The 1980 eruption of Mount St. Helens in southwest Washington State radically changed the physical and socioeconomic landscapes of the region. The eruption destroyed the summit of the volcano, sending large amounts of debris into the North Fork Toutle River and blocking the sole means of drainage from Spirit Lake 4 miles (6.4 km) north of Mount St. Helens (see Figure S.1). Rising lake levels could cause failure of the debris blockage, putting the downstream population of approximately 50,000 at risk of catastrophic flooding and mud flows. Furthermore, continued transport of sediment to the river from volcanic debris deposits surrounding the mountain reduces the flood carrying capacity of downstream river channels and leaves the population vulnerable to chronic flooding.
Engineering measures were implemented in the 1980s to manage both catastrophic and chronic risks associated with the debris blockage and sediment loads in the rivers. These included construction of a 1.56 mile (2.6 km) tunnel at Spirit Lake to drain the lake and control lake levels, a sediment retention structure (SRS) on the North Fork Toutle River approximately 8 miles (~13 km) downstream of Spirit Lake, and implementation of flood risk management measures, including levee upgrades in the lower Cowlitz River valley. River dredging has also been necessary to maintain navigation.
Engineering measures now in place, however, do not represent long-term solutions to the region’s risk management challenges. Because the Spirit Lake outflow tunnel serves as the only drainage for Spirit Lake, disruption of tunnel operations leaves the debris blockage vulnerable to breaching. The tunnel has required major repairs and is not operating optimally. Additional expensive repairs are necessary, and, as for any constructed facility, continued costly maintenance will be needed. Downstream, the SRS is close to reaching its sediment trapping capacity, and
plans to increase that capacity by raising the SRS spillway provide only short-term solutions to the sediment transport problem.
The legacy of the 1980 eruption and the prospect of future volcanic, seismic, and flood events mean that risk management in the Spirit Lake and Toutle River system will be challenging for decades to come. Future actions need to accommodate a much broader range of objectives and priorities of all interested and affected parties in the region. Meanwhile, the responsibilities for managing risk and the natural resources in the watershed are dispersed among federal, state, and local agencies of government with different and sometimes conflicting goals and authorities.
THE STUDY CHARGE
Inspections of the Spirit Lake outflow tunnel in 2014 indicating a need for millions of dollars in repairs to avoid failure led members of Congress to request that the U.S. Forest Service (USFS), the U.S. Army Corps of Engineers (likely), and the U.S. Geological Survey (USGS)—agencies with mandated responsibilities in the region—develop a long-term plan to manage Spirit Lake water levels. At the request of the USFS, the National Academies of Sciences, Engineering, and Medicine convened a committee to develop a decision framework to support the long-term management of risks related to the Spirit Lake and Toutle River system in light of the
different regional economic, cultural, and social priorities and the respective roles of federal, tribal, state, and local authorities, as well as other entities and groups in the region (referred to herein as interested and affected parties).
In addition to developing a decision framework, the committee was asked to consider the history and adequacy of characterization, monitoring, and management associated with the Spirit Lake debris blockage and outflow tunnel; to consider other efforts to control transport of water and sediment from the 1980 and later eruptions; and to suggest additional information needed to support implementation of the recommended decision framework. The committee was also asked to identify alternatives that might be considered for long-term management of water and sediment transport within the Spirit Lake and Toutle River system. The statement of task does not call for the committee to quantitatively examine the viability of long-term management alternatives. Instead, regional authorities, guided by the proposed decision framework, would perform detailed analyses later. The committee concluded that such an examination could include a quantitative risk assessment, benefit-cost analyses, and analyses of other data.
THE EVOLVING DECISION LANDSCAPE
Two types of long-term and system-wide risks need to be considered in the Spirit Lake and Toutle River region. First are the relatively high-probability, moderate-consequence risks associated with chronic flooding along the Toutle and Cowlitz Rivers. These could cause social and economic disruption in populated and commercial areas and are mostly the result of channel infill from the movement of sediment out of the Toutle River and into the Cowlitz River. Second is the likelihood of life loss and community destruction caused by catastrophic flooding and mudflows into populated areas along the Toutle River and the lower Cowlitz River. These risks are of relatively low probability but high consequence, likely the result of the destabilization of sediment above the SRS or due to a breach of the Spirit Lake debris blockage.
