decrease riparian inputs of
woody vegetation.
Johnson et al., 2003;
Townsend et al., 1997
3-3
-------
Table 3-1. continued.
Category
Expected climate change effects/sensitivities
References
Fish
Richness and
abundance
measures
May have initial increase in diversity as more warm-
water assemblages replace cool- or cold-water species.
Habitat availability is expected to be diminished by
altered flow regimes with an associated loss of diversity.
If barriers to dispersal limit community replacements,
richness also may decline. May also, for example, lose
spring spawners (e.g., some salmon species) due to
changes in timing of spring flows. Abundance of warm-
water species may increase, while coldwater species may
decrease.
Xenopoulos and Lodge,
2006; Xenopoulos et
al., 2005;Poffetal.,
2002; Grimm et al.,
1997; Hayhoe et al.,
2007; Wehrly et al.,
2003
Composition
measures
Expect fish community compositional changes resulting
from losses of cold- and/or cool-water fishes (e.g., brook
trout, dace and bleak), and increases in warm-water
fishes (e.g., chub and barbell).
Daufresne et al., 2003;
Mohseni et al., 2003;
Schindler, 2001;
Covichetal, 1997;
Moore et al., 1997;
Raheletal., 1996,
Eaton and Scheller,
1996
Tolerance/
intolerance
measures
Loss of temperature-sensitive cold- and cool-water
species will decrease intolerant measures, increase
tolerant measures.
Mohseni etal., 2003;
Moore et al., 1997;
Raheletal., 1996;
Eaton and Scheller,
1996
Feeding measures
Shift in food sources through attrition of lower trophic
levels will affect higher trophic levels, including top
carnivores.
Schindler etal., 2005;
Melacketal., 1997
Habitat measures
Breakdown of habitat features and connectivity fosters
hybridization and drift in species gene pool.
Matsubaraetal., 2001;
Heggenes and Roed,
2006
3-4
-------
(below), to the extent that they were addressed in this preliminary case study. Another aspect of
program adaptation is restructuring the analytical approach used to evaluate biological
monitoring data, detect impairment, and assess cause (see preliminary case study results
discussed in Section 5). These monitoring components—sampling strategy and analytical
approach—are clearly inter-related, and implicitly include components such as tracking changes
at reference locations through both altered sampling design and appropriate analyses.
Another component that is being considered for its potential contribution to tracking and
differentiating climate-change effects from other stressors is the categorization of biological
indicators based on differing sensitivities to these effects. These indicators include both
community metrics and population measures of individual sensitive species (see Table 3-1). In
concept, there would be analytical and interpretive advantages if at least some biological
indicators could be identified that are especially sensitive to particular conventional stressors but
insensitive to climate-change effects. Conversely, community metrics and individual taxa that
are specifically sensitive to climate change would be valuable in identifying and defining trends
at reference sites. These could be applied analytically to separate monitored biological responses
into components related to long-term, climate-change effects and other stressors. Such
separation is the major goal of efforts to adapt bioassessment programs to account for climate
change.
In practice, evidence gathered from the literature and the professional opinions of many
state/tribal bioassessment managers1 suggests that few, if any, biological indicators currently
used in bioassessment programs are likely to be insensitive to climate-change effects. This is
largely because climate change affects aquatic communities through the critical ecological
drivers of flow dynamics (hydrology) and water temperature. Thus, the modes of action of
climate-change effects and effects of other stressors are similar in many cases, and taxa that are
sensitive to conventional stressors are likely to be sensitive to climate change as well. Taxa
identified in the Workshop as being "potentially insensitive to climate change" were mainly
those species already characterized as being broadly tolerant, "weedy," and/or generalist species.
Beyond categorization of existing biological indicators as sensitive/insensitive to climate-
change effects, there are biological metrics that could be considered for incorporation into
bioassessment programs that are not currently measured on a routine basis in most existing
programs. Such "novel" indicators are considered specifically because of their sensitivity to
climate-change effects—most have been predicted or observed in the literature as biological
responses to directional climate change, especially increases in water temperature. Table 3-2
summarizes examples of such "novel" biological indicators.
1 Hamilton, A; Barbour, M; Gerritsen, J; et al. (2007) Introductory workshop on climate change effects on biological
indicators: workshop summary report.
3-5
-------
Table 3-2. Novel indicators that may be sensitive to climate change
Category
Phenology
Longer
growing
season
Life stage -
specific
Temperature
sensitivity
Metric
Early emergence of mayfly
species (also stonefly and
caddis species)
Early trout spawning in
warmer water
Accelerated development and
earlier breeding of the
amphipod Hyallela azteca
Increased algal productivity
Additional reproductive
periods of amphipod species
Altered sex ratios for certain
insects (e.g., trichopteran
Lepidostoma)
Smaller size at maturity and
reduced fecundity of
plecopteran Nenoura
trispinosa and amphipod
Hyallela azteca
Decreased salmon egg to fry
survival
Reduced size of sockeye
salmon
Increased growth rate of
juvenile salmon in Alaska
Decreased growth rate of trout
Comments
Indirect effects on timing of
salmonid feeding regime
In northern areas a response
to decreased ice cover and
increased light penetration
From increased temperature
Increased turbidity from
eroded sediment due to
increased precipitation
Reduced growth and
increased mortality in higher
temperatures as well as to
lower plankton productivity
References
Harper and
Peckarsky, 2006;
Briers et al., 2004;
Gregory et al., 2000;
McKee and
Atkinson, 2000
Cooney et al., 2005
Hogg et al., 1995
Flanagan et al., 2003
Hogg et al., 1995
Hogg and Williams,
1996
Turner and Williams,
2005; Hogg etal.,
1995
Melack et al., 1997
Melack et al., 1997
Schindler et al., 2005
Jensen et al., 2000
3-6
-------
Table 3-2. continued.
Category
Hydrologic
sensitivity
Metric
Decreased survival of eggs of
autumn-spawning salmon
(e.g., dolly varden, brook
trout, coho salmon)
Decreased fry survival of pink
and chum salmon due to
earlier (late winter to early
spring) peak flows
Differential mortality of
drought-intolerant mussel
species (e.g., Lampsilis
straminea claibornensis,
Villosa villosa, Lampsilis
subangulata)
Comments
Results in decreased
abundance of autumn-
spawning species, and/or
change in relative
composition between spring
and autumn spawners
Earlier emergence and
migration of pink and chum
salmon fry to estuaries at a
time when their food sources
have not developed
adequately
Results in changes in relative
abundance, extirpation of
vulnerable species
References
Gibson et al., 2005
Melack et al, 1997
Golladay et al., 2004
3-7
-------
One consideration that must be taken into account in the ongoing evaluation of potential
novel indicators and their role in adaptation of bioassessment programs is that many of these
metrics are more difficult or time- and resource-consuming to measure, especially on a routine
basis. Some of them also require sampling techniques and timing or frequency of sampling that
are quite different from the commonly applied bioassessment approaches. For example, the
process of measuring sizes of all individuals of one or more species of mayflies, stoneflies, or
caddisflies (representative EPT taxa) to establish size-class composition and evaluate reduction
in size of the last instars (i.e., the last nymphal stage just before emergence) and how this
changes over time to define climate-change effects; or similarly, the sampling of emerging adult
insects that would be needed to evaluate earlier emergence, are not commonly done. Another
consideration for future evaluation of novel indicators is their potential sensitivity to other
(conventional) stressors, in addition to their responsiveness to climate change. This will affect
how they might be incorporated into a monitoring design and analysis approach.
Having given a summary of climate-change effects, an overview of state bioassessment
and biocriteria programs, and a framework for considering the sensitivities of established and
novel biological indicators to climate change, preliminary consideration can be given to aspects
of possible vulnerabilities of biological assessment programs to climate change. This report first
addresses the sampling power needed to detect climate-change effects using current indicators.
Secondly, the report describes how climate change may affect reference and non-reference sites
differently. The report includes case studies using data from one program to illustrate these
preliminary results obtained from implementation of the proposed framework.
3-8
-------
4. CASE STUDY 1—ASSESSING TRENDS: THE POWER OF BIOLOGICAL
ASSESSMENTS TO DETECT CLIMATE CHANGE
The ability to account for climate change requires an understanding of how vulnerable
monitoring data are to climate change effects and how effectively differences that are a result of
climate change can be detected within existing monitoring programs. This case study describes
(1) how much sampling would be needed to distinguish expected levels of climate change
effects, and (2) how long it would take to detect climate change effects with a specified
probability of detection, given a particular monitoring framework. The information summarized
in this section highlights the approach, the key results, and the main conclusions of Case Study 1.
4.1. OBJECTIVES
The main objective of this case study is to evaluate one aspect of the vulnerability of
biological monitoring and biocriteria programs to climate change with respect to the effects on
ecological communities. This case study focuses on the ability of a typical bioassessment
program to detect expected climate change effects on one selected community component, taxon
richness. The focus is on two questions:
• How long must monitoring be conducted to have a fixed probability of detecting a change
in the mean native taxon richness of the reference site population?
• How long must monitoring be conducted to have a fixed probability of detecting a change
in mean native taxon richness for a particular site?
The first question is important because most states use reference populations as the basis
for constructing indices and deriving biocriteria. The second is important because many
individual sites are tracked for specific regulatory reasons (permitting, restoration, etc.).
4.2. ANALYSIS APPROACH
The questions in this study are approached by evaluating the ability, or power, of a
typical biological monitoring program to detect expected levels of change in a particular
biological attribute—in this case species richness. Statistical power, the ability to detect a real
effect, is a critical issue in designing monitoring programs and is expressed as a probability. The
more power a test has, the more likely one is to correctly infer that a real change has actually
occurred. In this case study, the power analysis approach is used to evaluate how much of a
change in taxon richness (the effect size or minimal detectable difference) can be detected in a
typical biomonitoring program. This detectable difference can then be compared to expected
4-1
-------
taxa loss from climate change to see how long (in years) the program must monitor to ensure
detection (at a set probability level) of a taxa loss signal resulting only from climate change.
4.2.1. Ability to Detect Change-Power Analysis
Power is defined as the probability of rejecting a false null hypothesis and is related to
type II error (|3), which is the probability of accepting a null hypothesis (no change) when it is
false (there is change). Thus, power = 1-|3. Power analysis requires several critical components:
• sample size (N)
• variability in an observed factor (taxon richness in this case) (a)
• effect size (how much of a change one wants to detect) (5)
• significance level (type I error, a)
For this application of power analysis, the desired level of power (1 -13) must be fixed as well.
This case study demonstrates how changing some of these components can increase or decrease
the ability to detect a climate change effect.
The equations for calculating effect size for comparing two paired population means can
be found in many statistical textbooks. The basic formula, assuming normally distributed
populations, is:
where Za and Zp are the z-scores (probability levels) for the desired type I (a) and type II (|3)
error rates.2
For this case study, variance (a2) is estimated using existing monitoring data for sites
sampled repeatedly over several years during an index period. Such data give an estimate of the
natural variability in biological condition through time, assuming minimal external changes.
Knowledge about what taxa changes (effect size) might be expected in response to climate
change is also needed. This value was derived from existing literature on taxa loss in relation to
temperature changes (see Section 4.2.3).
For each of the two assessment questions (see Section 4.1), different confidence levels
were investigated (i.e., a and |3 were varied). The effects of sample size were also investigated
2 Source: Snedecor and Cochran (1980).
4-2
-------
for both questions. To address the first question of detecting change in a reference population,
the number of sampling sites, N, was either fixed per year (at N = 40 for the reference
population) or cumulative (N increasing by 40 each year for the reference population) and
assumed a comparison of year one versus a cumulative pooled sample size through time. For the
second question (detecting a change at a particular site), N was either fixed at one of three
different levels (N = 5, 10, and 20) or N was cumulative, increasing by 5, 10, or 20 for the
respective analysis runs.
Variance estimates were derived from one particular state bioassessment data set (see
Section 4.2.2 below) and were then used for each region evaluated. Z scores were varied by
changing the type I and type II error rates to illustrate the effects of these choices on the ability to
detect a climate change effect. The outputs from these analyses are time series of taxa loss rates
predicted from climate change effects. These outputs are compared with minimal detectable
effect sizes to illustrate the length of time required to detect a climate change effect on taxon
richness under various conditions (taxa loss rates, temperature scenarios, and error rates).
4.2.2. The Maryland Biological Stream Survey Data Set
For this case study, variance (reflecting natural variability in biological condition over
time) is estimated using existing monitoring data from the Maryland Biological Stream Survey
(MBSS) (Boward et al., 1999; URL: http://www.dnr.state.md.us/streams/mbss/). The MBSS
program includes statewide monitoring of all watersheds using a multi-stage probability based
design with a 5-year rotating basin sampling approach. It also includes repeated annual sampling
at a series of 28 fixed reference ("sentinel") sites using the same sampling methods. It is this
repeat sampling of benthic macroinvertabrates and fish at sentinel sites that was used to estimate
population variance (a2) in taxon richness. While this case study compares expected effects
across various regions of the U.S., the MBSS derived variance was used for all regions as the
estimate of variance associated with biomonitoring.
Relative variability of taxon richness is assumed to be constant over time. However, this
is likely not true as both the mean and variation in biological condition of sites may change with
warming water temperatures. For simplicity, it is assumed that the variance in taxon richness
associated with the MBSS program can be extrapolated in time and across different regions.
MBSS data (Boward et al., 1999; URL: http://www.dnr.state.md.us/streams/mbss/) are
available with uniform collection methods for the period 1994-2004, with sentinel site data
available from 1998-2004. The 6-year period for which sentinel site data are available
fortuitously includes both a dry and a wet climate cycle. Sampling for the MBSS is conducted
during index periods—the spring reproduction/recruitment period for benthic macroinvertebrates
(March-early May), and the summer-fall low-flow period for fish (June-September). A wide
4-3
-------
range of physical, chemical, and habitat variables are measured and/or calculated in association
with the biological collections. These include water chemistry variables (e.g., temperature, pH,
DO concentration, various nutrient concentrations, conductivity); numerous physical habitat
variables (e.g., the Maryland physical habitat index (PHI), instream habitat condition, epifaunal
substrate, water velocity, water depth, embeddedness, shading, distance to road, riffle quality,
etc.); and land use characteristics (17 detailed categories aggregated to the larger categories of
agriculture, urban, water, wetland, barren, and forest).
Benthic macroinvertebrates are collected using D-nets, employing a multi-habitat
approach over a 75-m reach. A 100-organism subsample is processed for each sample.
Identifications are made to the genus level. Fish are collected using quantitative, double-pass
electrofishing in 75-meter stream segments, with a blocking net at the end of the segment. Fish
are identified to the species level (Boward et al., 1999).
4.2.3. Information on Taxa Loss Rates
This study focuses on climate change effects associated with temperature because (1) the
goal is to demonstrate a process for calculating the capacity of a program to detect change and
not to predict all the effects of climate change and (2) while few data exist on climate change
effects on aquatic assemblages in general, there are more data on temperature effects than
hydrologic effects.
Predicted macroinvertebrate taxa loss rates due to temperature increases were derived
from the literature based on observed changes in taxon richness associated with temperature
increases. For this study, native or expected taxon richness is considered rather than total
richness; species replacement is not considered. Total richness may not change if species
replacement rates are high. For example, it is possible that stenothermal species (those with a
narrow range of temperature tolerance) that are lost will be replaced over time by eurythermal
taxa (those with wide temperature tolerances). However, native taxa are expected to be lost from
many streams (e.g., Parmesan, 2006; Xenopolous et al., 2005; Moore et al., 1997), and native
taxon richness based on current climate will decrease.
Other ecological responses are expected but are not considered in this study: density
changes, range shifts, timing changes in important life history stages and phenology,
morphological changes, physiological changes, and behavioral changes, and gene frequencies
changes (Parmesan, 2006; Root et al., 2003; Walther et al., 2002; Hogg et al., 1998; Schindler,
1997). Taxon richness, a very common component metric evaluated in bioassessment programs
and incorporated in multimetric indices, is evaluated for signs of bioassessment program
vulnerability.
4-4
-------
Macroinvertebrate and fish taxa (i.e., genus or species, reflecting practical taxonomic
limitations) loss rates were obtained from literature reporting on climate-change associated
temperature effects on taxon richness (Daufresne et al., 2003), and on thermal discharge effects
(Gammon, 1973; Lehigh University, 1960) using linear projections (estimating average species
loss per unit of temperature change). Given the complexity of biotic interactions as well as
temperature effects, this is probably not an accurate representation of reality. In reality, taxa loss
rates are likely to occur episodically over time, for instance when particular species thresholds
are reached. For example, loss of keystone taxa may precipitate abrupt and dramatic changes on
stream communities as well as on stream processes (Flecker, 1996; Pringle et al., 1993; Power,
1990; Power et al., 1985). In addition, the amount of decrease and range of variation around
expected decreases may also be affected by differences in the size of the species pool, e.g.,
species-rich sites could lose more taxa, and/or show greater variation in loss rates. On the other
hand, small losses from locations with a naturally poor fauna may be ecologically more
significant. However, the simplifying assumption allows us to model changes into the future.
Based on the limited information currently available in the literature, the high and low native
taxa loss rates used are simply to bracket a range to estimate detection ability. The ecological
implications and relative importance of being able to detect changes in taxa will still vary
between sites.
Daufresne et al. (2003) observed a loss of 7 macroinvertebrate taxa in streams associated
with a 1.5°C increase over the period 1980-1999. This equals a loss rate of roughly 4.6 taxa per
°C and is being considered the high taxa loss rate. A second estimate was derived from the
literature associated with thermal discharge studies associated with the CWA 316 program.
