United States
Environmental Protection
Agency
EPA/600/R-11/058Fa | January 2012 | www.epa.gov
Vulnerability Assessments in Support
of the Climate Ready Estuaries Program:
A Novel Approach Using Expert Judgment
Volume I
Results for the San Francisco
Estuary Partnership
United States Environmental Protection Agency
Office of Research and Development, National Center for Environmental Assessment
-------
EPA/600/R-ll/058Fa
January 2012
Vulnerability Assessments in Support of the Climate Ready
Estuaries Program: A Novel Approach Using Expert
Judgment
Volume I
Results for the San Francisco Estuary Partnership
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
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
The San Francisco Estuary Partnership (SFEP), the San Francisco Bay Conservation and
Development Commission, and the Environmental Protection Agency (EPA) collaborated on an
ecological vulnerability assessment, using a novel methodology based on expert judgment, to
inform adaptation planning under EPA's Climate Ready Estuaries Program. An expert
elicitation-type exercise was designed to systematically elicit judgments from experts in a
workshop setting regarding climate change effects on two key ecosystem processes: sediment
retention in salt marshes and community interactions in mudflats. Specific goals were to assess
(1) the relative influences of physical and ecological variables that regulate each process,
(2) their relative sensitivities under current and future climate change scenarios, (3) the degree of
confidence about these relationships, and (4) implications for management. For each process, an
influence diagram was developed identifying key process variables and their interrelationships
(influences). Using a coding scheme, each expert characterized the type and sensitivity of each
influence under both current and future climate change scenarios. The experts also discussed the
relative impact of certain influences on the endpoints. This report shows how particular
pathways in such diagrams can be linked to management options in the context of planning
documents to identify opportunities for 'mainstreaming' adaptation.
Photo Credits (front cover):
Workshop (Amanda Babson), Godwit (Brian Currie), Western Sandpiper (George Jameson), and
Salt marsh (Ingrid Taylar)
Preferred Citation:
U.S. EPA (Environmental Protection Agency). (2012) Vulnerability Assessments in Support of the Climate Ready
Estuaries Program: A Novel Approach Using Expert Judgment, Volume I: Results for the San Francisco Estuary
Partnership. National Center for Environmental Assessment, Washington, DC; EPA/600/R-ll/058Fa. Available
online at http://www.epa.gov/ncea.
11
-------
CONTENTS
LIST OF TABLES vi
LIST OF FIGURES vii
LIST OF ABBREVIATIONS AND ACRONYMS viii
PREFACE ix
AUTHORS, CONTRIBUTORS AND REVIEWERS xii
ACKNOWLEDGMENTS xiii
EXECUTIVE SUMMARY xiv
1. INTRODUCTION 1-1
1.1. BACKGROUND 1-1
1.2. PURPOSE AND SCOPE 1-1
1.2.1. Purpose 1-1
1.2.2. Scope 1-2
1.3. ROADMAP FOR THE REPORT 1-4
2. EXPERT ELICITATION EXERCISE 2-1
2.1. JUSTIFICATION FOR METHOD 2-1
2.1.1. Definition and Uses 2-1
2.1.2. Novel Application 2-1
2.2. WORKSHOP DESIGN AND METHODOLOGY 2-2
2.2.1. Workshop Goals and Objectives 2-2
2.2.2. Approach and Methodology 2-2
2.2.2.1. Influence Diagrams 2-3
2.2.2.2. Climate Scenarios 2-6
2.2.2.3. Expert Facilitation 2-6
2.2.2.4. Coding Scheme and Exercise 2-9
2.2.2.5. Typologies for Understanding Influences and Sensitivities 2-11
2.2.2.6. Understanding Relative Impacts of Influences 2-13
2.2.2.7. Key Questions 2-13
2.3. RESULTS 2-14
2.3.1. Sediment Retention 2-15
2.3.1.1. Group Influence Diagram 2-15
2.3.1.2. Influence Types and Degrees 2-15
2.3.1.3. Influence Sensitivity 2-20
2.3.1.4. Relative Impact 2-23
2.3.1.5. Confidence 2-23
2.3.1.6. Interacting Influences 2-23
2.3.2. Community Interactions 2-27
2.3.2.1. Group Influence Diagram 2-27
2.3.2.2. Influence Types and Degrees 2-30
2.3.2.3. Influence Sensitivity 2-33
2.3.2.4. Relative Impact 2-36
2.3.2.5. Confidence 2-39
iii
-------
CONTENTS (continued)
2.3.2.6. Interacting Influences 2-39
2.4. DISCUSSION OF ADAPTATION STRATEGIES 2-39
2.4.1. Restoration and Conservation 2-42
2.4.2. Sediment Management 2-43
2.4.3. Planning and Monitoring 2-44
3. MAKING THE LINK TO MANAGEMENT 3-1
3.1. USING INFORMATION ON INFLUENCE TYPE and DEGREE,
SENSITIVITY AND RELATIVE IMPACT TO IDENTIFY KEY
MANAGEMENT PATHWAYS 3-1
3.1.1. Crosswalks: Influence Type and Degree, Sensitivity and Relative
Impact 3-1
3.1.1.1. Sediment Retention Crosswalk 3-1
3.1.1.2. Community Interactions Crosswalk 3-6
3.1.1.3. Information Gaps 3-11
3.1.2. Identifying Key Pathways for Management 3-13
3.2. TOP PATHWAYS AND IMPLICATIONS FOR ADAPTATION
PLANNING 3-15
3.2.1. Top Pathways and Associated Adaptation Options 3-16
3.2.1.1. Sediment Retention Top Pathways 3-16
3.2.1.2. Community Interactions Top Pathways 3-21
3.2.1.3. Top Pathway Caveats 3-24
3.2.2. Adaptation Planning 3-25
4. CONCLUSIONS 4-1
4.1. INSIGHTS FROM THE WORKSHOP EXERCISE 4-1
4.1.1. Group Influence Diagrams 4-1
4.1.2. Characterization of Influences 4-2
4.2. APPLICATION OF WORKSHOP RESULTS 4-4
4.2.1. Top Pathways for Management 4-4
4.2.2. Mainstreaming Adaptation into Planning 4-6
4.3. GENERAL CONCLUSIONS 4-7
4.3.1. Transferability of Results and Method 4-7
4.3.2. Utility of Method for Rapid Vulnerability Assessments 4-8
5. REFERENCES 5-1
APPENDIX A. DEVELOPMENTAL PROCESS FOR CLIMATE READY
ESTUARIES VULNERABILITY ASSESSMENT A-l
APPENDIX B. EXPERT ELICITATION WORKSHOP PREPARATION
AND IMPLEMENTATION B-l
IV
-------
CONTENTS (continued)
APPENDIX C. PARTICIPANT HANDOUT ON CLIMATE SCENARIOS C-l
APPENDIX D. PARTICIPANT HANDOUT ON CONFIDENCE D-l
-------
LIST OF TABLES
1-1. Breakout group participants for the expert elicitation workshop 1-4
2-1. Summary of Climate Scenario A ("lower-range" scenario) and Climate Scenario
B ("higher-range" scenario): averages for midcentury 2-7
2-2. Coding scheme used during the workshop exercise to characterize influences 2-9
2-3. Coding scheme used during the workshop exercise to characterize interactive
influences 2-10
2-4. Coding scheme used during the workshop exercise to characterize confidence 2-11
2-5. Sediment Retention variable definitions clarified during group discussion 2-16
2-6. Sediment Retention group influence judgments 2-17
2-7. Sediment Retention group confidence for influences with agreement 2-26
2-8. Sediment retention group interactive influences with agreement under current
conditions and Climate Scenarios A and B 2-26
2-9. Community interactions variable definitions clarified during group discussion 2-29
2-10. Community interactions group influence judgments 2-31
2-11. Community Interactions group confidence for influences with agreement 2-40
2-12. Adaptation strategies and associated top pathways for management 2-41
3-1. Sediment Retention group crosswalk for comparison of influence type and degree,
sensitivity and relative impact for current conditions and climate scenarios 3-2
3-2. Community Interactions group crosswalk for comparison of influence type and
degree, sensitivity and relative impact for current conditions and climate scenarios 3-7
VI
-------
LIST OF FIGURES
1-1. Vulnerability assessment process 1-6
2-1. The North Bay (San Pablo Bay) of the San Francisco Bay area, with expanded
detail showing China Camp State Park 2-3
2-2. Simplified influence diagram for sediment retention 2-5
2-3. Sediment Retention group influence diagram 2-16
2-4. Sediment Retention group summary influence diagram of sensitivities under
current conditions 2-21
2-5. Sediment Retention group summary influence diagrams of sensitivities: variance
across current conditions and two climate scenarios 2-22
2-6. Sediment Retention influences indicated as having high relative impact under
current conditions 2-24
2-7. Sediment Retention group influences indicated as having high relative impact
under climate scenarios 2-25
2-8. Community Interactions group influence diagram 2-28
2-9. Community Interactions group summary influence diagram of sensitivities under
current conditions 2-34
2-10. Community Interactions group summary influence diagrams of sensitivities:
variance across current conditions and two climate scenarios 2-3 5
2-11. Community Interactions influences indicated as having high relative impact under
current conditions 2-37
2-12. Community Interactions group influences indicated as having high relative impact
under climate scenarios 2-38
3-1. Sediment Retention example pathway 3-14
3-2. Top pathways for management of the Net Accretion/Erosion endpoint 3-17
3-3. Top pathways for management of the Shorebirds endpoint 3-18
vn
-------
LIST OF ABBREVIATIONS AND ACRONYMS
CRE Climate Ready Estuaries
NEP National Estuary Programs
SFEP San Francisco Estuary Partnership
BCDC Bay Conservation and Development Commission
H high
FIH high evidence/high agreement
HL high evidence/low agreement
LH low evidence/high agreement
LL low evidence/low agreement
L low
DO dissolved oxygen
CCMP Comprehensive Conservation and Management Plan
TMDL Total Maximum Daily Load
Vlll
-------
PREFACE
This report was prepared by the Global Change Research Program (GCRP) in the National
Center for Environmental Assessment of the Office of Research and Development at the
U.S. Environmental Protection Agency (EPA), in collaboration with the San Francisco Estuary
Partnership (SFEP) of the National Estuary Program (NEP) and the San Francisco Bay
Conservation and Development Commission (BCDC) of the state of California. The report
presents the results of a pilot, targeted, climate change vulnerability assessment for selected
ecosystem processes of San Francisco Bay mudflat and salt marsh ecosystems, using a new
methodology based on expert elicitation techniques. Both the place-based results and the
methodology itself are intended to serve not only the SFEP and the larger NEP community, but
also other natural resource managers who are interested in adapting to the impacts of climate
change on valued ecosystems.
This project originated from a confluence of two efforts. The first was a 2008 interagency
report led by the EPA GCRP under the auspices of the U.S. Climate Change Science Program,
entitled Adaptation Options for Climate-Sensitive Ecosystems and Resources, which laid out
general principles for evaluating vulnerabilities and identifying associated adaptation approaches
and called for application of these concepts in place-based, adaptation planning activities. In that
same year, EPA's Office of Water and Office of Air and Radiation launched a series of pilot
projects under a new, Climate Ready Estuaries (CRE) program designed to provide targeted
assistance to NEPs to assess climate change vulnerabilities and plan for adaptation. Based on the
complementarities of the two efforts, the EPA GCRP joined forces with the CRE program to
support two of its original pilot projects. These were collaborative vulnerability assessments
with the San Francisco Estuary Partnership (this Volume I report) and the Massachusetts Bays
Program (Volume II of this two-report set).
The San Francisco Estuary Partnership is a coalition of resource agencies, nonprofits,
citizens, and scientists working to protect, restore, and enhance water quality and fish and
wildlife habitat in and around the San Francisco Bay Delta Estuary. Through a comprehensive
conservation and management plant first developed in 1993 and a strategic plan updated
periodically, the SFEP's current strategic priorities are to: promote watershed stewardship;
support climate change resiliency; be a resource for elected officials, decision makers, and the
public in making decisions that benefit the Estuary; and develop green infrastructure leadership.
As a CRE pilot partner in 2008, the SFEP was afforded technical support to begin a process to
identify climate change vulnerabilities of its estuarine resources, develop adaptation plans and
begin to implement selected actions within these plans. This project is a first step in this process.
IX
-------
The San Francisco Bay Conservation and Development Commission is the
federally-designated state coastal management agency for the San Francisco Bay segment of the
California coastal zone. It is dedicated to the protection and enhancement of San Francisco Bay
and to the encouragement of the Bay's responsible use. As a member agency of the SFEP with
extensive technical expertise in vulnerability assessment and adaptation planning through its
climate change planning program, BCDC was brought on board at the outset as technical lead
and full collaborator on project development and implementation.
After a kickoff meeting with local experts and stakeholders, the SFEP/BCDC/EPA team
elected to focus the current vulnerability assessment on a narrow subset of key physical and
biological processes for salt marsh and mudflat ecosystems. The results of this pilot effort are
contained in this report and represent a proof of concept for a new type of assessment exercise
rather than a comprehensive vulnerability assessment for the whole estuary. Thus it focuses on a
deeper examination of the climate sensitivities of two selected processes—sediment dynamics
and wading bird feeding habitat—that are integral to functional salt marsh and mudflat systems.
Given the multi-disciplinary assessment objectives and the limitations on available data and
modeling tools, a new method based on expert elicitation was developed in order to capture the
current understanding of the climate sensitivities of these systems as a starting point for
adaptation, which will be an iterative process as new ecosystem processes are added to the
analysis and as our understanding of the climate and management impacts grows. We hope that
this report will be a useful starting point for adaptation action and a methodological basis for
future work on climate change vulnerability assessments for estuarine systems.
We would like to acknowledge the major contributions of ICF International, Inc.
throughout this project, including conceptualization, methodology development, workshop
support, and report production. We would also like to express our appreciation to John Wilson,
John Whitler and Michael Craghan (EPA Office of Water), Jeremy Martinich (EPA of Air and
Radiation), and the rest of the CRE team for their leadership, partnership and many useful
discussions. Many thanks go to the participants of the expert elicitation workshop for sharing
their knowledge and judgments. We also appreciate the substantive contributions of our external
and internal EPA reviewers. Finally, we would like to thank Mike Slimak, Anne Grambsch, and
all of the EPA Global Change Research Program staff for their advice and numerous and
significant inputs to this project.
-------
This final document reflects consideration of all comments received during a formal
external peer review and 30 day public review period on an External Review Draft posted
September 8 to October 11, 2011.
Jordan West & Amanda Babson
Global Change Research Program
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Judy Kelly
Director
San Francisco Estuary Partnership
U.S. National Estuary Program
Steve Goldbeck & Sara Polgar
Chief Deputy Director & Coastal Program Analyst
San Francisco Bay Conservation and Development Commission
State of California
XI
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS
The National Center for Environmental Assessment (NCEA) within EPA's Office of
Research and Development (ORD) was responsible for the preparation of this report. Portions of
this report were prepared by ICF Incorporated under EPA Contract Nos. GS-10F-0124J and
EP-C-09-009. Jordan M. West served as the EPA Work Assignment Manager, providing
overall direction and coordination of the project, as well as coauthor.
AUTHORS TECHNICAL SUPPORT
U.S. EPA ICF Incorporated
Jordan M. West Susan Asam Elizabeth Strange Katy Maher
Amanda Babson Brock Bernstein Peter Schultz Katharine Hayhoe
Phil Duffy
CONTRIBUTORS
San Francisco Estuary Partnership Team
Judy Kelly, San Francisco Estuary Partnership
Steve Goldbeck and Sara Polgar, Bay Conservation and Development Commission
Expert Workshop Participants
Dave Cacchione, U.S. Geological Survey Michelle Orr, Philip Williams & Associates
John Callaway, UC San Francisco Dave Schoellhamer, U.S. Geological Survey
Chris Enright, CA Dept. of Water Resources Stuart Siegel, Wetlands and Water Resources
Letitia Grenier, SF Estuary Institute Mark Stacey, UC Berkeley
Bruce Jaffe, U.S. Geological Survey Diana Stralberg, PRBO Conservation Science
Jessica Lacy, U.S. Geological Survey Lynne Trulio, San Jose State University
Lester McKee, San Francisco Estuary Institute Isa Woo, U.S. Geological Survey
REVIEWERS
The following U.S. EPA reviewers and external peer reviewers provided valuable
comments on earlier drafts of this document:
Ted DeWitt and Henry Lee, EPA Office of Research and Development
Tristan Peter-Contesse and Lester Yuan, EPA Office of Water
Luisa Valiella, EPA Region 9
Carl Hershner, College of William and Mary
Taewon Kim, Stanford University
Drew M. Talley, University of San Diego
Laura M. Valoppi, U.S. Geological Survey
xn
-------
ACKNOWLEDGMENTS
We would like to acknowledge the support of A. Grambsch and M. Slimak of EPA's
National Center for Environmental Assessment in the Office of Research and Development, and
especially our colleagues in EPA's Global Change Assessment Staff for many helpful
discussions during this project. Special thanks to C. Weaver for technical contributions.
xin
-------
EXECUTIVE SUMMARY
The San Francisco Bay estuary is highly vulnerable to climate-related changes including
increased water temperatures, changes in precipitation and winds, and sea level rise. Impacts
such as increased inundation of coastal wetlands, changes in water availability and quality, and
altered patterns of sedimentation and erosion are increasingly interacting with other human
stressors such as extractive water uses and land use changes. Thus it is essential that estuary
managers become 'climate-ready' by: assessing the vulnerability of natural resources to climate
change; considering strategic choices among adaptation strategies in the near term; and engaging
in longer term planning based on a range of plausible scenarios of future change. In an era of
shrinking budgets coupled with increasingly complex decision-making needs—often taking
place in a context of uncertainty and incomplete information—managing natural resources in the
face of climate change will be challenging. There is a need for assessment methods that take
advantage of existing scientific expertise to help identify robust adaptation strategies, weigh
difficult trade-offs, and justify strong action, all in a timely and efficient manner.
The purpose of this project was to carry out a pilot vulnerability assessment for the San
Francisco Estuary Partnership's (SFEP) natural resources using expert judgment, the results of
which could be linked to adaptation planning. To this aim, EPA's Office of Research and
Development collaborated with SFEP and the San Francisco Bay Conservation and Development
Commission on a novel expert elicitation exercise for 'rapid' vulnerability assessment. A trial
exercise was carried out during a two-day workshop in which two groups of seven experts each
focused on two key ecosystem processes: sediment retention in salt marshes and community
interactions of shorebirds with their predators and prey (see Figure ES-1). The exercise, which
was based on formal expert elicitation techniques but tailored specifically for qualitative analysis
of ecosystem processes, was designed to glean expert information on the sensitivities of
ecosystem process components under future climate scenarios. This was followed by group
discussions of the implications of the results for management in light of climate change, as well
as feedback on the exercise itself.
Sensitivities and Potential Adaptation Responses
Using the experts'judgments on the sensitivities of key ecosystem process components
to future climate conditions, it is possible to identify 'top pathways'for which there are
available adaptation options. After creating influence diagrams showing the relationships
among key process variables (see Figures ES-2 and ES-3), the experts generated information on
xiv
-------
Salt Marsh
Sediment Retention
Community Interactions:
Shorebirds
The balance between the processes of removal and Access of Western sandpiper and Marbled godwit
deposition ofsediment to mudflatprey
Figure ES-1. Selected ecosystem processes for the pilot vulnerability
assessment.
xv
-------
Sediment Flux Sediment Size
Key
'V' Increasing relative impact
Increasing sensitivity
Threshold
Figure ES-2. Top pathways for management of the Net Accretion/Erosion
endpoint. Colors are used to distinguish different pathways. Red symbols
highlight potential changes under future climate conditions.
xvi
-------
Sediment
Res us pension
and Deposition
Bed Sediment
Characteristics and
Quality
Key
Increasing relative impact
Increasing sensitivity
Threshold
Figure ES-3. Top pathways for management of the Shorebirds endpoint.
Colors are used to distinguish different pathways. Red symbols highlight
potential changes under future climate conditions.
xvn
-------
which relationships may show, under future climate change: (1) increasing relative impact on the
overall process; (2) increasing sensitivity; and (3) abrupt threshold changes. Based on the
amount of expert agreement on each relationship, it is possible to identify 'top pathways' of
interest for management. Three top pathways for each process are described below, with
accompanying discussion of adaptation options for management.
Sediment Retention Green pathway: Two relationships in this pathway (see Figure ES-2)
were indicated by the experts as having increasing relative impact on net accretion and erosion
under climate change. The direct effect of organic accumulation through below-ground biomass
production already has a high relative impact on the overall process, and this relative impact is
expected to increase under sea level rise associated with climate change. Likewise, the effect on
freshwater inflow of reservoir management is expected to be of increasingly high relative impact
under climate change as freshwater supplies become increasingly variable and human demand
continues to increase. Management options under this pathway include:
Managing reservoirs for steady, lower-volume releases to regulate salinity and favor
native marsh vegetative productivity
Investigating optimal timing of releases relative to the growing season
Prioritizing releases designed for salinity maintenance compared to high volume pulses to
support mineral sediment transport
'Stepping up' Spartina (invasive cordgrass) eradication programs since increased salinity
regimes favor this invasive species.
Sediment Retention Purple pathway: The climate-related shift in this pathway (see
Figure ES-2) involves an increase in the sensitivity of net mineral accumulation to changes in
sediment size. This is a direct relationship, with larger grain sizes favoring net mineral
accumulation since larger grains deposit more readily, are harder to resuspend and provide larger
building blocks for accretion. Increasing sensitivity of net mineral accumulation to sediment size
relates to the fact that sediment flux, the other determinant of net mineral accumulation, is
expected to continue to decrease because of continuing processes responsible for historical
declines from peak sediment inputs in the past and because of potential changes in wave-driven
erosive processes. Management options under this pathway include:
• Investigating how changes in land cover (including changes from impervious to
permeable pavement systems) may affect sediment size
xviii
-------
• Managing reservoirs for high volume pulses to increase transport of larger grain
sediments to marshes
• Adjusting policies that prevent coarse sediment from entering the Bay, such as changing
Total Maximum Daily Load requirements to allow an increase in sediment loads for
streams that do not support salmonids
• Engaging with flood control districts to recouple stream sediments to wetlands
Sediment Retention Blue pathway: In this pathway (see Figure ES-2), the experts
identified the potential for an abrupt threshold change in the effect of wind-generated waves on
sediment flux, from a direct to an inverse relationship of increasing relative impact on the overall
process. Under current conditions, wind-driven wave action has a net positive effect on sediment
flux onto salt marshes, as greater wave energy can mobilize and increase rates of sediment
transport from bays and adjacent mudflats deep into marsh systems. However, under future
climate conditions a threshold may be crossed because of a change in wave character as water
depth increases due to sea level rise. In deeper water, waves behave differently, with less wave
energy available for resuspension of bottom sediments and more energy delivered to the marsh
edge, leading to increased erosion. Management options under this pathway include:
• Monitoring wind, waves and sediment fluxes to detect the threshold shift when it occurs,
and in the meantime preparing a response plan for after the shift
• Building berms or restoring oyster reefs as protective barriers against wave energy
• Locating sites to deposit dredge materials with a goal of enhancing sediment
concentrations on mudflats adjacent to marshes
• Prioritizing development of new tools for reducing wave action on the front of marshes
Community Interactions Green pathway: Both relationships in this pathway (see
Figure ES-3) were indicated as having increasing relative impact on the shorebirds endpoint
under climate change. A strong direct effect of landscape mosaic (defined as a mixture of
habitats for secondary foraging, roosting, and cover from predators that support efficient use of
mudflat feeding habitat) already has a high relative impact on shorebirds; and this may increase
even further under climate change as mudflat habitats become scarcer and smaller in extent.
Likewise, the effect on landscape mosaic of restoration is expected to be of increasingly high
relative impact under climate change as individual habitats within the mosaic are differentially
xix
-------
impacted by temperature increases, altered precipitation patterns, and water diversions in a
context of continuing land use change. Management options under this pathway include:
• Assessing and mapping landscape mosaics to detect changes and support management at
the landscape scale
• Managing landscape mosaics through spatial planning designed to prioritize where and
how to restore which habitats, in order to ensure a continuum of wetland and upland
ecosystems which could migrate inland as sea level rises
• Including 'threshold landscapes' (those about to change from one set of dominant
processes to another, or from one state to another) in consideration for restoration
• Supporting legislation or incentives that encourage moving back or blocking of
development on lands where there is restoration potential now or in the future.
Community Interactions Purple pathway: This pathway (see Figure ES-3) shows a high
relative impact of mudflat prey populations on shorebirds. The abundance of prey per unit area
will become increasingly important (of increasing relative impact) under climate change as
spatial extent of mudflats shrink with sea level rise. Also, there is an abrupt threshold response
of the prey community itself to water quality (specifically, dissolved oxygen), from a direct to a
very strong direct effect under climate change. As decreases in dissolved oxygen occur with
climate change due to increased temperatures and/or eutrophication, prey communities may flag.
A critical threshold may occur in the future if dissolved oxygen reaches low enough levels to
cause prey populations to crash. Management options under this pathway include:
• Protecting water quality through integrated water resources management, including
stormwater management and rainwater-harvesting (which also benefits water
conservation)
• Using permeable rather than impervious surfaces to reduce runoff
• Restoring riparian zones to act as natural filters
Community Interactions Blue pathway: This pathway (see Figure ES-3) contains
two relationships that the experts identified as sensitive to climate change. The extent of mudflat
available for foraging (i.e., the number of hours per acre that mudflats are exposed and therefore
accessible) has a direct effect on shorebird populations, and this may become increasingly strong
xx
-------
as a threshold effect under climate change. This is because extent of mudflat may become
limiting as sea level rises, with available foraging habitat becoming too limited to support
shorebird populations. At the top of the pathway, there is a relationship of increasing sensitivity
of freshwater inflow to water management practices (specifically, reservoir management and
upstream operations). This effect will become increasingly strong as freshwater flows from
alternate sources such as precipitation and tributaries become more variable and/or scarce under
climate change. This relationship connects back down to the shorebirds endpoint through a
series of linked variables having to do with sediment supply, transport and effects on bathymetry
(which helps determine extent of mudflat). Management options under this pathway include:
Managing reservoir water releases to mobilize and transport sediments (e.g., through the
use of sediment maintenance flushing flows)
Improving upstream operations to ensure greater availability of water (for more frequent
and/or intense pulse releases)
Employing integrated water resources management, with an emphasis on shifting from
storage more toward conservation uses
Developing methods for moving coarse sediments into the bay (e.g., by strategically
locating dredge spoil sites to enhance sediment supplies to mudflats)
Based on the nature and timing of the sensitivity, some actions can be taken
immediately while others require monitoring and planning for multiple potential futures. In
the case of relationships that are well understood and for which there are management options
available, the nature of the expected climate-related shift has implications for when managers
may want to take action. In the case of relationships for which the expected climate-related shift
is toward increasing relative impact (and especially where the relationship is already of high
relative impact under current conditions), action can be taken immediately to put management
options into place for positive effects on those pathways. In the case of relationships for which a
change in sensitivity is possible under future climate scenarios, the expectation of increasing
sensitivity could be considered a 'notification' to managers to further study the relationship in
order to anticipate the degree and timing of the impending sensitivity and prepare best
management responses. Finally, thresholds are a particular challenge, as it is often impossible to
predict exactly when a threshold response will occur. In these cases it will be important to
monitor threshold variables to identify the shift when it occurs; in the meantime a manager might
act to keep the system 'below' the threshold as long as possible, while preparing a plan for what
xxi
-------
to do if an unavoidable shift occurs. After a shift occurs, a manager could decide to manage the
system differently in its new state, or take no action and instead shift priorities to other goals.
Adaptation Planning
Relating top pathways and associated adaptation options to existing management
activities is a path forward for action. The top pathways described above were used to identify
adaptation options that could be applied to sensitive ecosystem process components. A variety
of additional pathways and associated adaptation options can be further explored using the
detailed tables of judgments and lists of strategies provided in this report. The next step toward
adaptation planning is to connect the top pathways and adaptation options to existing
management activities and plans.
Under its current goals, SFEP is already undertaking a variety of activities that can be
related to these adaptation options, as described in its annual, midterm and long-term planning
documents. These include specific restoration, sediment management, monitoring and research
projects and strategies. The climate change sensitivities and potential adaptation strategies
identified in this report can be cross-referenced to these activities, goals and objectives to
identify where existing work can be adjusted to better support adaptation. Some examples of
such cross-referencing are provided as a starting point for more comprehensive adaptation
planning during future planning cycles. The intent is that the results of this assessment can be
used to prioritize increased investment in projects that take into account specific, known climate
sensitivities and make use of particular adaptation options that will be most effective.
Assessment results can also assist in priority-setting for long term research and monitoring
investment. Besides identifying well-understood relationships, the exercise also revealed gaps in
understanding of the system that indicate a need for further investigation of some sensitivities as
well as tailored projects to develop new management tools in response.
'Mainstreaming' climate change adaptation into ongoing, iterative planning processes
will increase the ability of managers to identify win-win options, weigh multiple trade-offs,
and prepare for long-term changes. For SFEP as well as other National Estuary Programs and
organizations with well established planning processes, there are benefits to 'mainstreaming'
(continuously integrating) adaptation into ongoing planning, rather than developing a stand-alone
adaptation plan. The objective is to start with actions that have multiple benefits, i.e., that
contribute to current management goals while also responding to climate change. For example,
starting with the separate preexisting plans in support of Baylands Ecosystem Habitat Goals,
Subtidal Habitat Goals and Upland Habitat Goals, projects could be designed to coordinate
xxn
-------
across goals and restore landscape mosaics that will support valued species such as shorebirds
not only today, but also under projected climate change (see Green pathway, Figure ES-2).
Since climate change has the potential to intensify and even create new trade-offs,
mainstreaming climate change into planning is also important for identifying and weighing
conflicts among adaptation options within the context of existing (and emerging) goals.
One example identified in this study is the simultaneous need to reduce sediments in salmon
stream habitats (under current SFEP goals) and increase coarse sediment transport to the Bay (as
indicated by the Purple pathway in Figure ES-2). Another example based on comparing
adaptation needs for two of the top pathways above is the trade-off between high volume pulses
to enable large grain sediment transport (see Purple pathway, Figure ES-2) and water availability
for steady, lower-volume releases to favor vegetative productivity (see Green pathway,
Figure ES-2).
Given the long-term nature of the climate change challenge, mainstreaming has an
additional advantage over a stand-alone plan in that it helps counteract the tendency to postpone
adaptation actions in the face of more immediate challenges. It often may be possible to adjust
current practices in ways that achieve adaptation while still fulfilling original goals.
Furthermore, thinking ahead as part of planning is essential for anticipating which of today's best
practices may become ineffective and even 'maladaptive' as sensitivities change and threshold
shifts occur under climate change. Once thresholds have been crossed or other unavoidable
changes of significance have occurred, some management goals may have to be revised.
Evaluation of Expert Judgment Approach
A novel methodology based on expert elicitation was developed and piloted as a tool for
'rapid assessment' of ecological sensitivities to climate change. The aim was to explore
whether it is possible to synthesize useful information from experts on key climate sensitivities
in the short time frame of a two-day workshop, using expert elicitation techniques. Expert
elicitation is a multidisciplinary process for using expert judgment to inform decision-making
when data are incomplete, uncertainties are large, and more than one model can explain available
data. The novel methodology introduced in this study is a modification of formal (usually
quantitative) expert elicitation that uses qualitative judgments in accordance with complex
ecological questions. Influence diagrams (showing the structure of causal relationships among
variables) were used successfully to capture the experts' collective understanding of the selected
ecosystem processes, under current conditions and under two scenarios of future climate change
for a midcentury time frame. A coding scheme was used by the experts to record their
judgments, with observational notes and group discussions used to gather additional information.
xxin
-------
The result was three categories of information based on the influence diagrams: (1) the
direction and strength of the relationships among variables, (2) the changing sensitivities of some
relationships to climate change (including potential threshold responses), and (3) the
relationships of highest relative impact on the process as a whole. When this wealth of
information is combined into a 'crosswalk' of all three categories, it is possible to identify top
pathways (see above) comprised of relatively well-under stood relationships that are sensitive to
climate change and for which management options are available. Managers are encouraged to
further 'mine' the tables for other key pathways applicable to their specific sites and to identify
potential research priorities based on information gaps.