Following the 1980 eruption, two principal considerations influenced management decisions: (1) the costs of possible management actions and (2) their impacts on the safety of downstream communities.
Decision making related to water and sediment transport in the region has tended to be linear: a responsible agency formulated a specific problem within its authority, analyzed options, and made a decision. Engagement with interested and affected parties consisted largely of public meetings held by the agency at certain points in the decision process to receive public comments. Although this process accomplishes some goals, it typically limits opportunities to explore the values and management ideas of other interested and affected parties, misses opportunities to identify joint gains, and can leave the excluded parties lacking trust in decisions made by those in authority. Decades past the initial response, values such as those related to ecological conditions and recreational benefits have gained currency in stakeholder perceptions. For example, prior to the eruption, the North Fork Toutle River valley was an important recreation area for fishers, hunters, and other users. Commonly discussed among local residents today are the impacts of management decisions on the possibility of recovering fish species, including salmon, steelhead, and the sea-run coastal cutthroat trout, all of which spawn naturally within river systems.
DATA AND ANALYTICAL NEEDS
Since 1980, natural and engineered processes have changed the Spirit Lake and Toutle River system. Engineering practice has evolved, as have concerns among interested and affected parties. Many data collection activities, however, such as groundwater monitoring within the debris blockage and most measurements of sediment sources and transport, stopped in the 1980s or 1990s, and few new data have been collected in response to changing priorities, such as those related to aquatic ecology. The information available to correctly inform long-term management of the region is outdated and incomplete. Physical characterization of components of the system needs to be updated. Monitoring capabilities and data collection programs need to be updated, and analytic capabilities
need to be reevaluated. Key data that could inform decisions need to be identified collaboratively.
Decisions related to the long-term management of Spirit Lake water levels need to be informed by a current characterization of the debris blockage damming the lake; the location and behavior of groundwater in the blockage; current meteorological trends; a quantified characterization of risks posed by volcanic activity on Spirit Lake water levels; and on the response of the blockage to local and regional seismic events. Recent insights regarding the likelihood of a Cascadia Seismic Zone earthquake affecting the Mount St. Helens vicinity warrant greater examination. Systematic site-specific seismic hazard studies to develop reliable estimates of anticipated ground shaking are needed (particularly at Spirit Lake and the SRS) but have not been performed although such analyses are routine for infrastructure around the country.
Recommendation: Agencies engaged in risk management in the Spirit Lake and Toutle River region should develop a coordinated and targeted monitoring system to track changes in factors that affect risk. Data and analyses should be shared and made available to all. (Chapter 4)
Among the more credible risk scenarios for Spirit Lake is the rapid lake level rise observed during periods when the tunnel is closed for repair. This represents an operational risk. Another operational risk could be failure of such engineered structures as the SRS or levees. Occupational safety and health risks associated with operating infrastructure (e.g., wood debris removal near the Spirit Lake tunnel intake or opening and closing the tunnel gates) represent another kind of operational risk. It appears that such operational factors have not been systematically considered in appraising risks associated with the Spirit Lake and Toutle River system.
Modern approaches to risk management are increasingly based on probabilistic risk analyses, which address the capability of a system to withstand extreme loads such as the demand caused by the probable max-
imum flood or a maximum credible earthquake. Probabilistic risk analysis is especially useful in appraising design and rehabilitation decisions and corresponding factors of safety.
Assessing operational risk involves considering the vagaries of weather, human operators, sensors, supervisory control and data acquisition (SCADA) systems, and other operational factors. It may be that the risks posed by the system derive in part from these operational factors rather than from natural hazards and extreme events alone.