Most of the 316(a) studies focused on fish effects, and, out of those studies, many were
physiological. One study (Lehigh University, 1960) included macroinvertebrate effects and this
study found a loss rate of approximately 1 taxon per °C over the range 22-28°C; this is the low
taxa loss rate. This represents a fairly high thermal range, but these studies were designed to
investigate effects of thermal effluent, not effects of climate change. Nevertheless, the results
are considered applicable. There are likely more 316(a) studies with invertebrate data, but these
individual studies, for the most part, are not published in standard scientific citation databases
and can be hard to locate (but see the Energy Citations Database
rhttp ://www. osti. gov/energycitations/index.j spl).
Predicted fish taxa loss rates were also considered from the thermal discharge literature.
A study of thermal effluent on the Wabash River found a loss rate of 3.6 fish taxa per °C increase
in temperature (Gammon, 1973). This may be on the high side for loss rates, but it was one of
the few data-based values found within a temperature range comparable to climate change
projections.
4-5
-------
4.2.4. Prediction of Expected Taxa Losses with Projected Temperature Increases
Estimates of taxa loss rates were coupled with projected temperature increases to model
the expected rate of taxa loss per year due to climate change. Projected temperature increases
due to climate change for each region of the U.S. were taken from the National Assessment
Synthesis Team (NAST) summary report (2001). NAST (2001) relied mainly on results from
two coupled atmosphere/ocean general circulation models (AO-GCMs) that were used to
estimate projected temperature increases for various regions of the U.S. (Table 4-1). Predicted
temperature increases by the year 2100 ranged between 2.3°C and 6.5°C for the Hadley and
Canadian models, respectively. Although the biological data are from the Mid-Atlantic region,
we also investigated how projected climate changes in other regions affected taxa loss rates.
Table 4-1. Average annual temperature increases expected by region of the
U.S. (NAST, 2001)
Region
Northeast/Mi d- Atl antic
Southeast
Midwest
Great Plains
West
Pacific Northwest
Average Annual Temperature (°C)
Increases by 2100
Min
2.6 (Hadley)
2.3 (Hadley)
3 (Hadley)
3 (Hadley)
4 (Hadley)
2.7 (by 2050) (Hadley)
Max
5 (Canadian)
5.5 (Canadian)
6 (Canadian)
6.5 (Canadian)
5.5 (Canadian)
3. 2 (by 2050) (Canadian)
Reported rates of temperature increase were linked with estimated rates of taxa losses to
model taxa losses per year due to climate change, incorporating both the low and high estimates
of each.
Linear projections of climate change effects are used in this case study as a basis for
estimating ability to detect climate-induced changes after various monitoring periods. It is likely
that climate will change in a non-linear fashion with periods of fast change followed by periods
of slower changes (IPCC, 2007). There is little way to predict this course, however, so the linear
assumption is the more conservative approach and is a common assumption used in the literature
(e.g., Najjaretal., 2000).
4-6
-------
4.3. KEY FINDINGS
4.3.1. How Long Must Monitoring be Conducted to have a Fixed Probability of Detecting
a Change in the Mean Native Taxon Richness of the Reference Site Population?
If a population of reference sites (N = 40) is sampled each year, an average
macroinvertebrate taxon richness in reference streams can be calculated. For comparing any two
samples of N = 40 sites, there is a fixed difference in mean taxon richness (effect size) at which
significance can be detected with a specified power. For a = P = 0.05 (95% confidence, 95%
power), and the Maryland data, the effect size is 4.5 taxa. Thus, to have a 95% probability of
detecting a significant (p < 0.05) taxa loss between 2 samples of 40 sites, requires a mean
difference of 4.5 taxa. At high taxa loss rates and under the higher estimate for warming in the
Northeast/Mid-Atlantic region, it will take 15 years to achieve a mean loss of 4.5 taxa
(Figure 4-1), assuming that (1) the same 40 sites are sampled each year; (2) samples from a site
are not treated as cumulative through time; and (3) the analysis uses type I and type II error rates
of 0.05. This value is derived by identifying the point where the effect size line (hatched)
crosses the taxa loss rate line (solid) (Figure 4-1).
Figure 4-1 illustrates a variety of scenarios representing different confidence levels and
either fixed or cumulative sample sizes. For example, relaxing the confidence level decreases
the time to detect a change. Increasing a and P from 0.05 to 0.20 (reducing both confidence and
power), reduces the time to achieve 80% probability of detecting a significant climate change
effect (p < 0.2) to approximately 8 years. If a 1 in 5 (rather than a 1 in 20) chance that
statistically significant results are due to random chance alone is acceptable, a taxa change
attributable to climate change could be detected in half the time. This is the type of trade-off that
is important for programs to consider.
Similarly, if samples taken across the reference population are treated as cumulative
estimates of the population condition (replicates of the reference condition), then the projected
climate change effect can be detected very quickly under the conditions of a high taxa loss rate
and high temperature increase. If this replication is temporal, i.e., if samples from consecutive
years are grouped, this also would increase the ability to detect the climate change effect.
However, there is an associated assumption that interannual variation is constant (i.e., that
successive years are comparable and can be grouped for analysis). This may be true on short
time scales, but it might be faulty over longer periods given climate change, which is
progressive. Combining samples into decadal (or shorter) groups (N = 400) may be more
defensible and would also result in detecting the climate change effect more quickly.
Using the same assumptions but altering the taxa loss rate to the lower taxa loss rate
(1 taxon per °C), the time to detect a climate change effect increases dramatically (Figure 4-2).
This is reasonable, given that subtle effects will be much harder to detect than strong effects.
4-7
-------
(0
X
(0
T3
|
CD
C
(0
CD
N
to
"o
UJ
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
High Taxa Loss Rate (4.6 per degree C)
1 Predicted Taxa Loss
(alpha = beta = 0.95) Fixed N
(alpha = beta = 0.95) Cumulative N
1 (alpha = beta = 0.90) Fixed N
(alpha = beta = 0.90) Cumulative N
1 (alpha = beta = 0.80) Fixed N
(alpha = beta = 0.80) Cumulative N
e.g., at fixed N=40 and a and b = 0.95
it will take 15 years to significantly
detect loss of taxa due to climate
change
0
20 40 60
Years
80
100
Figure 4-1. Effects of confidence level (a [a] and p [b]) on time to detect a
climate effect on macroinvertebrate taxa loss due to climatic warming at high
taxa loss rates in the Northeast/Mid-Atlantic U.S. Sample size (N) is either
fixed at 40 per year or is cumulative. This analysis was based on a high
estimate of global warming (5°C by 2100).
4-8
-------
(/>
o
CD
X
CD
•o
"o
"§
ol
•o
CD
CD
N
CO
"o
i i I
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
Low Taxa Loss Rate (1.0 per degree C)
Predicted Taxa Loss
(alpha = beta = 0.95) Fixed N
(alpha = beta = 0.95) Cumulative N
(alpha = beta = 0.90) Fixed N
(alpha = beta = 0.90) Cumulative N
(alpha = beta = 0.80) Fixed N
(alpha = beta = 0.80) Cumulative N
0
20 40 60
Years
80
100
Figure 4-2. Effects of confidence level (a [a] and p [b]) on time to detect a
climate effect on macroinvertebrate taxa loss due to climatic warming at low
taxa loss rates in the Northeast/Mid-Atlantic U.S. Sample size (N) is either
fixed at 40 per year or is cumulative. This analysis is based on a high
estimate of global warming (5°C by 2100).
4-9
-------
Uncertainty about the effects of warming water temperatures on taxa loss also influences the
ability to detect a climate change signal. The effects of confidence level and sample size are the
same under the lower taxa loss rates as under the high taxa loss rates. For example, under the
same fixed sample size (N = 40) and confidence level (95%), it would take approximately 70
years to detect a climate effect under the low taxa loss rate as opposed to 15 years under the high
taxa loss rate.
Lastly, there was only one defensible rate for fish taxa loss (3.6 per °C). This fish taxa
loss rate was used in the same models as those for the macroinvertebrate loss rate and indicates
similar effects of sample size and confidence level on time to detect a change. Under these
assumptions, it would take 10 to 20 years to achieve a fixed probability of detecting the loss of
fish taxa due to climate change in the reference site population using confidence levels of 0.80
and 0.95, respectively (see Figure 4-3).
Similar analyses were run for the lower temperature increase scenario for this region
(Table 4-1). Not surprisingly, if temperatures warm more slowly, they will have less of an effect
on taxa loss and it will take comparatively longer to detect a loss in average taxon richness in the
reference population (Table 4-2).
4.3.2. How Long Must Monitoring be Conducted to have a Fixed Probability of Detecting
a Change in the Mean Native Taxon Richness for a Particular Site?
This second question focuses on the ability to detect these same effects at a single site,
which could be a reach of stream or a watershed. In either case, the assumption is that replicate
samples are apportioned probabilistically across the site. The analysis specifically defines the
effect of increasing sample size.
For this question, three different sample sizes were investigated (N =5, 10, or 20) and
were treated as either fixed (non-additive over time) or cumulative. Only results for the
Northeast/Mid-Atlantic region under the maximum predicted temperature increase are shown,
although results are similar across regions, shifting only due to differences in the projected
temperature increases.
Whether for a watershed or a specific reach, increasing the sample size will shorten the
time required to detect an effect of climate change on taxon richness (see Figure 4-4). Many
biomonitoring programs may collect only one sample at a site per year; a means comparison
could be applied in this framework (use N = 1 in equation), but the differences would have to be
quite large to be significant, and this is not likely over the short term (e.g., between consecutive
years). Samples could be combined cumulatively over consecutive years to support testing, but
the same problem exists in combining consecutive years over a long time period for analysis: the
communities being sampled are probably changing over time due to climate change.
4-10
-------
5.00
S3 4.50
o
TO 4.00
X
H 3.50
Q)
"o
3.00
2.50
2.00
TO
CD 1.50
N
^ 1.00
o
' (alpha = beta = 0.95) Cumulative N
• • (alpha = beta = 0.90) Fixed N
• • (alpha = beta = 0.90) Cumulative N
— ' (alpha = beta = 0.80) Fixed N
~ ' (alpha = beta = 0.80) Cumulative N
0
20
40 60
Years
80
100
Figure 4-3. Effects of confidence level (a [a] and p [b]) on time to detect a
climate effect on fish taxa loss due to climatic warming in the Northeast/Mid-
Atlantic U.S. Sample size (N) is either fixed at 40 per year or is cumulative.
This analysis is based on a high estimate of global warming (5°C by 2100).
4-11
-------
Table 4-2. The time (years) to achieve a fixed probability of detecting a
statistically significant effect of temperature increases on macroinvertebrate
and fish taxa loss across different regions under maximum and minimum
temperature projections. These data are for question 1 and assume a fixed
sample size of N = 40 reference sites sampled each year. Data are shown for
different taxa loss rates and for different confidence levels.
Regions
Northeast/
Mid-Atlantic
Southeast
Midwest
Great
Plains
West
Pacific
Northwest
Maximum predicted temperature increase by 2100
Macroinvertebrates — high taxa loss rate
(4.6 per °C)
a=P = 0.95
a=P = 0.8
15
8
14
7
13
7
—
—
13
7
—
—
Macroinvertebrates — low taxa loss rate
(lper°C)
a=P = 0.95
a=P = 0.8
70
36
64
33
58
30
—
—
57
29
—
—
Fish taxa loss rate
(3.6per°C)
a=P = 0.95
a=P = 0.8
20
10
18
9
17
9
—
—
16
9
—
—
Minimum predicted temperature increase by 2100
Macroinvertebrates — high taxa loss rate
(4.6 per °C)
a=P = 0.95
a=P = 0.8
29
15
33
17
38
19
19
10
17
9
14
7
Macroinvertebrates — low taxa loss rate
(lper°C)
a=P = 0.95
a=P = 0.8
>100
69
>100
78
>100
89
88
45
79
41
64
33
Fish taxa loss rate (3.6 per °C)
a=P = 0.95
a=P = 0.8
38
19
42
22
49
25
25
13
22
12
18
9
4-12
-------
18.00
16.00
g 14.00
CD
c
CD
CD
N
C/)
"(J
s£
LU
12.00
£ 10.00
8.00
6.00
4.00
2.00
0.00
High Taxa Loss Rate (4.6 per degree C)
•Predicted Taxa Loss
•(alpha=0.95) Fixed N=5
(alpha=0.95) Fixed N=10
•(alpha=0.95) Fixed N=20
(alpha=0.95) Cumulative N=5
(alpha=0.95) Cumulative N=10
(alpha=0.95) Cumulative N=20
0
20 40 60
Years
80
100
Figure 4-4. Effects of sample size on time to detect a climate effect on
macroinvertebrate taxa loss due to climatic warming at high taxa loss rates
in the Northeast/Mid-Atlantic U.S. The confidence level is fixed at 0.95. This
analysis is based on a high estimate of global warming (5°C by 2100) and the
highest macroinvertebrate taxa loss rate.
4-13
-------
As before, relaxing assumptions about required power and confidence levels (e.g.,
decreasing a and |3 to 0.8) will decrease the duration of monitoring needed to be able to detect a
climate change effect of taxa loss at a particular site. Similarly, if macroinvertebrate taxa loss
rates are slower than the high taxa loss rate used for Figure 4-4, it would take much longer to
detect an effect (see Figure 4-5), all else being equal. Effects of sample size on fish taxa loss
rates are similar to those for macroinvertebrates (see Figure 4-6): Locations with higher rates of
climate change-associated temperature increases and/or higher rates of taxa loss responses would
require less monitoring time to detect (i.e., statistically demonstrate) an effect; the converse
(lower ranges of temperature increase and/or taxa loss) would increase the required monitoring
time.
4.4. KEY CONCLUSIONS
Results of this case study highlight considerations for monitoring programs in light of the
need to account for climate change. Increasing sample size, either by increasing the number of
reference sites sampled each year or increasing the number of samples taken per watershed, or
per reach for targeted studies, will increase the ability to discern a climate change effect using
biomonitoring data. Regions with lower rates of climate change and/or taxa loss rates will
require either a longer monitoring period or a larger sampling effort to detect climate change taxa
losses effectively. On the other hand, with lower rates of climate change, effects from other
regulated sources of perturbation may be reliably detectable for longer, although increases in
variability and degradation of signal-to-noise ratio will degrade ability to detect impairment to
some extent (see Section 5). Since greater variability in the data decreases ability to detect
differences in taxon richness due to climate change, region-specific estimates of data variance
are important for an evaluation of a particular monitoring program. In addition, factors within a
monitoring design that can control for predictable sources of variation, such as partitioning by
watershed or ecoregion, become important, as they would reduce (account for) natural sources of
variation and increase ability to reliably recognize climate change effects.
The choice of a probabilistic or targeted sampling protocol is an important monitoring
design issue, and will depend on the questions being asked. It also bears on the ability to detect
climate change effects. Probabilistic designs are good for asking questions about, for instance,
the average condition of streams or watersheds, including taxon richness, within a region. With
regard to climate change effects, probabilistic sampling across reference sites would be ideal for
defining condition but would require relatively large sample sizes to detect differences in
biological attributes such as taxon richness because of the greater variation in the data. In the
context of this case study, sample size and power are based on paired tests, which are much more
4-14
-------
18.00
o 16.00
_i
x 14.00
-a 12.00
| 10.00
^ 8.00
c
cc
0) 6.00
N
IB 4.00
E 2.00
0.00
Low Taxa Loss Rates (1 taxa per degree C)
•Predicted Taxa Loss
• (alpha=0.95) Fixed N=5
(alpha=0.95) Fixed N=10
• (alpha=0.95) Fixed N=20
• (alpha=0.95) Cumulative N=5
(alpha=0.95) Cumulative N=10
• (alpha=0.95) Cumulative N=20
r
0
20
40 60
Years
80
100
Figure 4-5. Effects of sample size on time to detect a climate effect on
macroinvertebrate taxa loss due to climatic warming at low
macroinvertebrate taxa loss rates in the Northeast/Mid-Atlantic U.S. The
confidence level is fixed at 0.95. This analysis is based on a high estimate of
global warming (5°C by 2100) and the highest macroinvertebrate taxa loss
rate.
4-15
-------
w
w
o
cc
X
cc
1-
T3
0
0
CL
T3
CC
0
N
-------
powerful than drawing new independent samples every year because site-to-site differences are
removed from the variance term, leaving only differences over time between sites.
Targeted site selection, however, is often needed to answer specific questions, including
site-specific questions such as whether a site is meeting its designated use or permit
requirements. Another question that benefits from targeted site selection is what the effect of a
specific land use is on stream condition, because of the benefits of targeting sampling locations
along a gradient of effects. This may be important for studying how land use will interact with
climate change to affect stream condition.
If a bioassessment program is going to encompass climate change monitoring, there are
several points to consider:
1. Protection of reference streams emerges as an important concept, especially considering
that reference sites will be used to gauge climate change effects as well as the relative
effects of climate change on other stressors.
2. Ongoing monitoring of reference sites becomes an even more important aspect of
program design, with more sampling sites in reference locations and/or greater frequency
of sampling increasing the ability to detect change.
3. The use of rotating designs (rotating sampling among basins over years so that a
complete cycle of sampling may take 5 or more years) is often employed by state
biomonitoring programs to optimize resources, because crews can stay within defined
areas, travel can be limited, and total numbers of samples collected and processed each
year is reduced by focusing on a subset of basins. This approach also means that
reference sites within any one basin will only be sampled once every several years,
increasing the time it will take to obtain replicate samples needed to define climate
change-associated trends.
4-17
-------
5. CASE STUDY 2—ACCOUNTING FOR TRENDS: BIOLOGICAL
ASSESSMENT IN THE PRESENCE OF CLIMATE CHANGE
Detection of biological impairment and identification of its causes are two principal
objectives of bioassessment. Climate change will affect these central objectives, especially the
ability to discern impairment by comparison to reference locations. The second case study
examines how to detect impairment under climate change, particularly the ability to differentiate
between reference conditions and locations of reduced biological condition and the ability to
assign cause to impaired conditions. This approach is a foundation for defining how monitoring
may have to be modified to incorporate climate change and how data can be analyzed to account
for climate change and remain viable. Appendix B provides some additional details of this case
study, in particular descriptive and supplemental results of correlations examining potential
relationships between environmental variables and stressors that might be used to reflect direct
climate change effects.