The expert elicitation exercise developed for this assessment has the potential to be
useful for other sites, processes and ecosystems. While an example North Bay site was used as
a means to focus the exercise, the variables that ended up in the final influence diagrams are
common enough that most of the results may transfer to the entire Bay for these particular
ecosystem processes. It is likely that the influence diagrams also could be transferred for use
with like ecosystem processes in other estuaries, with minor revisions for place-specific stressors
or other process variables; however the characterizations of variable relationships, sensitivity and
relative impact would have to be revised, particular to the location. Where information on
completely different processes is needed, the general methodology should be transferable to
other processes and ecosystems. The strengths of this method include the ability to capture more
recent knowledge than would be available from a literature review and more knowledge of the
type that is closely related to management. It is also effective at integrating across disciplines
and scales, which is particularly important for ecosystem and climate change assessments.
As a proof of concept for a new type of assessment exercise, this method and its results
come with a number of caveats. This was not a comprehensive vulnerability assessment for the
whole estuary, so prioritization based on these results should be considered in the broader
context of other vulnerable processes, ecosystems and goals. Given the complexity of these
systems and instances of uneven agreement among experts, actions based on the top pathways
should be taken with care, with each manager considering the applicability of the information to
his or her own specific system. Confidence estimates for individual judgments turned out to be
challenging, so improvements have been suggested for strengthening this aspect in future
assessments. There is also the potential to simplify the coding scheme based on what was
learned in this trial run, to improve efficiency and allow experts more time to fill in data gaps.
Regardless, the expert elicitation method developed for this study was well suited for achieving
the goals of this assessment, and in a time frame much shorter than would be required for more
traditional, detailed quantitative modeling. Having a well-supported and timely study to
substantiate new and existing ideas can position managers to justify the most appropriate
xxiv
-------
management options and priorities. It also can validate research priorities by highlighting known
research gaps. Overall, the method offers opportunities to capture and integrate the existing
collective knowledge of local experts, while pushing the boundaries to develop a new
understanding of the system and identify robust adaptation options in the face of climate change.
xxv
-------
1. INTRODUCTION
1.1. BACKGROUND
The San Francisco Bay estuary is highly vulnerable to the impacts of climate change.
Sea level rise, increased air and water temperatures, changes in precipitation, and changes in
storm climatology and winds are already causing increased inundation of coastal wetlands and
marshes, changes in water availability and quality, and altered patterns of sedimentation and
erosion (Scavia et al., 2002). These impacts are interacting with other anthropogenic stressors
such as extractive water uses and land use changes to make management of estuarine ecosystems
more challenging than ever. While there are many uncertainties regarding the nature of future
climate changes and the response of ecosystems to those changes, estuary managers can 'ready'
themselves by assessing the vulnerability of natural resources to climate change, making
strategic choices about how to implement adaptation strategies1 in the near term, and planning
for longer term management under a range of plausible scenarios of future change. It is the aim
of U.S. Environmental Protection Agency's (EPA's) Climate Ready Estuaries (CRE) Program to
assist National Estuary Programs (NEPs) in meeting such information and planning needs.
As part of the CRE Program, the San Francisco Estuary Partnership (SFEP), the San
Francisco Bay Conservation and Development Commission (BCDC), and EPA's Office of
Research and Development collaborated on the design and trial of a novel methodology for
conducting vulnerability assessments for sensitive ecosystems of the San Francisco Bay estuary.
The aim was to develop assessment capabilities using expert judgment to synthesize place-based
information on the potential implications of climate change for key ecosystem processes, in a
form that would enable managers to link the resulting information to adaptation planning.
1.2. PURPOSE AND SCOPE
1.2.1. Purpose
The purpose of this project was twofold: to conduct a vulnerability assessment using a
novel, expert judgment approach based on expert elicitation methods, and to analyze the
implications for adaptation planning. This was not a comprehensive vulnerability assessment for
the whole estuary but rather a proof of concept for a new type of assessment exercise, using
two key ecosystem processes of salt marsh and mudflat ecosystems as demonstration studies.
This was accomplished through a series of steps to: (1) identify key management goals and
throughout this report, "adaptation" refers to management adaptation rather than evolutionary adaptation.
Management adaptation refers to strategies for the management of ecosystems in the context of climate variability
and change (CCSP, 2008).
1-1
-------
ecosystem processes essential to meeting those goals; (2) create conceptual models of selected
ecosystem processes; (3) assess ecosystem process sensitivities to climate change; (4) consider
resulting vulnerabilities with respect to management goals; and (5) explore implications for
adaptation planning. Steps 1-2 were used to define the scope of the assessment, while steps 3-5
comprise the vulnerability assessment itself.
1.2.2. Scope
The scoping process began with a review of the SFEP Comprehensive Conservation and
Management Plan (CCMP) in order to select key management goals upon which to focus the
assessment. The key ecosystem-related goals selected by SFEP in consultation with BCDC and
EPA ORD were:
• Restore healthy estuarine habitat to the Bay-Delta, taking into consideration all
beneficial uses of Bay-Delta resources
• Protect and manage existing wetlands
• Restore and enhance the ecological productivity and habitat values of wetlands
Stem and reverse the decline of estuarine plants, fish and wildlife and the habitats on
which they depend; and
• Ensure the survival and recovery of listed and candidate threatened and endangered
species, as well as special status species.
After an information-sharing meeting with local experts to discuss the project and learn
about climate change impacts and adaptation work in the region, salt marshes and mudflats were
selected as focal ecosystems for the study. These systems were identified as highly relevant to
SFEP's management goals due to their ecological productivity, their habitat values for threatened
and endangered species, and their sensitivity to changes in climate-related variables such as sea
level rise and altered hydrology. For more detailed information on goal and ecosystem selection
processes, please see Appendix A.
The second step in the scoping process was the development of conceptual models to
understand the primary drivers and processes of salt marshes and mudflats. The conceptual
models were used to explore the linkages among key ecosystem processes within each
ecosystem, major stressors of concern, and climate drivers causing altered or new stressor
interactions. The models were refined to a set of five or six key ecosystem processes that are
essential to the maintenance of salt marsh and mudflat systems, as identified through literature
1-2
-------
review of salt marsh and mudflat conceptual models and climate change impacts. Based on these
general conceptual models, two specific processes of concern were selected for further analysis.
The purpose was to select good processes for piloting the method, but the choice does not imply
that these are necessarily the only important, or the most vulnerable, processes. The processes
were selected, in consultation with SFEP and BCDC staff, based on the criteria of being
identified by local experts as integral to ecosystem function, increasingly sensitive to climate
change, and sufficiently well-studied by the scientific community to provide the basis for a more
in-depth assessment.
The two processes selected for further analysis were sediment retention in salt marshes
and community interactions in mudflats (see Executive Summary Figure ES-1). Sediment
retention refers to the balance between the processes of removal and deposition of sediment onto
a salt marsh. The topic of community interactions in mudflats was narrowed to a tractable
"storyline" involving several interdependent species, which was selected based on interviews
with local experts who were asked to identify climate-sensitive interactions of interest. The
storyline selected was the relationship of two species of mudflat wading birds, the Marbled
Godwit and the Western Sandpiper, to their predators (Peregrin Falcons and Merlins) and prey
(invertebrate mudflat infauna). Expanded submodels were developed for each of the
two processes and served as the basis for designing the sensitivity analyses of the subsequent
assessment. For more detailed information on process selection and conceptual model
development, please see Appendix A.
The remaining steps of the assessment—the sensitivity analysis, vulnerability assessment,
and analysis of management implications—were accomplished through an expert
elicitation-style workshop, the results of which make up the core of this report. Expert elicitation
is a multidisciplinary process using expert judgment to inform decision-making when empirical
data are incomplete, uncertainties are large, more than one conceptual model can explain
available data, and technical judgments are required to assess assumptions. It takes advantage of
the vast amount of local knowledge that is available via regional experts who are familiar with
the state of the science for the system of interest. During a two-day workshop, a novel
application of the expert elicitation method was tested using two groups of seven expert
participants each. A list of the expert participants for each breakout group is provided in
Table 1-1. The experts were selected based on criteria that ensured extensive expertise in the
local system, broad coverage of multiple scientific disciplines, experience in both science and
management, and knowledge of both empirical and theoretical research (for additional
information on selection criteria and participant credentials, please see Appendix B). The
participants assessed the sensitivities of salt marsh sediment retention and mudflat community
interactions to climate- and nonclimate stressor interactions, with an eye toward informing
1-3
-------
adaptation. The methodology and results of this expert elicitation exercise are described in the
sections that follow.
Table 1-1. Breakout group participants for the expert elicitation workshop
(see Appendix B for further details on selection criteria and credentials)
Sediment retention group
Dave Cacchione
U.S. Geological Survey
John Callaway
University of California, San Francisco
Chris Enright
California Department of Water Resources
Bruce Jaffe
U.S. Geological Survey
Lester McKee
San Francisco Estuary Institute
Dave Schoellhamer
U.S. Geological Survey
Mark Stacey
University of California, Berkeley
Community interactions group
Letitia Grenier
San Francisco Estuary Institute
Jessica (Jessie) Lacy
U.S. Geological Survey
Michelle Orr
Philip Williams & Associates
Diana Stralberg
Point Reyes Bird Observatory Conservation Science
Stuart Siegel
Wetlands and Water Resources
Lynne Trulio
San Jose State University
Isa Woo
U.S. Geological Survey
1.3. ROADMAP FOR THE REPORT
This report presents a summary of the entire project, including goal selection and
conceptual modeling, the expert elicitation methodology, the results of the workshop, and
implications for management. Figure 1-1 provides a flow chart of the assessment process and
report structure.
Section 2 describes the expert elicitation exercise, including the approach, the exercise,
and the results. Section 3 provides an analysis of the results with respect to how they may be
used by estuary managers to understand ecosystem responses to climate change and engage in
adaptation planning. Section 4 provides key conclusions of the assessment. The appendices
provide additional detailed information on the activities conducted prior to and following the
workshop. Appendix A summarizes the goal selection and conceptual modeling processes used
for scoping the vulnerability assessment. Appendix B provides details on the expert elicitation
preworkshop preparations and postworkshop follow-up, including expert selection criteria,
preworkshop preparations by participants, and expert feedback. Appendix C and Appendix D
1-4
-------
contain detailed information that was provided to the participants on the development of climate
scenarios and the methodology for estimating confidence.
1-5
-------
Q.
Q.
u
l{
Selection of Key
ManagementGoals
Selection of Focal
Ecosystems
Selection of Ecosystem
Processes
Development of
Conceptual Models
Development of Climate
Scenarios
Development of Expert
j
OQ
x
'-a
Pre-Workshop
Selection of Workshop
Participants
Development of "Straw
Man" Influence
Diagrams
Pre-Workshop Briefing
and Homework
Assignment
Development of
Consolidated Influence
Diagrams
OQ
01 I
a.^
Workshop
Development of Group
Influence Diagrams
Expert Elicitation
Exercise and
Discussions
Discussion of Climate
Scenarios
Management
Discussion
{Diagrams, Agreement,
Relative Impact,
Confidence
-{_ Adaptation Strategies
,H J" Linking Information to
^ L Management
P^J Top Pathways for
^{
•« I
tn I Management
,_, I" Insightsfromthe
* L Workshop
-------
2. EXPERT ELICITATION EXERCISE
2.1. JUSTIFICATION FOR METHOD
2.1.1. Definition and Uses
Expert elicitation is a multidisciplinary process for obtaining the judgments of experts to
characterize uncertainty and fill data gaps where traditional scientific research is not feasible or
adequate data are not yet available. The goal of expert elicitation is to characterize each expert's
beliefs about relationships, quantities, events, or parameters of interest. The expert elicitation
process uses expert knowledge, synthesized with experiences and judgments, to produce
conclusions about the nature of, and confidence in, that knowledge. Experts derive judgments
from the available body of evidence, including a wide range of data and information ranging
from direct empirical evidence to theoretical insights.
Because EPA and other federal regulatory agencies are often required to make important
national decisions in the presence of uncertainty, EPA's Science Policy Council formed an
Expert Elicitation Task Force in April of 2005 to investigate how to conduct and use this method
to support EPA regulatory and nonregulatory analyses and decision-making. The result was an
Expert Elicitation Task Force White Paper that affirms the utility of using expert elicitation and
provides recommendations for expert elicitation "best practices" based on a review of the
literature and actual experience within EPA. The draft paper (see
http://www.epa.gov/spc/expertelicitation/index.htm) is currently under external peer review
through EPA's Science Advisory Board. The best practices outlined in the draft White Paper
formed the basis for the design of this project's expert elicitation-style workshop.
2.1.2. Novel Application
The specific elicitation exercise used in this assessment was custom-designed by Dr. Max
Henri on of Lumina Decision Systems, Inc. Dr. Henri on is a nationally-recognized authority on
decision analysis methods and tools, dealing with uncertainty in environmental risk assessment,
and expert elicitation (e.g., Morgan and Henrion, 1990; Henrion et al., 1991; Pradhan et al.,
1996). As a member of EPA's Expert Elicitation Task Force, he was uniquely qualified to assist
in designing a novel application of expert elicitation methods for use in a two-day workshop
format. Specifically, Dr. Henrion developed a qualitative coding scheme for expert judgments
about the sensitivity of ecosystem processes to physical and ecological variables, using
"influence diagrams" to depict the relationships among ecosystem process variables and external
drivers such as climate change. This new methodology, described in detail below, explores the
2-1
-------
utility of expert elicitation for conducting "rapid vulnerability assessments" for ecological
systems.
2.2. WORKSHOP DESIGN AND METHODOLOGY
2.2.1. Workshop Goals and Objectives
The overarching goals of the workshop were to: (1) improve the understanding of the
sensitivity of selected salt marsh sediment retention and mudflat community interactions
processes to the projected impacts of climate change; (2) improve the ability to identify
adaptation strategies that mitigate those impacts, given the uncertainties; and (3) demonstrate the
applicability of an expert elicitation approach to this type of analysis.
The workshop was held March 16-17, 2010, in San Francisco, California, at the BCDC
offices. During the workshop, experts were divided into two breakout groups to consider each
ecosystem process separately. The seven participants in each breakout group (see Table 1-1)
were asked to provide judgments about the ecosystem process under consideration by their
group. For each ecosystem process, the specific objectives were to: (1) characterize the relative
influences of physical and ecological variables that regulate the process; (2) assess the relative
sensitivity of the ecosystem process to key stressors under current conditions and future climate
scenarios; (3) assess the degree of confidence in judgments about these relationships; and
(4) relate the results of the exercise to adaptation planning through group discussions. Given the
range of habitats and issues in the entire San Francisco Bay area, the participants were asked to
consider the North Bay (see San Pablo Bay; Figure 2-1) when a more specific spatial scope
would be useful during the workshop exercise. In addition, an example site in the North Bay,
China Camp, was presented as a particular place upon which to focus when considering
management implications; however, issues and options that were not specific to China Camp
were also considered during group discussions.
For further details on workshop preparation and implementation, including selection
criteria for participants, please see Appendix B.
2.2.2. Approach and Methodology
According to protocols put forth in EPA's Expert Elicitation Task Force White Paper,
there are a variety of options for gathering and processing expert judgments. The specific
elicitation approach used in this workshop was one that asked experts to give their individual
judgments independently. This was done to reduce the tendency towards "group-think," i.e., the
tendency for many people to go along with the most vocal participant, even if s/he is not the
most knowledgeable. Since participants varied in their expertise about different aspects of the
2-2
-------
Figure 2-1. The North Bay (San Pablo Bay) of the San Francisco Bay area,
with expanded detail showing China Camp State Park.
system, they were encouraged to make adjustments to their judgments at any time based on any
deeper understanding gained during or after group discussions; however, consensus was not the
goal of the exercise. Rather, the aim was to look at the expert judgments in aggregate, while also
retaining information on variance in judgments. This approach is well-suited to the type of
qualitative judgments participants were asked to make at the workshop.
2.2.2.1. Influence Diagrams
Each breakout group participated in the development of an influence diagram of the
ecosystem process under consideration by their group. Decision analysts use influence diagrams
as a way to define the qualitative structure of causal relationships among variables that experts
believe are of greatest importance for understanding the problem being evaluated. Influence
2-3
-------
diagrams typically represent a subset of a larger, more detailed model such as the conceptual
models developed previously (see Appendix A).
A simplified influence diagram for sediment retention is provided in Figure 2-2. By
convention, the variables in an influence diagram are represented by rectangles (labeled boxes)
while arrows between the variables represent causal relationships, or "influences". Sequences of
arrows form pathways, all of which ultimately lead to the final variable, or endpoint, of concern.
In Figure 2-2, the endpoint that is being evaluated is sediment retention. Interactive effects of
multiple variables on each other, or on the endpoint, can occur where two "causal" variables both
influence (have arrows into) a common "response" variable. In Figure 2-2, an example
interaction is indicated by arrows B and C, where reservoir management and impervious cover
together could have an interactive effect on freshwater inflow.
In the case of community interactions, the diagram was constrained to a tractable number
of species of interest. It focused on the relationship of two species of mudflat wading birds, the
medium-bodied Marbled Godwit and the small-bodied Western Sandpiper, to their predators and
prey. Please see Appendix A for a more detailed explanation of this storyline.
While influence diagrams are widely used and relatively well-under stood, our proposed
use of qualitative degrees of influence is an innovation in expert elicitation. Typically, an expert
elicitation seeks to obtain expert judgments about uncertain quantities in the form of numerical
probability distributions. For the ecosystem processes considered during this workshop, there
were information, data and time limitations that made quantifying the influences as probability
distributions unrealistic. Instead, judgments were based on qualitative types (is the relationship
direct, or inverse?) and degrees (is the response small, or large?) of influences. The use of
qualitative degrees of influence provides much more detail than simply specifying causal
influences with arrows alone, but less specificity than required for quantified probabilities.
Prior to the workshop, the participants attended briefing calls in which they: learned
about the project plan; discussed background reading materials; and were presented with "straw
man" diagrams (see Appendix B) developed from the original conceptual models. They were
asked to review the diagrams and submit their own revised versions the week before the
workshop. Diagram submissions were combined into one consolidated draft diagram for each
group that served as the starting point for discussion at the workshop. The workshop itself began
with each group working together to refine their diagram into a "group diagram". The group
influence diagram was meant to distill the system to a tractable set of key variables and
influences, and as such it was not comprehensive. The groups were given complete freedom to
alter any part of the diagram, with the exception of the ecosystem process endpoint, as long as
they constrained the diagram to a total of no more than 15 boxes. At the same time, participants
were reminded to
2-4
-------
Altered Flows/Water
Demand:
Reservoir Management
Land Use/ Land Cover
Change:
Impervious Cover
Figure 2-2. Simplified influence diagram for sediment retention.
keep some of the top row stressor or management boxes, since these would serve as key linkages
back to management options. Participants were also encouraged to minimize the total number of
arrows in the diagram to include only the most key influences. The purpose was to capture the
key components and relationships of each ecosystem process in a concise form that could be
rapidly assessed in a workshop setting. Once the group diagrams were finalized, all of the
participants made their judgments using the same diagram throughout the remainder of the
workshop.
2-5
-------
2.2.2.2. Climate Scenarios
Dr. Katharine Hayhoe of Texas Tech University, an experienced climate scientist with an
extensive background in regional climate assessments, developed two climate change scenarios
for use in the expert elicitation exercise (see Appendix C for more detailed information on the
climate change scenarios). The scenarios represented two distinct but scientifically credible
climate futures for a midcentury (2035-2064) time period. (The midcentury time frame was
selected by the SFEP partners because of its suitability for adaptation planning.) The projections
were based on six leading climate models, using a lower emissions scenario (Climate
Scenario A) and a midhigh emissions scenario (Climate Scenario B) to generate values for
climate variables for use by the experts in making their judgments (see Table 2-1).
Under both climate change scenarios, California will retain its Mediterranean climate
(cool/wet winters and hot/dry summers) and continue to experience a high degree of variability
in precipitation with rising sea levels. By midcentury, the "higher-range" Climate Scenario B
(which includes higher emissions and a more sensitive climate) is projected to experience a
warmer and somewhat drier climate compared to the "lower-range" Climate Scenario A (with
lower emissions and a lesser impact on California's climate).
At the workshop, Dr. Hayhoe provided the participants with an overview of major
climate drivers and regional trends for California. She discussed five main sources of
uncertainty with climate projections, including: (1) the amount of future emissions; (2) the
degree to which the influence of global climate change on local climate is modified by local
factors; (3) the sensitivity of the climate system (as feedbacks are not well understood); (4) the
ability of climate models to simulate climate both globally and locally; and (5) the natural
variability of the climate system. Because of these factors, exact predictions of climate change
are not possible. However, uncertainty can be dealt with by using multiple scenarios to bracket a
range of plausible climate futures and identify key vulnerabilities in the system. In order to
consistently "bound" the consideration of future climate changes in the workshop exercise, the
participants were instructed to use the values provided under Climate Scenarios A and B (see
Table 2-1) to contextualize their judgments about future effects on the ecosystem processes
under consideration. For additional details on the climate scenarios, including data sources,
please see Appendix C.
2.2.2.3. Expert Facilitation
Due to the highly technical nature of the exercise, the complexity of the novel
methodology that was being used, and the ambitious time line for accomplishing multiple
outputs, it was essential that the workshop be run by skilled expert facilitators. These were
2-6
-------
Table 2-1. Summary of Climate Scenario A ("lower-range" scenario) and
Climate Scenario B ("higher-range" scenario): averages for midcentury
Temperature3
Precipitation
Sea Level
Annual average13
Average increase of
winter temperature0
Average increase of
summer temperature
Extreme heat daysd
Annual change"
Winter change
Heavy events
Total increase for
2050g
Hourly sea level
exceedances1
Storms/windJ
Snow pack change
Spring runoff
Seasonal changes in amount of
freshwater inflow to the bay from the
delta in 2060m
"Lower-range" scenario
+2.8°F(1.6°C)
+2.5°F(1.4°C)
+4.0°F(2.2°C)
+10 days/year
-4.5%
"Higher-range" scenario
+3.5°F(1.9°C)
+2.7°F(1.5°C)
+4.5°F(2.5°C)
+16 days/year
-7%
Reduced winter precipitationf
Decline in frequency of precipitation events (exceeding mm/day) but
not a clear signal in changes of precipitation intensity
+30 cm
1,343
+45 cmh
1,438
Tendency toward a decline in storms.k Projections suggest an
increased tendency for heightened sea level events to persist for more
hours. ENSO is not projected to increase in frequency or intensity.
For the Sacramento-San Joaquin watershed, April watershed-total
snow accumulation projected to drop by 64% by 20601
Spring runoff occurring earlier and reduced overall
October through February: inflow +20%
March through September: inflow -20%
aSince the 1920s, minimum and maximum daily temperature have been observed to have increased in California
with minimum temperature increasing at a greater rate accented by a small cooling trend in the summer (Cayan
et al., 2009). These averages are for 2035-2064 projections relative to a 1961 to 1990 baseline forBl and A2
emission scenarios.
bApproximate results using B1 and A2 emissions scenarios and three global climate models (PCM1, GFDL CM2.1,
HadCM3) (CEC, 2006).
These results are for Sacramento, California. This warming is projected to be more moderate along the coastline
(50 km from the coast) rising considerably inland (Cayan et al., 2009). These averages are for 2035-2064
projections relative to a!961 to 1990 baseline forBl and A2 emissions scenarios.
dExtreme heat days are defined as when the daily maximum temperature exceeds the 95th percentile of temperature
from the 1961-1990 historical averages of May-September days. 1961-1990 extreme heat days are approximately
8 days/year based on model runs. Results are provided by Cayan et al. (2009) using three climate models (CNRM
CM3, GFDL CM2.1, MICRO 3.2; with bias corrected spatial downscaling) forBl and A2 emissions scenarios.
Mid-century projections suggest hot daytime and nighttime temperatures increase in frequency, magnitude, and
duration (Cayan et al., 2009). Extreme warm temperatures in California, historically a July and August
phenomenon, will increase in frequency and magnitude likely beginning in June and may continue into September
(Hayhoe et al., 2004; Gershunov and Douville, 2008; Miller et al., 2008).
"Results are averaged across 6 GCMs using the grid point nearest to Sacramento (Cayan et al., 2009) for B1 and A2
emissions scenarios.
2-7
-------
Table 2-1. Summary of Climate Scenario A ("lower-range" scenario) and
Climate Scenario B ("higher-range" scenario): averages for midcentury
(continued)
fThese results are provided by CEC (2008).
8Sea level rise relative to 2,000 levels. This study applies Rahmstorf s methodology of estimating sea level rise as a
function of rising temperatures. This study assumes sea level rise along the coast to be the same as global estimates
given the observed rate of rise along the southern California coast has been about 17 to 20 cm per century similar to
that of global sea level rise (assume no future changes in other factors that affect relative sea level rise such as
changes in regional/local ocean circulation, ocean density, etc.) (Cayan et al., 2009). DMRS also provides
recommended 2,050 global sea level rise estimates relative to 1990 values: 11 cm (direct extrapolation of observed
increased during the 20th century), 20 cm (low-end value of Rahmstorf and approx midrange of IPCC TAR), 30 cm
(approx midrange of Rahmstorf and high-end of IPCC TAR); 41 cm (high end of Rahmstorf) (DMRS, 2007).
hThe total difference between mean range and spring range of 1.7 ft (50.3 cm) is slightly larger than the higher-range
scenario rise of 45 cm, based on the Point San Pedro tide station.
http://tidesandcurrents.noaa.gov/tideslO/tab2wcla.htmM128.
'The hourly sea level exceedance is defined as the maximum duration (hours) when San Francisco sea level exceeds
the 99.99th percentile level (140 cm above mean sea level) based on the GFDL climate change (A2) simulation
using the Rahmstorf sea level scheme averaged 2 to 4 hours increase for midcentury (Cayan et al., 2009).
JThese results are provided by Cayan et al. (2009).
kStorm is defined as sea level pressure (SLP) equaling or falling below 1,005 millibar.
'Results provided by the Bay-Delta watershed model driven by temperature projections from a parallel climate
model under a 'business-as-usual' scenario relative to 1995-2005 (precipitation is assumed to remain consistent
with today's observations) (Knowles and Cayan, 2004).
"This study does account for reservoirs, in-stream valley diversions, and in-Delta withdrawals and assumes no
future management adaptation or altered demand patterns (Knowles and Cayan, 2004).
chosen based on a number of criteria including: proven expertise in facilitating science-based
workshops; general knowledge of science behind estuary management (particularly wetlands
ecology); and experience working on national coastal issues and/or issues in the San Francisco
Bay region. The expert facilitators selected were Dr. Peter Schultz, Principal at ICF
International, and Dr. Brock Bernstein, independent consultant and President of the National
Fisheries Conservation Center. Dr. Schultz (who served as facilitator for the Sediment Retention
group) has served as the Director and Associate Director of the U.S. Global Change Research
Program Office, and has two decades of experience in climate and global change research,
management, decision support, and communication. Dr. Bernstein (who served as facilitator for
the Community Interactions group) is a marine ecologist with research experience in a range of
coastal and oceanic environments, including San Francisco Bay, and has worked on a wide
variety of management and policy issues.
Prior to the workshop, both facilitators attended training calls in which they were fully
briefed on the project background and conceptual models, the workshop goals and objectives,
and the expert elicitation exercise. Working together and with the SFEP/BCDC/EPA team, the
facilitators contributed to the refinement of the workshop agenda and improvements to the
workshop process.
2-8
-------
2.2.2.4. Coding Scheme and Exercise
Participants were asked to characterize each influence in their influence diagram
according to the coding scheme presented in Table 2-2, and to indicate their confidence in their
judgments using the confidence rankings described below (see next section). Influences were
characterized first under current conditions, and then under Climate Scenario A and Climate
Scenario B. The extent to which participants agreed in their judgments was variable across the
different influences. The rule that was adopted for determining agreement for each influence
was that a majority (four or more participants) had to have selected the same code. Majority
agreement among four or more participants was considered to indicate substantial agreement
across the group.
Table 2-2. Coding scheme used during the workshop exercise to characterize
influences. "Small" and "large" changes in variables are defined relative to the
current range of variation for each variable, with "small" indicating that the
variable is within its current range of variation and "large" indicating that the
variable has moved outside its current range of variation
Option
Type and degree of influence definition
0
No influence: we know that changes in X have no effect on changes in Y, holding all other variables
constant.
Unknown influence: we don't know whether an increase in X will increase, decrease, or have no effect on
Y.
Proportional increase: a large increase in X is likely to cause a large increase in Y. A small increase is
likely to cause a small increase.
Proportional decrease: a large decrease in X is likely to cause a large decrease in Y. A small decrease is
likely to cause a small decrease.
Inverse decrease: a small increase in X is likely to cause a small decrease in Y. A large increase in X is
likely to cause a large decrease in Y.
Inverse increase: a small decrease in X is likely to cause a small increase in Y. A large decrease in X is
likely to cause a large increase in Y.
A small increase in X is likely to cause a large increase in Y.
A small increase in X is likely to cause a large decrease in Y.
A large increase in X is likely to cause a small increase in Y.
A large increase in X is likely to cause a small decrease in Y.
10
A small decrease in X is likely to cause a large increase in Y.
11
A small decrease in X is likely to cause a large decrease in Y.
12
A large decrease in X is likely to cause a small increase in Y.
13
A large decrease in X is likely to cause a small decrease in Y.
2-9
-------
Participants were also asked to characterize interactive influences of their choosing (i.e.,
those they deemed important), under current conditions and under the climate change scenarios,
according to the coding scheme presented in Table 2-3. Since participants were given the option
to choose which interactive influences they considered significant and to provide judgments only
for those influences, and were limited by time, there were often interactions where only one or
two participants provided judgments. Only interactions scored by three or more participants
were examined in order to focus on interactions judged by several participants to be significant.
Three or more corresponding judgments were used to define agreement for interactive
influences.
Table 2-3. Coding scheme used during the workshop exercise to characterize
interactive influences
Interactive influence
Independence
Synergy
AND Gate
NOR Gate
Competition
Definition
The effect of X on Y is independent of Z (default situation)
The effect of X on Y increases with increase in Z
The effect of X on Y happens only with large Z
The effect of X on Y happens only with small Z
The effect of X on Y decreases with increase in Z
Finally, the participants were asked to assess their current level of scientific confidence in
their judgments for each influence or interactive influence using the confidence coding scheme
presented in Table 2-4. For each influence, each participant was asked to rate his confidence in
his judgment based on: (1) the amount of scientific evidence that is available in the scientific
community at large to support the judgment; and (2) the level of agreement/consensus in the
scientific community at large regarding the different lines of evidence that would support the
judgment. The coding options for "amount of evidence" were high (H) or low (L), based on
whether available information is abundant and well-studied and understood, versus sparse and
mostly experimental/theoretical. The coding options for "level of agreement" were H or L,
based on whether data, reports, and experience across the scientific community reflect a high or
low level of agreement about the influence. Thus it was possible to have four combinations of
evidence and agreement when assessing confidence: high evidence/high agreement (HH), high
evidence/low agreement (FIL), low evidence/high agreement (LH), and low evidence/low
agreement (LL). The rule for determining agreement in confidence was the same as described
above for influences: agreement was defined as a majority (four or more) of the same
2-10
-------
Table 2-4. Coding scheme used during the workshop exercise to characterize
confidence
Confidence
LH
LL
HH
HL
Definition
Low evidence, high agreement =
Low evidence, low agreement =
established but incomplete
speculative
High evidence, high agreement = well established
High evidence, low agreement =
competing explanations
categorization of confidence level. Similarly using the same rule as above for interactive
influences, agreement on confidence for interactive influences was defined as three or more of
the same categorization of confidence. For additional details on the method used to assess
confidence, please see Appendix D.
2.2.2.5. Typologies for Understanding Influences and Sensitivities
2.2.2.5.1. Type and degree of influence
The group's level of understanding of the different influences (arrows) in the influence
diagram can be gauged by the amount of agreement in participants' selection of influence codes.