Recommendation: Operational risk should be explicitly considered when evaluating alternatives for management. (Chapter 5)
THE DECISION FRAMEWORK
Given the uncertainties associated with potential moderate intensity and catastrophic events, as well as the analytic uncertainty associated with incomplete or outdated information, an analytic decision process that establishes risk management as an organizing principle is needed. But, given the competing values of interested and affected parties in the region; the lack of agreement on planning time frames; the overlapping but sometimes competing management responsibilities and authorities in the region; and the limited budgets of those authorities, that process needs to promote communication and trust among agencies and the public so that technical decisions effectively and satisfactorily incorporate the priorities of those interested and affected parties.
Recommendation: Adopt a deliberative and participator y decision-making process that includes technical considerations; balances competing safety, environmental, ecological, economic, and other objectives of participants; appropriately treats risk and uncertainty; and is informed by and responsive to public concerns. Dialogue among interested and affected parties and technical experts should be iterative, begin with the formulation of the problem, and continue throughout the decision process. (Chapter 6)
The multiple objectives of enhanced safety of downstream communities and the protection of the local and regional ecology and economic activities, per the statement of task, need to be integral to the decision-making process. A decision framework is defined herein as a model to guide the systematic processes for making choices in the face of complexity and uncertainty. It assists interested and affected parties in the region in addressing management issues. The “PrOACT” framework (Keeney, 1988) is one such model that includes the following steps: (1) clarify the decision Problem; (2) identify the decision Objectives and ways to measure them; (3) create a diverse set of Alternatives; (4) identify the Consequences; and (5) clarify the Trade-offs. These steps are similar to those in other frameworks.
Although originally intended for use by a single decision maker, the PrOACT framework has been modified for decisions made by multiple decision makers and applied successfully to other water management decisions of significant complexity. The committee recommends the PrOACT construct in this modified form because it is based on an analytical-deliberative process that relies on the results of scientific and engineering investigations and incorporates deliberation with representatives of the broader public throughout the decision process to both influence and be influenced by technical analysis. Second, the decision framework explicitly calls for use of decision analysis techniques to properly account for the multiple objectives and multiple values of interested and affected parties.
Step 1: Clarifying the Decision Problem in a Participatory Setting
A decision problem is defined as that issue or set of issues about which management decisions need to be made. Broadly stated, the decision problem in this case is to determine a long-term solution for managing water and sediment transport in the Spirit Lake and Toutle River system. An overall goal of the recommended decision framework is to search for and identify mutually supportable, effective, and defensible management alternatives. The process requires agreeing on the following elements:
- Who leads the process?
- Who is involved, and what are their roles?
- What types of solutions can be considered?
- What is the geographic scope under consideration?
- What is the time frame being considered for this decision problem?
Early in the decision process, the full range of interested and affected parties needs to be engaged at a depth sufficient for management decisions to be adequately informed by their concerns and values. Agencies may already include other interested and affected groups in community outreach, but their methods of inclusion are in need of reshaping.
Participants in the decision process may include, but are not limited to, agencies with authority or other interests in the area; those who experience the safety, economic, cultural, or life-quality impacts resulting from management decisions; and those with specialized knowledge related to potential management impacts. The number of people participating in focused discussions needs to represent the broad spectrum of interests of the region but also needs to be small enough (i.e., not more than 25 people) so that technical and socioeconomic trade-off discussions can be of sufficient depth to be meaningful and effective. In addition to this group, there must be a neutral support team that includes expertise in the technical and scientific fields of concern, decision analysis, stakeholder engagement, and group facilitation. This team is responsible for implementing the decision framework.
Recommendation: Broaden and deepen the participatory decision-making process from its early stages to include and assimilate the knowledge and interests of affected groups and parties whose safety, livelihoods, and quality of life are affected by management decisions. (Chapter 6)
IDENTIFYING A LEAD
No single agency in the region (e.g., the USFS or the USACE) has unilateral authority to make choices and funding decisions about management
across the system. Outcomes may be perceived as biased if a multiparty process is implemented primarily using the internal resources of any single agency. A framework implementer—a lead—needs to be identified that is responsible for understanding and applying the collaborative analytic decision-making process. This may be the agency with authority over the primary issue at hand, but agreement among interested and affected parties (including agencies with management authorities in the region) regarding the choice of the lead builds trust in the decision process. If the lead or the lead agency lacks the skills needed to provide high-quality and neutral support for decision making, those skills could be recruited externally.