5.1. OBJECTIVES
The case study examines the potential vulnerability of biomonitoring programs and
assessment methods to biological changes that result from climate change. This case study
addresses the following questions:
• How do we detect impairment under climate change?
• How does climate change affect our ability to identify causes of biological impairment?
• Are there analytical or monitoring design approaches that will allow managers to
effectively identify and manage stressors independently of climate change?
Climate change effects will likely drive the attributes of reference sites toward greater
similarity with impaired sites (i.e., decreased distance between the condition state of reference
and impaired). This decrease in effect (signal) may also be accompanied by increases in
variation (noise). A decrease in the signal-to-noise ratio would decrease the ability to detect
impairment. In addition to direct effects on site assessment, climate change effects may interact
with conventional stressors, further confounding the ability to discriminate stressor effects based
on reference/impaired site comparisons.
5.2. ANALYSIS APPROACH
The case study uses existing data, and by examining the associations of biological
attributes with proxy attributes of climate change, evaluates the potential effects and
5-1
-------
vulnerabilities of aquatic biomonitoring programs to climate change. Biological responses of
streams to various stressors were examined—with particular emphasis on hydrologic parameters
that may be influenced by climate change.
Detectable biological responses to climate change effects in streams that are important in
a bioassessment framework include the following:
• Southern taxa expanding their range northward
• Habitat change from increased winter/spring scour
• Loss of taxa sensitive to summer drought periods or higher temperatures (including
higher water temperature associated with drought)
• Improved conditions for invasive species, including disturbance regimes favoring
invasive species and warmer water temperatures allowing overwintering
• Change in number of reproductive periods leading to changes in timing of peak
abundance (possibly also tied to changes in phenology)
• In addition to the direct effects of temperature change, streams are also subject to
hydrologic changes from changed precipitation patterns and increased evapotranspiration
(e.g., Moore et al., 1997). Extreme stream flows reshape the stream habitat, and summer
low-flow events represent bottlenecks of both warm temperature and reduced habitat
(Moore et al., 1997; Poff and Ward, 1989). This analysis is focused on changes that
might occur in the Mid-Atlantic region, but the results may be generalized to similar
changes occurring in other regions
Data were partitioned into subsets defined by wet, normal, and dry periods, and
biological indicators of reference and impaired sites were examined. Several stressor-response
relationships were evaluated under the different climatic conditions. The intent was to estimate
probable minimum and maximum changes.
5.2.1. Datasets Evaluated
The MBSS data set was used to evaluate biologic responses to stressors under different
conditions. Section 4.2.2 describes attributes of this data set.
5.2.2. Metrics
Several invertebrate metrics were calculated in the Ecological Data Application System
(EDAS) database for the 1320 randomly located benthic samples in the Piedmont and Highlands
regions that were collected over the 10-year period (1994-2004) and analyzed as response
variables. These included total taxa (taxon richness), number of taxa in the insect families of
mayflies (Ephemeroptera), stoneflies (Plecoptera), and caddisflies (Trichoptera) (EPT taxa
5-2
-------
collectively), and the 2005 version of the Maryland benthic index of biotic integrity (B-IBI). For
fish, response variables examined include the Maryland fish IBI, total number (abundance) of
fish, and number of species offish (taxon richness). The main analytical focus was on 2 fish
indicators and 2 benthic macroinvertebrate indicators: the Maryland fish IBI score, and fish
taxon richness; and the Maryland B-IBI score, and the EPT taxon richness. The selected
indicators are all responsive, general indicators of stress but are not diagnostic of any particular
stressor.
The Maryland 305(b) evaluation of the status of waters of the state, which uses the MBSS
data in addition to other data sources, uses benthic and fish Indices of Biotic Integrity (IBIs) to
determine impairment status and attainment of uses (Maryland Department of the Environment,
2004; http://www.mde.state.md.us/assets/document/AppndxC2004-303d_Final.pdf). For a
single stream reach assessment, Maryland takes into account population-wide measurement
error. The approximate result is that if both indexes are >3.3, the stream segment is considered
unimpaired, and if either index is <2.7, the segment is impaired. Intermediate values are
considered to be potentially impaired but are still listed as supporting aquatic life uses.
5.2.3. Regional Data Partitions
The MBSS data were partitioned based on Maryland's classification into four ecoregions
(Coastal Plain, Eastern Piedmont, Cold-water Highlands, and Warm-water Highlands; see
Figure 5-1), to account for known sources of natural variation in both habitat (physical and
chemical) and biological data. The heavily developed Eastern Piedmont region, with a high level
of urbanization that represents an existing source of impairment, was targeted for evaluation.
Due to the level of development, the Eastern Piedmont region has relatively few reference areas.
In the original MBSS index development, the Piedmont and Highlands regions were deemed to
be biologically similar (Roth et al., 1998; Stribling et al., 1998). Sampling of more reference
sites showed that the Piedmont can be separated from the Highlands region (Southerland et al.,
2007, 2005). However, to have sufficient reference locations to support the analyses, the original
classification was used, recombining the Piedmont and Highlands sites for analyses requiring
identified reference sites.
5.2.4. Climate Data
The National Climatic Data Center (NCDC; www.ncdc.noaa.gov) makes available
several average monthly parameters, organized by state climatic regions. Although climate does
not follow state boundaries, it was convenient in this case because our biological data did follow
the state boundaries. We used data from two NCDC regions of Maryland: the Northern Central
Division (primarily Northern Piedmont ecoregion, and the Blue Ridge ecoregion within
5-3
-------
Highlands
Eastern Piedmont
Assessment Region Borders
Stations
u Coastal Plain
s Eastern Piedmont
T Highlands - Coldwater
s Highlands - Warmwater
/\/ Rivers
J Counties
| | Maryland
Coastal Plain
80
80
160 Miles
Figure 5-1. Maryland MBSS sampling stations showing regional divisions.
-------
Maryland), and the Appalachian Mountain Division (Central Appalachian Ridge and Valley
ecoregion). Maryland's Piedmont and warm-water mountain streams occur primarily in these
two climatic divisions. The Northern Central Division data were applied to the Eastern
Piedmont streams, and the Appalachian Division data to Highlands streams.
We estimated potential hydrologic effects of climate change by using the Palmer
Hydrologic Drought Index (PHDI) as a proxy for estimates of hydrologic changes due to climate
change. The PHDI is a monthly hydrological drought index used to assess long-term moisture
supply to water bodies (Karl, 1986) and is described in detail on the NCDC website
(http://www.ncdc.noaa.gov/oa/climate/onlineprod/drought/ readme.html). The index ranges
from -7 to +7, with negative values indicating dry spells, and positive values indicating wet
conditions. The PHDI takes into account water storage as soil and groundwater, and therefore is
more applicable to stream flow than the PHDI, which uses only temperature and rainfall
information (Karl, 1986). Figure 5-2 shows the 30-year distribution of the PHDI for the
Maryland Northern Central Division, which includes the Piedmont. The range of the PHDI
varies little from month to month, but the 30-year median value is positive during the
spring/summer macroinvertebrate sampling index period (>0 in March-May) and markedly
lower in late summer-early fall during the last half of the fish sampling index period (<-l in
S eptemb er-O ctob er).
MD Northern Central Division, 1970-1999
jz 4
O)
o
Hydrologic [
j o r>
35 "
E
ro
D- 4
i
i
[
_ T
1
2 3
1
1
J
L
I
r
[
t
1
5
1
j
6
N
1 i
/lont
i '
J
' i
h
i
i
5 J
I 1
3 1
1
1 [
0 1
1
1 1
I
[
2
• Median
D 25%-75%
X Non-Outlier Range
o Outliers
* Extremes
Figure 5-2. Monthly Palmer HDI for the 30-year period 1970-1999 (Source:
NCDC; http://www.ncdc.noaa.gov).
5-5
-------
In addition to calculating monthly PHDI, a variety of alternative running averages were
also calculated to account for possible lags in effects, and specifically for the time lag between
when droughts occur and directly impact the biota (summer/early fall) and the sampling period
(the spring index period): the previous 6-month average, the previous 12-month average, and the
previous summer PHDI. These alternative PHDI averaging methods did not yield results (i.e., in
developing wet/dry/average year comparison) different from those developed using the simple
monthly average; therefore average monthly PHDI was used in all analyses.
5.2.5. Hydrologic Attributes
Baker's flashiness index (Baker et al., 2004) was estimated for each stream. Flashiness is
a component of the hydrologic regime of streams, and, in general, is related to the frequency of
short-term changes in runoff associated with rainfall events, and how rapidly each event comes
and goes. Flashiness is generally considered to increase with increases in impervious cover
associated with urbanization and/or with land clearing for agriculture (Allan, 2004). It is both
responsive to urbanization as an existing stressor, and it also is expected to change in the future
in response to climate change projections of increased frequency and intensity of storms within
many regions of the U.S. Baker's index is calculated as the average of absolute values of daily
mean flow change divided by mean flow for the 2-day period. The maximum range is from 0
(absolutely constant flow) to 2 (alternating days of flow and no flow).
Daily flows were simulated for each site using the Flow Time Series Estimation tool
(FTSE; Tetra Tech, 2005). The model can be used to estimate daily flows for ungauged streams
based on multiple regressions using a smaller set of gauged streams. The main criterion for
proper functioning of the model is that there must be gauged stations relatively near to the
ungauged streams (e.g., within the same ecoregion) so that a standard is available for calibrating
the model. Estimates had been developed for a set of 764 streams in the Piedmont only (Barbour
et al., 2006); no set of appropriate gauged streams was available for the Appalachians.
5.2.6. Specific Analyses
5.2.6.1. Sensitivity of the System to Climate Change
As discussed in Sections 1 and 3, the primary hydrologic stressors associated with
climate change are changes in precipitation patterns combined with changes in temperature
regime, which will drive changes in hydrologic regime. The projected extent of changes in
temperature and precipitation varies regionally in the U.S.; therefore, so too will changes in the
magnitude, frequency, flashiness, and other patterns of runoff. The National Assessment of
climate change in the U.S. provides regional summaries of projected changes in the temperature
5-6
-------
and precipitation regimes of the major regions of the U.S. (NAST, 2001). Table B-l (in
Appendix B) summarizes these projections by region.
5.2.6.2. Stress-Responses
We established the ranges and variability of system response variables to climate change,
and described the response signatures to stressors. Due to the large number of parameters
available in the MBSS database, correlation analyses were used to identify stressor variables that
were most strongly related to response variables. Spearman Correlation coefficients were used
for this analysis; as a non-parametric test, there is no need to make assumptions about or test for
normal distributions for each variable. A locally weighted scatter plot smoothing (LOWESS)
line was used to illustrate the pattern of any relationship between the variables being correlated.
LOWESS smoothing was done in SystatlO by running along the x values and finding predicted
values from a weighted average of nearby y values. The surface is allowed to flex locally to
better fit the data. For the LOWESS, the degree to which the line or surface (tension) is allowed
to flex locally to fit the data to 0.5 was specified, meaning that half the points are included in the
running window. Graphs of key variables (scatter plots with LOWESS line) were used to
illustrate the relationships defined by correlation analysis, and to confirm that all relationships
reflected consistent data with no errors or false trends introduced by data entry errors, reporting
unit errors, or other inconsistencies. The subset of parameters showing the strongest
relationships were used for further exploration of stressor-response models.
A conditional probability approach (Paul and MacDonald, 2005) was used to examine
changes in the biological community along stressor gradients. A conditional probability
statement provides the likelihood (probability) of a predefined response, if the value of a
pollutant stressor (condition) is exceeded. Conditional probability is the probability of an event
when it is known that some other event has occurred. To estimate conditional probability of
impairment, we first define impairment as a specific value for a response variable (e.g., EPT <11
genera). The analysis addresses the following question: for a given threshold of a stressor, what
is the cumulative probability of impairment? For example, if total phosphorous concentration is
greater than 0.2 mg/L, what is the probability of biological impairment for each site under
consideration? All observed stressor values (in this example, all observed values of total
phosphorous) are used to develop a curve of conditional probability (Paul and MacDonald,
2005).
5.2.6.3. Effects of Climate Change
We used proxy estimates of climate (in the existing data) that are representative of
projected climate change, and examined the ability to detect biological impairment and stressor-
5-7
-------
response relationships. Our proxies of climate change were the estimates of wetter-than-normal
and drier-than-normal conditions in the PHDI for each sampling event. The MBSS data were
post-stratified into dry, normal, and wet conditions based on the index; selected stressor-response
relationships were then reexamined under the wet and dry scenarios. This evaluation was done
separately for reference sites (defined a priori in the MBSS), impaired sites (defined a priori in
the MBSS plus sites with 10% or more impervious surface), and intermediate sites (sites not
included in the impaired or reference groups).
As part of the analysis, it can be assumed that future biological responses to altered
hydrological conditions will be similar to responses to current natural variability, and that future
hydrologic changes will be comparable to extremes observed in the past 10 years. The
assumptions are probably reasonable in the near-term (i.e., 50 years), but become less reasonable
farther into the future.
5.3. KEY FINDINGS
5.3.1. Observed Stressor-Responses
Establishing definitive stressor-response relationships is a critical step in the Stressor
Identification (SI) process, and it is fundamental to identifying probable causes of impairment.
Numerous relationships were examined; Appendix B summarizes these results. Only a subset of
results that show some correlation and/or those that were considered potentially important but
showed no significant relationship are presented in this section.
The fish and benthic invertebrate response variables that showed the strongest responses
in these correlations were the fish index of biotic integrity (FIBI), fish taxon richness, total
number offish, the Maryland benthic IB I, total benthic taxon richness, and total EPT taxon
richness (see Appendix B). From these, 2 fish indicators were selected and 2 benthic
macroinvertebrate indicators were used to evaluate the Maryland Fish IBI score, and fish taxon
richness; and the Maryland Benthic IBI score (B-IBI), and EPT taxon richness.
5.3.1.1. Physical Habitat
Both fish and benthic macroinvertebrate measures were correlated with overall physical
habitat, as measured by the Maryland Physical Habitat Index (Paul et al., 2002) (Appendix B).
Fish taxon richness was not correlated with the habitat index, but the fish IBI and both
invertebrate indicators were strongly correlated, increasing with improved habitat score. Among
habitat components, the EPT taxa were also positively correlated with the embeddedness score,
reflecting a component of habitat (interstitial spaces in cobble substrate) utilized by these
organisms. Fish taxon richness was also very strongly correlated with total flow, but this was a
reflection of the effect of stream size.
5-8
-------
5.3.1.2. Hydrology
Both the fish and the benthic macroinvertebrate indicators were negatively associated
with Baker's flashiness index (see Appendix B). Below a flashiness index value of 0.5,
biological indicator values could be in the normal range, but above a flashiness of 0.6, most
biological values indicated impairment. Flashiness is affected by impervious surface, which in
the study area, indicates urban land use. The macroinvertebrate indicators declined with
impervious surface in a catchment, but the fish indicators did not (see Appendix B).
5.3.1.3. Water Quality
The invertebrate indicator EPT was associated with dissolved organic carbon (DOC),
total phosphorus (TP), and conductivity, with the number of EPT taxa declining as the stressors
increased. The strongest association was with conductivity. No other chemical water quality
measures were associated with either fish or benthos (DO was uniformly moderate to high in the
dataset, and there were too few observations of low DO to show any relationship).
5.3.1.4. Temperature
The associations of both the fish and benthic macroinvertebrate communities to water
temperature were examined. Fish observations in the data set had already been classified
according to expected warm-water and cold-water communities, using current and likely
sustainable distributions of brook trout to define cold-water streams in the region west of Evitts
Creek in western Maryland (Southerland et al., 2005). It is important to note that temperature
was measured in late summer and fall, at the same time that the fish assemblage was sampled.
Macroinvertebrates were sampled in spring, and temperature was not measured at that time.
Fish taxon richness increased with temperature in warm-water streams in both the
Piedmont and in the Appalachians, but there was no detectable relationship in the cold-water
streams (see Figure 5-3). EPT and total macroinvertebrate taxon richness (measured in early
spring) were reduced in the cold-water Highland streams where late summer temperatures
exceeded 18-20°C (see Figures 5-4a, b). There was no detectable relationship between
temperature and benthic macroinvertebrates in Piedmont streams (see Figures 5-4c, d).
5.3.2. Estimates of Climate Change Effects
5.3.2.1. Temperature
Increases in average regional temperature might have the result that some fraction of
cold- or cool-water streams change to warm-water conditions and biota. Global average air
temperatures are expected to increase by at least 2°C by 2100 (likely range 2°C to 4.5°C, likely
5-9
-------
CO
_
o
25
20
15
10
0
A •.* A
Warm Water
(Red)
Cold Water
(Blue)
10
15 20
Temperature (°C)
25
Figure 5-3. Fish richness vs. temperature in Highland reference streams.
Lines are LOWESS estimates.
5-10
-------
0
_
o
50
40
~1 I
• Cold Water
• x Warm Water
3 30
CO
.Q
0
0 20
'o
o
CD 10
0
10
15 20
Temperature (°C)
30
25
« 20
CO
0
c
.c
o
Q_
LU 10
• Cold Water
x Warm Water
x •
10
X X
15 20
Temperature (°C)
25
CO
U'J
0>
C
" 10 -
LL
LJJ
15 20
Temperature (°C)
25
LO
25
20
15
o
'l_
•5 10
10
* +
* •
15 20
Temperature (°C)
25
Figure 5-4. (a) Macroinvertebrate richness vs. temperature; (b) EPT
richness vs. temperature in Highland reference streams; (c) EPT vs.
temperature relation; and (d) fish richness vs. temperature relation in
reference sites in Piedmont streams. Lines are LOWESS estimates.