Sometimes participants agreed on the type of influence, but not necessarily the degree (strength)
of the influence. Codes 2-13 (see Table 2-2) represent different combinations of types and
degrees of influences that can be grouped according to the following typology:
Types:
Direct relationship (when X increases, Y increases) = Codes 2, 3, 6, 8, 11, 13
Inverse relationship (when X increases, Y decreases) = Codes 4, 5, 7, 9, 10, 12
Degrees:
Proportional response of Y to X = Codes 2-5
Disproportional response of Y to X = Codes 6-13
Codes can also be paired according to the same type and degree of influence, with the
only distinction being whether one is considering "X" to be increasing or decreasing. For
example 2/3 is a direct proportional influence, with 2 indicating when "X" increases, and
2-11
-------
3 indicating when "X" decreases, but in both cases "Y" is responding in a directly proportional
way. Six combinations of pairings are possible:
Pairings by type and degree of influence (where "X" can go up or down):
Direct proportional = 2/3
Inverse proportional = 4/5
Direct disproportional, strong response (xY) = 6/11
Direct disproportional, weak response (Xy) = 8/13
Inverse disproportional, strong response (xY) = 7/10
Inverse disproportional, weak response (Xy) = 9/12
In some cases, participants selected the same exact code, indicating that they had the
same understanding of the influence in terms of both type and degree. Or, sometimes
participants chose pairings such as 2/3 while their colleagues may only have noted a 2 or a 3; we
consider these cases to also indicate a correspondence in understanding of type and degree of
influence, since the only distinction was whether a participant was thinking of "X" as going up
or down (or both).
In another group of cases, there was agreement on the type of influence (i.e., whether X
affects Y directly or inversely), although there was lack of agreement on the degree of that
influence. These latter cases amount to an understanding of how X affects Y, just not the
magnitude. It may still be useful for management to know for which influences we at least have
some understanding of the type of response, even if we are not sure of the magnitude.
Finally, there were cases in which there was such a mixture of codes selected as to
indicate no agreement in either type or degree of influence. This indicated that, among this
group of experts, the influence was poorly understood or poorly defined.
2.2.2.5.2. Sensitivity
It is also possible to establish a typology for assessing the sensitivity of each influence
(i.e., how sensitive variable Y is to changes in X), especially with regard to how those may
change under the climate scenarios. Several codes can indicate the same level of sensitivity, so
the following groupings were used to indicate three levels of sensitivity:
2-12
-------
Low sensitivity = Codes 8-9 and 12-13
Intermediate sensitivity = Codes 2-5
High sensitivity = Codes 6-7 and 10-11
This typology was used to document all judgments, along with the following additional
categories of judgments:
No Influence = Code 0
Unknown influence = Code 1
None given = No judgment provided
Other = Response provided that does not fit into the coding scheme
2.2.2.6. Understanding Relative Impacts of Influences
While the coding scheme described above captures the nature of individual influences, it
is also of interest to identify which influences and interactions the participants perceived to have
the greatest relative impact on the ecosystem process endpoint. Here we define relative impact
as the combination of not only sensitivity but also how greatly the variable is changing relative to
other variables. There was no coding for this in the workshop exercise; rather, this concept was
an emergent property of group discussions that looked at the influence diagram as a whole and
identified influences of greatest relative impact in the context of the entire web of influences.
During group discussions that spanned both days of the workshop, information was gleaned as to
which influences participants perceived to have comparatively greater effects on the ecosystem
process endpoints, and whether this varied under the climate scenarios. These discussions were
captured in the workshop notes as well as in the influence diagrams, in which the participants
identified influences and interactions of highest relative impact (see Sections 2.3.1.4 and
2.3.2.4).
2.2.2.7. Key Questions
As described above, there are three categories of information that together comprise the
collective understanding of each ecosystem process as represented by its influence diagram:
(1) the type and degree of influence between variables, (2) the sensitivity of "response" variables
to changes in "affecting" variables, and (3) the relative impact of each variable on the ecosystem
process endpoint. For each of the three categories of information, the following key questions
are addressed.
2-13
-------
Types and Degrees of Influences:
• For which influences and interactions was there agreement in participants' judgments
(codes), and what were those codes?
• How did agreement on influences and interactions vary from current conditions to
Climate Scenario A and Climate Scenario B?
• For influences and interactions for which there was agreement in judgments, how did
confidence levels across the participants vary? Did this change under the climate
scenarios?
Sensitivity of Influences:
• For which influences and interactions was there greatest sensitivity and least
sensitivity in the response variable to changes in the "affecting" variable?
• Were there any influences or interactions where agreement on sensitivity across
participants increased or decreased under the climate scenarios?
Relative Impact of Influences:
• Which influences and interactions did the participants indicate have the greatest
relative impact on the ecosystem process endpoints?
• Were there any influences or interactions for which relative impact changed under the
climate scenarios?
Using the data from the coding exercise as well as information that emerged during group
discussions, these questions are explored in the results sections that follow.
2.3. RESULTS
Major outputs of the expert elicitation exercise included the group influence diagrams,
the judgments on influences (including interactive influences) along with their confidence
estimates, information on sensitivities (including thresholds), and characterizations of relative
impacts. For the purpose of this study, a threshold is defined (as per Groffman et al., 2006) as a
point at which there is an abrupt change in an ecosystem property (such as a flip in influence
type from direct to inverse), or where a small additional change in a driver produces a large
response (such as a shift from a proportionate to a disproportionately strong response of variable
Y to a change in variable X).
2-14
-------
2.3.1. Sediment Retention
2.3.1.1. Group Influence Diagram
Figure 2-3 shows the group diagram developed by the Sediment Retention group.
Variable definitions that were clarified by the participants during the construction of the diagram
are found in Table 2-5. Two main variables, Net Mineral Accumulation and Net Organic
Accumulation, influence the endpoint of the balance between Net Accretion and Erosion.
Organic and inorganic sediment accumulation processes are both influenced by Inundation
Regime, which is influenced by Relative Sea Level and Tides. There is a feedback loop from the
endpoint to Inundation Regime. The middle level in the diagram includes Tides, Relative Sea
Level, Freshwater Inflow, Sediment Flux, Sediment Size and Wind and Waves. Freshwater
Inflow and Inundation Regime are key factors influencing Net Organic Accumulation through
plant community composition and production. Of the management and stressor variables, three
are related to Water Resource Management: Delta Outflow, Reservoir Management and
Channelization. These influence a combination of Sediment Flux and Size and Freshwater
Inflow.
The 15-box constraint meant that the freshwater and sediment supply variables were not
split between Delta and tributary sources, even though much of the discussion on the diagram
highlighted the differences in those sources. Without separate variables differentiating between
local tributary and Delta freshwater inflow, Delta Outflow and Freshwater Inflow could be
considered to effectively act as a single variable. The fourth stressor variable is a Land Use and
Land Cover Change variable: Impervious Cover.
2.3.1.2. Influence Types and Degrees
2.3.1.2.1. Agreement
The influences upon which participants agreed with respect to type and degree help to
establish the nature of those relationships and indicate which are best understood. Table 2-6
presents these results for the Sediment Retention group.
In some cases, participants gave multiple codes for an arrow. When the multiple codes
represented one of the pairing types described above in Section 2.2.2.5 (e.g., 2/3), both codes are
shown, separated by a "/".
If multiple codes that do not fall into a pairing were given, both codes are shown,
separated by a symbol indicating the nature of the combination. In the first type of combination,
multiple codes with "X" going in the same direction (e.g., X is increasing in both codes) are
separated by a "A" symbol; and where these codes conflict and would make a difference in
2-15
-------
Water Resource
Management:
Water Resource
Management:
Reservoir
Water Resource
Management:
Land Use/ Land
Cover Change:
Vegetative Production
Net Organic
Accumulation
Figure 2-3. Sediment Retention group influence diagram.
Table 2-5. Sediment Retention variable definitions clarified during group
discussion
Variable
Land use/land change: impervious cover
Freshwater inflow
Sediment flux
Vegetative production: net organic accumulation
Definition agreed upon by group
Surfaces that reduce the ability of water to enter soil or
substrate
From local watersheds and the Delta, influence on Net
Organic Accumulation depends on total or mean flow,
influence on Sediment Flux depends on peak flow
Amount and rate
Net of plant production and decomposition
2-16
-------
Table 2-6. Sediment Retention group influence judgments. Columns A-Z
represent individual influences (arrows) in the influence diagram and rows
represent individual respondents: dark green = agreement on influence type and
degree, light green = agreement on type but not degree, gray = no agreement;
within columns, green numbers = same (majority) grouping of type (though
degree may be different), pink numbers = disagreement about type, red
outline = threshold response
CURRENT
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
SCENARIO A
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
SCENARIO B
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
A
2/3
3
2
2/3
2/3
A
2/3
2
3
2/3
A
2/3
2
3
2/3
B
4/5
4/5
4
9
4/5
4
2/3
B
4/5
4/5
4
9
4
2/3
B
4/5
4/5
4
9
4
1
C
4/5
4/5
7
9
4/5
4
4
C
4/5
4/5
7
9
4
4
9A7
C
4/5
4/5
7
9
4
4
9A7
D
4/5
4/5
4
9
4/5
4
4
D
4/5
4/5
4
9
4
4
4
D
4/5
4/5
4
9
4
4
4
E
0
E
E
8
F
2/3
4/5
4
9
8/13
8
9
F
-|
F
G
9/12
4/5
6
9
8/13
7
9
G
9/12
-|
6
9
8/13
7
9
G
9/12
1
6
9
8/13
7
9
H
4/5
4/5
7
1
2/3
8
9
H
4/5
4/5
7
1
2/3
8
9
H
4/5
4/5
7
1
2/3
8
9
I
I
I
2/3
J
2/3
2/3
6/11
2
6/11
2
6/11
J
8/13
2/3
2/3
2
6
2
6/11
J
8/13
2/3
2/3
2
6
6
6/11
K
8/13
2/3
8
8
8/13
8
2/3
K
8/13
8
8
8
8
2
2
K
8/13
8
8
8
8
8
2/3
L
2/3
2/3
6/11
2
2/3
2
2/3
L
8/13
2/3
6/11
2
3
2
2/3
L
8/13
2/3
6/11
2
3
6
2/3
M
2/3
2/3
6/11
2
8
2/3
M
2/3
2/3
6/11
2
8
2/3
M
2/3
2/3
6/11
2
2
2/3
2-17
-------
Table 2-6. Sediment Retention group influence judgments. Columns A-Z
represent individual influences (arrows) in the influence diagram and rows
represent individual respondents: dark green = agreement on influence type and
degree, light green = agreement on type but not degree, gray = no agreement;
within columns, green numbers = same (majority) grouping of type (though
degree may be different), pink numbers = disagreement about type, red
outline = threshold response (continued)
CURRENT
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
SCENARIO A
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
SCENARIO B
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
N
2/3
2/3
2/3
2
2/3
2
2/3
N
2/3
2/3
6
2
2
2
N
2/3
2/3
6
2
2
0
2
O
1
O
8/13
7
1
3
2
2/3
0
8/13
3
P
2/3
2/3
2/3
2
8/13
2
2/3
P
2/3
2/3
2/3
2A8
8/13
8
2/3
P
2/3
2/3
2/3
2A8
8/13
2
2/3
Q
2/3
2/3
2/3
2
2/3
2
Q
2/3
2/3
2/3
2
2/3
8
2
Q
2/3
2/3
2/3
2
2/3
2
2
R
6/11
2/3
3
6
2
2/3
R
6/11
2/3
3
6
2
2/3
R
6/11
2/3
3
6
1
6
2/3
S
9/12
6
2/3
2
6/11
8
2/3
S
9/12
6
2/3
2
6
8
2/3
S
9/12
6
2/3
2
6
2
0
T
1
2/3
1
4
6/11
2
2/3
T
1
T
1
U
2/3
2/3
8/13
2
6/11
2
2/3
U
2/3
2/3
2/3
2
6/11
8
8
U
2/3
2/3
2/3
2
6/11
8
V
2/3
2/3
2/3
2
8/13
2
2/3
V
2/3
2/3
2/3
2
8/13
8
2/3
V
2/3
2/3
2/3
2
8/13
2
2/3
W
9/12
2/3
2
8
6/11
3
7
W
9/12
6
2
8
6/11
2
W
6
2
8
6/11
6
6
X
4/5
4/5
9|10
4
6/11
3A4
X
4/5
4/5
9|10
4
4/5
X
4/5
4/5
9|10
4
6/11
8
0
Y
2/3
2/3
2
6/11
2
2/3
Y
2/3
2/3
2
6/11
2/3
Y
2/3
2/3
2
6/11
2/3
Z
2/3
2/3
6/11
2
2/3
2/3
Z
8/13
2/3
3
2A4
2
2
6
Z
8/13
2/3
3
2
6
2
2-18
-------
determining agreement, those cells were not counted. In cases where a reason was given for
multiple codes (such as when boxes had "lumping" problems and participants specified different
codes for different variables within the box), then the code that logically corresponded best to
other participants' codes (based on the notes column and other inferences) was used.
In the second type of combination, codes with "X" going in different directions (e.g., X is
increasing in one code and decreasing in the other) are separated by a "|". Since the response to
X can indeed be different depending on whether X is increasing or decreasing, these cells do not
represent a conflict but rather the opportunity to consider agreement in both the "X-up" and
"X-down" direction. In these cases it was possible to have agreement in one direction but not the
other.
The columns in Table 2-6 represent individual influences (arrows) in the group influence
diagram, and rows represent individual respondents. Dark green shaded columns indicate
agreement on both type and degree of influence; light green shaded columns indicate agreement
on type but not degree; gray shaded columns indicate no agreement. Within columns, numbers
in green are those that fall into the same (majority) grouping in terms of type of influence (even
though degree is different), while codes in pink indicate disagreement about type. Columns
outlined in red indicate threshold influences where there was either: (1) a change in type of
influence in the climate scenarios compared to current conditions (e.g., from a direct to an
inverse relationship), (2) a change in sensitivity (e.g., a change from a proportional to
disproportional response, or (3) an indication by multiple participants in their notes or in the
group discussions that the influence was likely a threshold relationship of type 1 or 2 above (but
for which they did not know in which scenario this would occur). In these cases the type and/or
degree of influence for the relationship would depend on a threshold, the exact location of which
may be uncertain.
There were 26 influences in total. Under current conditions, there was agreement on both
type and degree of influence for 61.5% of the influences, agreement on type but not degree for
27% and no agreement for 11.5%. Under Climate Scenario A, there was agreement on both type
and degree for 65.5% of the influences, agreement on type but not degree for 19% and no
agreement for 15.5%. Under Climate Scenario B, there was a drop to 58% of influences for
which there was agreement on both type and degree, 23% for which there was agreement on type
but not degree and 19% with no agreement.
2.3.1.2.2. Thresholds
Relationship N (Relative Sea Level on Inundation Regime) and relationship Z (Wind/
Waves on Sediment Flux) were identified to be threshold relationships under the climate
2-19
-------
scenarios. The threshold of relationship N is related to the marsh response to sea level rise, and
is tied to the rate of sea level rise. At the point in the inundation regime where the marsh is no
longer able to keep up with sea level rise, the marsh elevation will drop, thereafter experiencing a
different inundation regime. The threshold of relationship Z occurs where wind-driven waves
change from a source of sediment, adding to net vertical accretion, to a net negative impact
through erosion of the marsh edge. This occurs because, as water depth increases due to sea
level rise, the effect of wave energy on resuspension of bottom sediment will decrease while its
effect on marsh edge erosion will increase. In both of these cases the type or sensitivity of the
influence did not change across the scenarios (direct influence with intermediate sensitivity for
both), but the influences were indicated by participants to be important threshold relationships
through the discussion. One possible reason why these thresholds identified in the discussion did
not show up in the coding as changes in sensitivity is because participants did not know where
the threshold would occur, so they did not want to attach that estimate to a particular climate
scenario. Alternatively, it may be that there is a threshold that represents a state change that falls
within the range of natural variability, so this method was not sensitive enough to identify the
threshold. Relative sea level and wind and waves are both closely tied to climate drivers, making
relationships driven by them sensitive to climate change.
2.3.1.3. Influence Sensitivity
Figure 2-4 shows the sensitivity results using the influence diagram, indicating where
there is agreement under current conditions. The typology described in Section 2.2.2.5 was used
to code sensitivity, with an additional differentiation within the "no agreement" category. In all
"no agreement" cases, there was a mixture of codes for intermediate sensitivity along with low
and/or high sensitivity; if at least four participants provided judgments, and there were more high
sensitivity judgments than low sensitivity judgments, then the dashed arrow was colored orange
to indicate intermediate-to-high sensitivity. Under current conditions, 19 influences for which
there was agreement were categorized as intermediate sensitivity. Three influences were
categorized as low sensitivity, two of which originate from the variable Channelization. There
were no instances of agreement on influences with high sensitivity. There was no agreement on
sensitivity for four influences.
Figure 2-5 compares the sensitivities as in Figure 2-4, across the three scenarios. There
were no influences for which the sensitivity category changed between scenarios; the only
changes were between no agreement and a type of sensitivity. Under Climate Scenario A, all of
the same influences as those under current conditions were again categorized as intermediate
sensitivity, with the exception of both Freshwater Inflow and Inundation Regime on Net Organic
2-20
-------
Water Resource
Management:
Delta Outflow
Water Resource
Management:
Reservoir Management
Water Resource
Management:
Channelization
Land Use/Land Cover
Change:
Impervious Cover
Vegetative Production:
Net Organic
Accumulation
Key
Low sensitivity
Intermediate sensitivity
High sensitivity
Intermediate-to-high trend
No agreement
Figure 2-4. Sediment Retention group summary influence diagram of
sensitivities under current conditions.
2-21
-------
Current
Scenario A
Scenario B
Low sensitivity
Intermediate sensitivity
High sensitivity
Intermediate-to-high trend
No agreement
Figure 2-5. Sediment Retention group summary influence diagrams of
sensitivities: variance across current conditions and two climate scenarios.
Accumulation, for which there no longer was agreement. However, Inundation Regime on Net
Organic Accumulation showed a trend toward increasing sensitivity (orange arrow), as did
Inundation Regime on Wind/Waves. The same three influences as under current conditions were
categorized as low sensitivity for both Climate Scenario A and Climate Scenario B.
Under Climate Scenario B, one influence which previously had no agreement (but did
show an orange trend), Inundation Regime to Wind/Waves, increased in agreement, which
resulted in a categorization of high sensitivity. Three additional intermediate sensitivity
influences dropped below the standard of agreement: Wind/Waves on Sediment Size, Sediment
Size on Net Mineral Accumulation and the feedback from Net Erosion/Accretion on Inundation
Regime, such that the number of influences with no agreement on sensitivity increased to eight.
The disagreement shows a trend of some participants estimating increasing sensitivity, with
several of the influences characterized as a mix of intermediate and high sensitivity (orange
arrows) where there had once been agreement on intermediate sensitivity.
One reason for lack of agreement on changes in sensitivity across scenarios, as well as
lack of agreement within scenarios, may have been the degree of variability among participants
in their judgements. Overall, there was more variability among participants than across
scenarios for any given participant. There were no patterns across participants, such as
characterizing only increasing sensitivity. Further description, as well as figures depicting
variability in judgments across participants, can be found in Appendix B.
2-22
-------
2.3.1.4. Relative Impact
Figures 2-6 and 2-7 present the characterizations of relative impact between current and
future climate scenarios (the group's discussion did not differentiate between the two future
climate scenarios). Six influences were identified as having high relative impact under current
conditions. None of these are connected to the management options level within the diagram.
We have assumed that these same relationships are still of high impact under the climate
scenarios unless otherwise noted in the group discussion on climate change impacts. The
influences of Net Organic Accumulation on Net Erosion/Accretion and of Wind/Waves on
Sediment Flux were identified as having increasing impacts under the climate scenarios.
Two new influences, both on Freshwater Inflow and driven by variables in the management
options level, were identified as having high relative impact under the climate scenarios:
Reservoir Management and Channelization.
2.3.1.5. Confidence
The confidence results shown in Table 2-7 are provided for the Sediment Retention
influences for which there was agreement on type. The lack of agreement on confidence for
almost half of the judgments is a significant gap, limiting our ability to prioritize around
confidence judgments. Eight of the 11 influences for which there was agreement on confidence
across all three scenarios were scored as FtH. The influences of Reservoir Management on
Sediment Flux and Sediment Size, which were both categorized as HH under current conditions,
showed declining confidence under the climate scenarios, with scores of LH under Climate
Scenarios A and B. Relative Sea Level on Tides, scored as FtH under current conditions, also
showed declining confidence under the climate scenarios, with a score of LL under Climate
Scenario A and Climate Scenario B. The overall decrease in the total number of FtH judgments
from current conditions to the climate scenarios and the corresponding increase in the total
number of LL judgments show that influences become less well-understood, probably due to less
information being available about future climate conditions.
2.3.1.6. Interacting Influences
Table 2-8 presents the interactive influences upon which there was agreement for the
Sediment Retention group. The interactive influence columns indicate the type of interactive
influence and associated number of participants that chose that particular interactive influence
type. The confidence columns indicate the confidence judgment and associated number of
participants that chose that particular confidence score.
2-23
-------
Water Resource
Management:
Delta Outflow
Water Resource
Management:
Reservoir
Management
Water Resource
Management:
Channelization
Land Use /Land
Cover Change:
Impervious Cover
7v\ „ \ „ \
w
Tides
V
K
Relative Sea
Level
V J
S'
Freshwater Inflow
L
Sediment Flux
-~1
s '
SedimentSize
M
1
Vegetative Production:
Net Organic
Accumulation
Key
>
High Impact
Figure 2-6. Sediment Retention influences indicated as having high relative
impact under current conditions.
2-24
-------
Water Resource
Management:
Delta Outflow
Water Resource
Management:
Reservoir
Management
Water Resource
Management:
Channelization
Land Use /Land
Cover Change:
Impervious Cover
Vegetative Production:
Net Organic
Accumulation
Key
High impact
Increasing impact under climate scenarios
Figure 2-7. Sediment Retention group influences indicated as having high
relative impact under climate scenarios.
2-25
-------
Table 2-7. Sediment Retention group confidence for influences with agreement
CURRENT
SCENARIO A
SCENARIO B
A
HH
NA
NA
B
HH
HH
HH
C
HH
LH
LH
D
HH
LH
LH
F
NA
NA
NA
G
NA
LL
LL
H
NA
NA
NA
J
HH
HH
HH
K
HH
LL
LL
L
HH
NA
NA
M
HH
HH
HH
N
HH
HH
HH
0
NA
NA
NA
P
HH
HH
HH
Q
HH
HH
HH
R
HH
NA
NA
S
NA
NA
LL
T
NA
NA
NA
U
HH
HH
HH
V
HH
HH
HH
W
NA
NA
NA
X
NA
NA
NA
Y
NA
NA
NA
Z
HH
NA
NA
NA = no agreement; HH = high evidence, high agreement; HL = high evidence, low agreement; LH = low evidence, high agreement; LL = low evidence, low
agreement.
Table 2-8. Sediment retention group interactive influences with agreement under current conditions and
Climate Scenarios A and B
Interaction
M+N
P+Z
Q+R
Q+S
R+S
Variable X
Tides
Tides
Sediment flux
Sediment flux
Sediment size
On
on
on
on
on
on
Variable Y
Inundation
regime
Sediment flux
Net mineral
accumulation
Net mineral
accumulation
Net mineral
accumulation
With
with
with
with
with
with
Variable Z
Relative sea
level
Wind/waves
Sediment size
Inundation
regime
Inundation
regime
CURRENT
Interactive
influence
Synergy (3)
Synergy (3)
Synergy (5)
Synergy (5)
Synergy (3)
Confidence
NA
NA
HH(3)
HH(3)
NA
CLIMATE A
Interactive
influence
NA
Synergy (3)
Synergy (3)
Synergy (4)
NA
Confidence
NA
NA
NA
NA
NA
CLIMATE B
Interactive
influence
NA
Synergy (3)
Synergy (3)
Synergy (4)
NA
Confidence
NA
NA
NA
NA
NA
to
to
NA = no agreement; HH = high evidence, high agreement; HL = high evidence, low agreement; LH = low evidence, high agreement; LL = low evidence, low
agreement; () = number of respondents.
-------
Under current conditions, there were five interactive influences for which there was
agreement among participants in the Sediment Retention group. For each of these interactive
influences, Synergy was the type of influence chosen. Among these, there is a cluster of
multiple interactions between Inundation Regime, Sediment Flux and Sediment Size on Net
Mineral Accumulation. The other two interacting influences identified act on Inundation
Regime and Sediment Flux, so they are highly interconnected. There was only agreement on the
confidence for two of these interactive influences, both of which were scored as HH.
Under both Climate Scenario A and B, there was agreement on three of the previous
five synergistic interactive influences, with synergy again chosen as the type of interactive
influence. Again the same cluster of influences on Net Mineral Accumulation was identified.
There was no agreement on confidence for these interactive influences under either of the future
climate scenarios.
This lack of agreement on interacting influences was primarily due to not having many
influences with enough participants characterizing the same interacting influences. Of the 48
combinations of influences with interactions characterized by participants, only 10 could be
considered for agreement with at least three participants making a judgment; half of those had
three participants in agreement.
2.3.2. Community Interactions
2.3.2.1. Group Influence Diagram
Figure 2-8 shows the group diagram developed by the Community Interactions group.
Variable definitions that were developed by the participants during the construction of the
diagram are found in Table 2-9. Figure 2-8 shows a high degree of interconnectivity between
variables, especially among those directly influencing the endpoint. These variables are Extent
of Mudflat (and, therefore, extent of feeding habitat), Predators and Disturbance, Bed Sediment
Characteristics and Quality, Shorebird Prey Community and Landscape Mosaic (i.e., where
mudflats sit relative to other foraging and roosting habitats such as salt ponds). Many of the
variables encompass complex processes, which combine more than one key variable. Defining a
metric specific to such broad variables, including whether they are increasing or decreasing,
proved to be challenging. The possibility of differing assumptions about definitions among
participants complicates interpretation of the results. The variables indirectly affecting the
endpoint are primarily physical ones: Mudflat Bathymetry, Tides and Hydrodynamics, Sediment
Resuspension and Deposition, Wind/Waves, Water Quality, Freshwater Inflow and Sediment
Supply. The management and stressor variables are broad categories: Water Management,
Restoration and Land Use Change.
2-27
-------
IVarMeo
'A'esteiT Sarcpipe
Figure 2-8. Community Interactions group influence diagram.
2-28
-------
Table 2-9. Community interactions variable definitions clarified during
group discussion
Variable
Water management
Restoration
Land use change
Freshwater inflow
Sediment supply
Landscape mosaic
Wind/waves
Water quality
Inundation regime3
Sediment resuspension and deposition
Bed sediment characteristics and quality
Extent of mudflats (acre hours)
Predators and disturbance (anthropogenic)
Shorebird prey community
Shorebirds
Definition agreed upon by group
Reservoir management, upstream operations
Restoration and management of former Bay lands
Impervious surface, shoreline armoring, freshwater
demand, retaining sea level rise accommodation space
(land conservation to prevent development)
Annual hydrograph from local watersheds and the Delta
(includes winter storm frequency and intensity)
Total mass of sediment (physical material coming into the
system from local watersheds and the delta)
Includes ponds, diked wetlands, seasonal wetlands, muted
tidal wetlands and is spatially explicit (metric: amount of
energy needed per day; probability of mortality)
Wave power (spring and summer predominant winds,
storm events)
Nutrients, contaminants, salinity
Tides, bathymetry
Mass of sediment deposited or removed from mudflat
Grain size, bulk density, chemical contamination
Metric: acre hours
Predators: percentage shorebird population and numbers
taken; anthropogenic disturbance includes all human
activity in or adjacent to system that is affecting it (e.g.,
hiking, biking, recreational, commercial traffic, clamming)
Biomass, energetics
Winter abundance of shorebirds in San Francisco Bay
aOn Day 2, "inundation regime" was split into two variable boxes: "tides and hydrodynamics" and "mudflat
bathymetry", with the addition of accompanying arrows. Judgments for these new arrows under current conditions
were made before the group proceeded with judgments under the climate scenarios.
2-29
-------
2.3.2.2. Influence Types and Degrees
2.3.2.2.1. Agreement
Table 2-10 presents the results for the Community Interactions group. As in Table 2-6,
the columns in Table 2-10 represent individual influences (arrows) in the group influence
diagram, and rows represent individual respondents. Dark green shaded columns indicate
agreement on both type and degree of influence; light green shaded columns indicate agreement
on type but not degree; gray shaded columns indicate no agreement. Within columns, numbers
in green are those that fall into the same (majority) grouping in terms of type of influence (even
though degree is different), while codes in pink indicate disagreement about type. For further
explanation of table details, see Section 2.3.1.2.
There were 36 influences in total. The participants agreed on the type and degree of
influence for a smaller fraction of the total number of influences than the Sediment Retention
group did. Under current conditions, there was agreement on both type and degree for 50% of
the influences, agreement on type but not degree of influence for 14% and no agreement
for 36%. Under Climate Scenario A, this shifted to 33% of influences with agreement on both
type and degree, 25% with agreement on type but not degree and 42% with no agreement.
Under Climate Scenario B, influences with agreement continued to decline, with 31% for which
there was agreement on both type and degree, 25% for which there was agreement on type but
not degree and 44% with no agreement.
The larger number of influences for which there was no agreement under all scenarios
leaves a gap which makes it difficult to understand the type or degree of influence for these
relationships. This is partially due to a higher occurrence of no response given for the
Community Interactions group. It is not possible to differentiate between lack of response due to
insufficient time and disinclination to answer due to lack of knowledge about the influence.
2.3.2.2.2. Thresholds
Four relationships were identified as threshold relationships under the climate scenarios,
based on the notes and discussions. These were: Freshwater Inflow on Tides and
Hydrodynamics (Relationship K); Water Quality on Shorebird Prey Community (Relationship
S); Bed Sediment Characteristics and Quality on Shorebird Prey Community (Relationship BB);
and Extent of Mudflat on Shorebirds (Relationship DD). Relationship K was characterized as a
direct influence of uncertain degree; coding for degree was a mixture of weak, proportional and
strong influences with a slight trend toward increasingly strong influences through time. Some
participants indicated that winter increases in freshwater flow will be very important as mudflats
2-30
-------
Table 2-10. Community interactions group influence judgments. Columns A-KK
represent individual influences (arrows) in the influence diagram and rows represent
individual respondents: dark green = agreement on influence type and degree, light
green = agreement on type but not degree, gray = no agreement; within columns, green
numbers = same (majority) grouping of type (though degree may be different), pink
numbers = disagreement about type, red outline = threshold response
Current
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
Scenario A
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
Scenario B
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
A
3
2/3
3
4
2
2
A
A
3
B
3
2/3
3
4
2
B
3
3
B
C
2
c
9
9
2
8
2
2
C
9
9
2
8
2
2
D
4/5
4
5
4
9
D
4/5
7
3
7
4
9
D
4/5
7
3
4
9
E
1
7
2
3
6
6
6
E
1
7
6
11
6
2
7
E
1
7
6
11
6
2A6
7
F
8
3
2
2
4
7
2
F
8
F
8
G
G
G
2
H
3
H
3
H
I
2
2
2/3
2
2/3
2
I
2
3
11
2
2/3
I
2
7|3
11
2
2/3
J
J
1
J
1
K
2/3
2A8
2/3
2
8
8
K
2/3
2A6
6
13 A0
3A13
K
2/3
2A6
13 A0
3A13
L
2
2/3
2
2
L
1
L
1
M
M
M
N
N
1
N
1
O
6
6/11
6
2
6
2
O
6
6
6/11
6
2
3
6
0
6
6
6/11
6
2
11
6
P
2/3
8
2/3
8
8
2
P
2/3
8
2/3
6
8
P
2/3
6
2/3
8
Q
2
2
2/3
2
2
2
Q
2
2
8
2
2
2
Q
2
2
8
2
2
2
R
1 2
R
12
R
2-31
-------
Table 2-10. Community interactions group influence judgments. Columns A-KK
represent individual influences (arrows) in the influence diagram and rows represent
individual respondents: dark green = agreement on influence type and degree, light
green = agreement on type but not degree, gray = no agreement; within columns, green
numbers = same (majority) grouping of type (though degree may be different), pink
numbers = disagreement about type, red outline = threshold response (continued)
Current
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
Scenario A
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
Scenario B
Resp. 1
Resp. 2
Resp. 3
Resp. 4
Resp. 5
Resp. 6
Resp. 7
S
3
7
2/3
2
2
8
2
S
3
7
2
2
3
11
S
3
7
2
2
3
11
T
0
8/13
2
2/3
T
0
2/3
T
0
2/3
U
2/3
2
8/13
2
2/3
2
U
2/3
2
7
2
2/3
U
2/3
2
7
2
2/3
V
2/3
4
4/5
2
2
V
2/3
4
2
V
2/3
4
2
W
2/3
4
2/3
7
2
W
2/3
4
11
7
W
2/3
4
11
7
X
2/3
6A7
4|10
7
6
7
X
2/3
6A7
7
7
6
7|3
7
X
2/3
6A7
7
7
6
7
7
Y
2/3
2
2/3
2
6
Y
2/3
6
2/3
2
3
6
Y
2/3
6
2/3
2
3A11
2
Z
2/3
2
2/3
Z
2/3
2
Z
2/3
2
AA
4/5
5
4|10
10
3
4
AA
4/5
5
10
12
4
AA
4/5
10
10
12 A3
4
BB
2/3
7A4
8/13
8
2/3
3
3
BB
2/3
7A4
8
2/3
2A8
11
BB
2/3
7
8
2/3
2A8
11
CC
2/3
7A4
2/3
2
2/3
7|3
5A11
CC
2/3
7A4
2
2/3
2A8
11
CC
2/3
7
2
2/3
2A8
11
DD
2/3
2
6/11
2
2/3
6
6
DD
2/3
2
6/11
10
2/3
2A6
11
DD
2/3
6
6/11
10
2/3
6
11
EE
2/3
2
2/3
2
2/3
6
2
EE
2/3
2
2
2/3
2
11
EE
2/3
2
2
2/3
2
11
FF
4/5
7
9/12
9
4/5
8
7
FF
7
7
2
7
4/5
2
7
FF
7
7
6
7
4/5
2A6
7
GG
2/3
8
2/3A6/ll
2
2/3
2
2
GG
2/3
2
10
2/3
2
2
GG
2/3
6
10
2/3
2
2
HH
2/3
2
2/3
2
7
2
HH
2/3
2
2/3
2
7
11
HH
2/3
2
2/3
2
7
11
JJ
7
6/11
2
11
6
JJ
7
6/11
11
11
11
JJ
7
6/11
11
11
KK
3
2/3A6/ll
2
KK
3
6/11
2
KK
3
6/11
2
2-32
-------
reach a threshold of becoming subtidal; this would be especially true during high tides, where
flows could push a system above a threshold and create a large impact on inundation height.