Ideally, the lead would be a new system-level entity or a formal consortium of existing agencies. This would provide a central focus for congressional mandates and appropriations, ensure collaboration across agency and jurisdictional boundaries, and maintain continuous engagement by all interested and affected parties. Such an arrangement would likely require a number of congressional actions.
Recommendation: Create a system-level entity or consortium of agencies to lead a collaborative multiagency multi-jurisdictional effort that can plan, program, create incentives, and seek funding to implement management solutions focused on the entire Spirit Lake and Toutle River system. This effort should also be open and accountable to interested and affected parties involved in management decisions. (Chapter 6)
IDENTIFYING THE GEOGRAPHIC SCOPE
The Spirit Lake and Toutle River system is a physically and socioeconomically dynamic system that changes in response to natural and anthropogenic processes. Adequate long-term risk management of the system depends on recognition of the interconnections and interdependencies among subsystems, including engineered elements. Addressing risk in one part of the system (e.g., associated with sedimentation) can affect risk to other aspects of the system (e.g., associated with aquatic ecology). Responsibilities and concerns among interested and affected parties are
also connected in ways that become clear only with system-level analysis. Responsible agencies and other affected parties, however, tend to focus on their respective responsibilities and interests, on specific locations or features, and over short time frames. This divergence of interests is contrary to sound management. While the post-1980 eruption efforts by the USACE addressed flood mitigation and related sediment control options (e.g., the SRS, levee improvements), these individual solutions to system-wide problems were considered separately and rarely in consideration of other issues affecting the region. This pattern continues today for management of almost all elements of the system by all parties.
Recommendation: Engage in system-wide thinking when making decisions about management objectives, approaches, and alternatives for the Spirit Lake and Toutle River system. Depending on the issues being considered, the system may include the Cowlitz River or extend beyond it. (Chapter 6)
DEVELOPING COMMON UNDERSTANDING OF THE SYSTEM AND MANAGEMENT OPTIONS
Wise system management requires the development of shared knowledge and shared recognition of the visions, values, and objectives of key actors, but views on the nature and causes of problems in the Spirit Lake and Toutle River system diverge among interested and affected parties. Similarly, views on the feasibility and requirements of management alternatives diverge. This is exacerbated by fragmentation of information and expertise held by the various agencies given their respective agency missions. No single organization is responsible for investigating all aspects of the system. Various interested and affected parties sometimes use terminology or discuss concepts without appreciation of the ways others define those terms and concepts. For example, a surface outlet channel for Spirit Lake, which would allow passage of spawning fish (presumably with fish passage around the SRS ensured), has been described by some as representing a more “natural” management alternative for controlling lake levels than does
a tunnel. The hydraulic constraints of such an outlet, however, may require a steep and heavily reinforced spillway—far from “natural”—that may be inimical to fish passage.
Recommendation: Responsible agencies and other interested and affected parties should develop a common understanding of the Spirit Lake and Toutle River system, its features, hazards, and management alternatives. (Chapter 3)
CHOOSING A TIME HORIZON
Identifying a time horizon creates the potential for institutional and social conflict because different planning time frames may require different management strategies. Long time frames may result in avoidance of short-term solutions to immediate problems. They may focus on low-probability but catastrophic seismic or volcanic events that could overwhelm hydraulic infrastructure and make prior planning seem irrelevant. On the other hand, short time horizons may favor management alternatives that resolve existing problems, but they may preclude desirable capital-intensive projects. They may also result in understating the importance of high-consequence, low-probability events. Defining a time frame for risk management decisions is critical and should be explicit. Management time frames need to be reconsidered in light of short- and long-term risk, the finite engineering design life of infrastructure, and unanticipated events or conditions as well as in terms of the financial burdens left to future generations. The time frames need to be revisited during the decision process to determine their appropriateness as new information is gathered.