5-11
-------
average 3°C; IPCC, 2007). On the average, summertime air temperature increases are projected
to be less than wintertime increases (MacCracken et al., 2001), and the late-summer stream water
temperature increases are expected to be less than the average increase (Mohseni et al., 2003).
Based on these results for fish and invertebrate taxa in Mid-Atlantic streams, a net increase in
site-specific fish richness can be expected, as individual streams change from cold- or cool-water
conditions to warm-water. Fish taxon richness has previously been found to be higher in warm-
water habitats (Wehrly et al., 2003). In contrast, invertebrate taxa per site may decrease in
Highland streams that exceed 18°C due to climate change, but with no change in streams that
remain well below 18°C in late summer, suggesting that Highland streams macroinvertebrate
communities may be sensitive to climate change according to (future) temperature regime.
5.3.2.2. Hydrology
To examine the potential effects of changed rainfall and evapotranspiration patterns, the
existing data was divided into three groups: samples taken in relatively dry conditions, samples
taken in approximately normal conditions, and samples taken in relatively wet conditions.
"Dry," "Normal," and "Wet' were defined according to the distribution of the PHDI in the data
set, thus, the range of conditions from the recent past (from the month of sampling to the
preceding year) was used to obtain some insight into consequences of climate change. The range
of PHDI was from -4.24 to +4.75, with a median of+1.8. Although the total range was
symmetrical from extreme drought (< -4) to extreme wetness (> +4), there were more wet
months than dry months in the 10-year period. We defined 3 climatic groupings: Dry: PHDI < -
2.5 (N = 264); Normal:-1.1 3.5 (N = 353). These
groupings were selected to get substantial differences between wet and dry conditions, i.e., to
eliminate confounding effects of "moderately dry" and "moderately wet" conditions, and yet
have sufficient sample size in each of the hydrologic groups.
Figure 5-5 shows the Benthic IBI (B-IBI) scores of the three stream classes under the
three climatic conditions. Dry conditions are associated with greater variability of reference
sites, and a net degradation of median B-IBI score in both reference and intermediate sites. Wet
conditions are similarly associated with increased variability and a net decline in median B-IBI
score, but less so than in dry conditions. A comparison of reference sentinel site B-IBI (as well
as F-IBI) scores for the period 2000-2004, which included a wet year (2003) and a drought
period (2002-early 2003) did not show a notable variation between years among the Piedmont
and Highland sentinel sites (Prochaska, 2005). However, fewer sites and years are included in
that analysis, and variation around the mean is not evaluated. The EPT taxa metric showed the
same overall pattern (Figure 5-6): a slight net loss of median number of taxa in reference and
intermediate sites, and increased variability in reference sites. The macroinvertebrate
5-12
-------
CQ
O
CD
CD
6
5
4
3
2
1
n
1
" t
-
-
-
000
0
0
0
0
o
0
c
1
-
,
1 1
000
c
aeeec
soo «
>
DQOOOO Q5pD
«n
o
XEOoooooa
ccmxxo
a
000
I
c
9
D 8
DOD O —
) maaeaaaaai
) nn4m
L T
1 1
Ref Intermed Deg
Dry
CQ
o
CD
CQ
Ref Intermed Deg
Normal
CQ
o
CD
CQ
Ref Intermed Deg
Wet
Figure 5-5. Benthic IBI performance and climatic condition. The dry, normal and wet designations under each
of the three graphs refers to categorizations based on the PHDI.
-------
03
X
03
Q.
LU
20
15
10
i
T
oooo
ooooooo
oc xxxxxxx x>
ooooooo
oc »
O 0
OOOC WOO
O 0
I
Ref
Intermed
Dry
Deg
03
X
03
Q.
LU
20
15
10
o o
o o
o ooooocxx a
C 3OOOOOOC 3
oooo
oooo cxxx>
OOC XX)
Ref
Intermed
Normal
Deg
03
X
03
Q.
LU
20
15
10
Ref Intermed Deg
Wet
Figure 5-6. EPT performance and climatic condition.
-------
communities at degraded sites were low in EPT taxa and IBI scores, so changes of hydrological
condition did not affect them much. The Fish IBI was also similar (Figure 5-7) but showed
slightly greater effects under wet conditions than did macroinvertebrates: larger decline in
median reference score and larger reference variability. It should be noted that there are
differences in sample sizes between categories of climate condition, and, in particular, there are
fewer "dry" years represented. This is a consequence of limited (10 years) data combined with
the rotating-basin sampling scheme (i.e., only a subset of basins are sampled each year). Having
more data would improve these comparisons.
A quantitative measure of the efficacy of an index in discriminating between reference
and stressed sites is the Discrimination Efficiency (DE), which is calculated as the percent of
stressed sites with scores less than the 25th percentile of the reference sites (U.S. EPA, 1999).
DE is influenced both by the absolute difference between the reference and stressed site mean
scores, and the variability or spread of the scores. DEs under the scenarios described above are
given in Table 5-1. From this analysis, it appears that increased drought degrades reference sites
enough to reduce the ability to discriminate impaired from reference conditions for both the
benthic IBI and EPT taxon richness. Interestingly, the median value under both dry and wet
conditions was reduced compared to normal conditions in the intermediate sites, indicating a net
impairment from normal conditions. Also, the overall spread or variability of reference IBI
scores increased in both the wet and dry scenarios.
Benthic macroinvertebrates were sampled in spring, and fish were sampled in late
summer and fall. In wet years, the fish IBI showed much higher variability in reference sites,
reducing the discrimination efficiency (see Figure 5-7, Table 5-1). Late summer and fall are
slightly drier than other times of the year: the 30-year median of the PHDI during the fish
sampling index period is less than -1 (Figure 5-2). Thus, wet conditions during the fish
sampling period may represent a greater departure from a median expectation than do dry
conditions during the invertebrate sampling period. This may explain the increased variability of
the reference site fish IBI values under wet conditions than under dry conditions, and the
reduction of discrimination efficiency.
The pattern of extreme (wet or dry) hydrologic conditions both decreasing mean index
values at reference stations and increasing variability demonstrates a tendency for these surrogate
estimates of hydrologic changes associated with climate change to drive reference locations to be
more like impaired locations, and thus decrease the ability to discriminate between the two based
on biological indicator data.
The association of climatic condition on the relationships between EPT taxa and two
environmental stressors, conductivity and impervious surface, which had shown good stressor-
5-15
-------
5
4
3
2
1
n
-
-
-
-
fi
CO
O
0
-
COO
O
a
D
3XCC
ccoo
****** :m
XCOCCCOO D
BBB8
ccce
a
c
COS
»Jo
CCO C )
3 CO —
. 800000
ECO O
* U«J
i i
Ref Intermed Deg
Dry
a
5
m
._ 0
C/)
u_
9
1
n
-
"
J
c
a
L
D
00060
3
mono
aan
<*
E
SDOGOB
me
D
DOBDD
T
1
-
"
a
5
m
._ 0
C/5
LL
9
1
n
-
"
f
CD
O
1
1
A , '
^^^^
, ,
--
I I
i i
Ref Intermed Deg
Normal
Ref Intermed Deg
Wet
Figure 5-1. Fish IBI performance in three climatic conditions. The dry, normal and wet designations under
each of the three graphs refers to categorizations based on the PHDI.
-------
Table 5-1. Discrimination efficiencies of IBIs and EPT taxa under 3 climatic
conditions
Climatic condition
Base (current normal year)
Dry year
Wet year
Benthic IBI
100%
64%
98%
EPT Taxa
100%
78%
95%
Fish IBI
69%
60%
16%
response relationships, was examined further (Appendix Figure B-6). The full data set (both
Piedmont and Highland warm-water streams) was included in these analyses.
The plots of the stressor-response relationships from all three climate scenarios were
overlaid to determine whether changes in the response curves might be associated with the
climate change scenario, namely increasing drought or increasing storm events. Figure 5-8
shows the stress-response relationships (with linear regressions) between EPT taxon richness and
conductivity for the Piedmont region, and the conditional probability analysis. First, mean
number of EPT taxa is generally higher in the base condition, and reduced under wet conditions,
with little difference between base and dry conditions.
The conditional probability analysis (see Figure 5-8b) examined the probability of
impairment along the stressor gradient. EPT taxa <8 was defined as the threshold of impairment,
consistent with the threshold used by Maryland DNR in the Piedmont (Southerland et al., 2005).
Conditional probabilities of EPT impairment under base, wet and dry conditions, show that the
probability of impairment is higher under the wet scenario than under baseline conditions This is
not merely the result of reduced conductivity in wet years because the overall distribution of
conductivity in wet and normal years is almost identical (see Figure 5-9; CDF of conductivity).
Under dry conditions, the probability of impairment was greater at low conductivities, and less at
high conductivities, though the actual difference in numbers of EPT taxa were small.
The natural conductivities of streams in the region are generally low due to low buffering
capacities of the parent rocks and soils, with the exception of limestone-influenced streams in the
Great Valley, in smaller limestone valleys of the Ridge and Valley ecoregion, and marble
formations in the Piedmont ecoregion (Woods et al., 1999). Increased conductivity is
consistently and reliably associated with reduced stream biological condition throughout the
Appalachian region (Gerritsen and Zheng, unpublished data).
Figure 5-10 shows the relationship with impervious surface for the climate scenarios.
One of the consequences of urbanization is an increase in impervious area from roads, parking
5-17
-------
CO
CO
0
c
.c
o
CO
X
25
20
15
10
0.
LJJ
10 100 1000
Conductivity (^ S/cm)
T3
C
O
O
0.3
10
100
Conductivity
1000
Figure 5-8. (a) Relationship of EPT richness to conductivity under drought
(red), base (blue), and wet (black) conditions and (b) conditional probability
in impairment for the same three relationships.
5-18
-------
CD
M—
1.0
0.9
0.8
0.7
0.6
0.5
o
'o 0.4
CD
^ 0.3
0.2
0.1
0.0
10 100 1000
Conductivity (|i S/cm)
Figure 5-9. Conductivity CDFs—Piedmont. Drought—red, bast
wet—black.
-blue,
5-19
-------
05
05
CD
.
o
CD
X
CL
LU
25
20
15
10
0.01 0.10 1.00 10.00
Impervious surface (%)
CO
£ 1.0
CD
| 0.8
'CD
Q.
E
^ 0.6
CD
O
CL
"CD
o
o
O
0.4
0.2
b
I,.,i
0.01 0.10 1.00 10.00
Impervious surface (%)
Figure 5-10. (a) Relationship of EPT richness to impervious surface under
drought (red), base (blue), and wet (black) conditions, and (b) conditional
probability of impairment for the same three relationships.
5-20
-------
lots, and rooftops. Impervious surface increases the "flashiness" of streams, as well as being a
conduit for urban contaminants and pollutants. Overall, the base group (i.e., average hydrologic
conditions) had higher levels of EPT taxa than either the drought or the storm groups, but the
differences were subtle, a difference of approximately 1-2 taxa, and the differences were not
consistent. Drought conditions yield a higher risk of impairment with impervious surface, but
the change is marginal.
5.4. DISCUSSION AND KEY CONCLUSIONS
Several biological indicators and their associations with stressors have been examined
under scenarios of normal, relatively dry, and relatively wet conditions. These scenarios were
derived by partitioning a long-term data set from the Mid-Atlantic Piedmont and Appalachians
by moisture conditions estimated by the PHDI. Some caveats regarding the sampling design and
the partitioning include the following:
• Grab-sample temperature and water chemistry measurements did not coincide with the
benthic macroinvertebrate samples. Macroinvertebrates as well as nutrients were
sampled in spring (March-early May), coincident with the spring freshet; fish and water
quality (temperature, DO, conductivity, habitat, etc.) were sampled in late summer (June-
September), coincident with annual low water. Although the MBSS deploys continuous
temperature loggers during the summer index period (June 1-September 30), these data
were not available at the time the analyses for this report were conducted, and the
deployment still does not cover the benthic sampling index period. Different index
periods for the organisms would have resulted in different drought index estimates.
• The PHDI applied to the month and year a site was visited; all sites sampled in the same
month (e.g., March 1999) and NCDC district (e.g., Piedmont) would have the same PHDI
value.
• The climatic conditions we examined are all recent, from the period 1995-2005. Future
climate is expected to show a greater frequency of extreme conditions, but they have not
been linked to the frequency and magnitude of climate projections and models.
Differences in median values and distributions of several biological indicators associated
with dry, normal, or wet conditions were not observed; however, the associations may have been
due to an "unlucky" random sample and can not be ruled out—especially at the basin level. All
samples from a particular basin-year sampling would fall in the same dry-normal-wet category,
and there is no assurance that basins sampled in any one year are representative of the range of
stressor conditions throughout the region, especially with respect to urbanization.
In spite of these caveats, the results indicate the potential consequences of climate change
on bioassessment indicators. In dry and wet years, indicator variability increased markedly in
5-21
-------
reference sites and there were slight reductions in median indicator values. Consequently, there
was reduced ability to discriminate between reference and stressed sites under dry conditions
(especially for macroinvertebrates) and under wet conditions (especially for fish). Associations
of the indicators with stressors, which are used to develop stress-response relationships for SI
(Suter et al., 2002; Norton et al., 2002), may also change as evidenced by the apparent response
to conductivity; though there was little change in response to impervious surface.
5.4.1. Reference Conditions
These results illustrate the potential sensitivity of reference sites to climate change.
Reference sites in many regions of the country are not pristine, but are merely the "best
available" in the region. This is especially true for the eastern Piedmont ecoregion, which has
been settled, farmed, and industrialized since Colonial times. It is unlikely that there are any
sampled watersheds in the Piedmont of Maryland that are free of suburban development; the
average population density of HUC-8 accounting units in the Maryland Piedmont ranges from
111 to >400 persons per square kilometer (1990 census; U.S. EPA, 1997).
Moderately stressed reference sites may be more sensitive to slight increases in additional
stress due to climate change than truly minimally stressed reference sites (Stoddard et al., 2006).
Therefore, it would be important to identify minimally stressed reference sites if they exist, to
document reference site selection criteria, whether minimally stressed or not, and to monitor
reference sites to document changes over time.
5.4.2. Importance of Monitoring
To be able to account for the effects of climate change on biological indicators and on
stressor-response relationships, it will be necessary to monitor a set of fixed sites over time
("sentinel" sites), such that the same sites are revisited. Systematic changes in biological
attributes can only be attributed to climate change if other potential causes are eliminated or
accounted for, hence the need to have sentinel sites that span a wide range of other potential
stressors, and not just least-stressed reference sites.
Because climate change effects are pervasive, components of trends that are common to
all sentinel sites can be assumed to reflect climate change effects. If no other degradation was
occurring at reference sites, then the magnitude and variation in trends at reference sites could be
used directly to characterize the climate change component and account for that component
within trends observed at non-reference sites. However, assumptions of continued "pristine" (or
even steady) conditions at reference sites are unlikely over time, given population growth,
invasion of non-native species, expected encroachment of suburban and other land uses,
increased water withdrawals for human use, and other landscape-scale effects. Even if
5-22
-------
recommendations to protect reference sites are adopted, lack of contribution from landscape-
scale stressors would have to be verified in the process of estimating climate change-associated
trends.
Once trends common to all sentinel monitoring sites are defined, different components of
trends at non-reference sites can be considered potentially due to other stressors and evaluated
through the SI approach.
5.4.3. Analytical Methods
A question that arises is whether there are more robust or more powerful analytical
methods that can overcome the projected degradation in signal quality and discrimination ability.
Unfortunately, it is the quality of the information (signal to noise) that will degrade, and not the
analytical methods. If the information is degraded, then no amount of statistics can recover
something that no longer exists. Nevertheless, tracking time trends at both reference and non-
reference "sentinel" locations over time provides a framework for defining climate change-
associated trends and differentiating these from the effects of conventional stressors that are of
regulatory interest.
In view of the likelihood of ubiquitous biological degradation due to climate change, it
becomes increasingly important to protect reference sites from degradation. Application of the
Biological Condition Gradient (BCG) (a kind of universal measurement yardstick that will be
explain in Section 5.4.6) and Tiered Aquatic Life Uses (TALU) would establish a framework for
such protection (U.S. EPA, 2005) (see also Section 5.4.6). For example, one expected outcome
of defining TALUs is that states would adopt "high" and "exceptional" quality use classes along
the BCG, which would be above their current action threshold for "fishable/swimmable." Each
aquatic life use class would have biological criteria associated with it, which would allow
detection of degradation at reference sites at a stage substantially before the reference site would
be "impaired" under current definitions. Such a formalized process also provides for
implementation of particular management actions, such as identification of the cause of
impairment and implementation of corrective actions.
5.4.4. Stressor Identification
At least some associations of the indicators with stressors, which are used to develop
stressor-response relationships for SI (Suter et al., 2002; Norton et al., 2002), are expected to
change as hydrological conditions are altered by climate change. There was a marked change in
the stressor-response relationship between macroinvertebrates and conductivity under wetter
than usual conditions, which was associated with an increased probability of impairment.
However, almost no response was observed for impervious surface. SI may be similarly
5-23
-------
hampered by pervasive degradation and increased variability of all sites. If the conductivity
stressor-response in the wet condition is considered a typical scenario, then conductivity is
implicated in a smaller fraction of impairment (because the baseline frequency of impairment is
higher), yet the threshold water quality criterion for conductivity would also be lower. That is,
protection from degradation by conductivity may need to be tighter and set at a lower
conductivity than before the climate changed.
5.4.5. Biocriteria
Increased variability of reference sites as a consequence of climate change could decrease
the ability of states to detect impairment, if impairment thresholds are determined by a statistical
percentile of the indicator distribution in reference sites. Many states use a lower percentile of
the reference distribution as a numerical biocriterion for 305(b) assessment, for example, the 25th
percentile (Ohio EPA), or the 10th percentile (Maryland), or the 5th percentile (West Virginia). If
climate change causes the percentiles to drift downward, and the state reevaluates its water
quality criteria with new data, then the new criteria may set a lower bar, i.e., permit more
degradation to take place, before any kind of management is implemented (e.g., total maximum
daily load (TMDL) calculations). The potential drift of reference site condition due to climate
change illustrates the importance of establishing a universal measurement scale of biological
condition (e.g., the BCG) so that reference site drift can be identified as such.