Relationship S was characterized as direct proportional across the three scenarios. The
water quality aspect emphasized by the participants as a threshold was dissolved oxygen (DO).
They noted that small decreases in DO could have a large negative effect on mudflat prey
populations as a threshold is reached.
Relationship BB was direct proportional under current conditions, but there was no
agreement on degree under the climate scenarios. One participant indicated a change to a
disproportionately strong response through coding, but other participants did not change their
coding or left some blank cells, such that there was no majority agreement on degree under the
climate scenarios. However, participants' notes indicated that as habitat becomes more limited,
even small areas of poor habitat will have large effects on shorebirds. For this example as well
as the previous two threshold influences above, the participants chose to indicate the thresholds
through notes (rather than through coding) because they were not sure when (i.e., under which
climate scenario) the threshold was most likely to be reached.
Relationship DD was unique in being the only threshold influence that was identified
clearly through the coding exercise. Under current conditions the relationship was considered
direct proportional. Under Climate Scenario A there was agreement that it was still a direct
relationship, but there was no agreement on degree because there was a mixture of proportionate
and disproportionately strong codes. Under Climate Scenario B the conversion to agreement on
a direct disproportionately strong relationship was complete. This reflects the opinion of the
participants that as access to foraging habitat on mudflats becomes limiting due to sea level rise
and other factors, the effect on shorebird populations will become more extreme.
2.3.2.3. Influence Sensitivity
Figure 2-9 shows the sensitivity results using the influence diagram, indicating where
there is agreement under current conditions. The typology described in Section 2.2.2.5 was used
to code sensitivity, with an additional differentiation within the "no agreement" category. In all
"no agreement" cases, there was a mixture of codes for intermediate sensitivity along with low
and/or high sensitivity; if at least four participants provided judgments, and there were more high
sensitivity judgments than low sensitivity judgments, then the dashed arrow was colored orange
to indicate intermediate-to-high sensitivity. Under current conditions, 19 influences for which
there was agreement were categorized as intermediate sensitivity. Five influences were
categorized as high sensitivity: both the influence of Restoration and of Land Use Change on
Landscape Mosaic, Landscape Mosaic on the endpoint, and both the influence of Tides and
2-33
-------
Land UseChanae
Water Management
Sediment
Supply
Fresh water Inflow
Landscape
Mosaic
Wind/Waves
Water Quality
(Contaminants,
Salinity)
Tides and
Hydrodynamics
Bed Sediment
Sediment
Resuspensionand
Deposition
Characteristics/
Mudflat I KK
Bathymetry
Predators and
Disturbance
(Anthropogenic)
Extent of Mudflat
(Acre Hours)
Shorebird Prey
Community
Shore birds
Marbled Godwit
& Western
Sandpiper
Key
Intermediate sensitivity
High sensitivity
Intermediate-to-high trend
No agreement
Figure 2-9. Community Interactions group summary influence diagram of
sensitivities under current conditions.
2-34
-------
Hydrodynamics and of Mudflat Bathymetry on Extent of Mudflat. There was no agreement on
sensitivity for 12 influences. There were no instances of agreement on influences with low
sensitivity.
Figure 2-10 compares the sensitivities as in Figure 2-9, across the three scenarios. Under
Climate Scenario A, 10 influences for which there was agreement were categorized as
intermediate sensitivity. Five influences were categorized as high sensitivity: four of the same as
under current conditions, with a change the influence of Land Use Change on Landscape Mosaic
to no agreement and new agreement for the influence of Predators and Disturbance on the
endpoint. The number of influences with no agreement increased substantially to 21. Seven of
those are in disagreement because there is a combination of intermediate and high sensitivity
(orange arrows). This decrease in agreement reflects a trend of increasing sensitivity for some
participants, but not enough to shift to a new category. It could be indicative of either
disagreement about at what point such a shift would occur or differing assumptions about what
falls outside the current range of variability, which was left up to each participant to decide based
on their own knowledge and intuition.
Current
Scenario A
Scenario B
Key
High sensitivity
Intermediate-to-high trend
No agreement
Figure 2-10. Community Interactions group summary influence diagrams of
sensitivities: variance across current conditions and two climate scenarios.
Under Climate Scenario B, seven influences for which there was agreement were
categorized as intermediate sensitivity. Six influences were categorized as high sensitivity, with
2-35
-------
the addition of Extent of Mudflat on the endpoint, which had been intermediate under current
conditions. This is another way to identify a threshold, when there is a change in sensitivity to a
more sensitive category. The number of influences with no agreement increased again to 23;
however, for six of these the lack of agreement was due to a mixture of intermediate and high
sensitivity codes (orange arrows).
As with the Sediment Retention group, there was more variability in judgments among
participants than across scenarios for any given participant. The majority of changes in
sensitivity across the climate scenarios are of increasing sensitivity. Further description, as well
as figures depicting variability in judgments across participants, can be found in Appendix B.
2.3.2.4. Relative Impact
Figure 2-11 and 2-12 present the characterization of relative impact between current and
future climate scenarios (the group's discussion did not differentiate between the two future
climate scenarios). Relative impact was distinguished among the influences by indicating
primary, secondary or tertiary levels of relative impact. For the Community Interactions group,
the relative impacts of five influences were indicated as important under current conditions,
based on the discussion. The influences of Landscape Mosaic and of Extent of Mudflat on the
endpoint were both identified as having primary impacts. The influences of Predators and
Disturbance and of Shorebird Prey Community on the endpoint were identified as having
secondary impacts, and the influence of Bed Sediment Characteristics and Quality on the
endpoint was identified as having tertiary impact.
A total of 10 influences were indicated as having high relative impact under climate
change conditions for the Community Interactions group (see Figure 2-12). Three of the
influences indicated as having high relative impact under current conditions increased in relative
impact when considering future climate conditions: the influences of Landscape Mosaic
(restoration of which will be increasingly critical as mudflat habitat is threatened by sea level
rise), Predators and Disturbance (as predators will be able to prey on wading birds with greater
efficiency as they become concentrated on shrinking mudflats), and of Shorebird Prey
Community (density of which becomes increasingly important has mudflat extent decreases).
The relative impact of Extent of Mudflat on the endpoint stayed equally important.
2-36
-------
Key
Water
Management
Land UseChanee
Freshwater Inf ow
SedimentSupply
Landscape Mosaic
Wind/Waves K
Water Quality
(Contaminants,
Salinity)
Sediment
Resuspension
and Deposition
Bed Sediment
Characteristics
and Quality
Hydrodynamics
Predators and
Disturbance
(Anthropogenic)
Extent of Mudflat
(Acre Hours)
Shorebird Prey
Community
Marbled Godwit &
Western Sandpiper
-Primary impact
"Secondary impact
Tertiary impact
Figure 2-11. Community Interactions influences indicated as having high
relative impact under current conditions.
2-37
-------
Water
Manaeement
Land Use Change
Freshwater Inflow
SedimentSupply
Landscape Mosaic
Wind/Waves K
Water Quality
(Contaminants,
Salinity)
BedSediment
Characteristics
and Quality
Resuspension
and Deposition
Hydrodynamics
Predators and
Disturbance
(Anthropogenic)
Extent of Mudflat
(Acre Hours)
Shorebird Prey
Community
Marbled Godwit &
Western Sandpiper
Key
Relative impact remains the same under climate scenarios
Increased impact under climate scenarios
Figure 2-12. Community Interactions group influences indicated as having
high relative impact under climate scenarios.
2-38
-------
Five additional influences were indicated as having high relative impact under the climate
change scenarios. In addition, the influence of disease on the endpoint was identified as an
influence of emerging impact. Disease was not an original key variable in the influence diagram,
as variables were included based on importance under current conditions. This influence was not
scored, but was considered to be important by the participants. The influence of Bed Sediment
Characteristics on the endpoint was indicated as having high relative impact under current
conditions but not under the climate scenarios. It is unclear whether this was intentional or was
just not covered in the discussion of relative impact under future climate conditions.
2.3.2.5. Confidence
The confidence results shown in Table 2-11 are provided for the Community Interactions
influences for which there was agreement on type. The lack of agreement on confidence for
two-thirds of the judgments is a major gap, limiting our ability to prioritize around confidence
judgments. Five of the six influences that for which there was agreement on confidence across
all scenarios were scored as HH. The influence of Freshwater Inflow on Net Organic
Accumulation was scored as LH across all scenarios. The HH type of confidence was the most
used type of judgment. The dominant pattern on confidence across the climate scenarios was a
decrease in the number of influences on which there was agreement.
2.3.2.6. Interacting Influences
Under current conditions, there were no interactive influences for which there was
agreement among participants in the Community Interactions group. Likewise, under both
climate scenarios there were no interactive influences for which there was agreement on the type
of interactive influence. This lack of agreement was primarily due to not having many
influences with enough participants characterizing the same interacting influences. Of the
24 combinations of influences with interactions characterized by participants, only four had at
least three participants make any kind of judgment, which was the threshold for agreement, but
those were not ever in agreement on a type of interaction.
2.4. DISCUSSION OF ADAPTATION STRATEGIES
With background on strategic priorities provided by SFEP, the workshop participants
discussed the implications of the exercise results for management. Table 2-12 lists adaptation
strategies that emerged during the group discussions. The experts discussed a variety of general
adaptation strategies as well as some specific adaptation activities that would be responsive to
key potential climate-related changes identified through their judgments. The strategies fall into
2-39
-------
to
-k
o
Table 2-11. Community Interactions group confidence for influences with agreement
CURRENT
SCENARIO A
SCENARIO B
A
HH
HH
NA
B
HH
NA
NA
C
NA
NA
NA
D
LH
NA
NA
E
HH
NA
NA
F
NA
NA
NA
I
HH
NA
NA
K
NA
NA
NA
L
NA
NA
NA
0
LH
LH
LH
P
NA
NA
NA
Q
HH
NA
NA
S
NA
NA
NA
U
HH
HH
HH
X
HH
HH
HH
Y
NA
HH
HH
AA
NA
NA
NA
BB
LH
NA
NA
CC
HH
NA
NA
DD
HH
HH
HH
EE
HH
HH
HH
FF
NA
LH
LH
GG
HH
NA
NA
HH
NA
NA
NA
JJ
HH
HH
HH
NA = no agreement; HH = high evidence, high agreement; HL = high evidence, low agreement; LH = low evidence, high agreement; LL = low evidence, low
agreement.
-------
Table 2-12. Adaptation strategies and associated top pathways for
management (see Figures 3-3 and 3-4 for pathways)
Adaptation strategies
Start restoration soon to achieve functions of mature marshes, including attainment of threshold
elevations for organic accumulation, ahead of sea level rise
Plan for the temporal progression of habitats (e.g., by establishing habitats that will thrive under future
climate conditions)
Plan for the spatial progression of restoration (e.g., consider impacts of broaching Suisun Marsh
levees on downstream estuary restoration efforts)
Maintain adjacent transitional uplands to allow for local marsh migration
Move restoration focus from fringing marshes to where there is available space for multiple habitats
Create mosaics of habitats where there are opportunities for migration upslope
Plan restoration projects to provide connectivity
Sort sites with restoration potential based on where there is flexibility in management
Support resilience by restoring habitat complexity and facilitating high-energy parts of the system
such as tides, wind-driven waves, and freshwater flows
Develop policies that encourage removing or preventing barriers to marsh migration and discourage
new development on lands where there is restoration potential
Move highways and railroads that are barriers to marsh migration where there is otherwise space for
marsh expansion/migration
Preserve habitats that are unlikely to persist under future climate conditions as interim habitats until
alternate habitats that serve the same ecosystem functions can be established
Practice integrated water management, including water conservation, as a priority
If it is not possible to make maintaining marsh salinity a top priority for Delta freshwater storage
policies, plan for the restoration of tidal wetlands further up the estuary
Develop methods to move sediment into the bay, to keep pace vertically with sea level rise
Develop methods to reduce wave action on the front side of marshes
Adjust policies that prevent coarse sediment from entering the bay (e.g., for streams that don't support
salmonids, change policies to allow an increase in sediment load)
Involve authorities in flood control districts to recouple streams to wetlands
Monitor change at the landscape scale to assess management effectiveness
Develop rapid response plans for catastrophes (e.g., levee breaks), with the political and scientific
bases in place to respond properly
Pathways
SG, CG
SG, CG
SG, CG
CG
CG
CG
CG
CG
SB, SP, CG
CG
CG
CG
SG, SP, CB
SG, CG
SB, SP, CB,
CP
SB, CB
SG, SP, CG
SP, CG
SB, CG
SB, CG
SG = Sediment Retention Green pathway; SB = Sediment Retention Blue pathway; SP = Sediment Retention Purple
pathway; CG = Community Interactions Green pathway; CB = Community Interactions Blue pathway;
CP = Community Interactions Purple pathway
2-41
-------
several broad categories, including Restoration and Conservation, Sediment Management and
Planning and Monitoring.
2.4.1. Restoration and Conservation
Restoration was identified as a powerful management tool with a variety of specific
planning and prioritization considerations. The experts emphasized the urgency of implementing
restoration projects as an immediate priority, taking climate change impacts into consideration in
planning. For marsh restoration, the key is getting started early enough so that marshes can be
established before rates of sea level rise become too high. As a restored marsh matures, it is
better able to keep pace vertically through vegetative production. Similarly for other types of
restored habitats, they will be more resilient to changing climate conditions as they mature. The
other temporal issue for restoration planning was the need to plan for ecological succession,
building a dynamic landscape mosaic that includes habitats that will thrive under future climate
conditions.
Similar and related considerations apply to conservation strategies. One consideration for
conservation could be habitats that are well suited to future climate conditions. Though some
habitats may not survive climate change (e.g., where there are not long term opportunities for
migration), it may still be important to preserve and restore these habitats in the short and
medium term as interim habitats until alternate habitats that serve similar ecosystem functions
have been established.
Available space for habitat migration is a major consideration for both restoration and
conservation. On the conservation side, adjacent transitional uplands should be maintained to
allow for local habitat migration. Policy options may include regulation or incentives to
encourage relocation and to discourage development on lands where there is potential for
upslope habitat migration or restoration. The slope of the adjacent uplands is an important factor
in such conservation priorities, and the need for improved vertical data as a mapping priority was
highlighted in the workshop discussion. Such mapping would help to identify upland areas for
restoration adjacent to current healthy marshes, where migration of marshes is possible. The
experts especially emphasized the need to identify and prioritize wetland areas for restoration
where the adjacent uplands currently include complementary habitats that would contribute to a
complete landscape mosaic that could support valued species such as shorebirds that will be
stressed by climate change. Another key part of developing and conserving such landscape
mosaics is providing for connectivity between multiple habitats. Also underscored was the need
to focus restoration efforts in the North Bay because the shoreline of the South Bay is so
developed that it precludes the ability of marshes to migrate upland. Finally, many restoration
2-42
-------
efforts to date have involved fringing marshes. The focus of restoration could be expanded
beyond fringing marshes to larger areas where there is available space for multiple habitats.
A major spatial planning consideration for restoration is the need to consider the impacts
of each project on adjacent and downstream habitats and future restoration projects. In
particular, any project that broaches levees in Suisun Marsh will impact downstream sediment
budgets and adjacent hydrodynamics. Restoration projects could be coordinated so that projects
can be planned and timed to maximize success. Deciding when and where (e.g., how far
upstream) to focus restoration of different marsh types will also depend on changes in the
salinity gradient; conditions that suit freshwater and brackish marshes will be moving upstream
through time under climate change, unless maintaining marsh salinity becomes a priority for
Delta freshwater storage policies.
2.4.2. Sediment Management
Sediment management is already a priority within the region and will continue to be an
important focus for marsh management in the context of sea level rise and changing precipitation
patterns. The supply of sediments is declining as the estuary shifts from a system with larger
sediment loads as a result of past hydraulic mining to one that has a reduced sediment supply due
to the cessation of mining and an increase in tributery dams that trap sediment upstream (Wright
and Schoellhamer, 2004). Thus it will become increasingly important to support movement of
inorganic sediment into restoration sites in the near term, so that salt marshes can build to
threshold elevations for vegetation establishment and begin contributing organic sediment to
maintain themselves.
Changes in sediment supply will require local tributaries and Delta sources to be
managed differently. On the tributary side, there are opportunities to reconnect streams to
wetlands through flood control districts. An example of a specific option for increasing
inorganic sediment loads from local tributaries would be to for regional water boards to consider
adjusting sediment Total Maximum Daily Loads (TMDLs) to allow for increases in coarse
sediment loads in streams that do not support salmonids. On the Delta side, Integrated Water
Management will be increasingly important for maintaining sediment supply, and may prompt a
switch in priority from storage to conservation. In the Bay, dredge sediment reuse is an
opportunity to redistribute sediment to desired locations. Limiting factors on current use of this
technique were discussed by the workshop participants, including the need for best management
practices and funding.
While most of the discussion on sediment was about keeping pace vertically with sea
level rise, horizontal impacts through marsh edge erosion were also discussed. There is a need to
2-43
-------
develop ways to reduce wave action on the front sides of marshes. Protecting adjacent mudflats,
such as with berms, is one specific option.
2.4.3. Planning and Monitoring
The final category of adaptation strategies discussed at the workshop addressed planning
and monitoring. Many of the above recommendations are based on planning, including
prioritizing. The need to develop rapid response plans for catastrophes or contingency plans for
when thresholds are passed was emphasized. Preparing the political and funding conditions
necessary to implement such plans would be essential.
Monitoring will become increasingly important in order to detect when thresholds are
being crossed. The scales at which monitoring is focused will have to be adapted to changes in
restoration priorities. Monitoring at the landscape scale—especially for birds and other mobile
species that use multiple habitat types—will be a necessity in order to track potential thresholds.
This will likely require coordination among multiple agencies since many habitats and habitat
mosaics span jurisdictional boundaries. The current condition and extent of these habitats needs
to be monitored and understood now, as the current data are insufficient for a baseline at the
landscape level. Examining habitats at a larger scale will also be important for facilitating
species movements. It may become necessary to ensure that birds can move among ponds, tidal
marshes, and mudflats as conditions change. Some species (e.g., clapper rail, salt marsh harvest
mouse) may not be able to migrate from degrading areas on their own and may require
intervention.
The discussion of adaptation strategies described above was broad and free-ranging. The
next section will combine the analysis of the exercise results with the ideas in Table 2-12 to
discuss top pathways for management given climate change and identify specific adaptation
options in response.
2-44
-------
3. MAKING THE LINK TO MANAGEMENT
As detailed above, the workshop resulted in a large volume of information on the
sensitivities of the sediment retention and community interactions processes to stressor
interactions under current conditions and future climate scenarios. The next step lies in
organizing this information into a form which managers can use to identify influences of
particular importance upon which to focus management interventions and adaptation planning.
3.1. USING INFORMATION ON INFLUENCE TYPE AND DEGREE, SENSITIVITY
AND RELATIVE IMPACT TO IDENTIFY KEY MANAGEMENT PATHWAYS
In the workshop exercise and group discussions, the experts generated three categories of
information about the relationships in the influence diagrams: (1) the type and degree of each
influence; (2) the sensitivity of each influence (including thresholds); and (3) the high relative
impact of certain influences on the endpoints. All three categories of information should be
considered in concert when interpreting management implications. This can be done by
performing a "crosswalk" of all three categories of information in order to identify pathways of
particular interest that connect each endpoint (Net Accretion/Erosion or Shorebirds) to stressors
or drivers that can be addressed through particular management activities. The crosswalks as
well as example pathways are presented below.
3.1.1. Crosswalks: Influence Type and Degree, Sensitivity and Relative Impact
The crosswalks for Sediment Retention and Community Interactions are presented in
Table 3-1 and 3-2. For each influence, information on type and degree, sensitivity, and relative
impact is listed side-by-side, first for current conditions, followed by Climate Scenarios A and B.
This allows for easy comparison of all three categories of information, across all three scenarios.
The influences have also been rank-ordered based on the amount of information available for
each in terms of agreement on influence type, degree, sensitivity, relative impact and threshold
potential.
3.1.1.1. Sediment Retention Crosswalk
For Sediment Retention (see Table 3-1), there was agreement on both type and sensitivity
for the majority of influences. Especially when coupled with the designation of high relative
impact, certain influences emerge as being of special interest for management. These are
influences for which we have a good understanding of the nature of the relationships, their
3-1
-------
Table 3-1. Sediment Retention group crosswalk for comparison of influence type and degree, sensitivity and
relative impact for current conditions and climate scenarios. NA = no agreement; prop = proportional;
disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; X = high relative impact; i = increasing relative impact from current;
() = number of respondents; ranking column orders the influences according to completeness of information
Influence
N
Z
M
U
B
A
C
Variable X
Relative sea
level
Wind/waves
Tides
Net organic
accumulation
Water
resource
management:
reservoir
management
Water
resource
management:
delta outflow
Water
resource
management:
reservoir
management
On
on
on
on
on
on
on
on
Variable Y
Inundation
regime
Sediment flux
Inundation
regime
Net
accretion/erosion
Freshwater
inflow
Freshwater
inflow
Sediment flux
CURRENT
Influence
Direct
prop (7)
Direct
prop (5)
Direct
prop (4)
Direct
prop (5)
Inverse
prop (5)
Direct
prop (5)
Inverse
prop (5)
Sensitivity
1(7)
1(5)
1(4)
1(5)
1(6)
1(5)
1(5)
Relative
impact
X
X
X
X
CLIMATE A
Influence
Direct
prop (5)
Direct
prop (4)
Direct
prop (4)
Direct
prop (4)
Inverse
prop (4)
Direct
prop (4)
Inverse
prop (4)
Sensitivity
1(5)
1(5)
1(4)
1(4)
1(5)
1(4)
1(4)
Relative
impact
X
[threshold]
t
[threshold]
X
t
t
CLIMATE B
Influence
Direct
prop (5)
Direct
prop (4)
Direct
prop (5)
Direct
prop (4)
Inverse
prop (4)
Direct
prop (4)
Inverse
prop (4)
Sensitivity
1(5)
1(5)
1(5)
1(4)
1(5)
1(4)
1(4)
Relative
impact
X
[threshold]
1
[threshold]
X
1
t
Ranking
1
1
2
2
3
4
4
to
-------
Table 3-1. Sediment Retention group crosswalk for comparison of influence type and degree, sensitivity and
relative impact for current conditions and climate scenarios. NA = no agreement; prop = proportional;
disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; X = high relative impact; T" = increasing relative impact from current;
() = number of respondents; ranking column orders the influences according to completeness of information
(continued)
Influence
D
K
L
P
Q
V
Y
J
R
Variable X
Water
resource
management:
reservoir
management
Relative sea
level
Freshwater
inflow
Tides
Sediment flux
Net mineral
accumulation
Net organic
accumulation
Wind/waves
Sediment size
On
on
on
on
on
on
on
on
on
on
Variable Y
Sediment size
Tides
Sediment flux
Sediment flux
Net mineral
accumulation
Net
accretion/erosion
Net mineral
accumulation
Sediment size
Net mineral
accumulation
CURRENT
Influence
Inverse
prop (6)
Direct
disprop,
weak (5)
Direct
prop (6)
Direct
prop (6)
Direct
prop (6)
Direct
prop (6)
Direct
prop (5)
Direct
prop (4)
Direct
prop (4)
Sensitivity
1(6)
L(5)
1(6)
1(6)
1(7)
1(7)
1(5)
1(4)
1(4)
Relative
impact
CLIMATE A
Influence
Inverse
prop (6)
Direct
disprop,
weak (5)
Direct
prop (5)
Direct
prop (4)
Direct
prop (6)
Direct
prop (5)
Direct
prop (4)
Direct
prop (4)
Direct
prop (4)
Sensitivity
1(6)
L(5)
1(5)
1(4)
1(6)
1(6)
1(4)
1(4)
1(4)
Relative
impact
CLIMATE B
Influence
Inverse
prop (6)
Direct
disprop,
weak (6)
Direct
prop (4)
Direct
prop (5)
Direct
prop (7)
Direct
prop (6)
Direct
prop (4)
Direct (7)
Direct (6)
Sensitivity
1(6)
L(6)
1(4)
1(5)
1(7)
1(7)
1(4)
H-trend
H-trend
Relative
impact
Ranking
4
4
4
4
4
4
4
5
5
-------
Table 3-1. Sediment Retention group crosswalk for comparison of influence type and degree, sensitivity and
relative impact for current conditions and climate scenarios. NA = no agreement; prop = proportional;
disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; X = high relative impact; T" = increasing relative impact from current;
() = number of respondents; ranking column orders the influences according to completeness of information
(continued)
Influence
G
W
X
T
F
O
S
H
Variable X
Water
resource
management:
channelization
Inundation
regime
Net accretion
/erosion
Inundation
regime
Water
resource
management:
channelization
Freshwater
Inflow
Inundation
regime
Land use/land
cover change:
impervious
cover
On
on
on
on
on
on
on
on
on
Variable Y
Sediment size
Wind/waves
Inundation
regime
Net organic
accumulation
Sediment flux
Net organic
accumulation
Net mineral
accumulation
Sediment flux
CURRENT
Influence
Inverse
(5)
Direct (5)
Inverse
(4)
Direct (4)
Inverse
(4)
NA
Direct (6)
Inverse
(4)
Sensitivity
L(4)
NA
1(4)
1(4)
L(4)
1(4)
NA
NA
Relative
impact
X
X
CLIMATE A
Influence
Inverse
(5)
direct (6)
Inverse
prop (4)
NA
NA
Direct (4)
Direct (6)
Inverse
(4)
Sensitivity
L(4)
H-trend
1(4)
H-trend
L(4)
NA
NA
NA
Relative
impact
X
X
CLIMATE B
Influence
Inverse
(4)
Direct
disprop,
strong
(4)
Inverse
(4)
NA
NA
NA
Direct (5)
Inverse
(4)
Sensitivity
L(4)
H(4)
NA
NA
L(4)
NA
H-trend
NA
Relative
impact
X
X
Ranking
6
6
6
7
8
8
8
9
-------
Table 3-1. Sediment Retention group crosswalk for comparison of influence type and degree, sensitivity and
relative impact for current conditions and climate scenarios. NA = no agreement; prop = proportional;
disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; X = high relative impact; T" = increasing relative impact from current;
() = number of respondents; ranking column orders the influences according to completeness of information
(continued)
Influence
I
E
Variable X
Land use/land
cover change:
impervious
cover
Water
resource
management:
channelization
On
on
on
Variable Y
Sediment size
Freshwater
inflow
CURRENT
Influence
NA
NA
Sensitivity
1(4)
NA
Relative
impact
CLIMATE A
Influence
NA
NA
Sensitivity
1(4)
NA
Relative
impact
t
CLIMATE B
Influence
NA
NA
Sensitivity
1(4)
NA
Relative
impact
t
Ranking
9
10
-------
sensitivity to changes now and in the future, and their high relative impact on the endpoint of Net
Accretion/Erosion. Therefore these are the influences for which management interventions are
most likely to have the intended effects. Influences ranked one through three in Table 3-1 fall
into this category.
Even when not designated as highest relative impact, influences for which there was
agreement on type as well as sensitivity are equally important to consider. While not necessarily
of highest relative impact, they are well understood and sensitive to change, and may be linked
with other influences for important cumulative effects on the endpoint. Influences of rank four
and five, and also influence G, fall into this category (see Table 3-1). Meanwhile, lack of
agreement on one or more of the type and sensitivity categories indicates that more information
is needed to understand the particular influence. It does not imply that the relationship is not
potentially important, but rather that it is not well enough understood by this particular group of
experts for managers to be confident about the response to either climate change or management
interventions. The remaining nine influences fall into this group. Relationship E (Water
Resource Management: Channelization on Freshwater Inflow) and Relationship T (Inundation
Regime on Net Organic Accumulation) are interesting cases in that there was no agreement on
influence type or sensitivity, but there was agreement on high relative impact. These influences
were identified as having high or increasingly-high relative impact under the climate scenarios.
This indicates that the influences are not well understood, yet are considered by the experts to
have a high relative impact on the ecosystem process endpoint of Net Accretion/Erosion. In the
case of these influences as well as the remaining influences in this group, priorities for further
investigation (in the form of literature reviews to more deeply assess existing information,
followed by new research where understanding is confirmed to be lacking) could be based in part
on which of these influences are most critical to understand since they have a high relative
impact or have links to other influences of special importance to the endpoint.
3.1.1.2. Community Interactions Crosswalk
The Community Interactions crosswalk (see Table 3-2) also has some influences for
which there was agreement on both type and sensitivity. Especially when coupled with the
designation of high (or increasing) relative impact across the scenarios, these influences emerge
as being of special interest for management. These are influences for which we have a good
understanding of the nature of the relationships, their sensitivity to changes now and in the
future, and their high relative impact on the Shorebirds endpoint. Therefore these are the
influences for which management interventions are most likely to have the intended effects.
These influences include Relationships O, GG, Q, DD, EE, FF, and E.