Recommendation: Alternatives for managing the Spirit Lake and Toutle River system should be judged over both short and long time frames to ensure consideration of the range of the concerns of interested and affected parties. (Chapter 5)
Step 2: Identifying Decision Objectives
Once interested and affected parties are identified, the decision participant group is selected, and the team of experts that provides neutral support is in place, a set of decision-specific objectives can be clarified and structured. Decision objectives are the goals that matter to the participants of the deliberative process when comparing alternatives. They are always phrased as verbs—for example, to maximize economic well-being or to minimize adverse environmental impacts. Objectives become quantitatively defined once metrics are assigned for their measurement.
Objectives of all interested and affected parties need to be identified and the compiled list used as the basis for further deliberations among decision makers. Identifying objectives includes developing a common understanding of the underlying interests of decision participants. For example, an often-stated objective is to restore the “naturalness” of the system, but “naturalness” means different things to different people. Objectives related to such an ill-defined goal could be more specifically placed into such categories as increasing fish passage through the SRS and into Spirit Lake, increasing the “pristineness” of the area, or pursuing management solutions that require little human intervention.
Decision objectives serve as a basis for comparing alternatives. For example, the USACE developed a comprehensive plan in 1983 that compared management alternatives based on flood control (including both the risk of a catastrophic breakout of Spirit Lake and the risks of chronic floods), navigation, water quality, erosion, fish and wildlife, and maintenance of cultural resources. Decision group participants today may want to elaborate on that list to include ecosystem services (e.g., establishing an environmental landscape of the Toutle River), cost (both expected cost to implement and cost risk associated with potential nonperformance), and safety, including operational safety of workers inherent in the different alternatives.
Management alternatives could be studied in ways that allow for a better understanding of the relationships between objectives and alternatives—for example, through use of an objectives hierarchy, which
helps decision participants to better understand the relationships between or among specific goals. An objectives hierarchy is created by deciding which objectives represent the highest-level goals (e.g., minimize adverse impacts to the ecology of the system) and which are intermediate objectives that must be met to obtain the highest-level objectives (e.g., minimize impacts to anadromous fish, minimize impacts to large mammals, minimize impacts to waterfowl). Each of those subgoals can be further broken down into more subgoals, eventually representing the relationships of all the objectives identified by interested and affected parties.
Performance metrics will need to be established to give decision participants a means to quantify a desired objective outcome so that expected progress toward or away from that objective can be modeled. Some metrics may directly measure the consequences of interest in their own terms (e.g., maintenance cost can be measured in dollars). Other metrics are best stated as proxies (correlates) for the consequences of an alternative (e.g., acres or hectares of accessible fish spawning habitat as a proxy for fish abundance). Other objectives may be difficult to quantify directly or indirectly because of unobservable or hard to measure impacts. In such cases, scales may be constructed for the problem at hand, with each level of the scale defined with a succinct and relevant narrative agreed to by the decision makers.
Step 3: Creating a Diverse Set of Alternatives
The third step of the decision process addresses alternatives. The goal is to craft multiple and diverse sets of management alternatives that would address the collaboratively generated list of management objectives. Management alternatives need to be considered as region-wide strategies and in terms of how they affect different elements in the system (e.g., engineered infrastructure, capital works, operations of engineering works, emergency response plans, natural environment, socioeconomic elements) and in terms of all types of change. They reflect the decision objectives identified by the group of decision participants and clearly specify what actions would be needed in different parts of the system for the strategy to be implemented. A skilled facilitator and decision analyst may help decision participants
navigate through the objectives to avoid a stalling of deliberations. This process may use tools such as strategy tables (see Chapter 7) to create mental models for comparing individual actions within a strategy. Interdependent elements of the system need to be identified and linked (e.g., linking alternatives to encourage fish passage to Spirit Lake with actions to enhance passage beyond the SRS) and independent elements (i.e., those that do not affect reaching stated objectives) can be considered separately to simplify analysis. Alternatives that are dependent on system response with time (e.g., alternatives to address sediment buildup behind the SRS spillway with time) need to be adequately described.
Step 4: Identifying Consequences
In the vocabulary of the decision framework, “consequences” are the estimated impacts over time—both good and bad—of the various alternatives as determined using performance metrics. There may not be a common understanding of how physical processes interact with possible management alternatives. A deliberative and participatory approach that involves both interested and affected parties and technical experts in building “if-then” hypotheses and cause-and-effect relationships can aid those participants in generating a common understanding of how management options impact the issues of concern for participants.