5.4.6. Universal Scale to Measure Biological Condition
Acceptable biological condition is determined in many states from statistical properties of
a numerical index. Index values and criteria vary widely from state to state because of
differences among data sets used to develop the respective indexes. Furthermore, the criteria
"action level" often reflects substantial biological degradation from relatively undisturbed
conditions, such that the highest quality waters are not adequately protected. Results of this case
study demonstrate that biological responses to climate change may further confound assessment
and criteria for water management. To resolve these issues, panels of state and academic aquatic
biologists have proposed a conceptual model for a universal measurement scale of aquatic
biological condition—the BCG (Davies and Jackson, 2006).
The conceptual BCG model describes ecological changes that take place in flowing
waters with increased anthropogenic degradation, from pristine to degraded (Davies and Jackson,
2006). The BCG promotes consistency among agencies in the application of the CWA by
identifying tiers, or condition classes, that can be operationally defined in a consistent manner.
The model is intended to be broadly applicable to any kind of stream; the tiers are independent of
actual monitoring methods. Although the model promotes conceptual unification, it recognizes
5-24
-------
regional natural variability, and is not applied as a one-size-fits-all approach. The BCG is a
general description of change in aquatic communities, it is consistent with ecological theory, and
the BCG approach has been verified by aquatic biologists throughout the U.S. (Davies and
Jackson, 2006).
Calibration of the BCG to local conditions, and on a nationwide basis, would help
establish two baselines that would reduce the effects of confounding by climate change. The
first baseline is the description of pristine or nearly pristine conditions, Tier 1 of the BCG. In
many regions, the description of Tier 1 must rely on historical descriptions of fauna and
historical ranges of organisms (these may be available for fish, but rarely aquatic invertebrates),
on modeling approaches, on best professional judgment, or on sites available across political
boundaries (Stoddard et al., 2006). The second baseline is the description of the present-day
reference, or least stressed condition, before large-scale effects of climate change have occurred.
5-25
-------
6. RECOMMENDATIONS FOR U.S. EPA TO IMPLEMENT A FOUNDATION
FOR STATE/TRIBAL BIOASSESSMENT/BIOCRITERIA PROGRAMS
TO CONSIDER CLIMATE CHANGE
6.1. RECOMMENDATIONS FOR U.S. EPA
Results of the case study analyses to date, continued development and review of indicator
sensitivity classification, and discussion and input from state/tribal biocriteria managers at the
Workshop in March 2007 provide the basis for recommending the focus of ongoing and future
efforts to continue development and implementation of a framework for biological assessment
programs to account for climate change effects. Recommendations can be categorized as
technical requirements and resource requirements. Technical requirements focus on information
needed to better understand the interactions between expected effects of climate change and
biomonitoring program endpoints, additional technological support, and general policy support.
During the Workshop, some of these activities were identified as falling within the purview of
the U.S. EPA's ORD, and some in the purview of U.S. EPA's Office of Water (OW). These
identifications are made after each recommendation.
The ORD can
• Conduct further research through pilot studies (see Section 7) to determine the best
hydrologic and biological response indicators, to define biologically sensitive measures
to hydrologic changes, and to identify species traits responsive to climate change
(temperature, flow, sediment).
• Investigate how taxa replacement will affect biological indices used in state programs.
Determine the extent of change in the biological indices if specific metrics are changed.
• Develop and provide technical guidance regarding program adaptations and other
approaches needed to account for climate change in biological assessment programs,
including categorization of indicators (metrics), modification of monitoring designs, data
analysis approaches, etc., through guidance documents and/or website support.
• Fill gaps in knowledge and available modeling tools and outputs between regional
climate, hydrologic, and ecological models.
• Develop tools to make climate data available to other models (e.g., CADDIS).
The OW can
Include language in U.S. EPA grants, policies, etc. on climate change as a stressor for
monitoring and assessment programs, to establish climate change as an important
program focus.
6-1
-------
• Provide assistance to state bioassessment and resource management programs to integrate
the concept of climate change as a significant issue that should be accounted for in
assessing the condition of aquatic resources.
• Evaluate Water Quality Standards to be protective in the face of a changing condition
paradigm.
• Provide funding support for state/tribal water quality programs to assist in adaptations to
existing programs.
• Provide support for identification and sampling of reference sites, re-sampling of
reference sites, and more intensive characterization of reference and sentinel sites.
Together, the ORD and the OW can
• Conduct additional workshops to begin the process of evaluation and development of
recommendations for other aquatic ecosystems (e.g., large rivers, lakes, wetlands, coral
reefs, estuaries).
• Develop a nationwide database of state biological monitoring and assessment data to
support evaluation of national/ecoregional climate change trends and effects.
• Transfer technology for use of equipment, such as in situ temperature monitors, that
could be used to extend and enhance the value of monitoring data collected by state
programs with limited resources, including incorporation of processes and guidance.
• Provide technical support for data management tools (e.g., R code) to manage
temperature logger data and reduce it to useable metrics.
• Form partnerships across the U.S. EPA and other federal agencies on a comprehensive
climate change strategy to address mandates of CWA.
• Provide a summary of this meeting to U.S. EPA top management for information and
support for making informed decision-making.
States and tribes attending the Workshop also discussed resource limitations that impact
their ability to implement new and/or additional efforts related to revising and adapting their
existing programs to account for climate change. These limitations could be addressed through
• funding support
• personnel support
• priority setting for management actions
• assistance in developing and supporting a structure for sharing resources among agencies
to expand capacity
6-2
-------
6.2. RECOMMENDATIONS FOR STATES AND TRIBES
Even with constraints of limited resources, state and tribes participating in the Workshop
identified several potential program adaptations they considered feasible with resource and
technical assistance from the U.S. EPA. These actions include the following:
• Conduct regular and repeat reference site sampling.
• Consider strategies for maintenance and protection of reference sites and areas, including
identification of water bodies in the best condition.
• Evaluate the need to shift the sampling index period and/or expand sampling seasons.
• Establish sentinel sites for trend monitoring.
• Improve hydrological and temperature data collection.
• Retrieve historical data records to establish a basis for evaluating climate change.
• Incorporate traditional ecological knowledge, citizen monitoring, and phenological
knowledge in assessment of biomonitoring data.
• Continue the refinement of biocriteria programs to incorporate the Tiered Aquatic Life
Use (TALU) strategy.
• Accept moving target paradigm versus steady state model and adapt accordingly.
• Perform critical elements reviews of individual programs to identify relevant refinements.
• Engage in collaborative data and resource sharing to maximize limited resources.
6-3
-------
7. POTENTIAL NEXT STEPS
Several components of work, with somewhat different time frames, should be
contemplated to expand understanding of climate change effects on bioassessment programs and
develop a toolbox of appropriate responses.
1. Plan and conduct a national workshop for state and tribal WQ agency managers and
biologists on the next most appropriate ecosystem. There would be benefits to
developing and conducting a workshop for any of the remaining ecosystems of interest
with regard to bioassessment and biocriteria programs (large rivers, lakes, freshwater
wetlands, coastal wetlands, and estuaries). There are compelling reasons to consider
lakes and freshwater wetlands, including:
• Lakes have the next most well developed bioassessment and biocriteria programs,
and, together with flowing systems, they have the most TMDL issues with
biologically impaired waters; thus, integrating aspects of climate change with the
TMDL process is important here.
• Lakes are the next system of focus in the U. S. EPA's national assessment of
ecological condition of the Nation's aquatic resources.
• There is substantial overlap in state/tribal bioassessment/biocriteria scientists
and/or managers dealing with streams/rivers and lakes (often the same
individuals), providing an immediate opportunity to involve states and tribes that
were unable to participate in the first stream/river-oriented workshop, while still
expanding outreach to another ecosystem.
• Climate change effects in lakes, and freshwater wetlands will probably be similar
to effects in streams/rivers, but also would have different ecological responses
with different levels of importance (e.g., wetlands may be particularly susceptible
to droughts). Thus, this effort would build upon the experience and knowledge
developed in the Workshop and expand both the knowledge base as well as
consideration of ecosystem responses.
• Combining lakes and freshwater wetlands offers some efficiencies in
summarizing and discussing current information on climate change projections
and evidences that are most pertinent to inland freshwater non-flowing
ecosystems, despite ecological differences that certainly will have differences in
ecological processes and in responses that are most important.
• Inclusion of freshwater wetlands with lakes provides an opportunity to consider a
system much less advanced in the process of bioassessment/biocriteria program
development (wetlands) in an earlier time frame.
2. Implement a more in-depth assessment of climate change effects on stream and river
bioassessment programs in a detailed pilot study that would include selected states. It is
recommended that states in different parts of the U.S. be targeted to serve as regionally
distinct pilot studies. Ideally, as many as four states distributed regionally should be
7-1
-------
included, and, at minimum, two states should be included to account both for regional
(ecological) variations and for differences between bioassessment programs to at least
some extent. Some important considerations for including states in this pilot study are
• Regional distribution, preferably representing a spectrum of very different
geographical and ecological areas.
• Continuity and temporal duration of data sets available (ideally at least 20 years)
with comparable collection and analytical methods that would support rigorous
long-term analyses.
• Willingness of state personnel to be involved and interactive throughout the
analysis process—to maximize effective consideration of state- and location-
specific issues.
• Information from multiple basins or watersheds is typically needed to characterize
the breadth of variation in stressors and responses. An analysis approach should
be developed that includes several major aspects.
• Evaluation of all the specific metrics and the composite indices used in the state.
• Consideration and incorporation of the ecological traits of the species included in
the state database (classification by ecological traits and sensitivities may already
exist for the state databases likely to be utilized, especially if they completed
development of biological indices).
• Use of the long-term data sets to investigate and document existing evidence of
climate change.
• Compilation of thermal tolerance information for fish and invertebrates as a
resource to support predictions of probable climate change effects.
• Evaluation of the sensitivity of component metrics and biological indices to
climate change effects, possibly including recalibration of indices (and/or the
index development process) to identify components that may be more robust.
Analyze how biological indices can be modified to detect or exclude climate
change effects; investigate how taxa loss, replacement, and other predicted
responses will affect multimetric and other biological indices.
• Evaluation of index sampling periods, including the possible need to shift or
expand recommended sampling periods to better account for climate change
effects.
• Incorporation of the BCG and TALU into the analysis framework, to evaluate, for
example, how climate change degrades reference sites over time between tiers
above the CWA "fishable/swimmable" threshold, how this progressively impacts
detection of impairment and identification of stressors, and how reference
locations can be classified and protected.
• Use of ecological, habitat, and climatological data to characterize climate changes
and resulting changes in biological structure and function, especially in reference
sites or other benchmark for assessment of condition. May introduce targeted
7-2
-------
species/communities changes to the data to mimic climate change responses for
"future" analyses, based on documented projections for local/regional climate
effects and knowledge of species traits and sensitivities. Relate findings to state
WQ standards and designated uses as an example of confounding factors for
assessing and determining impairment.
• If scope of effort allows, evaluation of some novel indicators/metrics identified in
the framework based on extant research reported in the literature. Consideration
should be given to the feasibility of long-term, spatially distributed measurements
that could be made within the framework of a monitoring program; and to
robustness and interpretability of results with regard to climate change effects and
other stressors.
• Beyond the physical and chemical habitat data and biological data typically
collected in bioassessment programs, it will be very important to have
comprehensive climatological data corresponding to the regions being analyzed.
Projection of precipitation data to all sampling locations may be important. More
specifically, it may be important to be able to develop site-specific hydrologic
projections. In the preliminary case studies, the PHDI was used to project
possible effects of dry years and wet years to establish a proxy for projected
climate effects of increased summer droughts and increased precipitation. An
alternative would be to develop site-specific hydrological estimates to correspond
to sampled biological data. The calibrated FTSE model can be used to estimate
high and low flow conditions for a specific site and a specific time period, to
estimate hydrologic conditions associated with a given sampling event. Such
hydrologic projections produced by the model could be informative in estimating
the effects of dry periods, or of numerous storm events, and in projecting future
climate changes.
• Having sufficiently detailed climate change projections for the states that will be
evaluated also is of great importance. It is clear from the Workshop just
conducted that detailed regional downscaling from GCMs are possible, and that
the technical approaches for developing these are improving. It was also clear
that such regionally specific modeling is not accomplished for all areas. An effort
will be needed to determine the nature of modeling results available for each
state/region considered in the pilot study, and to interact with the appropriate
climate modeling scientists to understand the status of these results and obtain
needed outputs.
Plan a special JNABS issue and special workshop/session at the ASLO/NABS
conference in 2010 on the effects of climate change on biological indicators. This would
provide a scientific forum to articulate the known science of the effects of climate change
on biological indicators. This publication/special session is a follow-on to an earlier
ASLO/NABS collaboration held in 1998.
• Special publication series in the Journal of the North American Benthological
Society would bring together international scientists working on the concept of
climate change upon aquatic ecosystems, particularly biological indicators.
7-3
-------
Ideally, the papers would be published prior to the joint congress of the two
societies in June 2010.
• Special session devoted to climate change would be held at the joint congress and
would be highlighted as a key theme of the congress. The international scientists
in the publication series would be the featured speakers in the session at the
ASLO/NABS conference to be held in Santa Fe, New Mexico in June 2010.
Work across U.S. EPA and state programs to develop a national database compiling all
available state/tribal bioassessment data to support regional and national-scale evaluation
of climate change status and trends. To consider this strategy, it is suggested that
development of a national database compiling all available state bioassessment data be
considered to support regional and national-scale evaluation of climate change status and
trends. At least two frameworks exist, which should be considered for adaptation to this
purpose.
• Oracle-based Ecological Data Application System (EDAS), an extension and
improvement over Access-based EDAS that is already used by many states. This
is a purpose-tailored database for bioassessment data, which accommodates
physical, chemical and habitat data, and biological data for multiple assemblages
including detailed taxonomic review and manipulation. In addition (and
importantly), it includes built-in analyses that support all the steps in
bioassessment, metric evaluation and index calculations and development.
• WQX, the replacement for STORET, is being designed to accommodate existing
state bioassessment data, but is not quite ready to house the volumes of state
ecological data. The existing accessibility to all states is an advantage of this
option. A disadvantage is the lack of associated bioassessment-specific analytical
capability. This lack could be addressed relatively easily by developing an
analysis front-end (from existing resources to a great extent). The handling of
taxonomic data in WQX is potentially another disadvantage that may be more
difficult to address.
• The effort to establish a national data base with acceptable quality control,
comparable data (considering taxonomy, reporting units, collection and analytical
methods, sampling index periods, and many other factors) would be substantial.
Analyses would be relatively simple once this was accomplished. It may be (and
perhaps is likely) that not all state data would be adequate for inclusion, and
certainly there will be large differences in spatial coverage, and especially in
chronological longevity of the data sets.
7-4
-------
8. CONCLUSIONS
The review of the literature on climate change effects on aquatic ecosystems shows that it
is likely that changes are already occurring (see Section 1). Although current sampling schemes
used by bioassessment programs are not explicitly designed to detect climate change effects, it is
possible to use the data for this purpose. The case studies presented in this report demonstrate
this capability. While the first case study focuses on the length of time it would take to detect a
specific effect due to climate change under a variety of scenarios, it is important to remember
that the aquatic systems being surveyed are probably already somewhere on the trajectory toward
a detectable effect. Recent climate change reports underscore this point that systems are not at
time zero with respect to effects (IPCC, 2007).
Existing and ongoing climate change effects have impacts within bioassessment
programs that affect how benchmarks are set and how expectations for acceptable conditions are
anchored. Monitored reference conditions now reflect temporally changing conditions, and the
results from case study two underscore the importance of monitoring reference sites.
Characterizing climate change as an additional but global stressor must be accounted for within
monitoring designs, analytical approaches, and assessment frameworks. Ultimately, efficacy of
the current programmatic approach to definition of acceptable and/or desirable conditions and
assessment of the need for regulatory intervention in the management of water resources requires
an understanding of all significant influences on the systems being assessed and regulated. It is
critically important to be able to distinguish between multiple stressors, and this is done through
the acquisition of high-quality bioassessment and other ecological data. This, in part, guarantees
the integrity of regulatory decisions through appropriate program adaptations.
3-1
-------
REFERENCES
Allan, JD. (2004) Landscapes and riverscapes: the influence of land use on stream ecosystems. Annu Rev Ecol Evol
Syst 35:257-284.
Bale, JS; Masters, GJ; Hodkinson, ID; et al. (2002) Herbivory in global climate change research: direct effects of
rising temperatures on insect herbivores. Glob Change Biol 8:1-16.
Barbour, MT; Gerritsen, J. (2006) Key features of bioassessment development in the United States of America.
pp. 351-366. In: . Ziglio, G; Siligardi, M; Flaim, G; eds. Biological monitoring of rivers: applications and
perspectives. Chichester, England: John Wiley and Sons, Ltd.. 469 pgs.
Barbour, MT; Stribling, JB; Karr, JR. (1995) Multimetric approach for establishing biocriteria and measuring
biological condition; pp. 63-77. In: Davis, WS; Simon, TP; eds. Biological assessment and criteria. Tools for water
resource planning and decision making. Boca Raton, FL: Lewis Publishers.
Barbour, MT; Swietlik, WF; Jackson; SK; et al. (2000) Measuring the attainment of biological integrity in the
U.S.A.: a critical element of ecological integrity. Hydrobiologia 422/423:453-464.
Barbour, MT; Paul, MJ; Bressler, DW; et al. (2006) Bioassessment: a tool for managing aquatic life uses for urban
streams. Water Environ Res Found 01 -WSM-3.