3-6
-------
Table 3-2. Community Interactions group crosswalk for comparison of influence type and degree, sensitivity
and relative impact for current conditions and climate scenarios. NA = no agreement; Prop = proportional;
Disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; t = increasing relative impact from current; () = number of
respondents; ranking column orders the influences according to completeness of information
Influence
0
GG
Q
S
DD
EE
U
Y
Variable X
Landscape
Mosaic
Shorebird prey
community
Wind/waves
Water quality
Extent of
mudflat
Extent of
mudflat
Tides and
Hydrodynamics
Sediment
resuspension /
deposition
On
on
on
on
on
on
on
on
on
Variable Y
Shorebirds
Shorebirds
Sediment
resuspension/
deposition
Shorebird prey
community
Shorebirds
Shorebird prey
community
Sediment
resuspension/d
eposition
Extent of
mudflat
CURRENT
Influence
Direct
disprop,
strong (4)
Direct
prop (5)
Direct
prop (6)
Direct
prop (5)
Direct
prop (4)
Direct
prop (6)
Direct
prop (5)
Direct
prop (4)
Sensitivity
H(4)
1(6)
1(6)
1(6)
1(4)
1(6)
1(5)
1(4)
Relative
impact
Primary
Secondary
Primary
CLIMATE A
Influence
Direct
disprop,
strong (5)
Direct prop
(5)
Direct prop
(5)
Direct prop
(4)
Direct (6)
Direct prop
(5)
Direct prop
(4)
Direct prop
(4)
Sensitivity
H(5)
1(5)
1(5)
1(4)
H-trend
1(5)
1(4)
1(4)
Relative
impact
t
t
t
[threshold]
Primary
1
CLIMATE B
Influence
Direct
disprop,
strong (6)
Direct prop
(4)
Direct prop
(5)
Direct prop
(4)
Direct
disprop,
strong (4)
Direct prop
(5)
Direct prop
(4)
Direct prop
(4)
Sensitivity
H(6)
1(4)
1(5)
1(4)
H(5)
1(5)
1(4)
1(4)
Relative
impact
t
t
t
[threshold]
Primary
t
Ranking
1
1
2
2
2
2
3
3
-------
Table 3-2. Community Interactions group crosswalk for comparison of influence type and degree, sensitivity
and relative impact for current conditions and climate scenarios. NA = no agreement; Prop = proportional;
Disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; t = increasing relative impact from current; () = number of
respondents; ranking column orders the influences according to completeness of information (continued)
Influence
FF
HH
BB
E
I
X
AA
D
CC
JJ
Variable X
Predators and
disturbance
Sediment
resuspension/
deposition
Bed sediment
characteristics
Restoration
Freshwater
inflow
Tides and
hydrodynamics
Extent of
mudflat
Restoration
Bed sediment
characteristics
Mudflat
bathymetry
On
on
on
on
on
on
on
on
on
on
on
Variable Y
Shorebirds
Mudflat
bathymetry
Shorebirds
Landscape
mosaic
Sediment
supply
Extent of
mudflat
Predators and
disturbance
Sediment
supply
Shorebird prey
community
Extent of
mudflat
CURRENT
Influence
Inverse
(6)
Direct
prop (5)
Direct
prop (4)
Direct (5)
Direct
prop (6)
NA
Inverse
prop (4)
Inverse
prop (4)
Direct
prop (5)
Direct (4)
Sensitivity
NA
1(5)
1(4)
H(4)
1(6)
H(4)
1(4)
1(5)
1(4)
H(4)
Relative
impact
Secondary
Tertiary
CLIMATE A
Influence
Inverse
disprop,
strong (4)
Direct prop
(4)
Direct (5)
Direct (4)
Direct prop
(4)
Inverse
disprop,
strong (4)
Inverse (5)
Inverse (5)
Direct (5)
Direct
disprop,
strong (4)
Sensitivity
H(4)
1(4)
NA
H(5)
1(4)
H(5)
NA
H-trend
H-trend
H(5)
Relative
impact
t
[threshold]
t
t
CLIMATE B
Influence
Inverse
disprop,
strong (4)
Direct prop
(4)
Direct (5)
Direct (4)
Direct (4)
Inverse
disprop,
strong (4)
Inverse (4)
Inverse (4)
Direct (5)
NA
Sensitivity
H(5)
1(4)
H-trend
H(5)
NA
H(6)
H-trend
NA
NA
H(4)
Relative
impact
t
[threshold]
t
t
Ranking
3
3
4
5
5
5
5
6
6
6
oo
-------
Table 3-2. Community Interactions group crosswalk for comparison of influence type and degree, sensitivity
and relative impact for current conditions and climate scenarios. NA = no agreement; Prop = proportional;
Disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; t = increasing relative impact from current; () = number of
respondents; ranking column orders the influences according to completeness of information (continued)
Influence
A
B
K
F
G
L
P
W
C
N
Variable X
Water
management
Water
management
Freshwater
inflow
Land use
change
Land use
change
Sediment
supply
Wind/waves
Sediment
resuspension/
deposition
Restoration
Landscape
mosaic
On
on
on
on
on
on
on
on
on
on
on
Variable Y
Freshwater
inflow
Sediment
supply
Tides and
hydrodynamics
Sediment
supply
Landscape
mosaic
Sediment
resuspension/
deposition
Tides and
hydrodynamics
Bed sediment
characteristics
Tides and
hydrodynamics
Predators and
disturbance
CURRENT
Influence
Direct
prop (5)
Direct
prop (4)
Direct (6)
Direct
prop (4)
NA
Direct
prop (4)
Direct (6)
NA
NA
NA
Sensitivity
1(7)
1(6)
NA
1(5)
H(4)
1(4)
NA
1(4)
NA
NA
Relative
impact
CLIMATE A
Influence
NA
NA
Direct (5)
NA
NA
NA
Direct (5)
NA
Direct (4)
NA
Sensitivity
1(5)
1(4)
NA
H-trend
H-trend
NA
NA
H-trend
NA
H-trend
Relative
impact
t
[threshold]
CLIMATE B
Influence
NA
NA
Direct (4)
NA
NA
NA
Direct (4)
NA
Direct (4)
NA
Sensitivity
H-trend
H-trend
NA
NA
H-trend
NA
NA
H-trend
NA
H-trend
Relative
impact
t
[threshold]
Ranking
7
7
7
8
9
9
9
9
10
10
-------
Table 3-2. Community Interactions group crosswalk for comparison of influence type and degree, sensitivity
and relative impact for current conditions and climate scenarios. NA = no agreement; Prop = proportional;
Disprop = disproportional; L = low sensitivity; I = intermediate sensitivity; H = high sensitivity; H-trend = no
agreement but trending toward high sensitivity; t = increasing relative impact from current; () = number of
respondents; ranking column orders the influences according to completeness of information (continued)
Influence
V
H
J
M
R
T
Z
KK
Variable X
Bed sediment
characteristics
Land use
change
Freshwater
inflow
Sediment
supply
Water quality
Sediment
resuspension/
deposition
Extent of
mudflat
Mudflat
bathymetry
On
on
on
on
on
on
on
on
on
Variable Y
Sediment
resuspension/
deposition
Water quality
Water quality
Bed sediment
characteristics
Bed sediment
characteristics
Tides and
hydrodynamics
Sediment
resuspension/
deposition
Sediment
resuspension/
deposition
CURRENT
Influence
NA
NA
NA
NA
NA
NA
NA
NA
Sensitivity
1(5)
NA
NA
NA
NA
NA
NA
NA
Relative
impact
CLIMATE A
Influence
NA
NA
NA
NA
NA
NA
NA
NA
Sensitivity
NA
NA
NA
NA
NA
NA
NA
NA
Relative
impact
CLIMATE B
Influence
NA
NA
NA
NA
NA
NA
NA
NA
Sensitivity
NA
NA
NA
NA
NA
NA
NA
NA
Relative
impact
Ranking
11
12
12
12
12
12
12
12
-------
Even when not designated as highest relative impact, influences for which there was
agreement on type as well as sensitivity are equally important to consider. While not necessarily
of highest relative impact, they are well understood and sensitive to change, and may be linked
with other influences for important cumulative effects on the endpoint. Influences for which
there was agreement on both type and degree across at least two of the three scenarios include
Relationships S, U, Y, HH, I, X and JJ.
Meanwhile, lack of agreement on one or more of the type and sensitivity categories
indicates that more information is needed on the particular influence. Again, it does not imply
that the relationship is not potentially important, but rather that it is not well enough understood
by this particular group of experts such that more information is needed. The remaining
influences all lacked agreement in type and/or sensitivity for at least two of the three scenarios.
Relationship K (Freshwater Inflow on Tides and Hydrodynamics) and Relationship AA (Extent
of Mudflat on Predators and Disturbance) are cases for which there was no agreement on
sensitivity across the climate scenarios, but there was agreement on high relative impact. These
influences were identified as having increasingly-high relative impact under the climate
scenarios (with Relationship K being designated a threshold response). This indicates that
although the influences are not fully understood, they are considered by the experts to have a
high relative impact on the Shorebirds endpoint. In the case of influences in this group, priorities
for further investigation (through literature reviews and further research where needed) could be
based in part on which of these influences are most critical to understand since they have a high
relative impact or have links to other influences of special importance to the endpoint.
It is notable that the community interactions influence diagram had a larger proportion of
influences that were not well understood compared to the sediment retention group. There were
eight influences for which there was no agreement on any of the three categories of information,
for any of the three scenarios. The larger number of influences in the Community Interactions
diagram (36 compared to 26 for Sediment Retention) reflects the complexity of modeling both
physical (e.g., sediment supply) and biological (e.g., shorebirds and their prey) components of
the community interactions process, and may have contributed to less agreement among the
participants, especially across scenarios.
3.1.1.3. Information Gaps
3.1.1.3.1. Crosswalks
Patterns of information gaps in the crosswalk tables varied for Sediment Retention (see
Table 3-1) compared to Community Interactions (see Table 3-2). Influences for Sediment
Retention were relatively well understood across type and degree and sensitivity categories of
3-11
-------
information. However, in quite a few cases, even though there was agreement on type there was
not agreement on degree. Where there was agreement under current conditions, the agreement
tended to carry across the climate scenarios as well. For Community Interactions, there was far
less agreement overall about the nature of the influences, with a greater number of gaps in
influence type, degree and sensitivity. Also, compared to Sediment Retention there were more
cases where agreement that was present under current conditions was lost under the climate
scenarios, indicating a greater uncertainty about how the influences might behave in the future.
Such information gaps—especially involving influences in otherwise well-understood
pathways that link to rich opportunities for management—could be used to prioritize targeted
literature reviews and/or scientific research that focuses on key process components of interest.
Another method for sorting through and prioritizing "nonagreement" influences for further study
might be to start from the perspective of management opportunities. Managers could look at
their most tractable and effective management levers currently available, and trace pathways
from those down to the endpoint of interest, as a means of identifying and selecting priority
influences for research. Examples of promising pathways are presented below.
3.1.1.3.2. Confidence
Confidence estimates were not included in the crosswalk tables because of extensive
information gaps in the form of missing estimates. It is possible that this was partly due to time
limitations as participants prioritized characterizing the influences before marking confidence.
However, another problem that may have led to gaps was that the confidence exercise did
not take into account specialty areas of participant knowledge. Due to the complex and
interdisciplinary nature of the influence diagrams and the individual specialties of the
participants, some participants may have been asked to make judgments on influences for which
they felt they had insufficient expertise. In some cases they may have elected to leave those cells
blank.
Even where confidence estimates were entered, there is cause for caution in interpreting
the information. Discussions during and after the exercise revealed some confusion about the
definitions of evidence and agreement (as per the confidence handout in Appendix D). In
particular, there may have been a misunderstanding related to equating agreement alone (even
where there was minimal evidence) with full confidence; that is, a large amount of agreement,
even where little information (data) was available, may have been misconstrued as highest (HH)
confidence in some cases.
Thus, the large number of missing cells for confidence could have been due to one or
more of the following: (1) lack of time; (2) inability to judge confidence in certain influences due
3-12
-------
to lack of expertise; and (3) confusion about the confidence definitions and coding scheme.
These problems could be corrected in subsequent workshops through preworkshop trainings to
clarify the coding scheme, provision of a code to allow participants to indicate lack of expertise
as a reason for leaving a cell blank, and additional time to complete the exercise.
3.1.2. Identifying Key Pathways for Management
Using the crosswalk tables (see Tables 3-1 and 3-2), it is possible to identify influences
that are well understood, become more sensitive, and have a greater relative impact under future
climate scenarios. By combining a series of such influences into a pathway to the endpoint, we
can begin to identify key responses and changes in variables of interest to management. A
"pathway" is defined as a series of connected variables and their influences, beginning with a
driver or stressor variable and ending at the endpoint. The purpose is to be able to apply
management interventions in order to impact the endpoint. "Management levers" are those
variables for which it is possible to intervene with management options; the clearest connections
to management options are for the top level variables that are drivers or stressors. When
multiple management levers are available for a pathway, the one that was more completely
characterized or that had potential changes under the climate scenarios was selected. An
example pathway from the Sediment Retention process (see Figure 3-1) is described here, to
show the process by which these types of pathways can be identified. This will be followed in
the next chapter by summary diagrams showing the top three pathways for each process, along
with discussion of specific management options.
In the Sediment Retention process, the pathway of Reservoir Management to Freshwater
Inflow (Relationship B) to Net Organic Accumulation (Relationship O) to the endpoint
(Relationship U) is a relatively direct route to the endpoint of Net Accretion/Erosion (see
Figure 3-1). For type and degree of influence, Relationship B was characterized as being an
inverse proportional influence under all scenarios. For sensitivity, Relationship B was
characterized as having intermediate sensitivity under all scenarios. In terms of relative impact,
Relationship B was indicated as an influence with increasing impact under the climate change
scenarios.
Relationship O had less agreement. For type and degree of influence, Relationship O was
characterized as being a direct influence under Climate Scenario A (not shown in Figure 3-1),
but there was no agreement on type and degree under current conditions or Climate Scenario B.
For sensitivity, Relationship O was characterized as having intermediate sensitivity under current
> 1O
5-13
-------
Influence Type
Sensitivity
Relative Impact
Current
Future
— Direct
__ Inverse
— Noagreement
Thickness denotes degree: all are proportional
^^_ Intermediate sensitivity
Noagreement
High impact
Increasing impact under climate scenarios
Lowerimpact
Figure 3-1. Sediment Retention example pathway. Future = Climate Scenario B.
-------
conditions, but there was no agreement under the climate scenarios. In terms of relative impact,
Relationship O was indicated as having high impact under both current and future scenarios. An
area for further investigation would be the source of disagreement on the influence type for
current conditions versus Climate Scenario B. Is there a potential for a threshold since multiple
participants changed their characterization of type between each scenario? Or was the
disagreement based on differences in definitions between participants, as the influence could act
differently based on considering a change in timing or volume of flow? In order to use Reservoir
Management to impact Net Accretion, it will be necessary to understand the nature of the
influence of Freshwater Inflow.
Relationship U was characterized as a direct proportional influence under all scenarios. It
has intermediate sensitivity under all scenarios. In terms of relative impact, it was identified as
having high impact under current conditions and increasing impact under the climate scenarios.
When examined using all three categories of information, this is a relatively well
understood pathway for which there was agreement on type for two of the three influences.
Where there is agreement on sensitivity along this pathway, the influences are characterized as
intermediate sensitivity, which indicates that they would be responsive to management actions.
Each influence has high relative impact, with both Relationships B and U having increased
impact under the climate change scenarios. To manage along this pathway, an increased
understanding of Relationship O would be important. Understanding the specifics of how the
timing or volume of freshwater inflow influences the aboveground and belowground organic
processes that lead to net accumulation would complete the pathway so managers can utilize
reservoir management to increase net accretion.
3.2. TOP PATHWAYS AND IMPLICATIONS FOR ADAPTATION PLANNING
Section 3.1.2 above has used an example to demonstrate how the results of the expert
elicitation exercise can be used to help identify key pathways for management. This method of
identifying well-understood pathways that can be traced from endpoints of concern to
management levers is a useful way to explore the implications of the workshop results for
adaptation planning. In some cases it may be possible to identify management actions for
immediate implementation, i.e., where there is sufficient understanding of the relationships
among the variables as well as their sensitivities to act with relative confidence in the effects of
management interventions. Additional pathways of interest can be identified through further
examination of the crosswalk tables (see Tables 3-1 and 3-2), using amount of information for
which there was agreement (to identify current best-understood influences) as well instances of
climate thresholds (indicating potential climate-induced shifts) to identify "top pathways" of
3-15
-------
interest for management. This section describes three top pathways for the Sediment Retention
and Community Interactions processes, as well potential adaptation responses. This is followed
by a brief review of SFEP planning documents and discussion of where adaptation activities
relating to the top pathways could be linked into these existing plans and strategies.
3.2.1. Top Pathways and Associated Adaptation Options
Three top pathways for each process are presented in Figure 3-2 (Sediment Retention)
and Figure 3-3 (Community Interactions). For ease of viewing, each pathway is highlighted by a
color (green, purple or blue), and influences that undergo changes under the climate scenarios are
highlighted with red boxes indicating the nature of the change. Dashed lines indicate
inconsistent agreement among participants under at least one scenario. The order in which the
pathways are presented below is not an indication of order of importance. These are all
management pathways with notable potential for addressing the climate sensitivities identified.
3.2.1.1. Sediment Retention Top Pathways
3.2.1.1.1. Green pathway
The Sediment Retention example pathway described in Section 3.1.2.1 above is
elaborated upon here as the Green top pathway (see Figure 3-2). Starting with the Net
Accretion/Erosion endpoint and working "up" the diagram, a major determinant of the net
balance between accretion and erosion is the contribution of organic accumulation by way of
below ground biomass production. Organic accumulation has a higher relative impact on the
endpoint than mineral accumulation, and this relative impact is expected to increase under
climate change. This is because as marshes are challenged to adjust to sea level rise, vegetative
processes can respond by increasing below ground biomass productivity, and this may become
increasingly important in the context of historical (and continuing) decreases in mineral sediment
supplies.
At the next level up the pathway, net organic accumulation is directly affected by
freshwater inflow due to salinity effects. As freshwater flows decrease, salinity increases,
favoring more salt tolerant, but less productive plants in the community. Agreement on the
nature of this relationship was not consistent across the climate scenarios, with some participants
indicating the potential for an inverse relationship depending on whether or not they were
considering changes in species composition. If invasive Spartina alterniflora, which is favored
and has higher productivity than native species under higher salinity regimes, is allowed to
become established, then the influence of freshwater inflow could be an inverse relationship in
which increased salinity favored vegetative production. However, for the purposes of this
3-16
-------
Direct > Inverse
Effect (Threshold)**
Sediment Flux Sediment Size
Figure 3-2. Top pathways for management of the Net Accretion/Erosion
endpoint. Blue, green and purple colors are used to distinguish different
pathways. Red boxes highlight changes under future climate conditions. *
indicates high relative impact under current conditions. A indicates increasing
relative impact under future conditions. A direct to inverse threshold occurs
where there is a direct effect under current conditions that may shift to an inverse
effect under future climate conditions. Dashed lines indicate inconsistent
agreement across scenarios.
5-17
-------
Water
anagement
Land Use Change
Increasing
Direct Effect
Landscape
Mosaic
reshwater Inflow
Sediment Supply
Water Quality
(Contaminants,
Salinity)
Direct > Strong
Direct Effect
(Threshold)
Sediment
Resuspension
and Deposition
Bed Sediment
haracteristicsand
Quality
Hydrodynamics
Mudflat
Bathymetry
Predatorsand
Disturbance
(Anthropogenic)
Strong Direct
Effect. 1°A
Strong Direct Effect
Shorebird Prey
Community
Extent of Mudflat
(Acre Hours)
Direct > Strong Direct
Effect (Threshold), 1°
Direct Effect, 20/N
Marbled Godwit&
Western Sandpiper
Figure 3-3. Top pathways for management of the Shorebirds endpoint.
Blue, green and purple colors are used to distinguish different pathways.
Red boxes highlight changes under future climate conditions. 1° and 2° indicate
primary and secondary relative impact under current conditions, respectively. A
indicates increasing relative impact under future conditions. A direct to strong
direct threshold occurs where there is a direct effect under current conditions that
may shift to a very strong direct effect under future climate conditions. Dashed
lines indicate inconsistent agreement across scenarios.
5-18
-------
discussion we are assuming that the goal is to preserve the native salt marsh habitat; hence we
are considering management of the direct relationship.
Finally, freshwater inflow is acutely affected by reservoir management (see top level of
pathway). Here, reservoir management refers to the practice of storing and diverting freshwater
supplies for flood control, agriculture and other human uses. Thus, an increase in reservoir
management typically results in an overall decrease in availability of freshwater flows to salt
marshes. This inverse effect on the volume, speed and seasonal availability of freshwater flows
is expected to be of increasingly high relative impact under climate change as freshwater
supplies become increasingly variable in a context of increasing human demand.
The management implications for adaptation given the relationships within this pathway
are that managing reservoirs for downstream salinity regulation in favor of native marsh
vegetative productivity will require more steady, lower-volume releases. The strategy for such
releases relative to the growing season requires further investigation of the effects of timing
versus volume of freshwater inflow on net organic accumulation, on a site-specific basis. Given
the high relative impact of net organic accumulation compared to net mineral accumulation, a
priority could be placed on releases designed for salinity maintenance compared to high volume
pulses to support mineral sediment transport (discussed in the Purple pathway below). In
addition, since increased salinity regimes favor invasive Spartina, if the goal is to preserve native
salt marsh, another important management consideration will be the need to continue or even
"step up" invasive species eradication programs.
3.2.1.1.2. Purple pathway
Starting with the Net Accretion/Erosion endpoint and working up the Purple pathway
(see Figure 3-2), a second major determinant of the balance between accretion and erosion is the
contribution of mineral accumulation. Net mineral accumulation has a direct effect—in
conjunction with organic accumulation—on net accretion.
The next level up the Purple pathway represents the effect on mineral accumulation of
sediment size. This is another direct relationship, with larger grain sizes favoring mineral
accumulation, since larger grains deposit more readily, are harder to resuspend and provide
larger building blocks for accretion. An important effect of future climate change may be an
increasing sensitivity of net mineral accumulation to sediment size. This increase is partly
because sediment flux, the other determinant of net mineral accumulation, is expected to
continue to decrease, not only because of continuing processes responsible for historical declines
from peak sediment inputs in the past, but also because of potential changes in wave driven
erosive processes (see Blue pathway description below).
3-19
-------
At the top level of the diagram, the Purple pathway links sediment size to
two management levers: impervious cover and reservoir management. It is likely that changes in
impervious cover will have important effects on sediment size, but our understanding of these
effects is incomplete. The influence of impervious surfaces on runoff, and resulting impacts on
sediment size, is highly dependent on the location of impervious surface relative to other land
cover, topography and proximity to water bodies. Workshop participants noted that the effect of
impervious surfaces on runoff is much greater than the expected effect of climate driven changes
in precipitation and resulting runoff patterns. Meanwhile, the effect of reservoir management on
sediment size is better understood. As flow volumes and speeds are reduced, there is a negative
impact on transport of large grains.
Management implications for adaptation based on this pathway vary. In the case of the
impervious cover management lever, further investigation of the effects of impervious cover on
sediment size is greatly needed in order to identify appropriate management strategies. Basic
information on how changes in land cover (including changes from impervious to permeable
pavement systems) may affect sediment size will be critical as land use change and development
continue to increase into the future.
With regard to reservoir management, high volume releases will increase transport of
larger grain sediments to marshes. However, there is a trade-off to consider between high
volume pulses to enable large grain sediment transport and water availability for steady,
lower-volume releases to favor vegetative production as discussed for the Green pathway (see
above).
Other options for addressing sediment size through management could include adjusting
policies that prevent coarse sediment from entering the Bay. This could include a change in
TMDL requirements for streams that do not support salmonids, to allow an increase in sediment
loads. Engaging with flood control districts is another possible avenue for recoupling stream
sediments to wetlands. Current flood control priorities may not take into account the future
benefits of downstream wetlands as climate change and associated sea level rise begin to
increase the occurrence of flooding due to changes in upstream tidal pulses. Also there is an
opportunity to take advantage of strategies that maximize sediment transport to wetlands, with
the additional benefit to flood control districts of decreasing their dredging needs.
3.2.1.1.3. Blue pathway
Like the Purple pathway, the Blue pathway (see Figure 3-2) starts with the direct effect
on net accretion of net mineral accumulation, but from there it diverges. The next step up this
pathway concerns the direct relationship of net mineral accumulation to sediment flux, which is
5-20
-------
an expression of the rate of sediment supply. All else being equal, increases in sediment flux
result in increased net mineral accumulation.
An important driver of sediment flux is wind driven wave action, and the nature of this
relationship will potentially change under future climate conditions. Under current conditions,
wave action has a net positive effect on sediment flux onto salt marshes, as greater wave energy
can mobilize and increase rates of sediment transport from bays and adjacent mudflats deep into
marsh systems. However, under future climate conditions a threshold may be crossed whereby
the role of waves as a sediment source will decrease and the erosive effect of waves will
increase, leading to a shift from a direct to an inverse relationship. This threshold is caused by a
potential change in wave character as water depth increases due to sea level rise. In deeper water
waves behave differently, with less wave energy available for resuspension of bottom sediments
and more energy delivered to the marsh edge, leading to increased erosion.
While managing wind driven waves may not immediately appear straightforward as a
management lever, given the importance of this potential threshold it is necessary to think
broadly about adaptation options for this pathway. It would be valuable to monitor wind, waves
and sediment fluxes in marshes in order to detect the threshold shift when it occurs. Ideally a
response plan would be prepared in advance of such a shift, with the necessary public and
management backing and resources in place to implement the plan when needed. Current tools
for reducing wave action on the front sides of marshes are limited; such tools need to be further
developed and tested in areas where waves are currently having an inverse influence on sediment
flux. Existing tools include building berms or restoring oyster reefs as protective barriers against
wave energy. Depending on the depth of adjacent water, such barriers could either be designed
to reduce wave energy on mudflats or on the marshes themselves. Another adaptation option
might be to strategically locate sites to deposit dredge materials with a goal of enhancing
sediment concentrations on mudflats adjacent to marshes. This could serve the dual purpose of
both increasing sediment fluxes and compensating for changes in water depth above mudflats.
3.2.1.2. Community Interactions Top Pathways
3.2.1.2.1. Green pathway
In this pathway (see Figure 3-3), starting with the Shorebirds endpoint and working "up"
the diagram, shorebirds are best able to effectively use mudflat foraging habitats only if
landscape mosaics—which are defined as a mixture of habitats for secondary foraging, roosting,
and cover from predators—are available in close proximity. Therefore, the presence of such
mosaics, including ponds, diked wetlands, seasonal wetlands, and muted tidal wetlands, has a
strong positive effect and a high relative impact on shorebird populations. The high relative
3-21
-------
impact of this influence is likely to increase even further under climate change as mudflat
habitats become scarcer and smaller in extent.
At the next level up the pathway, landscape mosaic is directly affected by restoration.
Workshop participants noted that small changes in restoration can have the potential for large
positive effects on landscape mosaic. The relative impact of restoration on landscape mosaic is
likely to increase under climate change as temperatures increase, precipitation patterns change,
and water diversions continue to escalate. These effects also will be exacerbated by ongoing and
increasing land use changes (especially development), further raising the importance of
restoration as a mitigating force.
Management options for adaptation given the relationships in this pathway center on
restoration as a key and increasingly critical management activity. Workshop participants noted
that there is little monitoring of landscape mosaics currently underway, so there is almost no
information on the current status of, or rates of change in, mosaic habitats. Assessment and
mapping is needed to detect changes and manage at the landscape scale for shorebirds and other
mobile species that use multiple habitats.
Given the implications under the climate change scenarios, managers might place a
priority on "stepping up" management of landscape mosaics through spatial planning designed to
prioritize where and how to restore which habitats. The goal would be to create a continuum of
wetland and upland ecosystems, across a range of salinity regimes, which could migrate inland
as sea level rises. Workshop participants noted that managers might also include "threshold
landscapes" (those that are about to change from one set of dominant processes to another, or
from one state to another) in consideration of good landscapes for restoration. As part of spatial
planning, restoration should be focused on where there are good opportunities for restoration,
where there can be flexibility in management, and where migration inland is possible. Decision
makers could even be urged to consider legislation or incentives that encourage moving back or
blocking of development on lands where there is restoration potential now or in the future.
3.2.1.2.2. Blue pathway
Starting with the Shorebirds endpoint and working up the Blue pathway (see Figure 3-3),
another key factor affecting shorebird populations is the extent of mudflat available for foraging
(i.e., the acre-hours that mudflats are exposed and therefore accessible). Extent of mudlflat has a
direct effect of high relative impact on shorebird populations. This direct effect will become
increasingly strong, and is considered a threshold effect, under climate change as extent of
mudflat may become limiting as sea level rises, with available foraging habitat becoming too
limited to support shorebird populations.
3-22
-------
Continuing to work up the pathway, a major determinant of extent of mudflat is mudflat
bathymetry, which has a strong direct effect on the extent of mudflat exposed at low tide.
Mudflat bathymetry is itself directly affected by sediment resuspension and deposition processes,
which are turn directly affected by sediment supply. A key source of sediment supply is the
direct effect of freshwater inflows.
An important management lever appears in the form of the sensitivity of freshwater
inflow to water management. Workshop participants noted that water management practices
(specifically, reservoir management and upstream operations) have an important direct effect on
freshwater inflows. This effect will become increasingly strong under climate change. This is
because the sensitivity of freshwater inflow to water management may increase in the future as
freshwater flows from alternate sources such as precipitation and tributaries become more
variable and/or scarce, and as the entire watershed becomes even more highly managed.
Implications for adaptation based on this pathway currently center around management of
water releases and sediment supply. Reservoir management of water releases in order to
mobilize and transport sediments (e.g., through the use of sediment maintenance flushing flows)
could become an increasingly important option; in light of this, continuing to improve upstream
operations in order to ensure greater availability of water (to allow more frequent and/or more
intense pulse releases) is warranted. Participants noted that an increasingly important technique
will be integrated water resources management, with an emphasis on shifting from storage more
toward conservation uses.
Attention to management options for directly enhancing sediment supply is also needed.
This is especially the case since sediment supply will become increasingly sensitive to
restoration as well as water management, with participants noting in their discussions a potential
synergistic interaction between the two. Methods for moving sediment (especially coarse
sediments) into the bay need to be developed. Adaptation options might include strategically
locating dredge spoil sites to enhance sediment supplies to mudflats.
3.2.1.2.3. Purple pathway
Starting with the Shorebirds endpoint, the Purple pathway (see Figure 3-3) begins with a
link to the shorebird prey community. Shorebird communities rely on abundant mudflat prey
populations, which therefore have a positive effect—of high relative impact—on shorebird
populations. The abundance of prey per unit area will become increasingly important (of
increasing relative impact) under climate change as spatial extent of mudflats shrink with sea
level rise. This effect will be magnified if secondary feeding habitats in the landscape mosaic
(see Green pathway) continue to be lost due to development and other pressures.
o oo
3-23
-------
Working up the next step of the pathway, shorebird communities are strongly affected by
water quality. Specifically, sufficient levels of DO have a direct positive effect on prey
communities. Since this is a direct relationship, this means that as decreases in DO occur with
climate change due to increased temperatures and/or eutrophication, prey communities may flag.
A threshold may occur in the future if DO reaches low enough levels to cause prey populations
to crash.
With regard to management implications, the Purple pathway provides further
justification for continued prioritization of water quality management. Water quality is already a
major concern in the watershed; and for shorebirds, a water quality aspect of special concern
may be DO. Oxygen depletion usually results from high rates of microbial and/or algal
respiration that exceed the capacity of the water body to replenish oxygen through phytoplankton
photosynthesis and diffusion from the air. Excessive inputs of organic material and nutrients, for
example from poorly treated sewage discharges or from agricultural activities, can accelerate
respiration rates and trigger localized and regional oxygen depletion. Protection and
improvement of water quality will be critical through integrated water resources management,
which could include stormwater management and rainwater-harvesting (which benefits water
conservation as well).
Other improvements to water quality could occur through land use decisions (e.g., use of
permeable surfaces to reduce runoff) and restoration decisions (e.g., restoration of riparian zones
as natural filters). The importance of land use decisions was emphasized by the participants,
who discussed a synergistic effect of freshwater inflow on water quality that is dependent on
land use change. In summary, water quality is an existing concern for which there are already a
variety of management options available; and because the affects of water quality on mudflat
prey populations are known and expected to increase with climate change, this is further
confirmation of the importance of implementing such options, in conjunction with public
education and outreach to explain and justify support for their use.
3.2.1.3. Top Pathway Caveats
Above we have described three pathways that scored as especially promising for
successful management application in light of the information provided by the particular group
of experts at this workshop. Given the complexity of these systems and instances of uneven
agreement among participants, actions based on these pathways should be considered with care.
A different set of pathways could be chosen based on additional meaningful criteria that are
site-specific and specific to individual managers' expertise. Based on their own knowledge of
their sites and/or input from different experts, managers are encouraged to examine the potential
3-24
-------
for additional top pathways for their own particular systems by examining the crosswalk tables
and applying their unique knowledge.
While top pathways based on the expert knowledge from this workshop are useful, it is
also important to look at gaps in the crosswalk tables where some influences did not show
agreement in type, degree and/or sensitivity. Lack of agreement does not necessarily mean there
is no information available; often the experts did not agree based on competing evidence, or as a
result of limitations of the expert elicitation process. Where there are gaps in otherwise-strong
pathways for management, further research—in the form of literature searches, data mining, or
original research if needed—could be highly valuable.