Analysis of possible consequences needs to include consideration of the range of uncertainties associated with the management alternative itself and those inherent in the Spirit Lake and Toutle River system. Capturing risk and uncertainty is a technical exercise that may also be value laden. Participants’ attitudes toward risk may be diverse and complicated. Iterative, structured dialogue with interested and affected parties throughout the process allows attitudes toward risk to be captured in a consistent way, understood by all, and incorporated appropriately into the decision process. The resulting deeper insight related to uncertain consequences informs later trade-off discussions when choosing among alternatives.
In order to capture the risk of a breakout of Spirit Lake, for example, more work to aggregate different sources of risk into one overall measure
of risk is needed. This would allow alternatives intended to address other (perhaps, non-flood-related) objectives to be adequately assessed for how they might also affect the risk of a catastrophic breakout. The results would then be translated into a useful measure of “risk of catastrophic flood” to help participants compare options and understand trade-offs.
Comparing alternatives using quantitative performance metrics defined by the decision group is important for understanding the alignment between the consequences of alternatives and the stated objectives. The development of detailed consequence tables (see Chapter 8) based on those metrics helps focus value-laden discussions on key trade-offs and minimizes deliberation over what may be inconsequential technical issues.
Step 5: Clarifying the Trade-Offs
Identifying and closely considering trade-offs (i.e., compromises) is the last step of the decision process. Getting to this step, however, may require an iterative revisiting of previous steps. The overall purpose of the decision process is not to find some objectively defined optimal solution, but rather to find a good solution that is supportable at some level by all the decision participants. In most cases, this support hinges on participants’ awareness and acceptance of various trade-offs. Some anticipated trade-offs could revolve around downstream sedimentation versus a more “natural” drainage system; cost versus catastrophic flood risk; sediment retention versus anadromous fish abundance; fish populations downstream versus fish populations upstream of the SRS; and short-term versus long-term actions and consequences.
A well-implemented decision process should help participants balance competing objectives in searching for a mutually acceptable solution. Complex trade-offs that involve multiple conflicting objectives and multiple alternatives may be addressed with decision analysis techniques that focus consideration on key value trade-offs, perhaps through quantitative ranking and weighting methods.
THE FIRST APPLICATION OF THE DECISION PROCESS: MANAGING SPIRIT LAKE WATER LEVELS
It is likely that the first attempt to apply this decision framework will be related to decisions regarding management of water levels in Spirit Lake. These currently fluctuate seasonally approximately 11 feet (~3.4 m), and it is assumed by regional experts the lake will breach its blockage if water levels rise another 26 feet (~8 m). The repeated need for repairs on the outflow tunnel controlling lake levels has led to a recent, largely non-quantitative potential failure modes analysis (PFMA)—based largely on professional judgment—of management alternatives that were first considered shortly after the 1980 eruption. These include major rehabilitation of the Spirit Lake outlet, the creation of a permanent pumping facility, installation of a buried conduit through the debris blockage, and digging a riverine channel across the debris blockage. Though the committee was neither asked to evaluate the PFMA nor was it provided direct access to the PFMA, committee members nevertheless concluded that there is a substantive knowledge gap regarding practical design issues that needs to be resolved before alternatives can be usefully compared by decision group participants.
As decision participants consider long-term management of Spirit Lake, they may want to consider a broader and bolder set of alternatives. Options for consideration could include, for example, lowering lake levels, draining the lake, installing a second modern drainage tunnel on a different alignment, or constructing a dry spillway as a backup outlet. A second tunnel would allow unconstrained rehabilitation of the existing tunnel and provide redundancy in the control of Spirit Lake and may also open the possibility for more flexible long-term management. A second tunnel might also allow restoration of a more natural drainage through the debris blockage more than a reinforced engineered outflow channel would as currently envisioned. The viability of such options is best quantified through an analytical-deliberative process as outlined in the report.
This page intentionally left blank.