Baker, DB; Richards, RP; Loftus, TT; et al. (2004) A new flashiness index: characteristics and applications to
Midwestern rivers and streams. J Am Water Resour Assoc April 2004:503-522.
Barnett, TP; Adams, JC; Lettenmaier, DP. (2005) Potential impacts of a warming climate on water availability in
snow-dominated regions. Nature 438(17):303-309.
Barron, E. (2001) Chapter 4 Potential consequences of climate variability and change for the Northeastern United
States. In: Climate change impacts on the United States: the potential consequences of climate variability and
change (Foundation report). A report of the National Assessment Team for the U.S. Global Change Research
Program. Cambridge University Press.
Bertreaux, D; Reale, D; McAdam, AG; et al. (2004) Keeping pace with fast climate change: can artic life count on
evolution? Integr Comp Biol 44(2):140-151.
Boward, DM, Kazyak, PF; Stranko, SA; et al. (1999) From the mountains to the sea: the State of Maryland's
freshwater streams. U.S. EPA (Environmental Protection Agency). 903-R-99-023. Maryland Department of Natural
Resources, Monitoring and Nontidal Assessment Division, Annapolis, MD.
Bradley, DC; Ormerod, SJ. (2001) Community persistence among stream invertebrates tracks the North Atlantic
oscillation. J Anim Ecol 70:987-996.
Bradshaw, WE; Holzapfel, CM. (2006) Evolutionary response to rapid climate change. Science 312:1477-1478.
Briers, RA; Gee, JHR; Geoghegan, R. (2004) Effects of the North Atlantic oscillation on growth and phenology of
stream insects. Ecography 27:811-817.
British Columbia Ministry of Water, Land and Air Protection (2002) Environmental Trends in British Columbia
(2002) State of Environment Reporting, Biennial Report. ISSN 1481-7284.
Burger, R; Lynch, M. (1995) Evolution and extinction in a changing environment: a qualitative-genetic analysis.
Evolution49(l):151-163.
R-l
-------
Burgmer, T; Hillebrand, H; Pfenninger, M. (2007) Effects of climate-driven temperature changes on the diversity of
freshwater macroinvertebrates. Oecologia 151:93-103.
Caissie, D. (2006) The thermal regime of rivers: a review. Freshwat Biol 51:1389-1406.
Carpenter, SR; Fisher, SG; Grimm, NB; et al. (1992) Global change and fresh-water ecosystems. Ann Rev Ecol
Syst 23:119-139.
Cooney, SJ, Covich, AP; Lukacs, PM; et al. (2005) Modeling global warming scenarios in greenback cutthroat trout
(Oncorhynchus clarki stomias) streams: implications for species recovery. West N Am Naturalist 65:371-381.
Covich, AP; Fritz, SC; Lamb, PJ; et al. (1997) Potential effects of climate change on aquatic ecosystems of the Great
Plains of North America. Hydrol Process 11:993-1021.
Daufresne, M, Roger, MC; Capra, H; et al.; (2003) Long term changes within the invertebrate and fish communities
of the upper Rhone River: effects of climate factors. Glob Change Biol 10:124-140.
Davies, SP; Jackson, SK (2006) The biological condition gradient: a descriptive model for interpreting change in
aquatic ecosystems. Ecol Appl 16:1251-1266.
Dettinger, MD, Cayan, DR; Meyer, MK; et al. (2004) Simulated hydrologic responses to climate variation and
changes in the Merced, Carson, and American River basins, Sierra Nevada, California, 1900-2099. Clim Change
62:283-317.
Diffenbaugh, NS; Pal, JS; Trapp, PJ; et al. (2005) Fine-scale processes regulate the response of extreme events to
global climate change. Proc Nat Acad Sci 102(44): 15774-15778.
Dodds, WK; Welch, EB. (2000) Establishing nutrient criteria in streams. J N Am Benthol Soc 19:186-196.
Durance, I; Ormerod, SJ. (2007) Climate change effects on upland stream macroinvertebrates over a 25-year period.
Glob Change Biol 13:942-957.
Eaton, JG; Scheller, RM. (1996) Effects of climate warming on fish habitat in streams of the United States. Limnol
Oceanogr41(5):1109-1115.
Fausch, KD; Schrader, LH. (1987) Use of the index of biotic integrity to evaluate the effects of habitat, flow, and
water quality on fish communities in three Colorado Front Range streams. Final Report to the Kodak-Colorado
Division and the Cities of Fort Collins, Loveland, Greeley, Longmont, and Windsor. Department of Fishery and
Wildlife Biology, Colorado State University, Fort Collins, CO.
Fisher, AE; Barren, B; Yarnal, CG; et al. (1997) Climate change impacts in the Mid-Atlantic region, a workshop
report. The Pennsylvania State University, sponsored by U.S. EPA/ORD/GCRP cooperative agreement no. CR
826554-01 and the Office of Policy cooperative agreement no. CR 824369-01.
Flanagan, KM; McCauley, E; Wrona, F; et al. (2003) Climate change: the potential for latitudinal effects on algal
biomass in aquatic ecosystems. Can J Fish Aquat Sci 60:635-639.
Flato, GM; Boer, GJ; Lee, WG; et al. (2000) The Canadian Centre for Climate Modeling and Analysis global
coupled model and its climate. Clim Dynam 16:451-467.
Flecker, AS. (1996) Ecosystem engineering by a dominant detritivore in a diverse tropical ecosystem. Ecology
77:1845-1854.
R-2
-------
Frey, DG. (1977) Biological integrity of water - an historic approach; pp. 127-140. In: Ballentine, RK; Gaurraia,
LJ; eds. The Integrity of Water. Proceedings of a Symposium, March 10-12, 1975. U.S. Environmental Protection
Agency, Washington, DC.
Gafner, K; Robinson, CT. (2007) Nutrient enrichment influences the response of stream macroinvertebrates to
disturbance. J N Am Benthol Soc 26:92-102.
Gammon, JR (1973) Completion Report: the effect of thermal inputs on the populations of fish and
macroinvertebrates in the Wabash River. West Lafayette, IN: Purdue University Water Resources Research Center.
Gedney, N; Cox, PM; Betts, RA; et al. (2006) Detection of a direct carbon dioxide effect in continental river runoff
records. Nature 439:835-838.
Gerritsen, J, Zheng, L. (2001) Unpublished data. Analyses for West Virginia Department of Environmental
Protection, Division of Water and Waste Management. Owings Mill, MD: Tetra Tech, Inc.
Gibson, CA; Meyer, JL; Poff, NL; et al. 2005. Flow regime alterations under changing climate in two river basins:
Implications for freshwater ecosystems. River Res Appl 21:849-864.
Golladay, SW; Gagnon, PG; Kearns, M; et al. 2004. Response of freshwater mussel assemblages
(Bivalvia:Unionidae) to a record drought in the Gulf Coast Plain of southwestern Georgia. J N Am Benthol Soc
23(3):494-506.
Gregory, JS; Beesley, SS; Van Kirk, RW. (2000) Effect of springtime water temperature on the time of emergence
and size of Pteronarcys californica in the Henry's Fork catchment, Idaho, U.S.A. Freshwat Biol 45:75-83.
Grimm, NB; Chicon, A; Dahm, CN; et al. 1997. Sensitivity of aquatic ecosystems to climatic and anthropogenic
changes: the basin and range, American southwest and Mexico. Hydrobiol Proc 11:1023-1041.
Hall, LW; Scott, MC; Killen, WD. (1996) Development of biological indicators based on fish assemblages in
Maryland coastal plain streams. Annapolis, MD: Maryland Department of Natural Resources, Chesapeake Bay and
Watershed Programs. CBWP-MANTA-EA-96-1.
Hansen, J; Sato, M; Rudey, R; et al. (2006) Global temperature change. Proc Nat Acad Sci 103(39):14288-14293.
Harper, MP; Peckarsky, BL. (2006) Emergence cues of a mayfly in a high-altitude stream ecosystem: Potential
response to climate change. Ecol Appl 16:612-621.
Hawkins, CP; Hogue, JN; Decker, LM; et al. (1997) Channel morphology, water temperature, and assemblage
structure of stream insects. J N Am Benthol Soc 16(4):728-749.
Hayhoe, K; Wake, CP; Huntington, TG; et al. (2007) Past and future changes in climate and hydrological indicators
in the U.S. Northeast. Clim Dynam 28:381-407.
Heggenes, J; Roeds, KH. (2006) Do dams increase genetic diversity in brown trout (Salmo trutta)?
Microgeographic differentiation in a fragmented river. Ecol Freshwat Fish 15(4):366-375.
Hilderbrand, RH. (2003) The roles of carrying capacity, immigration, and population synchrony on persistence of
stream-resident cutthroat trout. Biol Conserv 110(2):257-266.
Hogg, ID; Williams, DD. (1996) Response of stream invertebrates to a global-warming thermal regime: an
ecosystem-level manipulation. Ecology 77(2):395-407.
Hogg, ID; Williams, DD; Eadie, JM; et al. (1995) The consequences of global warming for stream invertebrates - a
field simulation. J Therm Biol 20:199-206.
R-3
-------
Hogg, ID; Eadie, JM; De Lafontaine, Y. (1998) Atmospheric change and the diversity of aquatic invertebrates: are
we missing the boat? Environ Monit Assess 49:291-301.
Hughes, RM; Gammon, JR. (1987) Longitudinal changes in fish assemblages and water quality in the Willamette
River, Oregon. Trans Am Fish Soc 116(2): 196-209.
IPCC (Intergovernmental Panel on Climate Change). (2001) Climate change 2001: the scientific basis. In:
Houghton, JT; Ding, Y; Griggs, DJ; et al.; eds. Intergovernmental Panel on Climate Change: Working Group I.
Cambridge, UK: Cambridge University Press.
IPCC (Intergovernmental Panel on Climate Change). (2007) Climate change 2007: the physical science basis. In:
Solomon, S; Qin, D; Manning, M; et al.; eds. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press.
Jensen, AJ; Forseth, T; Johnsen, BO. (2000) Latitudinal variation in growth of young brown trout Salmo trutta. J
Animal Ecol 69:1010-1020.
Johns, TC; Carnell, RE; Crossley, JF; et al. (1997) The second Hadley Center coupled ocean-atmospheric GCM:
model description, spinup, and validation. ClimDynam 13:103-134.
Johnson, LB; Breneman, DH; Richards, C. (2003) Macroinvertebrate community structure and function associated
with large wood in low gradient streams. River Res Appl 19(3): 199-218.
Karl, TR. (1986) The sensitivity of the Palmer Drought Severity Index and Palmer's Z-index to their calibration
coefficients including potential evapotranspiration. J Clim Appl Meteorl 25:77-86.
Karr, JR; Dudley, DR. (1981) Ecological perspectives on water quality goals. Environ Manage 5:55-68.
Karr, JR; Fausch, KD; Angermeier, PL; et al. (1986) Assessing biological integrity in running waters: A method and
its rationale. Special publication 5. Illinois Natural History Survey.
Knowles, N; Cayan, DR. (2002) Potential effects of global warming on the Sacramento/San Joaquin watershed and
the San Francisco estuary. Geophys Res Lett 29(18):38-l-38-4.
Lehigh University. (1960) Research project on effects of condenser discharge water on aquatic life. Bethlehem, PA:
Institute of Research, Lehigh University.
Leonard, PM; Orth, DJ. (1986) Application and testing of an index of biotic integrity in small, coolwater streams.
Trans Am Fish Soc 115:401-414.
Lyons, J. (1992) Using the index of biotic integrity (IBI) to measure environmental quality in warmwater streams of
Wisconsin. General Technical Report, NC-149. St. Paul, MN: U.S. Department of Agriculture, Forest Service.
Lyons, J; Wang, L; Simonson, TD. (1996) Development and validation of an index of biotic integrity for coldwater
streams in Wisconsin. N Am J Fish Manage 16:241-256.
MacCracken, M; Barron, E; Easterling, D; et al. (2001) Chapter 1 Scenarios for climate variability and change. In:
Climate change impacts on the United States - the potential consequences of climate variability and change
(Foundation report). A report of the National Assessment Team for the U.S. Global Change Research Program.
Cambridge, UK: Cambridge University Press.
Magnuson, JJ; Webster, KE; Assel, RA; et al. (1997) Potential effects of climate changes on aquatic systems:
Laurentian Great Lakes and Precambrian Shield region. HydrolProc 11:825-871.
R-4
-------
Maryland Department of the Environment. (2004) 2004 List of impaired surface waters [303(d) List] and integrated
assessment of water quality in Maryland. Annapolis, MD. Available online at
http://www.mde.state.md.us/Programs/WaterPrograms/TMDL/Marvland%20303%20dlist/final 2004 303dlist.asp.
Masters, Z; Petersen, I; Hildrew, AG; et al. (2007) Insect dispersal does not limit the biological recovery of streams
from acidification. AquatConserv 17(4):375-383.
Matsubara, H; Sakai, H; Iwata, A. (2001) A river metapopulation structure of a Japanese freshwater gobie,
Odontobutis obscura, deduced from allozyme genetic indices. Environ Biol Fish 61(3):285-294.
Matthews, WJ. (1998) Patterns in freshwater fish ecology. New York, NY: Chapman & Hall.
McFarlane, NA; Boer, GJ; Blanchett, JP; et al. (1992) The Canadian Climate Centre second-generation general
circulation model and its equilibrium climate. J Climate 5:1013-1044.
McKee, D; Atkinson, D. (2000) The influence of climate change scenarios on populations of the mayfly Cloeon
dipterum. Hydrobiologia 441(l):55-62.
Melack, JM; Dozier, J; Goldman, CR; et al. (1997) Effects of climate change on inland waters of the Pacific Coastal
Mountains and Western Great Basin of North America. Hydrol Process 11:971-992.
Milewski, C. (2001) Local and systemic controls on fish and fish habitat in South Dakota rivers and streams:
Implications for management. PhD dissertation. South Dakota State University, Brookings, SD.
Miller, DL; Leonard, PM; Hughes, RM; et al. (1988) Regional applications of an Index of Biotic Integrity for use in
water resource management. Fisheries 13(5): 12-20.
Milly, PCD; Dunne, KA; Vecchia, AV. (2005) Global pattern of trends in streamflow and water availability in a
changing climate. Nature 438(17):347-350.
Mohseni, O, Stephan; HG; Eaton, JG. (2003) Global warming and potential changes in fish habitat in U.S. streams.
Climactic Change 59:389-409.
Moore, MV; Pace, ML; Mather, JR; etal. (1997) Potential effects of climate change on freshwater ecosystems of the
New England/Mid-Atlantic region. Hydrol Proc 11:925-947.
Moyle, PB; Brown, LR; Herbold, B. (1986) Final report on development and preliminary tests of indices of biotic
integrity for California. Final report to the U.S. Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
NAST (National Assessment Synthesis Team). (2001) Climate change impacts on the United States - The potential
consequences of climate variability and change (Foundation Report). A report of for the U.S. Global Change
Research Program. Cambridge University Press. Cambridge, UK. ISBN 0-521-00075-0.
Najjar, RG; Walker, H; Anderson, P; et al. (2000) The potential impacts of climate change on the Mid-Atlantic
Coastal Region. Climate Res 14:219-233.
Neff, R; Chang, H; Knight, CG; et al. (2000) Impact of climate variation and change on Mid-Atlantic region
hydrology and water resources. Climate Res 14:207-218.
Norton, SB; Cormier, SM; Smith, M; et al. (2002) Predicting levels of stress from biological assessment data:
empirical models from the eastern corn belt plains, Ohio, U.S.A. Environ Toxicol Chem 21(6): 1168-1175.
Ohio EPA (Environmental Protection Agency). (1987) Biological criteria for the protection of aquatic life: volumes
I-III. Ohio Environmental Protection Agency, Columbus, OH.
R-5
-------
Parmesan, C. (2006) Ecological and evolutionary responses to recent climate change. Annu Rev Ecol Evol Syst
37:637-669.
Paul, JF; McDonald, ME. (2005) Development of empirical, geographically specific water quality criteria: A
conditional probability analysis approach. J Am Water Resour Assoc 121 1-1223.
Paul, MJ; Stribling, JB; Klauda, R; et al. (2003) A physical habitat index for freshwater wadeable streams in
Maryland. Prepared by Tetra Tech, Inc., Owings Mills, MD for Maryland Department of Natural Resources.
Maryland DNR Chesapeake Bay and Watershed Programs, Monitoring and Non-Tidal Assessment CBWP-
MANTA-EA-03-4
Poff, LN; Allan; JD. (1995) Functional organization of stream fish assemblages in relation to hydrologic variability.
Ecology 76(2):606-627.
Poff, LN; Ward, JD. (1989) Implications of streamflow variability and predictability for lotic community structure:
A regional analysis of streamflow patterns. Can J Fish Aquat Sci 46: 1805-1818.
Poff, LN; Tokar, S; Johnson, P. (1996) Stream hydrological and ecological responses to climate change assessed
with an artificial neural network. Limnol Oceanogr 41(5):857-863.
Poff, LN; Brinson, MM; Day, JW, Jr. (2002) Aquatic Ecosystems and global climate change: potential impacts on
inland freshwater and coastal wetland ecosystems in the United States. Prepared for the Pew Center on Global
Climate Change. 44 pp.
Polsky, C; Allard, J; Currit, N; et al. (2000) The Mid-Atlantic region and its climate - past, present, and future.
Climate Res 14: 161-173.
Power, ME. (1990) Effects offish in river food webs. Science 250:81 1-814.
Power, ME; Matthews, WJ; Stewart, AJ. (1985) Grazing minnows, piscivorous bass and stream algae: Dynamics of
a strong interaction. Ecology 66:1448-1456.
Pringle, CM; Blake, GA; Covich, AP; et al. (1993) Effects of omnivorous shrimp in a montane tropical stream:
Sediment removal, disturbance of sessile invertebrates and enhancement of understory algal biomass. Oecologia
93:1-11.