A final consideration is that these diagrams were developed considering the current
condition of each system. At this workshop there was a variable (disease for the Community
Interactions group) that arose as critical to the system under future conditions, even though it had
not been included in the diagram under current conditions. Individual managers should
identify—and consider management options for—any other such variables that may be specific
to their system or site.
3.2.2. Adaptation Planning
There are multiple ways to go about climate change adaptation planning, including
integrating adaptation into existing plans, or developing a stand-alone adaptation plan. This
report focuses on the planning options for SFEP, which as an NEP has several key management
plans. The SFEP management plans discussed here are used to demonstrate the type of
adaptation planning that can be done to address these particular issues. Other organizations can
use their particular planning documents to apply the same approach.
SFEP's planning documents include a CCMP, which articulates long range goals and
objectives, a Strategic Plan for midterm objectives, and an annual Work Plan that lays out
short-term actions to implement the goals and objectives. Each of these plans addresses climate
change on some level, so it makes sense to use the results of this study to continue integrating
climate change into each of these planning scales. In this section we provide some links between
SFEP's plans and the top pathways and management options discussed above; this set of
examples is not comprehensive, but rather is meant to illustrate how the results of this study can
be used to inform adaptation planning.
One management strategy outlined in the 1993 CCMP that pertains to the Green
Sediment Retention pathway (see Figure 3-2) is Action AR-4.1: "Adopt water quality and flow
standards and operational requirements designed to halt and reverse the decline of indigenous
and desirable nonindigenous estuarine biota" (SFEP, 1993). The relevance of the Green
3-25
-------
Sediment Retention pathway to this action is that reservoir management activities and their
effects on freshwater inflow are considered key to sustaining net organic accumulation through
this pathway. While animal species have been the primary focus of this action, it also applies to
managing for salt tolerance of plant species with resulting effects on plant productivity and net
organic accumulation.
Another relevant action, from the revised 2007 CCMP, is Action DW-1.1: "Conduct
studies, research, modeling, and analysis of sediment processes and trends to more thoroughly
understand sediment transport in San Francisco Bay, particularly in light of sea level rise and
changing sediment inputs from the Delta and major tributaries" (SFEP, 2007). This mandate will
be important for informing management along both the Blue and Purple Sediment Retention
pathways (see Figure 3-2) as well as the Blue Community Interactions pathway (see Figure 3-3).
For the Blue Sediment Retention pathway, the priority research would be monitoring of the
potential threshold effect of waves on sediment flux. For the Purple Sediment Retention
pathway, information is needed to understand the relationship between impervious cover and
sediment. Such understanding could inform where and how to manage impervious cover. In
addition, further investigation is needed into reservoir management options that provide pulse
events which can increase the supply of large grain sediment. Very similar questions on how
water management methods affect sediment supply come up for the Blue Community
Interactions pathway.
A strategy from the revised 2007 CCMP that relates to the Green Community
Interactions pathway (see Figure 3-3) is Action WT-4.1: "Identify, convert, or restore
nonwetlands to wetlands or riparian" (SFEP, 2007). While mudflats may not have been the
wetlands originally intended for this action, considering mudflats within the landscape mosaic is
advisable while implementing strategies of where to restore nonwetlands to wetlands (see also
Table 2-12).
In the 2010-2012 Strategic Plan (SFEP, 2010), one of the four key objectives for
focusing the implementation of the CCMP is Objective 2: "Support Estuary resilience in the face
of climate change". Under Subobjective 2.3 is "Promote climate adaptation strategies and
policies that encourage protection and restoration of Estuary health and reduce damage to the
ecosystem". The workshop results of this study can be used to provide specific areas of focus for
that objective and for mainstreaming adaptation into applicable actions under the other
objectives.
Meanwhile, under Objective 1: "Promote integrated watershed stewardship," there is
Subobjective 1.2: "Assist development of regional goals projects and management plans (i.e.,
Habitat Goals, Subtidal Habitat Goals, Upland Habitat Goals, regional sediment plans)". The
abovementioned Goals projects enumerating Baylands Ecosystem Habitat Goals (Monroe et al.,
3-26
-------
1999), Upland Habitat Goals (Weiss et al., 2007; 2008; 2010), and Subtidal Habitat Goals
(BCDC, 2010) are highly relevant to the Green Community Interactions pathway (see
Figure 3-3). Developing projects that coordinate across goals to connect habitat types could
serve as a strategy for rebuilding landscape mosaics through restoration projects. The original
Baylands Goals report did not take climate change into consideration, but the two newer projects
consider climate change as a factor.
Many of the current projects in the 2010-2011 Annual Work Plan are examples of
management options that potentially could be informed by the results of this study. For instance,
sediment reduction is an objective or component of 14 of the projects. For both the Blue
Community Interactions pathway (see Figure 3-3) and the Purple Sediment Retention pathway
(see Figure 3-2), one management option cited in Table 2-12 is to adjust policies that keep coarse
sediment from reaching the Bay. Many of the current projects are based on meeting TMDL
restrictions for salmonid streams, which is a competing goal with developing methods to increase
sediment supplies into the Bay. Some current projects focus on reducing fine sediment;
management practices that target fine sediment only (while allowing coarse sediment) could be
developed or expanded to other projects where appropriate. Tools such as those in the
"Watershed Scale Mapping of Project Results: Linking On-the-Ground Results to Measurable
Regional Outcomes" designed to assist in stopping downstream sediment migration could be
adapted to prioritize salmonid streams while allowing increased sediment transport in other
streams. Such a tool would distinguish where recommended practices should target fine
sediment while maintaining a supply of coarse sediment, versus those streams where it is
important to reduce all sediment sizes. Another current project, "Innovative Wetland Adaptation
Techniques in Lower Corte Madera Creek Watershed", will provide results that can be used for
one of the recommended actions in Table 2-12: "Develop methods to reduce wind/wave action
on the front side of marshes". The project includes measurements for wind-wave propagation
and attenuation in the marsh as well as developing best practices for flood control.
Within each plan are a variety of additional opportunities for incorporating the workshop
results. The examples offered here are intended to demonstrate the links, but are not
comprehensive. In addition to the adaptation of current management projects and strategies, this
study has identified sensitivities that may require the development of entirely new management
options. Planning for future projects should identify opportunities to fills those needs and test
new methods. Conflicting goals due to trade-offs—such as the simultaneous need to reduce
sediments in salmonid habitats, but increase coarse sediment transport to the Bay—will become
increasingly problematic with climate change. Indeed, in some cases such trade-offs may
necessitate reevaluation of habitat goals, and even the application of a "triage" approach to
prioritize certain habitats over others in the system.
3-27
-------
4. CONCLUSIONS
This report has described the results of a vulnerability assessment aimed at synthesizing
place-based information on the potential implications of climate change for key ecosystem
processes, with the intent of enabling managers to undertake adaptation planning. The
assessment involved identification of key management goals and ecosystem processes,
conceptual modeling of those processes, a climate change sensitivity analysis in a workshop
setting, and discussions/analysis of the potential applicability of the results for adaptation. The
workshop exercise—an expert elicitation sensitivity analysis combined with management
discussions—tested a novel approach for conducting "rapid vulnerability assessments" for
ecological systems. The sections that follow discuss general observations, insights, and
conclusions that emerged from the workshop exercise, from the analyses of management
implications, and from our assessment of the methodology's utility for potential use in other
locations/ecosystems.
4.1. INSIGHTS FROM THE WORKSHOP EXERCISE
4.1.1. Group Influence Diagrams
Figure 2-3 and 2-8 were developed by the workshop participants based on edits to straw
man diagrams prior to the workshop, followed by group discussions and refinement of a final
group diagram during the workshop. While the main purpose of the group influence diagrams
was to establish a framework for the subsequent sensitivity analysis, these diagrams represent
key outputs in and of themselves. The construction of the diagrams proved an interesting group
exercise in building a highly constrained representation of a complex system, with only the most
critical elements and interrelationships included. The iterative process of distillation into basic
diagrams by the two interdisciplinary teams of experts resulted in some interesting differences in
the Sediment Retention and Community Interactions diagrams.
The Sediment Retention group focused on the physical components of sediment
processes as the highest priority factors influencing the balance of salt marsh accretion and
erosion in their diagram, with less focus on biological factors. There appeared to be good
familiarity with each piece of the diagram across all members of the group; this allowed them to
be specific in defining (and hence envisioning the effects of) management-related variables
(levers), which may have contributed to the high amount of agreement in judgments during the
subsequent coding exercise. The participants reported that given more time they would have
distinguished between delta and local tributary sediment and freshwater inputs, but were forced
to lump these due to the 15-variable constraint and time limitations. Nevertheless, they were
able to agree on an acceptable influence diagram for the exercise, with a tractable number of
4-1
-------
unidirectional influence arrows and a few large feedback loops to handle important bidirectional
influences.
The Community Interactions group was also successful in agreeing on an acceptable
influence diagram for the exercise. However, their diagram was more complex, with a greater
mixture of both physical sediment processes (which maintain mudflats) and biological processes
(which affect shorebirds and their prey). In fact, a set of variables similar to the Sediment
Retention group's diagram was nested within the Community Interactions diagram; but then the
Community Interactions diagram was expanded further to also include a set of variables
representing the three biological communities (shorebirds, prey, and predators). Due to this
greater complexity, the Community Interactions group was forced to use two more boxes, a
larger number of influence arrows, and four bidirectional arrows, resulting in nine more
influence arrows than the Sediment Retention group. The complexity also led to less time for
defining the management levers specifically, making it more difficult to judge their effects.
These factors—combined with varying expertise in the specialty areas of sediment processes
versus ecological processes in this interdisciplinary group—may have contributed to the greater
numbers of gaps in agreement in the subsequent exercise.
While the two groups had different experiences and challenges in building their influence
diagrams, both groups were effective in generating a useful representation of their ecosystem
process for the sensitivity exercise. Participants reported that the highly constrained
diagram-building procedure challenged them to focus on the most key elements of the system
while still maintaining a sufficiently realistic model for sensitivity analysis. Designing the
diagrams while considering current conditions, then applying climate scenarios to the same
diagrams during the sensitivity exercise, worked smoothly. The one exception was a disease
variable in the Community Interactions diagram that was not in the original diagram, but which
emerged as critical to the Community Interactions process under the climate scenarios. This and
other complications could be avoided in future workshops by providing the participants with an
opportunity for one more revision of the diagrams after being briefed on the climate change
scenarios. This would allow them to account for how future climate might raise additional
variables for priority consideration in the diagrams.
4.1.2. Characterization of Influences
One technique for ensuring the effectiveness of expert elicitation is to break down the
problem (i.e., what are the climate change sensitivities of the selected ecosystem processes) into
a set of distinct questions that clearly and explicitly define parameters and relationships of
interest (see EPA's white paper at http://www.epa.gov/spc/expertelicitation/index.htm). This
was accomplished by way of a systematized coding exercise—using the influence diagrams as a
4-2
-------
framework—in which the experts made a series of judgments about individual components of
the system, in order to better understand the system as a whole. For each individual influence
arrow in the diagram, each expert was asked to characterize the effect of variable "X" on the
response variable "Y", including their confidence in that judgment. In future applications of this
method, the complexity of the coding scheme could be greatly reduced by condensing the
original 13 codes (see Table 2-2) down to the six typologies described in Section 2.2.2.5. This
would reduce confusion and increase efficiency of the exercise. Nevertheless, using the pilot
coding scheme, the experts were able to provide characterizations of all relationships in each
process, and based on these results, some general observations of interest have emerged.
Participant notes and discussions revealed that for both processes, while there are many
intermediate (and some high) sensitivity relationships among variables that are useful to be
aware of for management, it was difficult to detect changes in sensitivities across the scenarios
based on this method. Under the climate scenarios, one influence for the Sediment Retention
group became highly sensitive and four others showed a trend (but no majority agreement)
toward greater sensitivity, but most of the sensitivities remained intermediate. For the
Community Interactions group, under the climate change scenarios there were seven influences
that trended toward increasing sensitivity, but without majority agreement, while the majority of
influences remained intermediate in sensitivity, or lost agreement. It was noted that natural
variation in most of the variables is large enough that changes generated by the climate scenarios
would not be enough to move the variables outside their current range of variation. In other
words, a response in variable "Y" would need to be outside the normal range of variation in
order to clearly detect a sensitivity threshold change by way of the coding scheme that was used.
Only one threshold (see Relationship DD, Table 2-10) was indicated through coding in the
Community Interactions group.
Yet outside of the coding exercise, there were indications based on participant notes and
discussions that additional potential threshold relationships do exist. Identifying thresholds is
challenging because while there may be general recognition of the potential for certain threshold
effects, it can be very difficult to identify where and when a threshold may occur. Multiple
potential thresholds were identified in both processes, through one of two ways. In some cases,
participants tried to indicate thresholds with their sensitivity codes, but did so by including
two codes under each of the scenarios to signal uncertainty as to when the threshold might occur.
Others did not indicate the threshold with their codes at all because they were not sure whether
the climate scenarios represented a big enough change to cause threshold exceedance. In these
cases, the thresholds indicated in Table 2-6 and were ultimately identified through the
participants' notes and discussions as relationships that could change dramatically at some point
that is currently difficult to define.
4-3
-------
Another way of identifying relationships of particular interest for management is to
examine the relative impact of certain influences in the context of the whole process. For both
processes, under current conditions the influences identified as having primary impact tended to
include variables closer to the endpoints (see Figure 2-6 and 2-11) compared to relationships that
emerged as high or increasingly-high impact under future climate (see Figure 2-7 and 2-12).
This implies that variables related to management levers may become increasingly important as
climate changes.
Finally, characterization of interactions and confidence were also included in the
sensitivity exercise, with mixed results. Trying to consider interactive effects of multiple
variables moves the exercise to a much greater level of complexity. The number of possible
pairwise interactions in the influence diagrams was very large, and the challenge of
understanding combinations of effects could become very complicated. Thus the participants
were not asked to attempt every possible pairwise combination, but rather were asked to indicate
which interactions "leapt out" as well understood and important. Of course, even looking at all
pairwise interactions would be a vast oversimplification because variables interact in greater
multiples than just pairs. Nevertheless, while there were only a few pairwise interactions
identified by enough participants to stand out, at least these relationships are sufficiently well
understood to merit consideration in management planning. With regard to confidence, the
exercise made a good start of acknowledging the need to gauge confidence in the judgments and
providing a systematic way for doing so; however, the large number data gaps indicate that there
were difficulties with this part of the methodology. Potential reasons for these difficulties, as
well as potential improvements, have been discussed in Section 3.1.1.3. Both interactions and
confidence are concepts that need further refinement and better estimation methods before they
can be effectively interpreted for management planning.
4.2. APPLICATION OF WORKSHOP RESULTS
4.2.1. Top Pathways for Management
When using the workshop results, it is essential to examine all three types of
information—influence type, sensitivity, and relative impact—when thinking about management
applications. For some questions, one type of information may be useful individually, but
because there are gaps and limitations within each type of information, a more complete
management picture can be built using all three types together. It is helpful to focus on
influences that are well understood, become more sensitive, and have a greater impact under
future climate scenarios. In some cases, it is possible to connect a series of influences that meet
these criteria to identify a path between the endpoint and a management lever. We have
presented what we consider to be three top pathways for management (see Figures 3-2 and 3-3)
4-4
-------
for each process based on the information currently available from the workshop results. These
delineate relatively well-understood relationships that are climate sensitive and for which there
are consequent implications for management adaptation.
The climate-related changes of interest in the top pathways are of three main types:
(1) changes in relative impact under climate change; (2) changes in sensitivity under climate
change; and (3) threshold shifts under climate change. In the case of the influences for which
relative impact is likely to increase under one or both future climate scenarios, and especially
where relative impact is already high under current conditions, action could be taken
immediately. These are influences for which there is sufficient understanding and opportunity to
connect to management options that favor desirable outcomes, with increasing relative impact on
the process as a whole as climate change continues. In the case of influences for which an
increase in sensitivity is expected under climate change, there is still time to further study and
anticipate the degree and timing of the sensitivity and to prepare best management responses.
An expectation of increasing sensitivity could be considered a notification to monitor and plan
for when and how management practices can be adjusted to account for the impending change.
Finally, in the case of thresholds, there is often a strong expectation that a threshold shift is
likely, but usually a great deal of uncertainty as to exactly when the threshold will be crossed.
Monitoring of threshold variables is needed so that managers will be alerted immediately to the
shift when it occurs. In the meantime, actions can be taken to attempt to prevent the shift by
keeping the system "below" the threshold as long as possible, while preparing a plan for what to
do if an unavoidable shift occurs. After a shift occurs, managers should have a plan as to how
they will manage the system differently in its new state, or whether they will take no action and
instead shift their priorities to other goals.
It is important to note at this point that each pathway also sits in the context of other
influences with which there could be key interactions, so there may be opportunities for
management options beyond those most directly evident from the main pathways. Also, in the
case of other management pathways for which there are currently information gaps based on the
workshop results, it is vital to remember that lack of agreement does not mean zero
understanding of influences or zero degree of sensitivity. Closer inspection can show that the
agreement may be split between intermediate and high sensitivity, so the understanding that the
sensitivity of the influence is important may be obscured by the distinctions between categories.
It is of note that for influences for which there was agreement, the variation among participants
was greater than that between scenarios. This could be due to a number of reasons: a limited
range between the two midcentury climate scenarios; the number of assumptions each participant
was required to make individually for each judgment; and the interdisciplinary and complex
nature of the questions. This is an indication that these types of questions do not lend themselves
4-5
-------
to consulting a single expert, but rather require the combined judgments of a group of experts to
complete the full picture. This also highlights the need for caution against relying solely on
combined (agreement) information: the nature of the variation across participants is also
important to consider. Equally important is the application of local expertise when further
examining the results of this study; the local manager is the best expert on his or her unique
system and should thus apply an appropriate filter when making final interpretations and
decisions based on these results.
4.2.2. Mainstreaming Adaptation into Planning
The vulnerability assessment results for the two ecosystem processes presented here are a
big first step in the climate change adaptation planning process. We have given examples of
ways to tie the vulnerability assessment results to potential management options as a starting
point, but incorporating adaptation fully into management planning will require a more
systematic and comprehensive process. Planning is an iterative process, especially for climate
change adaptation, which is still a nascent field. Due to this iterative nature, the planning
recommendations presented here are based on mainstreaming (continuously integrating)
planning into existing planning mechanisms and documents, rather than developing a
comprehensive, stand-alone adaptation plan. For SFEP, nearer-term planning includes a
multiyear Strategic Plan and an Annual Work Plan, both of which provide ways to insert specific
management options into projects that are currently underway. In future plans, new projects that
specifically incorporate climate adaptation priorities can be added. Repeating vulnerability
assessments—once management options have been tested through project
implementation—should be part of the iterative process. This is consistent with adaptive
management approaches that emphasize "learning by doing", by way of concrete steps to test a
range of management choices, monitor and evaluate outcomes, incorporate learning into future
decisions and regularly revisit and revise goals (Boesch, 2006). Finally, this study only covered
two ecosystem processes and did not attempt to evaluate relative vulnerability or resilience
across different ecosystem processes. The vulnerabilities of additional ecosystems, processes
and goals will need to be assessed, taking into account what was most useful in the results of this
study for adaptation.
Thresholds remain a major unknown, and while much can be done to improve our
understanding of factors affecting thresholds, some may only be revealed after they have been
crossed. Thus it would be advisable for monitoring plans to be put into place to track indicators
of state changes. Contingency plans for management actions once a system has changed states
could be developed, as well as contingency planning for ways to respond to catastrophic events
such as levee failures or earthquakes. Successful implementation of contingency responses will
4-6
-------
require that the political and scientific base be put into place now for responding properly
following catastrophes or threshold changes. In the meantime, when prioritizing implementation
of adaptation actions, it is advisable to start with win-win options that contribute to current
management goals and efforts while also responding to current and future climate change. For
example, working now to proactively move highways and railroads that are barriers to marsh
migration where there is otherwise space for migration is not only advantageous for marshes, but
will also prevent damage and disruption to human transportation and infrastructure as inundation
from sea level rise continues. Likewise, the practice of working with authorities in flood control
districts to recouple streams to wetlands will not only benefit wetlands but will also improve
natural flood control services.
Looking beyond the win-win options, many other actions will force managers to confront
trade-offs that will require difficult policy decisions. One example highlighted in Sections 2.4.2
and 3.2.2.1 is the trade-off between increasing coarse sediment supply from tributaries, which
comes into conflict with current sediment reduction efforts for salmonid streams. While a first
step is to set up different best practices for salmonid and nonsalmonid streams, beyond that there
may come a decision point when it is no longer possible to meet both goals, so a choice between
the two conflicting goals will be necessary. As climate change progresses, there are likely to be
more trade-offs, often between short and long term goals. Mainstreaming adaptation planning
will provide a better chance of foreseeing conflicts between long and short term goals and
identifying opportunities to build support for hard decisions and creative solutions.
4.3. GENERAL CONCLUSIONS
4.3.1. Transferability of Results and Method
The results of this study were developed for two specific ecosystem processes within
two ecosystems. Therefore the question arises as to how transferable the results may be. The
sensitivities examined in this study are specific to sediment retention in salt marshes and
community interactions of shorebirds in mudflats, so the characterizations of influence type,
sensitivity and relative impact cannot be transferred directly to other ecosystems and do not
apply to different processes within these ecosystems. However, an example site was used as a
way to focus the exercise and was chosen as a representative example of intact ecosystems, thus
the results could be transferable to other North Bay locations in which the same ecosystem
processes are present. The variables that ended up in the group influence diagrams are general
enough that most of the results may transfer to the entire Bay, with only a few specific enough to
only apply to the North Bay. In addition, it is likely that the influence diagrams could also be
transferred for use with like ecosystems in other estuaries, with minor revisions for place-specific
4-7
-------
stressors or other process variables. The characterizations of influence type, sensitivity and
relative impact would have to be revisited, particular to that location.
Where the specific results are not transferable, the methodological process is certainly
transferable to other processes, ecosystems and locations. The methodology used for this
assessment—an analysis of key ecosystem processes through expert elicitation—is a useful
framework for understanding the current state of knowledge and research. The experts in this
study were able to share their combined understanding of key processes and how they are
expected to respond to climate change. The expert elicitation process also helped to identify
where key gaps in understanding exist, what type of research is necessary, and how management
should proceed. This methodology is transferrable in that the process used to compile, distill,
and assess key information can be replicated. Expert elicitation is used in many fields of study
and has been demonstrated here to be useful in understanding localized climate change impacts.
Experts can think integratively across studies and disciplines and often have access to more
current research and data than is currently available or published. As the climate change
research is constantly evolving, this type of process is useful for synthesizing the most current
information available. However, as climate change research is constantly changing, new
information and research will need to be integrated concurrent with management decisions.
4.3.2. Utility of Method for Rapid Vulnerability Assessments
Given that the method is transferable, the question of utility arises: in what cases is this
method advantageous? This method could be repeated elsewhere as a "rapid" vulnerability
assessment, with opportunity for some of the improvements that have been suggested for some
of the limitations. By rapid, we mean assessments that can be carried out within six months to a
year, as opposed to assessments based on detailed quantitative modeling that can take many
years. Another advantage is that this method is able to capture more recent knowledge than
would be available from a literature review. It is also better able to capture more knowledge of
the type that is closely related to management, which is less frequently published than scientific
studies. Finally, the information is more integrated across disciplines and scales and is designed
to better match the scale of adaptation decisions. In some cases new insights about management
effectiveness may arise while in other cases existing understanding may be validated. Having a
well supported study to substantiate new and existing ideas can position managers to justify the
most appropriate management options and priorities. It also can validate research priorities by
highlighting known research gaps.
The disadvantages are that this method is designed to focus only on a specific piece of the
system, compared to initial assessments that often rely on surveying the system more
comprehensively (though less deeply), often through literature reviews. The amount of caution
4-8
-------
required to properly interpret the results is another disadvantage, given multiple limitations and
caveats. The method is not intended as a consensus exercise, and the large number of influences
without agreement present challenges to either fill those research gaps to improve agreement or
to manage around limited information. In addition, this is only one group of experts, and another
group could reach different conclusions. Group selection is critical to making sure appropriate
areas of expertise and conflicting views on the system are represented. This is another reason
why in addition to looking for areas of agreement, the results of individual judgments should also
be examined. At the same time, since no participant can have complete expertise in every facet
of a system, it is also important that participants have the opportunity to confer amongst
themselves and adjust their judgments based on what they learn from each other.
Nevertheless, the expert elicitation method developed for this study was well suited for
achieving the purpose and goals of the assessment. In addition to achieving the workshop goals,
several unexpected benefits emerged from the workshop. Participants reported that the
combination of the development of the influence diagrams with systematic judgments facilitated
thinking about the system and questions of climate change vulnerability in a different way than
they had previously. Several expressed an intention to explore adapting the method for use in
other workshop or classroom settings. Many participants found that the multidisciplinary
interactions with colleagues were a valuable, personal learning experience and that the group
together generated new insights about the system and links to management that may not have
been seen by individuals. In short, the method demonstrated in this project offers opportunities
to capture and integrate the existing collective knowledge of local experts, while pushing the
boundaries to develop a new understanding of the system and management options in the face of
insufficient data and deep uncertainty about future climate.
4-9
-------
5. REFERENCES
BCDC (San Francisco Bay Conservation and Development Commission). (2010) San Francisco Bay Subtidal
Habitats Goals Project, Conservation Planning for the Submerged Areas of the Bay. San Francisco Bay
Conservation and Development Commission, the California State Coastal Conservancy, National Oceanic and
Atmospheric Administration, and the San Francisco Estuary Partnership. Available online at
http://www.sfbavsubtidal.org/PDFS/Full%20Report.pdf.
Boesch, DF. (2006) Scientific requirements for ecosystem-based management in the restoration of Chesapeake Bay
and Coastal Louisiana. Ecol Eng 26:6-26.
Cayan, D; Tyree, M; Dettinger, M; et al. (2009) Climate change scenarios and sea level rise estimates for the
California 2008 Climate Change Scenarios Assessment. California Climate Change Center. CEC-500-2009-014-D.
http://www.energy.ca.gov/2009publications/CEC-500-2009-014/CEC-500-2009-014-D.PDF
CCSP (Climate Change Science Program). (2008) Preliminary review of adaptation options for climate-sensitive
ecosystems and resources. Synthesis and Assessment Product 4.4. A report by the U.S. Climate Change Science
Program and the Subcommittee on Global Change Research. Julius, SH; J.M. West, JM; eds. Baron, JS; Griffith, B;
Joyce, LA; et al (Authors). U.S. Environmental Protection Agency, Washington, DC, USA, 873 pp.
http://downloads.climatescience.gov/sap/sap4-4/sap4-4-fmal-report-all.pdf
CEC (California Energy Commission). (2006) Our changing climate: assessing the risks to California. A Summary
Report from the California Climate Change Center. http://meteora.ucsd.edu/cap/pdffiles/CA climate Scenarios.pdf
CEC (California Energy Commission). (2008) The future is now: An update on climate change science, impacts, and
response options for California. A report from the California Climate Change Center California Energy
Commission's Public Interest Energy Research Program. http://www.energy.ca.gov/2008publications/CEC-500-
2008-077/CEC-500-2008-077.PDF
GalbraithH; Jones, R; Park, R; et al. (2005) Global climate change and sea level rise: potential losses of intertidal
habitat for shorebirds. USDA Forest Service Gen. Tech. Rep. PSW-GTR-191. Available online at
http://research.fit.edu/sealevelriselibrarv/documents/doc mgr/470/Global Potential Losses of Shorebird Habitats
- Galbraith et al. 2005 .pdf
Gershunov, A; Douville, H. (2008) Extensive summer hot and cold extremes under current and possible future
climatic conditions: Europe and North America. In: Diaz, H; Murnane, R; eds. Climate Extremes and Society.
Cambridge, UK: Cambridge University Press; pp 74-98.
Groffman, PM; Baron, JS; Blett, T; et al. (2006) Ecological thresholds: The key to successful environmental
management or an important concept with no practical application? Ecosystems 9 (1): 1-13, DOI: 10.1007/sl0021-
003-0142-z
Hay hoe, K; Cayan, D; Field, CB; et al. (2004) Emissions pathways, climate change, and impacts on California. Proc
Nat Acad Sci 101(34):12422-12427.
Henrion, M; Breese, JS; Horvitz, EJ. (1991) Decision analysis and expert systems. Al Magazine 12(4): 64-91.
Available online at http://www.aaai.org/ojs/index.php/aimagazine/article/view/919/837
Jaffe, BE; Smith, RE; Torresan, LZ. (1998) Sedimentation and bathymetric change in San Pablo Bay: 1856-1983.
OPEN-FILE REPORT 98-759. Version 1.0. Available online at http://geopubs.wr.usgs.gov/open-file/of98-
759/of98-759.pdf.
5-1
-------
Kneib, R; Simenstad, C; Nobriga, M; et al. (2008) Tidal marsh conceptual model. Sacramento (CA): Delta Regional
Ecosystem Restoration Implementation Plan. Available online at
nrm.dfg. ca.gov/FileHandler.as hx?DocumentID=10093
Knowles, N; Cayan, D. (2004) Elevational dependence of projected hydrologic changes in the San Francisco estuary
and watershed. Clim Change 62:319-336. Available online at http://tenaya.ucsd.edu/~dettinge/noah_cc.pdf
Miller, NL; Hayhoe, K; Jin, J; et al. (2008) Climate, extreme heat, and electricity demand in California. J Appl
Meteorol Climatol 47:1834-1844. Available online at
http://journals.ametsoc.org/doi/pdf/10.1175/2007JAMC1480.1
Monroe, M; Olofson, PR; Collins, JN; et al. (1999) Baylands ecosystem habitat goals. A report of habitat
recommendations prepared by the San Francisco Bay Area Wetlands Ecosystem Goals Project. Oakland, CA: San
Francisco Bay Regional Water Quality Control Board. Available online at http://www.sfei.Org/node/2123.
Morgan, MG; Henrion, M. (1990) Uncertainty: A guide to dealing with uncertainty in quantitative risk and policy
analysis. New York: Cambridge University Press.
Orr, M; Crooks, S; Williams, PB. (2003) Will restored tidal marshes be sustainable? San Francisco Estuary
Watershed Sci 1(1): Article 1. Available online at http://www.escholarship.org/uc/item/8hj3d20t.
Page, GW; Whitacre, DF. (1975) Raptor predation on wintering shorebirds. Condor 77:73-78. Available online at
http://www.jstor.org/stable/1366760?seq=3
Pradhan, M; Henrion, M; Provan, G; et al. (1996) The sensitivity of belief networks to imprecise probabilities: an
experimental investigation. Artificial Intell 85(l-2):363-397.
Reed, DJ. (1995) The response of coastal marshes to sea-level rise: survival or submergence? Earth Surf Proc
Landforms 20(l):39-48.
Ruhl, CA; Schoellhamer, DH. (2004) Spatial and temporal variability of suspended-sediment concentrations in a
shallow estuarine environment. San Francisco Estuary Watershed Sci 2(2): Article 1. Available online at
http ://repositories. cdlib. org/j mie/sfews/vol2/iss2/art 1.
SFEIT (San Francisco Estuary Indicators Team). (2008) Assessment framework as a tool for integrating and
communicating watershed health indicators for the San Francisco estuary. Technical Report #1. Submitted by the
San Francisco Estuary Indicators Team to the California Department of Water Resources, September 30, 2008.
Scavia, D; Field, JC; Boesch, DF; et al. (2002) Climate change impacts on U.S. coastal and marine ecosystems.
Estuaries 25(2): 14-164.
Schoellhamer, D; Wright, S; Drexler; J; et al. (2007) Sedimentation conceptual model. Sacramento, (CA): Delta
Regional Ecosystem Restoration Implementation Plan. Available online at
http ://science .calwater.ca. gov/pdf/drerip/DRERIP_sediment_conceptual_model_fmal_ 111307 .pdf
SFEP (San Francisco Estuary Partnership). (1993) Comprehensive conservation and management plan.
SFEP (San Francisco Estuary Partnership). (2007) Comprehensive conservation and management plan. Available
online at http://www.cakex.org/sites/default/files/San%20Francisco%20Estuary%20Project.pdf
SFEP (San Francisco Estuary Partnership). (2010) Strategic plan for 2010-2012. Available online at
http://www.sfestuary.org/userfiles/file/SFEPstratplan.pdf.