Prochaska, AP. (2005) Maryland biological stream survey, volume 1 1 : Sentinel site network. Maryland Department
of Natural Resources, Montioring and Non-tidal Assessment Division. CBWP-MANTA-EA-05-8.
Rahel, FJ; Keleher, CJ; Anderson, JL. (1996) Potential habitat loss and population fragmentation for coldwater fish
in the North Platte River drainage of the Rocky Mountains: response to climate warming. Limnol Oceanogr
Rahmstorf, S; Cazenave, A; Church, JA; et al. (2007) Recent climate observations compared to projections. Science
316:709.
Resh, V; Rosenberg, D. (1984) The ecology of aquatic insects. New York, NY: Praeger.
Richter, BD; Baumgartner, JV; Powell, J; et al. (1996) A method for assessing hydrologic alteration within
ecosystems. Conserv Biol 10(4): 1 163-1 174.
Rogers, CE; McCarty, JP. (2000) Climate change and ecosystems of the Mid- Atlantic region. Climate Res.
14:235-244.
R-6
-------
Root, TL; Price, JT; Hall, KR; et al. (2003) Fingerprints of global warming on wild animals and plants. Nature
421:57-60.
Rosemond, AD; Mulholl, PJ; Elwood, JW. (1993) Top-down and bottom-up control of stream periphyton: effects of
nutrients and herbivores. Ecology 74:1264-1280.
Rosenberg, D; Resh, V. (1993) Freshwater biomonitoring andbenthic macroinvertebrates. Chapman Hall
Publishers.
Roth, NE; Southerland, MT; Chaillou, JC; et al. (1997) Maryland biological stream survey: Ecological status of non-
tidal streams in six basins sampled in 1995. Maryland Department of Natural Resources, Chesapeake Bay and
Watershed Programs, Monitoring and Non-tidal Assessment, Annapolis, Maryland. CBWP-MANTA-EA-97-2.
Roth, NE; Southerland, MT; Chaillou, JC; et al. (1998) Maryland biological stream survey: Development of a fish
index of biotic integrity. Environ Manag Assess 51:89-106.
Schindler, DW (1997) Widespread effects of climatic warming on freshwater ecosystems in North America. Hydrol
Proc 11:1043-1067.
Schindler, DW. (2001) The cumulative effects of climate warming andother human stresses on Canadian
freshwaters in the new millennium. Can J Fish Aquat Sci. 58:18-29.
Schindler, DW; Bayley, SE; Parker, BR; et al. (1996) The effects of climatic warming on the properties of boreal
lakes and streams at the Experimental Lakes Area, northwestern Ontario. Limnol Oceanogr 41(50):1004-1017.
Schindler, DW; Rogers, DE; Scheuerell, MD; et al. (2005) Effects of changing climate on zooplankton and juvenile
sockeye salmon growth in southwestern Alaska. Ecology 86:198-209.
Seager, R; Ting, M; Held, I.; et al. (2007) Model projections of an imminent transition to a more arid climate in
Southwestern North America. Science 316:1181-1184.
Simon, TP. (1991) Development of ecoregion expectations for the index of biotic integrity (IBI) Central Corn Belt
Plain. U.S. Environmental Protection Agency, Region V, Chicago, IL; EPA 905/9-91/025.
Simon, TP; Lyons, J. (1995) Application of the index of biotic integrity to evaluate water resource integrity in
freshwater ecosystems. In: Davis, WS; Simon, TP; eds. Biological assessment and criteria: Tools for water
resource planning and decision making. Boca Raton, FL: Lewis Publishers, pp. 245-262.
Smith, JB. (2004) A synthesis of potential climate changes impacts on the U.S. Prepared for the Pew Center on
Global Climate Change.
Snedecor, GW; Cochran, WG. (1980) Statistical methods. Ames, IA: Iowa State University Press.
Southerland, MT; Rogers, GM; Kline, MJ; et al. (2005) New biological indicators to better assess the condition of
Maryland Streams. In Maryland Biological Stream Survey 2000-2004, Volume XVI. Prepared for the Maryland
Department of Natural Resources Monitoring and Non-Tidal Assessment Division, Annapolis, Maryland. CBWP-
MANTA-EA-05-13.
Southerland, MT; Rogers, GM; Kline, MJ; et al. (2007) Improving biological indicators to better assess the
condition of streams. Ecol Indicat 7:751-767'.
Steedman, RJ (1988) Modification and assessment of an index of biotic integrity to quantify stream quality in
southern Ontario. Can J Fish Aquat Sci 45:492-501.
R-7
-------
Stephan, HG: Preudhomme, EB. (1993) Stream temperature estimation from air temperature. Water Resour Bull
29:27-45.
Stoddard, JL; Larsen, DP; Hawkins, CP; et al. (2006) Setting expectations for the ecological condition of streams:
the concept of reference conditions. Ecol Appl 16:1267-1276.
Stribling, JB; Jessup, BK; White, JS; etal. (1998) Development of a benthic index of biotic integrity for Maryland
streams. Report no. CBWP-EA-98-3. Maryland Department of Natural Resources, Monitoring and Non-Tidal
Assessment Division. Annapolis, MD 21401.
Suter, GW; Norton, SB; Cormier, SM. (2002) A methodology for inferring the causes of observed impairments in
aquatic ecosystems. Environ Toxicol Chem 21(6):1101-1111.
Sweeney, BW; Vannote, RL. (1978) Size variation and the distribution of hemimetabolous insects: two thermal
equilibrium hypotheses. Science 200(4340):444-446.
Tetra Tech, Inc. (2005) Flow time series estimation tool. Users manual. Water Environmental Research Foundation.
WERF01-WSM-3
Townsend, C; Doledec, S; Scarsbrook, M. (1997) Species traits in relation to temporal and spatial heterogeneity in
streams: a test of habitat templet theory. Freshwater Biol 37(2):3670387.
Tuchman, NC; Wetzel, RG; Rier, ST; et al. (2002) Elevated atmospheric CO2 lowers leaf litter nutritional quality for
stream ecosystem food webs. Global Change Biol 8:163-170.
Turner, D; Williams, DD. (2005) Sexual dimorphism and the influence of artificial elevated temperatures on body
size in the imago ofNemoura trispinosa (Plecoptera : Nemouridae). Aquat Insects 27:243-252.
Tyedmers, P; Ward, B. (2001) A review of the impacts of climate change on EC's freshwater fish resources and
possible management responses. Fisheries Centre Research Report 9(7): 1-15. Vancouver, B.C.: Fisheries Centre,
University of British Columbia.
U.S.EPA (Environmental Protection Agency). (1989) Rapid bioassessment protocols for use in streams and rivers:
benthic macroinvertebrates and fish. U.S. Environmental Protection Agency, Office of Water, Assessment and
Watershed Protection Division, Washington, DC; EPA/444/4-89-001.
U.S.EPA (Environmental Protection Agency). (1992) Biological populations as indicators of environmental change.
Office of Policy, Planning and Evaluation, U.S. Environmental Protection Agency, Washington, DC.
U.S.EPA (Environmental Protection Agency). (1996) Biological criteria: Technical guidance for streams and rivers.
Office of Science and Technology, U.S. Environmental Protection Agency, Washington, DC; EPA 822-B-96-001
U.S. EPA (Environmental Protection Agency). (1997) An ecological assessment of the United States mid-Atlantic
region. U.S. Environmental Protection Agency, Office of Research and Development, Washington, DC;
EPA/600/R-97/130.
U.S. EPA (Environmental Protection Agency). (1998) Lake and reservoir bioassessment and biocriteria. Technical
guidance document. Gerritsen, J; Carlson, R; Charles, DL; et al. U.S. Environmental Protection Agency, Office of
Water, Washington, DC; EPA/841-B-98-007.
U.S. EPA (Environmental Protection Agency). (1999) Rapid Bioassessment Protocols for Use in Streams and
Wadeable Rivers: Periphyton, Benthic Macroinvertebrates and Fish. 2nd Ed. U.S. Environmental Protection
Agency, Office of Water, Washington, DC; EPA/841-B-99-002.
R-8
-------
U.S. EPA (Environmental Protection Agency). (2000) Stressor identification guidance document. U.S.
Environmental Protection Agency, Office of Water and Office of Research and Development Washington, DC;
EPA-822-B-00-025. Final Report.
U.S. EPA (Environmental Protection Agency). (2002) Summary of biological assessment programs and biocriteria
development for states, tribes, territories, and interstate commissions: streams and wadeable rivers. U.S.
Environmental Protection Agency, Washington, DC; EPA-822-R-02-048.
U.S. EPA (Environmental Protection Agency). (2005) Use of biological information to better define designated
aquatic life uses in state and tribal water quality standards: Tiered Aquatic Life Uses. Draft. U.S. Environmental
Protection Agency, Office of Water, Washington DC; EPA-822-R-05-001.
Vannote, RL; Sweeney, BW. (1980) Geographic analysis of thermal equilibria: a conceptual model for evaluating
the effects of natural and modified thermal regimes on aqualtic insect communities. Am Natural 115(5):667-695.
Walther, GR; Post, E; Convey, P; et al. (2002) Ecological responses to recent climate change. Nature 416:389-395.
Watanabe, NC; More, I; Yoshitaka, I. (1999) Effect of water temperature on the mass emergence of the mayfly,
Ephoron shigae, in a Japanese river (Ephemeroptera: Polymitarcyidae). Freshwat Biol 41:537-541.
Wehrly, KE; Wiley, MJ; Seelbach, PW. (2003) Classifying regional variation in thermal regime based on stream
fish community patterns. Trans Am Fish Soc 132:18-38.
Wilby RL. (2006) When and where might climate change be detectable in UK river flows? Geophys Res Lett
33(19), Article Number L19407, doi:10.1029/2006GL027552.
Woods, AJ; Omernik, JM; Brown DD. (1999) Level III and Level IV Ecoregions of Delaware, Maryland,
Pennsylvania, Virginia and West Virginia. Washington, DC: U.S. Environmental Protection Agency, Western
Ecology Division. Available at http://www.epa.gov/wed/pages/ecoregions/reg3_eco.htm.
Xenopoulos, MA; Lodge, DM. (2006) Going with the flow: Using species-discharge relationships to forecast losses
in fish biodiversity. Ecology 87(8): 1907-1914.
Xenopoulos, MA; Lodge, DM; Alcamo, J; et al. (2005) Scenarios of freshwater fish extinctions from climate change
and water withdrawal. Glob Change Biol 11:1557-1564.
Yoder, CO; Rankin, ET. (2005) Changes in fish Assemblage status in Ohio's nonwadeable rivers and streams over
two decades. Am Fish Soc Symp 45:000-000 (31 pp).
R-9
-------
APPENDIX A
REGIONAL PATTERNS OF CLIMATE CHANGE PROJECTIONS AND
CONSEQUENCES FOR RIVERS AND STREAMS
-------
A.l. CLIMATE CHANGE AND FUTURE PROJECTIONS
The rate of global warming has increased over the last century; the linear average over
the last 50 years (0.13°C per decade) has almost doubled the linear rate over the last hundred
years (0.074°C per decade) (IPCC, 2007). The current rate is estimated at about 0.2°C per
decade (IPCC, 2007; Rahmstorf et al, 2007; Hansen et al, 2006) and may increase in the future.
There also have been widespread changes in precipitation. Frequency of heavy precipitation
events is also predicted to increase over most areas, as is the frequency of droughts (IPCC,
2007).
Climate change will continue and temperatures will potentially increase in the future
(IPCC, 2007). General projections for the year 2100 include global average temperature
increases of 1.1-2.9°C (from the lowest emissions scenario) to 2.4-6.4°C (from the highest
emissions scenario). Increases in precipitation are predicted, with a higher percentage of total
precipitation occurring in more frequent and intense storms. Other projections include more
precipitation in winter and less precipitation in summer; more winter precipitation as rain instead
of snow; earlier snow-melt; earlier ice-off in rivers and lakes; and longer periods of low flow and
more frequent droughts in summer (Hayhoe et al., 2007; IPCC, 2007; Barnett et al., 2005; Fisher
et al., 1997). Changes in temperature and precipitation will have regional differences that will be
important for assessing ecological effects.
A.1.1. REGIONAL PATTERNS
Several points must be considered to understand how the existing and future climate
changes are likely to affect aquatic ecosystems, specifically streams and rivers. Ecosystems do
not respond to global averages but to regional and local patterns (Walther et al., 2002). Regional
patterns of climate change are affected by factors that include atmospheric circulation patterns,
topography, land use, and region-specific feedbacks (Hayhoe et al., 2007), and they are more
difficult to project because of these localized factors. Within the U.S., regional projections for
future temperature increases are variable among models, but almost all regions project greater
temperature increases in winter than summer (NAST, 2001). Table A-l presents a summary of
regional climate change projections from the National Assessment and Synthesis Team (NAST)
(2001). Increased frequencies of extreme hot days (and decreases in extreme cold days) are also
projected throughout the U.S., with the greatest increase projected for the Southwest (Seager et
al., 2007). Other notable increases are projected for high-elevation areas of California, central
Utah, central Idaho, and the Appalachian Mountains (Diffenbaugh et al., 2005).
A-2
-------
Table A-l. Summary of regional climate projections (NAST, 2001). Averages and/or ranges for Hadley (H) and
Canadian (C) model projections to 2100 compared to 1961-90.
Temperature
Precipitation
Average
annual
Winter min
Winter
max
Summer
min
Summer
max
Average
annual
Extreme
events
Northeast3
2.6-5°C increase
4 to 2-7°C
1.6-3 to 3-6.6°C
3°C
1.3to4-6°C
Small increase to 5-10%
decrease (C); up to 25% increase
(H)
Small increase in count and
strength of storms (H); decrease
in count for most of region
except mid-Atlantic (C)
Southeast11
2.3-5.5°C
increase
10% decrease
(C) to 20%
increase (H)
Midwest0
2-6°C increase
(mins increase
more than max)
Increase 20^0%
by 2100 in upper
Midwest, decrease
up to 20% Ohio
Valley (C );
average
precipitation
increases
everywhere, up to
20-40% (H)
Increase in mean
associated with
increase in
frequency and
intensity of heavy
events
Great Plains'1
Increase >3°C
(both models)
5-7°C (winter
average)
4-5°C (summer
average)
13% increase
(both models)
Likely increase,
esp. in southern
GPs
West6
4.5-6°C
increase
4.9-7. 1°C
(winter average)
4.3-4.6°C
(summer
average)
Double winter
precipitation in
CA, decrease in
parts of Rockies
(both models);
summer
precipitation no
change (C ) to
decrease (H).
More extreme
wet and dry
years (both
models)
Pacific
Northwest'
3°C increase
by 2050
4.7-5.9°C
(winter
average)
4.1-4.6°C
(summer
average)
Increase of a
few to 20%
(average 10%)
(H); increase
0-50%
(average 30%)
(C)
Winters wetter
and warmer,
increase in
precipitation
in heavy
storms
>
-------
Table A-l. continued.
Droughts
Runoff
Magnitude
Timing
Northeast3
Less (H) to more (C) drought
More high flow events in winter
(+80% in northern areas under
higher emissions).
Lowest weekly flow projected to
decrease -10% by 2 100.
For a 2.5°C increase in average
temperature and 17.5% (17.5
cm) increase in precipitation,
flow at the mouth of the
Susquehanna will increase 24%
+ 13% (1 1.8 + 6.7 cm) (Najjar et
al, 2000). Neff et al. (2000)
projects decrease of 4% to
increase of 24%. Moore et al.
(1997) estimates decrease of
2 1-31% by 2 100
Peak (spring) flows advancing
10 to >14 days by 2100.
Low summer flows extended
almost 1 -month under higher
emissions
Southeast11
Slightly drier
(C) to more
precipitation
in long term
(H)
Midwest0
Small increase in
soil moisture (H);
small decrease (C)
Overall decrease
in river levels (i.e.,
decrease in
runoff?)
Great Plains'1
Decrease in soil
moisture in
large part of
region
Great Lakes
area runoff
decrease up to
32% (Magnuson
et al., 1999)
West6
Summer runoff
less by 20% of
annual total
(Knowles and
Cayan, 2002)
Pacific
Northwest'
>
-------
There are substantial regional and model differences in projections for precipitation
changes. The biggest average increases are projected for the Pacific Northwest and Midwest
(10-30% by 2100) and the Northeast (up to 25% by 2100), with smaller increases in the Great
Plains (13% by 2100), and variable projections for the Southeast (10% decrease to 20% increase
by 2100) and the West (e.g., doubling of winter precipitation over California, but decreased
precipitation over some parts of the Rockies) (NAST, 2001). Most of the increase in
precipitation is projected to occur during the winter, with more frequent and/or more intense
storm events, and with more winter precipitation as rain instead of snow (e.g., Polsky et al,
2000; Magnuson et al., 1997).
Most models project increases in evapotranspiration—due to increased temperature rather
than increased summer precipitation—leading to a net decrease in soil moisture and a greater
likelihood of late-summer drought (NAST, 2001).
A.1.2. HYDROLOGY
There are also secondary drivers that are important in structuring aquatic ecosystems that
will be altered by climatic changes, especially hydrologic regimes (e.g., Poff and Ward, 1989;
Richter et al., 1996). Projected hydrologic changes, driven by climate-associated changes in
temperature and precipitation, include changes in the magnitude, timing, frequency, and duration
of various flow events. Together these projected hydrologic changes will result in redistributing
stream flow (Hayhoe et al., 2007).
In North America, projected changes in average stream flow range from an increase of
10-40% (at high latitudes) to a decrease of about 10-30% (in mid-latitude western North
America) by 2050 (Milly et al., 2005). Consistent with this large-scale pattern, Hayhoe et al.
(2007) predict an increase in stream flow in the northeastern U.S., ranging from and increase of
9-18% in the southwest part of the region to an increase of 11-27% in the northeast part of the
region. In an earlier study, in the U.S. Northeast, however, average stream flow was projected to
decrease by an average of 21-31%, reflecting a temperature-associated increase in
evapotranspiration that will exceed small net changes in precipitation (increases in winter and
spring and decreases in summer and fall) (Moore et al., 1997). Hayhoe et al. (2007) attributes
these differences to their use of updated forcing scenarios that include projected increases in
winter precipitation. Patterns of stream flow also are projected to change in the Northeast, with
increases in stream flow occurring mainly in the winter and spring, but with lower stream flow in
the summer and fall (Hayhoe et al., 2007). Climate changes in the Northeast may also include
more intense thunderstorms, especially in the summer. If this happens, it could result in greater
variability and "flashiness" of stream flows (Moore et al., 1997). In the Great Lakes Basin,
A-5
-------
average basin runoff is projected mainly to decrease up to 32% in response to increased
temperatures, despite precipitation increases (Magnuson et al, 1997).
In snow-pack dominated regions, the combination of warming temperatures, a shift
toward less winter precipitation as snow, and snow-melt occurring earlier will transition the peak
runoff from spring to late-winter/early spring (Barnett et al., 2005). Predicted shifts in peak
runoff (anywhere from about two weeks to one month earlier) by the end of the century are
anticipated (Dettinger et al., 2004; Hayhoe et al., 2007).
Rain-dominated streams are expected to be especially responsive to altered precipitation
patterns, with runoff, flow variability, and flood frequency responding directly to changes in
precipitation. These streams may also respond to increased variability in precipitation, impacting
flood frequency and variation in flow; and to increasing temperatures, causing decreases in
runoff (Poff at al., 1996). In a Mid-Atlantic perennial flow (rain-dominated) stream, with a 25%
modeled increases in precipitation, Poff et al. (1996) predicts mean flow (runoff) to increase
about 10-15%, flow variability to increase about 20-25%, and flood frequency to more than
double (from 1.1 to 2.5 floods per year). Doubling the coefficient of variation of precipitation
had an even more dramatic effect on these hydrologic characteristics.
In the Mid-Atlantic region, using the Susquehanna River as a model, Neff et al. (2000)
estimated annual changes in stream flow by 2100. Estimated changes range from -4%, based on
the Canadian Climate Center (CCC) model (Flato et al., 2000; McFarlane et al., 1992) to +24%,
based on the Hadley model (Johns et al., 1997). Both models project increased stream flows
during the winter, with peak stream flow occurring up to 1 month earlier.
Even with net projected increases in annual precipitation for many regions of the U.S.,
increased durations of low flows and increased frequency of summer droughts are also projected.
In California's Sacramento/San Joaquin Basin, Knowles and Cayan (2002) predict a 20% loss in
the amount of annual stream flow occurring during the summer (April-July) by 2100. Dettinger
et al. (2004) made a similar prediction for the Merced, American, and Carson Rivers in
California, including a decrease in summertime low flows and reduced soil moisture.
Finally, the increased temperatures will lengthen the growing season, as well as increase
evapotranspiration during the warm months. These could have the net effect of reducing
groundwater recharge of streams, which would increase the severity of summer dry periods
independent of rainfall. This could, however, be mediated to some extent by the increased CC>2
concentrations, which reduce evapotranspiration (e.g., Gedney et al., 2006).
A.1.3. WATER TEMPERATURE
In addition to hydrologic alterations, changes in stream temperature are also expected.
The IPCC report projects a continued increase in global air temperature at approximately 0.2°C
A-6
-------
each decade (IPCC, 2007). Although these are global averages, regional models also support
these projections. For example, Mid-Atlantic regional models project that average air
temperature will increase 2.6-5.0°C by 2100 (Polsky et al, 2000; Barren, 2001). These
temperature changes will increase the maximum, average, and minimum stream temperatures as
well as the number of degree days and the rate of degree day accrual (Note: degree days are the
cumulative sum of average daily temperatures above a baseline. For example, if the baseline is
10°C, then one day with an average temperature of 12°C contributes 2°C days).
Though a relationship between increasing air and water temperatures is expected, the
magnitude and seasonal patterns of changes in stream and river water temperatures are likely to
vary regionally. These regional differences will be due to water source influences (surface
versus ground water), watershed characteristics, and season. Stephan and Preudhomme (1993)
estimated weekly average water temperatures in °C (excluding the ice cover period) to be a factor
of 0.86 times the weekly average air temperatures for 11 streams in the Mississippi River Basin.
Eaton and Scheller (1996) used this same linear relationship between air temperature and stream
water temperature to estimate probable loss offish habitat due to global warming projections.
However, Mohseni et al. (2003) claim that the relationship between air and water temperatures is
better explained by an S-curve such that at higher air temperatures, stream temperature increases
level off due to evaporative cooling. In the Upper Rhone River, Daufresne et al. (2003) showed
a clear, though non-linear, correspondence between long-term increases in annual average air
temperature (increase of about 1.0°C, 1979-1999) and average annual water temperature
(increase of about 0.6°C, 1979-1999). Annual patterns were similar, but average annual water
temperature did not correspond with average air temperature perfectly, suggesting possible
influences of other factors such as annual variations in flow conditions or snow melt. In a review
of the thermal regime of rivers, Caissie (2006) showed that thermal regime is strongly influenced
by meteorology, river conditions, and geographic setting.
A.1.4. HABITAT
Stream hydrologic patterns control habitat stability, channel formation and maintenance
(Poff et al., 1996), and they define composition, structure, and functioning of aquatic
assemblages (Richter et al., 1996, Poff and Allan, 1995). Changes in hydrologic pattern,
especially flood frequency, episodic runoff events frequency and intensity (including
"flashiness"), peak runoff magnitude, and flow total, will alter stream habitat and its dynamics.
Flow dynamics not only influence sediment supply and transport and, therefore, channel form,
but water volume also influences the amount of available habitat and water quality. Seasonal
patterns of flow magnitude, duration and frequency of runoff events, and other parameters
strongly influence the types of species that can inhabit an area (Poff et al., 2002). As a result,
A-7
-------
over time, regional changes in hydrologic regime are expected to modify habitat, species
composition, and ecological interactions.
In the Southwest, Grimm et al. (1997) concluded that neither hydrologic nor climate
models were sufficiently developed at that time to predict magnitude and direction of future
climate changes. Instead, they identified stream and river surface flows in the arid Southwest as
being particularly vulnerable to even small changes in precipitation, with even modest decreases
in precipitation potentially causing large decreases in stream flow. In addition, increased
temperatures could increase the likelihood of winter/early spring precipitation as rain instead of
snow and could increase the likelihood of severe episodic flooding, while increased temperatures
may also be associated with increased drought conditions during the summer. Associated habitat
changes could include changes in riparian vegetation, shifts from perennial to intermittent flow,
loss of aquatic habitat, and alterations in nutrient retention and instream production (Grimm et
al., 1997). More recent evaluation of the numerous climate model outputs available provide
consistent projections for existing and future decreases in annual precipitation minus evaporation
(increased aridity), especially during the winter (Seager et al., 2007), which will have substantial
negative impacts on availability and condition of stream and river aquatic habitat.
A.1.5. POLLUTANT BEHAVIOR
Stream water quality is expected to respond to changes in runoff magnitude and timing.
Reduced flow in summer combined with increased temperatures will likely decrease dissolved
oxygen (DO) concentrations, while increased storm frequency and magnitude are likely to
increase introduction of silt and pollutants (Poff et al., 2002). In the Mid-Atlantic region,
increased stream flow in the winter and the spring is expected to degrade water quality due to
increased inputs of nutrients, sediments, and toxicants (Neff et al., 2000; Rogers and McCarty,
2000). For example, nitrate loads have a high, positive correlation with stream flow in this area,
r2= 0.8, (Neff et al., 2000). However, nitrate loads are projected to decrease in July and August
associated with projected decreases in stream flow, which could, in part, ameliorate low (DO)
conditions. In boreal streams in northwestern Ontario, Schindler et al. (1996) reported
substantially reduced runoff during warm, dry periods in the 1970s and '80s that coincided in
magnitude with similar increases in temperature and decreases in precipitation that are projected
for future climate change in that region. There was also a substantial reduction in stream export
of phosphorus in association with these periods, though not in nitrogen export. In the Northeast,
overall drier conditions and reduced stream flow are expected to increase watershed retention of
non-point source nutrients, to reduce nutrient runoff, and to reduce erosion; increased
thunderstorm intensity could increase episodic erosion and nutrient loading (Moore et al., 1997).
These region-specific examples help define general expectations (i.e., for all regions) for
A-8
-------
pollutant loading and other water quality (e.g., DO) changes that should be expected in
association with climate-driven changes in temperature and runoff.
A-9
-------
APPENDIX B
DESCRIPTIVE AND SUPPLEMENTAL RESULTS FOR CASE STUDY 2:
BIOLOGICAL ASSESSMENT IN THE PRESENCE OF CLIMATE CHANGE
-------
B.I. RESULTS OF CORRELATION ANALYSES
The fish and benthic invertebrate response variables that showed the strongest responses
in these correlations were the 2005 version of Maryland's fish index of biotic integrity (IB I), fish
taxa richness, total number offish, the 2005 version of Maryland's B-IBI, total benthic taxa
richness, and total EPT taxa richness (Southerland et al, 2005). A series of these stressor
response relationships are shown below, and are used for further exploration of stressor-response
models.
Relationships between flashiness and a variety of environmental variables were
investigated to identify variables that are potential stressors related to flashiness. Flashiness is a
hydrologic characteristic that can be reflective of stream and surrounding watershed alterations,
and might also be responsive to climate change. Figure B-l shows the relationships between
Baker's flashiness index scores (Baker et al., 2004) and eight other environmental parameters
tested (dissolved oxygen, phosphate concentration, physical habitat index, conductivity,
embeddedness, percent urban, impervious surface, and total phosphorus). In this and subsequent
figures, the solid line is the locally weighted scatter plot smoothing line (LOWESS). LOWESS
smoothing was done in SystatlO by running along the x values and finding predicted values from
a weighted average of nearby y values. The surface is allowed to flex locally to better fit the
data. For the LOWES S, the degree to which the line or surface (tension) is allowed to flex
locally to fit the data to 0.5 was specified, meaning that half the points are included in the
running window. The strongest relationships were between percent urban land use and Baker's
flashiness index score, and between impervious surface and flashiness. Both of these are factors
contribute to alterations in watershed runoff that result in greater "flashiness." Some other
variables that are closely associated with runoff, such as nutrient concentrations, had relatively
weak relationships with flashiness (e.g., total phosphorus and total organic carbon).
Figure B-2 shows the relationships between physical habitat index (PHI) and
macroinvertebrate and fish IBI scores in the MBSS data. Overall, as habitat condition improved
(PHI increased), the fish and benthic IBIs increased.
Comparisons were made between three fish response variables and a suite of
environmental parameters. Figure B-3 shows the relationships between fish IBI scores and
dissolved oxygen (DO), instream habitat, temperature, channel flow, flashiness, and impervious
surface. Fish IBI scores increased with increasing DO, habitat score, and channel flow. Fish IBI
declined with increased impervious surface and flashiness. The lack of relationship between fish
IBI and temperature is at least, in part, related to the factors that collections are made only during
a seasonal (summer/fall) index period, and that these analyses are being conducted on a
combination of only two very closely related ecoregions.
B-2
-------
15
en
to
O
10
01000
Q0100
Q0010
00001
00
Bakers RasHress Index
0.2 04 0.6 0.8
Bakei's Rashiness Index
1.0 1.2
0.2 0.4 0.6 0.8
Bakei's Rashiness Index
1.0 1.2
-_ 1000
100
10
00
0.2 0.4 06 0.8
Bakei's Rashiness Index
1.0 1.2
td
oo
V,
Embeddedne
90
80
70
60
50
40
30
20
10
0
...... . ...
:|p^:
A , ,*
0 0.2 0.4 Q6 0.8 1.0 1.
Bakei's Rashiress Index
1 9
X
03
~o
— DR
03
C
^ 0.6
to
D:
-™ 0.4
ffl
" 0.2
2 QC
:
• ««
n 0.10 1.00
%Urten
10.00
1.000
0100
0010
Q01 0.10 1.00 1QOO
Irrpeiviousness sufiace
0001
.*•-?.
•.% t*'***' "
00 0.2 0.4 06 0.8 1.0 1.2
Bakei's Rashiness Index
Figure B-l. Flashiness vs. environmental variables; the solid line is the locally weighted scatter plot smoothing
line (LOWESS).
-------
td
05
• • •• •••••••
• •• • ••• I
I I
I I
20 30 40 50 60 70 80 90 100 110
Physical habitat index
05
• • • m
I I
I I
20 30 40 50 60 70 80 90 100 110
Physical habitat index
Figure B-2. Index comparison; the solid line is the locally weighted scatter plot smoothing line
(LOWESS).
-------
td
5 10
Dssolved Qcygen (mg/L)
15
10 20 30
Temperature (°C)
5
4
3
2
1
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Baker's Rashness Index 0.01
• •••••• ••§ •• ••• ••
40
M • M»
0.10 1.00 10.00
Imperious surface
Dni| i i i i
y> ^ ^* y> ^ ^
Row
Figure B-3. Fish vs. environment; the solid line is the locally weighted scatter plot smoothing line (LOWESS).
-------
Figure B-4 shows the relationships between abundance offish and the same suite of
environmental variables (DO, habitat score, temperature, channel flow, flashiness, impervious
surface). Fish abundance increased with increasing DO, though the relationship is very weak.
Fish abundance also increased with increasing habitat score, PHI, and channel flow. Fish
abundance declined with increasing impervious surface, though again in a very weak
relationship. There was no meaningful relationship between fish abundance and flashiness.
The number offish species present (richness) also increased with increasing instream
habitat score and with flow (Figure B-5). Fish richness increased with increasing DO, but with a
weaker relationship. Fish species richness declined with increased impervious surface, but
showed no real relationship with flashiness.
For the benthic macroinvertebrate community, species richness increased very slightly
with increasing PHI, and decreased slightly with increasing flashiness and impervious surface
(Figure B-6). However, it appeared that total benthic taxa richness was not a strong response
variable with most parameters. In comparison, EPT taxa richness showed reasonable stressor-
response relationship with several environmental parameters (Figure B-7). The number of EPT
taxa increased with increasing PHI, and decreased with increasing DOC, phosphorus,
conductivity, embeddedness, Baker's flashiness index score, and impervious surface. EPT taxa
had strong relationships with all these variables and were, therefore, selected as a primary
response variable to examine potential climate change effects.
Table B-l summarizes the hydrologic parameters of greatest interest in the analysis. Low
flow events, high flow events, and Baker's flashiness were estimated by the FTSE model. Low
flow and high flow events are the number of events during a year below the 25th and above the
75th percentiles, respectively, of the area-weighted mean discharge of all streams.
B-6
-------
40
td
to 30
0}
0)
c
_c
o
20
10
0
0 5 10
Dissolved Ckygen (mg/L)
40
30
0)
c
.
o
CD 20
x
.2
C/)
10
o
40
15
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Eater's Rashness Index
w 30
C/)
0)
C
.c
o
CD 20
X
CD
10
0
20 30 40 50 60 70 80 90 100 110
Physical Habitat Score
40
to 30
C/)
0)
C
_c
o
120
JS
10:-
0.01 0.10 1.00 10.00
Impervious surface (%)
0-x \>
Row
Figure B-4. Fish richness vs. others; the solid line is the locally weighted scatter plot smoothing line (LOWESS).
-------
0 5 10 15
Dissolved Organic Carbon (mg/L)
0.0001 0.0010 0.0100
Phosphorus (mg/L)
0.1000
.g
TO
50
40
30
^ 20
TO
-5
10
0
20 30 40 50 60 70 80 90 100 110
Physical habitat index
td
oo
.o
TO
TO
•5
50
40
30
20
10
15 25
Temperature (°C)
35
.o
TO
TO
•5
50
40
30
20
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Baker Flashness Index
.o
ro
TO
•5
0.01 0.10 1.00 10.00
Impervious surface (%)
Figure B-5. EPT taxa vs. environmental 2; the solid line is the locally weighted scatter plot smoothing line
(LOWESS).
-------
0 5 10
Dssdved Qganic Caibcn (rng'L)
30
_
20
10
Q0010 Q0100
Phosphoius (rrgl)
30
10
Q1000
20 30 40 50 60 70 80 90 100 110
Physical habitat index
100 1000
Ccndidivity(jjS/crr)
30
td
JS
t
10
30
0 102D30405060708090100
Enteddectiess
10
00
0.2 0.4 06 08
Baker Rashness Index
1.0 1.2
010 1.00
Inpervious suface (%)
10.00
30
™ in
0) A)
10
20 30 40 50 60 70 80 SO 100 110
Physical Habitat Index
Figure B-6. EPT and environmental variables; the solid line is the locally weighted scatter plot smoothing line
(LOWESS).
-------
CO
CO
03
_C
O
'i_
CD
0.
LU
30
20
10
• M • ••• •••••• I
-0.003 -0.002 -0.001
Fall
0.000
0.000 0.001 0.002 0.003 0.004 0.005 0.006
Rise
Figure B-7. EPT and hydrology; the solid line is the locally weighted scatter
plot smoothing line (LOWESS).
Table B-l. Summary of hydrologic parameters used in analyses
Number of cases
Minimum
Maximum
Median
Mean
Standard Deviation
Baker's Flashiness Index
764 (streams)
0.132
1.121
0.362
0.385
0.179
Palmer Hydrologic Drought Index
15 (months)
-4.24
4.75
1.02
0.819
2.56
B-10
-------
vvEPA
United States
Environmental Protection
Agency
National Center for Environmental Assessment
Office of Research and Development
Washington, DC 20460
Official Business
Penalty for Private Use
$300
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERM IT NO. G-35
------- |