Stenzel, LE; Hickey, CM; Kjelmyr, JE; et al. (2002) Abundance and distribution of shorebirds in the San Francisco
Bay area. Western Birds 33:69-98.
5-2
-------
Weiss, SB; Schaefer, N; Branciforte, R. (2007) Study design and methodology for review by the Peer Review Panel.
A project of the Bay Area Open Space Council. Available online at
http://www.uplandhabitatgoals.org/upload/files/20100517114132 Studv%20Design%20and%20Methodologv%20f
or%20Peer%20Review.%20Mav%202007.pdf.
Weiss, SB; Schaefer, N; Branciforte, R. (2008) Methodology and vegetation representation analysis. A project of the
Bay Area Open Space Council. Available online at
http://www.uplandhabitatgoals.org/upload/files/20100909121856 Methodologv%20and%20Veg%20Rep%20Analv
sis%20v2.pdf.
Weiss, SB; Schaefer, N; Branciforte, R. (2010) Riparian/fish focus team report. A project of the Bay Area Open
Space Council. Available online at
http://www.uplandhabitatgoals.org/upload/files/20100908211440_Draft%20RFFT%20Report%20090110doc.pdf.
Wright, SA; Schoellhamer, DH. (2004) Trends in the sediment yield of the Sacramento River, California, 1957-
2001. San Francisco Estuary Watershed Sci 2:2. Available online at
http ://repositories. cdlib. org/j mie/sfews/vol2/iss2/art2.
Ydenberg, RC; Butler, RW; Lank, DB; et al. (2004) Western sandpipers have altered migration tactics as peregrine
falcon populations have recovered. Proc R Soc LondB 271(1545): 1263-1269. Available online at
http://www.sfu.ca/biology/wildberg/ydenbergetal2004PRSB.pdf
5-3
-------
APPENDIX A. DEVELOPMENTAL PROCESS FOR
CLIMATE READY ESTUARIES VULNERABILITY ASSESSMENT
A.l. SELECT KEY GOALS, ECOSYSTEMS, AND ECOSYSTEM PROCESSES
The SFEP partners participated in several discussions and meetings to outline
management priorities, key resources to consult, and other considerations for selecting key goals
for the assessment. As a starting point, SFEP's Comprehensive Conservation and Management
Plan (CCMP) (SFEP, 1993; SFEP, 2007) was examined and discussed to select four to six key
management goals as a focus of the assessment. These goals would help to further refine the
analysis to specific ecosystems, ecosystem processes, stressors of concern, and indicators for
measuring changes in the ecosystem. Selected management goals included:
• Restore healthy estuarine habitat to the Bay-Delta, taking into consideration all
beneficial uses of Bay-Delta resources
• Stem and reverse the decline of estuarine plants, fish and wildlife and the habitats on
which they depend
• Ensure the survival and recovery of listed and candidate threatened and endangered
species, as well as special status species
• Protect and manage existing wetlands
• Restore and enhance the ecological productivity and habitat values of wetlands
Following an October 2008 kickoff meeting with over 30 local experts to share
information on climate change impacts and adaptation work in the region, salt marshes and
mudflats were selected as two wetland habitats of focus for the project. These systems were
identified by the local experts as highly relevant to SFEP's management goals due to their
ecological productivity and their habitat values for threatened and endangered species, and are
deemed highly sensitive to changes in climate-related variables such as sea level rise and altered
hydrology. In order to explore the linkages among such climate-related variables, their
interactions with nonclimate stressors of concern, and the key ecosystem processes that maintain
the systems, a conceptual model was developed for each ecosystem type.
A.2. CONCEPTUAL MODELS
The conceptual models were intended to serve as a framework for further analysis in the
vulnerability assessment. The models depicted likely pathways by which climate drivers may
A-l
-------
directly or indirectly affect interacting stressors that impact ecosystem processes. The process is
intended as iterative, as we learn from exploring the first two ecosystem processes, next steps
can involve focusing on additional ecosystem processes, or for repeating a similar analysis for
additional habitats. The development of the conceptual model has also served to help with the
scoping process; we began with a comprehensive list of ecosystem processes and indicators and
then chose those more important to and best representing healthy salt marsh functioning. The
total number of possible ecosystem processes was narrowed down to five or six key ones for the
ecosystem. The models also included a similar number of variables that may serve as indicators
for the status of these endpoints. Ecosystem processes and indicators were identified in
discussions among SFEP, BCDC, and EPA ORD, as well as through examination of the Delta
Regional Ecosystem Restoration Implementation Plan's conceptual models developed by the
CALFED Bay-Delta Program (Schoellhamer et al., 2007; Kneib et al., 2008). To ensure
consistency with current research, these indicators were cross-walked with indicators developed
for the Watershed Assessment Framework, which were being incorporated into a revision of
SFEP's CCMP (San Francisco Estuary Indicators Team, 2008).
Stressor interactions are stressors that work together to affect ecosystem functioning.
These included both nonclimate and climate-related influences that stress salt marsh and mudflat
ecosystems. Preexisting stressors and stressor interactions were identified during the
development of salt marsh and mudflat conceptual models, and impacts of these stressors of
concern were identified using the SFEP Comprehensive Conservation and Management Plan.
Climate drivers are climate variables that may impact ecosystem processes directly (e.g.,
raise water temperature) or indirectly (e.g., cause changes in nutrient inputs). The climate
drivers relevant to salt marshes and mudflats were identified by first examining climate drivers
for estuarine systems outlined in Synthesis and Assessment Product 4.4: Preliminary review of
adaptation options for climate-sensitive ecosystems and resources (CCSP, 2008), followed by
extensive discussions among the SFEP partners. The climate drivers were then mapped to the
key processes of each ecosystem, either directly or through interactions with pre-existing
stressors. These pathways provided the basis for the development of the conceptual models.
The pathways included are intended as a heuristic, without distinguishing between the
magnitudes between them. It is not possible to include all possible system components and
connections between them. General models are first presented, and then additional detail for
individual ecosystem processes is described in the two submodels.
A-2
-------
A.2.1. General Models
A.2.1.1. Salt Marshes
The general model for salt marshes in presented in Figure A-l. Climate drivers in the salt
marsh conceptual model include: changes in air temperature, changes in precipitation, sea level
rise, and changes in storm climatology and wind. Changes in air temperature refers to the
variation from the climatological mean surface air temperature in a particular region. Changes in
precipitation refers to variation from the climatological mean of the amount, intensity, frequency
and type of rainfall, snowfall and other forms of frozen or liquid water falling from clouds in a
particular region, changes refer to both the form and flow of precipitation. Sea level rise is
defined as "relative sea-level rise," the change in sea level relative to the elevation of the
adjacent land, which can also subside or rise due to natural and human induced factors. Relative
sea-level changes include both global sea-level rise and changes in the vertical elevation of the
land surface. Changes in storm climatology and wind refers to the variation from the
climatological mean of the frequency, intensity and duration of extreme events (such as
hurricanes, heavy precipitation events, drought, heat waves, etc.) and the changes in the direction
and timing of the dominant seasonal winds.
Stressor interactions within the salt marsh conceptual model include: changes in water
temperature, changes in salinity, sedimentation and erosion, flooding, invasive species, other
human uses, land use/land cover change, contaminants, and altered flows/water demand.
Changes in water temperature refers to variation in the climatological mean surface water
temperature in a particular region. Changes in salinity are measured by changes in the location
along the estuary of different salinity zones (e.g., polyhaline, mesohaline, and oligohaline), or
changes in vertical stratification based on salinity. Sedimentation and erosion includes the
transport, deposition, and removal of soil and rock by weathering, mass wasting, and the action
of streams, waves, winds and underground water. Flooding is defined as an excess of water that
does not recharge ground water beyond time frames typical for watersheds due to high
precipitation events, storm surge, or infrastructure damage. Invasive species are alien species
(species not native to a particular ecosystem) whose introduction causes, or is likely to cause,
economic or environmental harm or harm to human health. Other human uses is a catch all
category based on the CCMP which includes the use of the marsh and surrounding area for
activities such as fishing, shipping and ports, dredging, transportation projects, sand mining,
recreational use, marinas, and industrial uses that may impact the marsh. Land use/land cover
change is defined as the current use of marsh and human-induced changes to the marsh or
surrounding land, including wetland alteration and expansion of the built environment.
Contaminants include material that creates a hazard to the ecosystem by impairing water quality,
A-3
-------
Climate Drivers
Changes in Air
Temperature
Changes in
Precipitation
Sea Level Rise
Changes in Storm
Climatology and
Wind
Invasive
Species
>
Changes in Water
Temperature
nd Use/Land
CbiverChange
Community
Interactions
Primary
Productivity
Water
Purification
Water
Retention
Nutrient
Cycling
Sediment
Retention
Biodiversity
Species
Population
Size
Water Quality
Standards
Freshwater
Inflow
Extent of
Aquatic
Habitat
Sediment
Quantity
Figure A-l. Salt Marsh Conceptual Model.
-------
poisoning or through the spread of disease (e.g., mercury, selenium, polychlorinated biphenyls
(PCBs), dichlorodiphenyltrichloroethane (DDT), chlordane, dieldrin, dioxin, trash and debris,
and acid mine drainage). Altered flows/water demand includes upstream water diversions for
agricultural, industrial, or urban uses that change the natural flow of freshwater and sediment
into the marsh, including leveeing, diking, damming, filling, or channeling.
Ecosystem processes in the salt marsh conceptual model include: community
interactions, primary productivity, sediment retention, water retention, nutrient cycling, and
water purification. Community interactions is defined as the interrelations among species within
the ecosystem. Primary productivity is the production of energy by plants and phytoplankton
within the entire system. Sediment retention is the balance between the processes of removal
and deposition of suspended sediment. Water retention is defined as the capability to buffer
against flooding. Nutrient cycling is the process of transfer of nutrients between organisms and
the water. Water purification is defined as the removal of pollutants and harmful
microorganisms.
Indicators within the salt marsh conceptual model include: species population size, water
quality standards, freshwater inflow, sediment quantity, extent of aquatic habitat, and
biodiversity. Species population size is defined as the number of similar organisms residing in a
defined place at a certain time, including threatened and endangered species, native species, and
invasive species. Water quality standards are provisions of State or Federal law which consist of
designated uses for waters of the United States, and water quality criteria for such waters based
upon such uses. Criteria address the values for water quality indicators (e.g., water temperature,
salinity, water contaminant exposure, biological thresholds for water contamination, nutrient
concentrations, water toxicity) that are required to support designated uses. Freshwater inflow is
the amount of freshwater inflow to the estuary from the watershed. Sediment quantity is defined
as suspended sediment concentration. Extent of aquatic habitat is defined as the area of all
contiguous, vegetated salt and brackish wetland, or mean width of marsh (may be divided into
low or high marsh or by dominant species). Biodiversity is the presence and abundance of
different species types (e.g., fish, birds, submerged aquatic vegetation (SAV)).
The salt marsh conceptual model focuses on a limited number of ecosystem processes
that are key to the habitat and region. In some instances, a component of the system may fill
roles at multiple levels, and the model does not represent all possible roles a particular
component may fill. The model does not take the cumulative effects of climate stressors or
tipping points/critical thresholds into account. The model does not include ocean acidification as
a climate driver, as current understanding of salt marshes indicate it as secondary compared to
the other stressors.
A-5
-------
A.2.1.2. Mud Flats
The general model for mudflats is presented in Figure A-2. Climate drivers in the
mudflat conceptual model include: changes in air temperature, changes in precipitation, sea level
rise, and changes in storm climatology and wind. Changes in air temperature refers to the
variation from the climatological mean surface air temperature in a particular region. Changes in
precipitation refers to variation from the climatological mean of the amount, intensity, frequency
and type of rainfall, snowfall and other forms of frozen or liquid water falling from clouds in a
particular region, changes refer to both the form and flow of precipitation. Sea level rise is
defined as "relative sea-level rise," the change in sea level relative to the elevation of the
adjacent land, which can also subside or rise due to natural and human induced factors. Relative
sea-level changes include both global sea-level rise and changes in the vertical elevation of the
land surface. Changes in storm climatology and wind refers to the variation from the
climatological mean of the frequency, intensity and duration of extreme events (such as
hurricanes, heavy precipitation events, drought, heat waves, etc.) and the changes in the direction
and timing of the dominant seasonal winds.
Stressor interactions within the mudflat conceptual model include: changes in water
temperature, changes in salinity, sedimentation and erosion, flooding, invasive species, other
human uses, and contaminants. Changes in water temperature refers to variation in the
climatological mean surface water temperature in a particular region. Changes in salinity are
measured by changes in the location along the estuary of different salinity zones (e.g.,
polyhaline, mesohaline, and oligohaline), or changes in vertical stratification based on salinity.
Sedimentation and erosion includes the transport, deposition, and removal of soil and rock by
weathering, mass wasting, and the action of streams, waves, winds and underground water.
Flooding is defined as an excess of water that does not recharge ground water beyond time
frames typical for watersheds due to high precipitation events, storm surge, or infrastructure
damage. Invasive species are alien species (species not native to a particular ecosystem) whose
introduction causes, or is likely to cause, economic or environmental harm or harm to human
health. Other human uses includes the use of the marsh and surrounding area for activities such
as fishing, shipping and ports, dredging, transportation projects, sand mining, recreational use,
marinas, and industrial uses that may impact the marsh. Contaminants include material that
creates a hazard to the ecosystem by impairing water quality, poisoning or through the spread of
disease (e.g., mercury, selenium, PCBs, DDT, chlordane, dieldrin, dioxin, trash and debris, and
acid mine drainage).
A-6
-------
Climate Drivers
Changes in Air
Temperature
Changes in
Precipitation
Sea Level Rise
Changes in Storm
Climatology and
Wind
Flooding
>
Changes in Water
Temperature
Contaminants
Community
Interactions
Primary
Production
Biomass
Key Species
Nutrient
Cycling
Indicators
Biodiversity
Species Population
Size
Extent of Aquatic
Habitat
WaterQuality
Standards
Figure A-2. Mudflat Conceptual Model.
-------
Ecosystem processes in the mudflat conceptual model include: community interactions,
primary productivity, biomass, nutrient cycling, and key species. Community interactions is
defined as the interrelations among species within the ecosystem. Primary productivity is the
production of energy by plants and phytoplankton, within the entire system. Biomass is the total
mass of biological material within the system or within a particular category or group. Nutrient
cycling is the process of transfer of nutrients between organisms and the water. Key species are
species which serve as a foundation for other species or fill a similar pivotal role for the rest of
the ecosystem (e.g., ecosystem engineers). Indicators within the mudflat conceptual model
include: species population size, water quality standards, extent of aquatic habitat, and
biodiversity. Species population size is defined as the number of similar organisms residing in a
defined place at a certain time, including threatened and endangered species, native species, and
invasive species. Water quality standards are provisions of State or Federal law which consist of
designated uses for waters of the United States, and water quality criteria for such waters based
upon such uses. Criteria address the values for water quality indicators (e.g., water temperature,
salinity, water contaminant exposure, biological thresholds for water contamination, nutrient
concentrations, water toxicity) that are required to support designated uses. Extent of aquatic
habitat is defined as the area of all contiguous, vegetated salt and brackish wetland, or mean
width of marsh (may be divided into low or high marsh or by dominant species). Biodiversity is
the presence and abundance of different species types (e.g., fish, birds, SAV).
The mudflat conceptual model focuses on a limited number of ecosystem processes that
are key to the habitat and region. In some instances, a component of the system may fill roles at
multiple levels, and the model does not represent all possible roles a particular component may
fill. The model does not take the cumulative effects of climate stressors or tipping points/critical
thresholds into account. The model does not include ocean acidification as a climate driver, as
current understanding of salt marshes indicate it as secondary compared to the other stressors.
A.2.2. Submodels
Following the development of the general ecosystem models, one ecosystem process
within each model was chosen to move to the specifics for an individual ecosystem process. The
purpose was to select good processes for piloting the method, but the choice does not imply that
these are necessarily the only important, or the most vulnerable, processes. Sediment retention
was identified as a key salt marsh process because of the importance of sediment supply to allow
for marsh development and growth. In the Bay, sediment supply has been declining due to
changes in human activities and the use of the land and waterways (Jaffe et al., 1998; Wright and
Schoellhamer, 2004). SFEP, BCDC, and other regional partners have done extensive work on
A-8
-------
examining changes in sediment and how these changes may be influenced by changes in climate.
This provided the basis for the development of the sediment retention submodel.
Community Interactions was chosen as the second ecosystem process of focus. To select
a specific well-constrained "storyline" of interactions between two-four species for this process,
ICF and EPA consulted with SFEP, BCDC, and regional experts on key sensitivities for this
process within the Bay system. The shorebird and mudflats interaction was selected for further
study because of the priority status of key species and the climate sensitivities that structure the
system. As mudflats may be one of the habitat types most vulnerable to climate changes
(especially sea level rise) the interactions among wading shorebirds were identified as a key
process for study. This provided the basis for the development of the community interactions
submodel.
A.2.2.1. Sediment Retention
The sediment retention submodel is presented in Figure A-3. It focuses on the balance
between the processes of removal and deposition of sediment onto a salt marsh and the resultant
ability of the marsh to persist in the face of climate change. The accumulation of sediments and
marsh vertical accretion result from interactions among tidal imports, vegetation dynamics, and
depositional processes. Freshwater flow and coastal storms transport and deposit sediments onto
the marsh surface, and the roots and stems of marsh vegetation retain sediment that would
otherwise be carried away from the marsh by wind and waves. Over time, the accumulation of
dead and dying organic matter produces peat, and the combination of peat accumulation and
sediment deposition gradually builds up the marsh surface. Ultimately, it is the balance between
marsh vertical accretion and sea level rise that determines whether a tidal marsh will persist in
the face of rising seas or will convert to tidal flats or open water (Reed, 1995).
In the San Francisco Estuary, there is an annual cycle of sediment deposition and
resuspension that begins when freshwater flow from the Delta in winter carries pulses of
sediment to the bay. Most of this new sediment is deposited in shallow areas and where tidal
velocities are lower. In spring and summer suspended sediment concentrations increase again as
a result of wind-wave resuspension of bottom sediments (Ruhl and Schoellhamer, 2004).
Sediment supply will play an increasingly critical role in this process. The supply of sediments
is declining as the estuary completes the shift from a system with larger sediment loads as a
result of past hydraulic mining to one that has a reduced sediment supply due to the cessation of
mining and an increase in tributary dams that trap sediment upstream (Jaffe et al., 1998; Wright
and Schoellhamer, 2004).
A-9
-------
Climate Drivers
>
o
Indicators
I
Changes in Air
Temperature
Changes in
Precipitation
Changes in Storm
Climatology and Wind
Altered Flows/Water Demand
Upstream water diversions and
controls:
• Change in peak (max or min
flow/volume
• Change in flow variability
• Change in flow frequency and
duration
Sediment
Retention
Sea Level Rise
Sedimentation
OtherHumanllses
Invasive Species
Spartina alterniflom
Eelgrass, Mussels,
Mitten crabs:
o%coveror
redgingand dredge disposal:
requency, location and extent
of dredging
Boating/shipping:
Frequencyand degree of wake
disturbance
Land Use/Land CoverChange
Increase in impervious cover:
•Areaor%change in impervious cover
•Land conversion from agriculture to
urban
•Areaor%change in land use
classification
Freshwater
inflow
[ Change in X2 location (Km)
Streamflow (daily flow in cubic
•J ftper second)
Net Delta outflow (daily flow in
I cubicfeetpersecond)
SedimentQuantity
Marsh elevation (feet)
Suspended sediment
concentration (mg/L)
Sediment deposition (cubic
metersor metric tons annually)]
Species Population
Size
Native speciescoverfe.g.,
pickleweed,% cover or area)
Invasive species cover (e.g.,
Spartina alterniflom, % cover
or area [acres])
Extent of Aquatic
Habitat
Acreage (total acres or
J hectares using aerial data)
Change in extent
I
Figure A-3. Sediment Retention submodel.
-------
A number of key climate variables (air temperature, precipitation, storm climatology and
wind, and sea level rise) and interacting human stressors (e.g., altered flows, dredging/dredge
disposal, land use/land cover changes) may impact this process, either directly or indirectly.
Increases in winter storms and in strong wind-driven waves may increase erosion of uplands,
increasing sediment availability. The North Bay has become somewhat erosional because of an
altered balance between riverine sediments and sediment transport. Storms and storm surges
may also carry more sediment away from marshes or promote resuspension of bottom sediments,
leading to increased suspended sediment concentrations. Sediment deposition and retention on
the marsh surface will ultimately depend on marsh geomorphology and surface vegetation (Orr
et al., 2003).
A.2.2.2. Community Interactions
The community interactions submodel is presented in Figure A-4. This submodel
focuses on community interactions between two species of mudflat wading birds, the Marbled
Godwit and the Western Sandpiper, and their predators and prey. Inundation and sediment
regimes influence not only mudflat extent, but also mudflat trophic dynamics (Takekawa et al.,
2006a). The trophic structure of North Bay mudflats includes invertebrates within mudflat
sediments, shorebirds that feed on mudflat infauna (Stenzel et al., 2002), and Peregrine Falcons
and Merlins, which prey on shorebirds (Page and Whitacre, 1975; Ydenberg et al., 2004).
A number of key climate variables and interacting human stressors (altered flows,
dredging, land use/land cover changes) may impact these trophic interactions, directly or
indirectly. Depending on sediment supply, increased inundation from sea level rise may drown
mudflats, while an increase in winds and wave action from more frequent and intense storms
may change sediment deposition patterns. Because suspended sediment concentrations are
sensitive to the extent and elevation of mudflats, as mudflat elevations decrease, suspended
sediment concentrations may decrease over time (Orr et al., 2003). If sediment deposition does
not keep pace with sea level rise, mudflat invertebrates will become less available for shorebirds.
There is limited information on shorebird diets, making it difficult to group shorebirds by
prey type, but shorebirds can be distinguished based on the depth at which they probe into the
sediment for prey. Short-legged shorebirds that are shallow probers, represented by Western
Sandpiper, forage in the top layer of sediments (<3 cm) and will lose foraging habitat first. But
eventually mudflat invertebrates will also become inaccessible to long-legged deep probers,
represented by Marbled Godwit, which penetrate up to 8 cm into the substratum (Takekawa
et al., 2006b).
A-ll
-------
CM mate Drivers
Changes in Air Changes in
Temperature ^ Precipitation
Changes in Salinity
Location of mesohaline/oligohaline transition
Frequencyor%oftime mesohaline
Sea Level Rise
Changes in Storm
Climatology and Wind
Invasive Species
Sport/no alterniflora colonization
Sedimentation and Erosion
Floodin
%Time inundated
Watershed infrastructure
and management practi
OtherHumanUses
Contaminants
Dredgin
Boat Wake
Existing exposures
New sources
>
Marbled
sodwit
Community Interactions
Peregrine falcon, Merlin
Western
sandpiper
~8cm\
deep \
Invertebrates:
Polychaetes
** Bivalves
Amphipods
Crabs
—f
/<3cm
* deep
Indicators
Biodiversity
Species
Population Size
Extent of
Aquatic Habitat
Figure A-4. Community Interactions submodel.
-------
A.3. CONCLUSIONS
The analysis of available data for potential indicators and of existing models indicated
that there was insufficient information available on metrics for the indicators to answer the
sensitivity questions of this assessment using quantitative modeling. However, it was also
evident that a vast amount of information local knowledge was available through consultation
with regional experts in the processes of interest. This led to the development of the expert
elicitation workshop approach described in Section 2 of this report. The workshop was meant to
serve as an opportunity to supplement current knowledge based on background research and
examine potential changes that may occur due to climate influences. The conceptual diagrams
described above provided the basis for the development of the initial influence diagrams used at
the workshop (as described in Section 2 of this report) as well as context for how these
ecosystem processes of focus fit with the rest of the ecosystem.
A-13
-------
A.4. REFERENCES
CCSP (Climate Change Science Program). (2008) Preliminary review of adaptation options for climate-sensitive
ecosystems and resources. Synthesis and Assessment Product 4.4. A report by the U.S. Climate Change Science
Program and the Subcommittee on Global Change Research. Julius, SH; J.M. West, JM; eds. Baron, JS; Griffith, B;
Joyce, LA; et al (Authors). U.S. Environmental Protection Agency, Washington, DC, USA, 873 pp.
http://downloads.climatescience.gov/sap/sap4-4/sap4-4-fmal-report-all.pdf
Jaffe, BE; Smith, RE; Torresan, LZ. (1998) Sedimentation and bathymetric change in San Pablo Bay: 1856-1983.
OPEN-FILE REPORT 98-759. Version 1.0. Available online at http://geopubs.wr.usgs.gov/open-file/of98-
759M98-759.pdf.
Kneib, R; Simenstad, C; Nobriga, M; et al. (2008) Tidal marsh conceptual model. Sacramento (CA): Delta Regional
Ecosystem Restoration Implementation Plan. Available online at
nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=10093
Orr, M; Crooks, S; Williams, PB. (2003) Will restored tidal marshes be sustainable? San Francisco Estuary
Watershed Sci 1(1): Article 1. Available online at http://www.escholarship.org/uc/item/8hi3d20t.
Page, GW; Whitacre, DF. (1975) Raptor predation on wintering shorebirds. Condor 77:73-78. Available online at
http://www.jstor.org/stable/1366760?seq=3
Reed, DJ. (1995) The response of coastal marshes to sea-level rise: survival or submergence? Earth Surf Proc
Landforms 20(l):39-48.
Ruhl, CA; Schoellhamer, DH. (2004) Spatial and temporal variability of suspended-sediment concentrations in a
shallow estuarine environment. San Francisco Estuary Watershed Sci 2(2):Article 1. Available online at
http ://repositories. cdlib. org/j mie/sfews/vol2/iss2/art 1.
SFEIT (San Francisco Estuary Indicators Team). (2008) Assessment framework as a tool for integrating and
communicating watershed health indicators for the San Francisco estuary. Technical Report #1. Submitted by the
San Francisco Estuary Indicators Team to the California Department of Water Resources, September 30, 2008.
Schoellhamer, D; Wright, S; Drexler; J; et al. (2007) Sedimentation conceptual model. Sacramento, (CA): Delta
Regional Ecosystem Restoration Implementation Plan. Available online at
http ://science .calwater.ca. gov/pdf/drerip/DRERIP_sediment_conceptual_model_fmal_ 111307 .pdf
SFEP (San Francisco Estuary Partnership). (1993) Comprehensive conservation and management plan.
SFEP (San Francisco Estuary Partnership). (2007) Comprehensive conservation and management plan. Available
online at http://www.cakex.org/sites/default/files/San%20Francisco%20Estuary%20Project.pdf
Stenzel, LE; Hickey, CM; Kjelmyr, JE; et al. (2002) Abundance and distribution of shorebirds in the San Francisco
Bay area. Western Birds 33:69-98.
Takekawa, JY; Miles, AK; Schoellhamer, DH; et al. (2006a) Trophic structure and avian communities across a
salinity gradient in evaporation ponds of the San Francisco Bay estuary. Hydrobiologia 567:307-327.
Takekawa, JY; Woo, I; Spautz, H; et al. (2006b) Environmental threats to tidal-marsh vertebrates of the San
Francisco Bay estuary. Stud Avian Biol 32: 176 197. Available online at
http://www.sfei.org/bioinvasions/Reports/No489 06 BIWT EnviroThreatstoVertebrates.pdf.
Wright, SA; Schoellhamer, DH. (2004) Trends in the sediment yield of the Sacramento River, California, 1957-
2001. San Francisco Estuary Watershed Sci 2:2. Available online at
http ://repositories. cdlib. org/j mie/sfews/vol2/iss2/art2.
A-14
-------
Ydenberg, RC; Butler, RW; Lank, DB; et al. (2004) Western sandpipers have altered migration tactics as peregrine
falcon populations have recovered. Proc R Soc Lond B 271(1545): 1263-1269. Available online at
http://www.sfu.ca/biology/wildberg/ydenbergetal2004PRSB.pdf
A-15
-------
APPENDIX B. EXPERT ELICITATION WORKSHOP PREPARATION
AND IMPLEMENTATION
B.I. PREWORKSHOP
B.I.I. Selecting Workshop Participants
The SFEP partners developed a list of criteria for selecting highly qualified local experts
who spanned the range of disciplines, science and management continuum, and empirical versus
theoretical research experience needed to collectively characterize the ecosystem processes under
consideration. Criteria for selecting participants included:
Demonstrated understanding of the body of literature with regard to sediment
retention OR community interactions (depending on which breakout group), as
evidenced by academic framing, research, and publications
Demonstrated ability to think of uncertainty in qualitative terms
Knowledge of science behind estuary management, as evidenced by academic
training, research, and publications
Knowledge of estuary management issues as evidenced by academic training,
research, and publications
Past work in SFEP region
Past work with salt marsh development/sediment retention processes (the balance of
sediment supply versus loss) OR mudflat development/community interactions
(interactions of shorebirds and their predators and prey), depending on the candidate's
proposed breakout group
These criteria were considered in developing a list of 15-17 qualified candidates for each
breakout group. Candidates were then contacted to determine their availability and interest in
testing a new method for vulnerability assessment. From this larger pool, a group of
seven experts was selected for each breakout group. According to EPA's Expert Elicitation Task
Force White Paper (http://www.epa.gov/spc/expertelicitation/index.htm), a review of the
literature indicates that 90% of successful expert elicitations use between 3 and 11 experts, with
B-l
-------
a law of diminishing returns in having a group larger than six. For this study, workshop
participants included the following individuals:
Sediment Retention Breakout Group:
Dave Cacchione, U.S. Geological Survey
John Callaway, UC San Francisco
Chris Enright, CA Department of Water Resources
Bruce Jaffe, U.S. Geological Survey
Lester McKee, San Francisco Estuary Institute
Dave Schoellhamer, U.S. Geological Survey
Mark Stacey, UC Berkeley
Community Interactions Breakout Group:
Letitia Grenier, San Francisco Estuary Institute
Jessica (Jessie) Lacy, U.S. Geological Survey
Michelle Orr, Philip Williams & Associates
Diana Stralberg, PRBO Conservation Science
Stuart Siegel, Wetlands and Water Resources
Lynne Trulio, San Jose State University
Isa Woo, U.S. Geological Survey
The expertise of each of the individual participants contributed to the interdisciplinary
complexity of the group. Experts were selected from the management and adaptation research
communities, and represented federal and state government agencies, research and consulting
organizations, and academia. The credentials for each of the participants, including past and
current work and research and areas of expertise, are summarized for the Sediment Retention
group in Table B-l, and for the Community Interactions group in Table B-2.
B.1.2. "Straw Man" Influence Diagrams
An initial "straw man" influence diagram (see Figure B-l and B-2) for each breakout
group was developed by ICF, EPA ORD, SFEP, and BCDC prior to the workshop based on the
B-2
-------
Table B-l. Sediment Retention breakout group participants, affiliations, and
qualifications
Name
Dave Cacchione
John Callaway
Chris Enright
Bruce Jaffe
Lester McKee
Dave Schoellhamer
Mark Stacey
Affiliation
U.S. Geological Survey
University of California,
San Francisco
California Department of
Water Resources
U.S. Geological Survey
San Francisco Estuary
Institute
U.S. Geological Survey
University of California,
Berkeley
Qualifications
Emeritus oceanographer for USGS. Research on sediment
transport, ocean-bottom boundary layers, erosion, wave
effects in San Francisco Bay area. Expertise in sediment
processes and wave impacts on coastal areas.
Research on wetland ecology and restoration in San
Francisco Bay. Expertise in wetland restoration, wetland
plant ecology, and sediment dynamics.
Chief Water Resources Engineer for Suisun Marsh Branch
of California Department of Water Resources. Expertise in
water resources planning, management, and sediment
dynamics.
Research on historical sedimentation and geomorphic
evolution of the San Francisco Estuary. Expertise in
sediment transport.
Research on transport, transformation, and loadings of
sediments, nutrients and contaminants in San Francisco
Bay area watersheds. Expertise in sediment transport,
hydrology, and nutrients.
Research on suspended-sediment transport in San Francisco
Bay and Delta. Expertise in estuarine physics, sediment
transport, and hydrology.
Research on transport and mixing in estuarine and coastal
environments. Expertise in sediment transport and
environmental fluid mechanics.
B-3
-------
Table B-2. Community Interactions breakout group participants,
affiliations, and qualifications
Name
Letitia Grenier
Jessica (Jessie) Lacy
Michelle Orr
Diana Stralberg
Stuart Siegel
Lynne Trulio
Isa Woo
Affiliation
San Francisco Estuary
Institute
U.S. Geological Survey
Philip Williams &
Associates
Point Reyes Bird
Observatory Conservation
Science
Wetlands and Water
Resources
San Jose State University
U.S. Geological Survey
Qualifications
Research on tidal marsh food web structure, song sparrow
fitness and behavior, monitoring of biota in the South Bay
Salt Ponds. Expertise in tidal marsh ecology.
Research on interaction between aquatic vegetation and
hydrodynamics. Expertise in sediment transport, estuarine
hydrodynamics, and aquatic ecosystems.
Water resources engineer involved with coastal marsh
geomorphology, hydraulic and sediment transport
modeling, and tidal channel dynamics. Expertise in
wetland restoration planning and design.
Research on modeling avian distributional responses to
climate, vegetation, and land use patterns. Expertise in
landscape ecology and avian species.
Consulting on wetlands technical and regulatory issues in
the San Francisco Bay area. Expertise in wetland and
aquatic ecology, wetland restoration and management.
Research on tidal salt marsh restoration and wildlife
management in the San Francisco Bay. Expertise in tidal
marsh ecology and restoration.
Research on tidal marsh foodwebs, trophic interactions, and
wetland restoration. Expertise in wetland restoration and
management.
B-4
-------
Altered Flows/
Water Demand:
Delta Pumping
Altered Flows/
Water Demand:
Reservoir
Management
Other Human Uses:
Dredging
Land Use/ Land Cover
Change:
Impervious Cover
Land Use/ Land Cover
Change:
Shoreline Armoring
Sediment
Deposition /
Retention
Figure B-l. Sediment Retention "straw man" influence diagram.
B-5
-------
Altered Flows/
Water Demand:
Delta Pumping
Altered Flows/
Water Demand:
Reservoir
Management
Other Human Uses:
Dredging
Land Use / Land Cover
Change:
Impervious Cover
Land Use / Land Cover
Change:
Shoreline Armoring
Sediment
Deposition /
Retention
Shorebird Predators
Peregrine Falcons,
Merlins
Water Depth to
Mudflat Surf ace
Shore birds
Ratio Deep Probers (Marbled
Godwit) to Shallow Probers
(Western Sandpiper)
Shorebird Prey
Polychaetes, Bivalves,
Amphipods
Figure B-2. Community Interactions "straw man" influence diagram.
B-6
-------
more detailed salt marsh and mudflat conceptual models and sediment retention and community
interactions submodels developed previously (see Appendix A). The "straw man" influence
diagrams differed from the more comprehensive conceptual models in that they focused on only
those elements of the model that participants believe are most critical for understanding
responses of the ecosystem process to the human and climate stressors under consideration. The
"straw man" influence diagrams were used in the preworkshop briefing and homework
assignment in order to further refine the sediment retention and community interactions influence
diagrams.
B.1.3. Preworkshop Briefing and Homework Assignment
Participants participated in a preworkshop briefing call and a homework assignment that
would be used to develop consolidated influence diagrams to be used at the workshop. The
preworkshop briefing call was held on March 2, 2010. This call gave participants a briefing on
the background of the project, work to date, the purpose of the workshop, and an overview of the
homework assignment. Part of the background material presented was information on an
example site for participants to consider when more spatial specificity would be useful during the
workshop exercise. China Camp, a site on the southwest shore of San Pablo Bay, was chosen
because it includes large wetland areas in a transitional salinity zone with intact adjacent
habitats.
The homework assignment asked participants to review a number of items: (1) selected
articles relevant to the ecosystem process breakout group to which they were assigned (for the
Sediment Retention breakout group: Orr et al. [2003]; Ruhl and Schoellhamer [2004]; and
Wright and Schoellhamer [2004]; for the Community Interactions breakout group: Galbraith
et al. [2005]; Takekawa et al. [2006b]; Page and Whitacre [1975]; and Stenzel et al. [2002]);
(2) conceptual models of the ecosystem and ecosystem process to which they were assigned; and
(3) the draft influence diagram for the ecosystem process to which they were assigned.
Participants were asked to review the draft influence diagram and provide recommendations on
what should be added or removed. Participants were asked to add or subtract variables or
relationships until the preliminary influence diagram matched their understanding of the process.
We asked participants to include no more than 10-15 variables in the diagram in order to keep it
focused on the highest priority influences. We also asked participants to focus on current
conditions (including current climate) when reviewing and commenting on the diagram.
B-7
-------
Participants were asked to provide a quantitative definition for each variable, a metric for
measuring the variable, and a range of values for the metric. Participants were also asked to
assign values to the metrics they selected. This could include actual measured values
(e.g., 35 km3 of inflow) as well as a range of values (e.g., 5 to 50 km3 of inflow).
B.I.4. Consolidated Influence Diagrams
The preliminary diagram for each breakout group was revised prior the workshop based
on the participants' homework responses. The process involved examining the participants'
responses and constructing a tally of the variables used and influences (arrows) included.
Variables and influences that were most frequent across all responses were included in the
consolidated influence diagrams. For the both the Sediment Retention and Community
Interactions groups, all but one of the participants provided comments on the preliminary
influence diagram. Based on the responses from the participants, consolidated influence
diagrams were developed for the workshop.
B.2. WORKSHOP
B.2.1. Group Influence Diagrams
Group influence diagrams were developed during the first day of the workshop. Within
their breakout groups, the participants discussed how the consolidated influence diagrams should
be refined for use as a final "group" influence diagram. The participants added, removed, or
redefined variables based on a group discussion. The group diagrams were to become the basis
for the expert elicitation exercise of assigning judgments about influences among variables. The
Sediment Retention and Community Interactions group influence diagrams are provided in
Section 2.
B.2.2. Introduction to Climate Scenarios and Confidence
The participants received two handouts designed to orient them to the climate scenarios
and to the methodology for assessing confidence. The first handout contained a summary of
Climate Scenarios A and B, which was used by the participants in assessing the sensitivity of salt
marshes and mudflats across a range of plausible scenarios of climate change. It explained the
development of two climate futures in a midcentury (2035-2064) time frame. Participants used
B-8
-------
these scenarios on Day 2 to make new judgments compared to their judgments under "current
conditions" on Day 2. The full climate scenarios handout can be found in Appendix C.
The second handout presented explanatory information and a coding scheme for use by
the participants in assessing their confidence in each of their judgments under both current
conditions and under Climate Scenarios A and B. The full handout may be found in
Appendix D.
B.2.3. Coding Exercise
Following the development of the group influence diagrams, participants were asked to
make their individual judgments on the diagram using the coding scheme. As described in
Section 2, the participants used the coding scheme to make judgments on the following: (1) type
and degree of influence for each relationship included in the influence diagram; (2) the
associated confidence for each influence judgment; (3) type of interactive influences for
relationships of their own choosing; and (4) the associated confidence for each interactive
influence judgment. These judgments were done for current conditions (on the first day of the
workshop), and Climate Scenario and Climate Scenario (on the second day of the workshop).
Example handouts that participants used to make their judgments are provided in Tables B-3,
B-4, andB-5.
B.2.4. Variation Across Participants in Sensitivity Judgments
For both the Sediment Retention and Community Interactions groups, variability among
participants in their judgments contributed to lack of agreement on sensitivities for some
influences. Figure B-3 presents the full range of variation among participants of the Sediment
Retention group by showing the same trio of figures as shown in Figure 2-5, but broken out for
each individual participant. Looking across all the participants, there was more variability
between participants than across scenarios for any given participant. There were no patterns
across participants, such as characterizing only increasing sensitivity. The changes across the
scenarios made by Participant 3 were of only increasing sensitivity, Participant 1 only had
one change to decreasing sensitivity, and Participants 2, 6 and 7 had both increases and
decreases, sometimes across the scenarios for one influence. Partipants 4 and 5 made no changes
in sensitivity across the climate scenarios.
For the Community Interactions group, Figure B-4 presents the full range of variation
among participants by showing the same trio of figures as those shown in Figure 2-10, but
B-9
-------
Table B-3. Example of expert elicitation handout for influences under current conditions (Sediment Retention
group)
Instructions:
Please assess the effect of X on Yby selecting the appropriate "degree of influence" and its associated "confidence".
Relationship A
Relationship B
Relationship C
Relationship D
Relationship £
Relationship F
Relationship G
Variable X
Water resource
management: delta
outflow
Water resource
management:
reservoir management
Water resource
management:
reservoir management
Water resource
management:
reservoir management
Water resource
management:
channelization
Water resource
management:
channelization
Water resource
management:
channelization
on
on
on
on
on
on
on
Variable Y
Freshwater inflow
Freshwater inflow
Sediment flux
Sediment size
Freshwater inflow
Sediment flux
Sediment size
Degree of influence
(Please select 0-13)
Confidence
(LH, LL, HH, HL)
Notes
td
I
o
-------
Table B-4. Example of expert elicitation handout for influences under climate scenarios (Community
Interactions group)
Instructions: Please assess the effect of X on Y by selecting the appropriate "degree of influence" and its associated "confidence".
Relationship A
Relationship B
Relationship C
Relationship D
Relationship £
Relationship F
Relationship G
Variable X
Water
management
Water
management
Restoration
Restoration
Restoration
Land use
change
Land use
change
On
On
On
On
On
On
On
Variable Y
Freshwater
inflow
Sediment
supply
Tides and
hydrodynamics
Sediment
supply
Landscape
mosaic
Sediment
supply
Landscape
mosaic
Climate Scenario A
Degree of influence
(Please select 0-13)
Confidence
(LH, LL, HH,
HL)
Climate Scenario B
Degree of
influence
(Please select
0-13)
Confidence
(LH, LL, HH,
HL)
Notes
td
-------
Table B-5. Example of expert elicitation handout for interactive influences under climate scenarios (Sediment
Retention group)
Instructions: Please assess the effect of X on Y with Z by selecting the appropriate "interactive influence" and its associated "confidence".
Example 1:
Relationship
A+B
Example 2:
Relationship
Q+R
Variable X
Water
resource
management:
delta outflow
Sediment flux
On
On
On
Variable Y
Freshwater
inflow
Net mineral
accumulatio
n
With
With
With
Variable Z
Water
resource
management:
reservoir
management
Sediment size
Climate scenario A
Interactive
influence
Confidence
(LH, LL,
HH, HL)
Climate scenario B
Interactive
influence
Confidence
(LH, LL,
HH, HL)
Notes
td
-------
Current
Scenario A
Scenario B
Participant 1
Participant 2
Participant 3
Participant 4
Key
Low sensitivity
Intermediate sensitivity
High sensitivity
No influence
Unknown
+•
No answer
Figure B-3. Sediment Retention influence diagrams of sensitivities: variance
across participants.
B-13
-------
Current
Scenario A
Scenario B
Participant 5
Participant 6
Participant 7
Key
Low sensitivity
Intermediate sensitivity
High sensitivity
No influence
Unknown
».
No answer
Figure B-3. Sediment Retention influence diagrams of sensitivities: variance
across participants, (continued)
B-14
-------
Current
Scenario A
Scenario B
Participant 1
Participant 2
Participant 3
Participant 4
Key
Low sensitivity
Intermediate sensitivity
High sensitivity
No influence
Unknown
+.
No answer
Figure B-4. Community Interactions influence diagrams of sensitivities:
variance across participants.
B-15
-------
Current
Scenario A
Scenario B
Participant 5
Participant 6
Participant 7
Key
Low sensitivity
Intermediate sensitivity
High sensitivity
No influence
Unknown
».
No answer
Figure B-4. Community Interactions influence diagrams of sensitivities:
variance across participants, (continued)
B-16
-------
broken out for each individual participant. Looking across all the participants, we see that there
is again more variability between participants than across scenarios for any given participant.
The majority of changes in sensitivity type across the climate scenarios are of increasing
sensitivity. The changes across the scenarios made by Participants 1, 2 and 4 are of only
increasing sensitivity; Participant 5 only had one change, to decreasing sensitivity; Participants 3,
6 and 7 had both increases and decreases, but more of the former.
B.2.5. Exercise Discussions and Report-outs
After participants made their individual judgments on the influence diagram using the
coding exercise, the participants reconvened in their breakout groups for a group discussion.
Participants discussed their reactions to the exercise and how it was structured, individual
judgments on type and degree of influence, individual judgments on confidence, key issues and
gaps in understanding. This group discussion often helped to clarify issues that participants may
have had in understanding the coding scheme or influences that they may have lacked clarity on.
Based on this group discussion, the facilitator helped the participants to identify some key
points that emerged. These key points addressed issues such as key influences, important
pathways, thresholds, significant changes associated with climate change, management
implications, etc. One of the participants from each breakout group presented these key points to
the larger group to summarize the discussion. Following the discussion, participants were given
time to revisit their individual judgments.
B.2.6. Discussion of Management Implications
Following the breakout group discussions and exercise of making individual judgments,
participants gathered in the larger group to discuss management implications. This discussion
would help SFEP and BCDC to examine some of the key issues that emerged from the expert
elicitation exercise and translating those issues into action. The facilitator led the discussion by
asking participants to consider how climate stressors might impact the estuary across a range of
management scenarios. The discussion also explored research and data needs, suggestions for
water and sediment management, and fundamental shifts in management that may be necessary.
B.3. POSTWORKSHOP
B.3.1. Review of Workshop Report
A report was developed subsequent to the workshop documenting key outputs in
two sections: key results and workshop discussions. This report provides a documentation of all
of the participant materials, including: participant guidance documents, participant homework
B-17
-------
responses, handouts and other materials used at the workshop, and individual participant
judgments. Key points that emerged during the breakout group and larger group discussions are
summarized, as well as the discussion on management implications. Participants were asked to
review this report and provide any comments. These comments were incorporated into a final
workshop report, which is available upon request from the authors.
B.3.2. Synthesis of Results
A synthesis of results was developed in order to analyze the participants' individual
judgments made at the workshop. The synthesis reviews the objectives of conducting the expert
elicitation workshop and identifies key questions that the synthesis of judgments seeks to answer.
It reviews the coding schemes used by participants during the workshop and summarizes a
coding typology that was used to group codes to characterize types and degrees of influences and
sensitivities. Finally, it describes the methodology for analyzing the available judgments and
presents key results in the form of tables and figures. The contents of this synthesis comprise
much of the substance of the results sections of this report.
B.3.3. Review of Draft Report
The workshop report and preliminary results reports were used to develop this technical
report to present the synthesis results and place them in the larger context of the implications for
management and SFEP's capacity to respond. The draft report was revised based on an internal
review by EPA scientists. The report is now under public and expert peer review. Following
this review, a final report will be developed that responds to the public and peer-review
comments. An additional report that focuses on lessons learned across the two assessments for
SFEP and the Massachusetts Bays Program will also be developed.
B-18
-------
B.4. REFERENCES
GalbraithH; Jones, R; Park, R; et al. (2005) Global climate change and sea level rise: potential losses of intertidal
habitat for shorebirds. USD A Forest Service Gen. Tech. Rep. PSW-GTR-191. Available online at
http://research.fit.edu/sealevelriselibrarv/documents/doc mgr/470/Global Potential Losses of Shorebird Habitats
-_Galbraith_et_al._2005 .pdf
Orr, M; Crooks, S; Williams, PB. (2003) Will restored tidal marshes be sustainable? San Francisco Estuary
Watershed Sci 1(1): Article 1. Available online at http://www.escholarship.org/uc/item/8hj3d20t.
Page, GW; Whitacre, DF. (1975) Raptor predation on wintering shorebirds. Condor 77:73-78. Available online at
http://www.jstor.org/stable/1366760?seq=3
Ruhl, CA; Schoellhamer, DH. (2004) Spatial and temporal variability of suspended-sediment concentrations in a
shallow estuarine environment. San Francisco Estuary Watershed Sci 2(2):Article 1. Available online at
http ://repositories. cdlib. org/j mie/sfews/vol2/iss2/art 1.
Stenzel, LE; Hickey, CM; Kjelmyr, JE; et al. (2002) Abundance and distribution of shorebirds in the San Francisco
Bay area. Western Birds 33:69-98.
Takekawa, JY; Woo, I; Spautz, H; et al. (2006b) Environmental threats to tidal-marsh vertebrates of the San
Francisco Bay estuary. StudAvian Biol 32: 176 197. Available online at
http://www.sfei.org/bioinvasions/Reports/No489_06_BI_WT_EnviroThreatsto Vertebrates.pdf.
Wright, SA; Schoellhamer, DH. (2004) Trends in the sediment yield of the Sacramento River, California, 1957-
2001. San Francisco Estuary Watershed Sci 2:2. Available online at
http ://repositories. cdlib. org/i mie/sfews/vol2/iss2/art2.
B-19
-------
APPENDIX C. PARTICIPANT HANDOUT ON CLIMATE SCENARIOS
SFEP Workshop Climate Scenarios
This handout is intended to assist participants in assessing the sensitivity of salt marshes and
mudflats across a range of plausible scenarios of climate change. It provides the details of two
distinct but scientifically credible climate futures for a mid-century (2035-2064) time period.
Participants will use these scenarios in revisiting their assessments of influence completed on the
first day.
Two Climate Change Scenarios: "Lower-Range" and "Higher-Range"3
Under both climate change scenarios, California will retain its Mediterranean climate (cool/wet
winters and hot/dry summers) and continue to experience a high degree of variability in
precipitation with rising sea levels. By mid-century, the "higher-range" scenario (including
higher emissions and a more sensitive climate) is projected to experience a warmer and
somewhat drier climate compared to the "lower-range" scenario (with lower emissions and a
lesser impact on California's climate).
Development of the Climate Scenarios
The two bounding scenarios were developed from a collective group of studies in large part
funded by the California Energy Commission (CEC) under the mandate of the Governor's
Biennial Climate Change Report. A majority of the climate projections presented here were
developed by Cayan et al. (2009), based on projections from 6 leading climate models.4 These
models were selected based on their reasonable representation of historical simulation of
seasonal precipitation, seasonal temperature, the variability of annual precipitation, and El
Nino/Southern Oscillation (ENSO). All models were run with both a lower emission scenario
(Bl SRES) and a mid-high emission scenario (A2 SRES) to capture a range of plausible future
emissions trajectories. The "lower-range" and "higher-range" temperature and precipitation
scenarios for 2035-2064 compared to 1961-1990 baseline conditions are based on these climate
model simulations, for the SRES Bl (lower) and SRES A2 (higher) scenarios, respectively.
Regional projections were developed by statistical downscaling.5
For a given U.S. coastal location, relative sea level rise may differ from global estimates due to a
number of factors such as changes in local ocean circulation, ocean density, vertical land motion,
2 These two futures are designed to capture a large part of the uncertainty inherent to future projections that is the
result of two key factors: (1) the amount of future emissions of greenhouse gases from human activities that are
driving global change, and (2) the ability of scientists to simulate the response of the Earth's climate system to those
emissions.
3 The usage of the terms "lower-range" and "higher-range" refers to the scenarios provided in this handout and are
not intended to reflect the lowest and highest possible futures.
4 U.S. NOAA's Geophysical Fluid Dynamics Laboratory (GFDL) CM2.1; the National Center for Atmospheric
Research (NCAR) Parallel Climate Model (PCM); the National Center for Atmospheric Research (NCAR)
Community Climate System Model (CCSM); the Max Plank Institute ECHAM5/MPI-OM; the Center for Climate
System Research of the University of Tokyo MIROC 3.2 medium-resolution model; and the French Centre National
de Recherches Meteorologiques (CNRM) models.
5 Statistical downscaling methodology includes constructed analogues, bias correction and spatial downscaling of the
results from each of the 6 climate models.
C-l
-------
erosion/sedimentation, gravitational effects, etc. Relative sea level rise in California has
demonstrated similar rates of rise compared to global estimates (Cayan et al. 2008). Many
California studies recommend using projections of global sea level rise estimates, which assumes
relative sea levels continue to rise at the same rate as projected global sea level rise. The "lower-
range" sea level rise estimate is provided as a mid-range of Rahmstorf (2007) and high-end of
IPCC TAR. The "higher-range" sea level rise estimate is provided as the high estimate of
Rahmstorf (2007).
Summary of Climate Scenarios: Averages for Mid-Century
Temperature6
Precipitation
Annual Average7
Average Increase of
Winter Temperature8
Average Increase of
Summer
Temperature0
Extreme Heat Days
Annual Change
Winter change
Heavy Events
"Lower-Range"
Scenario
+2.8°F (1.6°C)
+2.5°F(1.4°C)
+4.0°F(2.2°C)
+10 days/year
-4.5%
"Higher-Range"
Scenario
+3.5°F (1.9 °C)
+2.7°F (1.5 °C)
+4.5°F (2.5 °C)
+16 days/year
-7%
Reduced winter precipitation11
Decline in frequency of precipitation events
(exceeding 3mm/day) but not a clear signal in changes
of precipitation intensity
Since the 1920s, minimum and maximum daily temperature have been observed to have increased in California
with minimum temperature increasing at a greater rate accented by a small cooling trend in the summer (Cayan et al.
2009). These averages are for 2035-2064 projections relative to a 1961 to 1990 baseline forBl and A2 emission
scenarios.
7 Approximate results using B1 and A2 emissions scenarios and three global climate models (PCM1, GFDL CM2.1,
HadCMS) (CEC 2006).
8 These results are for Sacramento, California. This warming is projected to be more moderate along the coastline
(50 km from the coast) rising considerably inland (Cayan et al. 2009). These averages are for 2035-2064 projections
relative to a!961 to 1990 baseline forBl and A2 emissions scenarios.
9 Extreme heat days are defined as when the daily maximum temperature exceeds the 95th percentile of temperature
from the 1961-1990 historical averages of May-September days. 1961-1990 extreme heat days are approximately 8
days/year based on model runs. Results are provided by Cayan et al. (2009) using three climate models (CNRM
CMS, GFDL CM2.1, MICRO 3.2; with bias corrected spatial downscaling) forBl and A2 emissions scenarios.
Mid-century projections suggest hot daytime and nighttime temperatures increase in frequency, magnitude, and
duration (Cayan et al. 2009). Extreme warm temperatures in California, historically a July and August phenomenon,
will increase in frequency and magnitude likely beginning in June and may continue into September (Hayhoe et al.
2004; Gershunov and Douville 008; Miller et al. 2008).
1 ° Results are averaged across 6 GCMs using the grid point nearest to Sacramento (Cayan et al. 2009) for B1 and A2
emissions scenarios.
11 These results are provided by CEC (2008).
C-2
-------
Sea Level
Storms/Wind15
Total Increase for
205012
Hourly Sea Level
Exceedances14
"Lower-Range"
Scenario
+30 cm
1343
"Higher-Range"
Scenario
+45 cm13
1438
Tendency toward a decline in storms. Projections suggest an increased
tendency for heightened sea level events to persist for more hours.
ENSO is not projected to increase in frequency or intensity.
What else do these changes mean for our system?
Snow Pack Change
Spring Runoff
Seasonal Changes in Amount
of Freshwater Inflow to the
Bay from the Delta in 2060 18
For the Sacramento-San Joaquin watershed, April watershed-
total snow accumulation projected to drop by 64% by 2060. 1?
Spring runoff occurring earlier and reduced overall
October through February: inflow +20%
March through September: inflow -20%
Where can I find additional information?
California Climate Change Research Center
http ://www. climatechange. ca. gov/research/index.html
Union of Concerned Scientists
www. climatechoices. org/ca
Sea level rise relative to 2000 levels. This study applies Rahmstorf s methodology of estimating sea level rise as a
function of rising temperatures. This study assumes sea level rise along the coast to be the same as global estimates
given the observed rate of rise along the southern California coast has been about 17 to 20 cm per century similar to
that of global sea level rise (assume no future changes in other factors that affect relative sea level rise such as
changes in regional/local ocean circulation, ocean density, etc.) (Cayan et al. 2009). DMRS also provides
recommended 2050 global sea level rise estimates relative to 1990 values: 11 cm (direct extrapolation of observed
increased during the 20th century), 20 cm (low-end value of Rahmstorf and approx mid-range of IPCC TAR), 30 cm
(approx mid-range of Rahmstorf and high-end of IPCC TAR); 41 cm (high end of Rahmstorf) (DMRS 2007).
13 The total difference between mean range and spring range of 1.7 ft (50.3 cm) is slightly larger than the higher-
range scenario rise of 45 cm, based on the Point San Pedro tide station.
http ://tidesandcurrents. noaa. gov/tides 10/tab2wc 1 a. html# 128
14 The hourly sea level exceedance is defined as the maximum duration (hours) when San Francisco sea level
exceeds the 99.99th % level (140 cm above mean sea level) based on the GFDL climate change (A2) simulation
using the Rahmstorf sea level scheme averaged 2 to 4 hours increase for mid-century (Cayan et al. 2009).
15 These results are provided by Cayan et al. (2009).
16 Storm is defined as sea level pressure (SLP) equaling or falling below 1005 millibar (mb).
17 Results provided by the Bay-Delta watershed model driven by temperature projections from a parallel climate
model under a 'business-as-usual' scenario relative to 1995-2005 (precipitation is assumed to remain consistent with
today's observations) (Knowles and Cayan 2004).
18 This study does account for reservoirs, in-stream valley diversions, and in-Delta withdrawals and assumes no
future management adaptation or altered demand patterns (Knowles and Cayan 2004).
C-3
-------
Works Cited
California Energy Commission (CEC). 2006. Our Changing Climate: Assessing the Risks to
California. A Summary Report from the California Climate Change Center.
California Energy Commission (CEC). 2008. The Future is Now: An Update on Climate Change
Science, Impacts, and Response Options for California. A report from the California Climate
Change Center California Energy Commission's Public Interest Energy Research Program.
Cayan, D., P. Bromirski, K. Hayhoe, M. Tyree, M. Dettinger, & R. Flick. 2008. Climate change
projections of sea level extremes along the California coast. Climatic Change 87(0), 57-73. DOI:
10.1007/sl0584-007-9376-7.
Cayan, D., M. Tyree, M. Dettinger, H. Hidalgo, T. Das, E. Maurer, P. Bromirski, N. Graham,
and R. Flick. 2009. Climate Change Scenarios and Sea Level Rise Estimates for the California
2008 Climate Change Scenarios Assessment. California Climate Change Center. CEC-500-
2009-014-D.
Delta Risk Management Study (DRMS) 2007. Technical Memorandum: Delta Risk
Management Study (DRMS) Phase 1.
Gershunov A., and H. Douville. 2008. Extensive summer hot and cold extremes under current
and possible future climatic conditions: Europe and North America. In: H. Diaz, and R. Murnane
(Eds.), Climate Extremes and Society. Cambridge University Press.
Hayhoe K., D. Cayan, C. B. Field, P. C. Frumhoff, E. P. Maurer, N. L. Miller, S. C. Moser, S. H.
Schneider, K. N. Cahill, E. E. Cleland, L. Dale, R. Drapek, R. M. Hanemann, L. S. Kalkstein, J.
Lenihan, C. K. Lunch, R. P. Neilson, S. C. Sheridan, and J. H. Verville. 2004. "Emissions
pathways, climate change, and impacts on California." PNAS 101(34): 12422-12427. Aug. 24;
Epub Aug. 16, 2004.
Knowles, N. and D. Cayan. 2004. Elevational Dependence of Projected Hydrologic changes in
the San Francisco Estuary and Watershed. Climate Change 62: 319-336.
Miller, N. L., K. Hayhoe, J. Jin, and M. Auffhammer. 2008. "Climate, Extreme Heat, and
Electricity Demand in California." Journal of Applied Meteorology and Climatology
47:1834-1844.
C-4
-------
APPENDIX D. PARTICIPANT HANDOUT ON CONFIDENCE
Method for Assessing Confidence in Expert Judgments
Characterization of uncertainty is a critical component of assessment science. Thus this
workshop exercise includes a component in which the expert participants will assess their current
level of scientific confidence in each influence for which they are making a judgment. The aim is
to provide information on not only degrees of influence among variables, but also the degree of
uncertainty associated with each judgment, given the current state of knowledge in the scientific
community.
The design of this analysis is derived from general guidance on uncertainty from recent large
assessment efforts such as those of the Intergovernmental Panel on Climate Change (IPCC) and
the U.S. Climate Change Science Program (CCSP) [e.g., see Moss and Schneider, 2000; IPCC,
2004; IPCC, 2005; CCSP, 2008; CCSP, 2009]. One fundamental principle is the distinction
between uncertainty expressed in terms of "likelihood" of an outcome versus "level of
confidence" in the science underlying the finding. Likelihood is relevant when assessing the
chance of defined future occurrence or outcome, and involves assigning numerical probabilities
to qualifiers such as "probable," "possible," "likely," "unlikely" (CCSP 2009). In contrast, level
of confidence refers to the (qualitative) degree of belief within the scientific community that
knowledge, models, and analyses are accurate, based on the available evidence and the degree of
consensus in its interpretation. We are taking this latter approach.
Each expert is asked to rate his/her confidence in each judgment about degree of influence based
on: (1) the amount of scientific evidence that is available to support the judgment; and (2) the
level of agreement/consensus in the expert community regarding the different lines of evidence
that would support the judgment. These confidence attributes are further described below:
High/low amount of evidence: Is the judgment based on information that is well-studied and
understood, or mostly experimental or theoretical and not well-studied? Does your experience in
the field, your analyses of data, and your understanding of the literature indicate that there is a
high or low amount of information on this influence? Sources of evidence - in order of relative
importance - include: 1) peer-reviewed literature; 2) grey literature; 3) data sets; 4) personal
observations and personal communications.
High/low amount of agreement: Do the studies and reports across the scientific community, as
well as your own experience in the field or analyzing data, reflect a high degree of agreement
about the influence, or do they lead to competing interpretations?
Based on the above, levels of confidence in judgments can be sorted into four general categories:
• Well established = high evidence/high agreement (HH);
• Competing explanations = high evidence/low agreement (HL);
• Established but incomplete = low evidence/high agreement (LH);
• Speculative = low evidence/low agreement (LL).
D-l
-------
References
CCSP, 2008: Preliminary review of adaptation options for climate-sensitive ecosystems and
resources. A Report by the U.S. Climate Change Science Program and the Subcommittee on
Global Change Research. [Julius, S.H., J.M. West (eds.), J.S. Baron, B. Griffith, L.A. Joyce, P.
Kareiva, B.D. Keller, M.A. Palmer, C.H. Peterson, and J.M. Scott (Authors)]. U.S.
Environmental Protection Agency, Washington, DC, USA, 873 pp.
CCSP, 2009: Best practice approaches for characterizing, communicating and incorporating
scientific uncertainty in climate decision making. A Report by the U.S. Climate Change Science
Program and the Subcommittee on Global Change Research. [Morgan, M.G., H. Dowlatabadi,
M. Henrion, D. Keith, R. Lempert, S. McBride, M. Small, and T. Wilbanks (Authors)].
Washington, DC, 156 pp.
IPCC, 2004: Workshop on Describing Scientific Uncertainties in Climate Change to Support
Analysis of Risk and of Options. Working Group I Technical Support Unit, Boulder, Colorado
[Manning, M., M. Petit, D. Easterling, J. Murphy, A. Patwardhan, H.-H. Rogner, R. Swart and
G. Yohe (eds.)], May 11-13, 2004, National University of Ireland, Maynooth, Co. Kildare,
Ireland, 146 pp. Available at:
http://ipcc-wgl.ucar.edu/meeting/URW/product/URW Report v2.pdf
IPCC, 2005: Guidance Notes for Lead Authors of the IPCC Fourth Assessment Report on
Addressing Uncertainties, 4 pp. Available at: www.ipcc.ch/pdf/assessment-report/ar4/wgl/ar4-
uncertaintyguidancenote.pdf.
Moss, R. and S.H. Schneider, 2000: Uncertainties in the IPCC TAR: Recommendations to lead
authors for more consistent assessment and reporting. In: Guidance Papers on the Cross Cutting
Issues of the Third Assessment Report of the IPCC [Pachauri, R., T. Taniguchi, K. Tanaka
(eds.)], World Meteorological Organisation, Geneva, Switzerland, 33-51.
D-2
------- |