Adaptive management is a formal, systematic, and rigorous program of learning from the outcomes of management actions, accommodating change, and thereby improving management (Holling, 1978; NRC, 2003). It has been recommended as part of the solution to many environmental problems (e.g., NRC, 2004a), and it is quite appropriately an important part of the draft BDCP. Adaptive management was developed in response to the difficulty of predicting the outcome of management alternatives in natural systems, because of the many uncertainties involved. Current models, typically used for formulating restoration plans, often lack predictive power. Adaptive management, at least in theory, provides resource managers with an iterative strategy to deal with uncertainties and use science, with a heavy emphasis on monitoring, for planning, implementation, and assessment of restoration efforts (Williams et al., 2009). The BDCP has correctly recognized the importance of adaptive management in its various conservation measures and its developers should be commended for emphasizing this aspect of the plan.
Despite numerous attempts to develop and implement adaptive environmental management strategies, many of them have not been successful (Gregory et al., 2006; Walters, 2007). Walters (2007) concluded that most of more than 100 adaptive management efforts worldwide have failed primarily because of institutional problems that include lack of resources necessary for expanded monitoring; unwillingness of decision makers to admit and embrace uncertainties in making policy choices; and lack of leadership in implementation. Thus many issues affecting the successful implementation of adaptive management programs are attributable to the context of how they are applied and not necessarily to the approach itself (Gregory et al., 2006). In addition, the aims of adaptive management often conflict with institutional and political preferences for known and predictable outcomes (e.g., Richardson, 2010) and the uncertain and variable nature of natural systems (e.g. Pine et al., 2009). The high cost of adaptive management, and the large number of factors involved also often hinder its application and success (Lee, 1999; NRC, 2003). Thus, adaptive management, although often recommended, is not a silver bullet and it is not easy, quick, or inexpensive to implement.
In addition to the above difficulties, Doremus (forthcoming) has advocated an analysis of conditions to determine whether adaptive management is an appropriate strategy before it is undertaken. This is good advice, and by implication it could be followed as a method of evaluating existing adaptive management programs. Doremus argues that three conditions favor the use of adaptive
management: the existence of information gaps, good prospects for learning at an appropriate time scale compared to management decisions, and opportunities for adjustment. This panel has not performed a formal analysis of the BDCP’s situation in regard to these three conditions, and is not aware of any such analysis, but it does draw some preliminary conclusions. Clearly, the first condition (the presence of information gaps) exists, and the second condition (good prospects for learning) seems likely to exist if the program is designed well. The third condition (opportunities for adjustments) is more problematic. There are pressures for management guarantees; for example, the draft BDCP makes clear that one of its aims is a reliable water supply, and Sagouspe (2010) points out that the Planning Agreement that led to the BDCP provides assurances that “no additional restrictions on the use of land, water, or financial resources” beyond the agreed-on amounts will be required without the agreement of the water users (c.f. Richardson, 2010, cited above). Such agreements on their face seem to reduce opportunities for adjustments, although they do not necessarily preclude them altogether.
All of the above considerations lead as well to a reminder of the need for clear goals, cited in many appraisals of adaptive management (e.g., Milon et al., 1998), and this returns the panel to its earlier concern, namely, that the goals of the BDCP are multiple and not clearly integrated with each other. Despite all of the above challenges, there often is no better option for implementing management regimes, and thus the panel concludes that the use of adaptive management is appropriate for the BDCP.
In light of the above, this panel further concludes that the BDCP needs to address these difficult problems and integrate conservation measures into the adaptive management strategy before there can be confidence in the adaptive management program. In addition, an important step in adaptive management that is often given less attention than the others is the need for a mechanism to incorporate the information gained into management decision-making (e.g., NRC, 2003, 2006, 2008). This matter is critical;it also was raised by the Bay Delta Conservation Plan Independent Science Advisors (draft BDCP, Appendix G) and is discussed further below.
In 2009, the BDCP’s developers engaged a group of Independent Science Advisors to provide expertise on approaches to adaptive management in the BDCP (draft BDCP, Appendix G-3). Their advice has been incorporated into the adaptive management program presented in Section 3.7 of the draft BDCP. The Independent Science Advisors’ report to the BDCP Steering Committee identified key missing elements in the available documentation at the time, including the formal setting of goals based on problems; more effective use of conceptual or simulation models; a properly designed monitoring strategy to evaluate the effectiveness of conservation measures; and more effective assessment, synthesis, and assimilation of information collected during the implementation. Further, their report recommended an adaptive management framework for the BDCP (Bay Delta Conservation Plan Independent Science Advisors’ Report on Adaptive Management,2009, Figure 1, p. 3). The panel concludes that the Inde-
pendent Science Advisors have provided a logical framework and guidance for the development and implementation of an appropriate adaptive management program for the BDCP.
Much of the information on the adaptive management program is contained in Section 3.7 of the draft BDCP. A brief description of the management of the adaptive management program is presented in Section 7.35. Identification of uncertainties, a critical step in any adaptive management program, is discussed under each of the Conservation Measures (Section 3.4) and adaptive management considerations are shown in Table 3-20, which is part of Section 3.6, Monitoring and Research Program. Because the details of the adaptive management program are fragmented and occur throughout the BDCP without clear linkages of critical components in one section of the document, it is difficult to obtain an overall assessment of the promise of the adaptive management program. The information is not sufficient to demonstrate that the adaptive management plan is properly designed and follows the guidelines provided by the Independent Science Advisors.
Although the adaptive management framework provided by the Independent Science Advisors recommended a logical, stepwise approach for flow of information (Bay Delta Conservation Plan Independent Science Advisors’ Report on Adaptive Management, 2009, Figure 1, p. 3), the adaptive management framework shown in Figure 3-63 of the BDCP (also shown in Appendix E of this report) is significantly different and is missing some key elements. It is not clear how the monitoring and “targeted research” programs were designed using goals and objectives, desired outcomes, and performance metrics to select and evaluate steps outlined in the Independent Science Advisors’ report. More important, clearly defined uncertainties at various scales starting with the ecosystem level are not presented adequately in the BDCP. In particular, the role of models is not clearly identified in the adaptive management framework, except in Figure 3-63. Box 5b of that figure simply suggests a refinement of models without identifying them. Also, the BDCP does not make clear whether adaptive management applies to broad, ecosystem goals or narrower goals related to specific natural communities or specific conservation measures, or both. Without this distinction and a clear discussion of the role of adaptive management at the ecosystem level, the draft BDCP does not provide assurance that it will successfully use adaptive management to make adjustments during the planning, design, and operational stages of the project.
The Independent Science Advisors correctly pointed out the need for an emphasis on when and where the active versus passive approaches should be used during the design phase. A passive approach is used when the projects are irreversible in nature, as in the case of a dual conveyance facility whereas an active approach involves experiments to test competing hypotheses in cases of significant uncertainties in ecosystem response. The BDCP lacks details of the types of adaptive management approaches and the specifics of the experimental testing that would be conducted to reduce uncertainties. Passive adaptive management is used when there is a high confidence regarding the anticipated eco-
system response, often predicted by reliable models. However, the BDCP does not explicitly rationalize the particular selections in the adaptive management framework, for example, with regard to proposed creation of wetlands, levee restoration, and conveyance options.
The lack of detail about the adaptive management program’s details makes evaluating it difficult. Many details of adaptive management are needed to perform a thoughtful review of it, and in some cases, those details emerge only as the plan is implemented. For these reasons, the panel is unable to provide a detailed review of the adaptive management plan at this stage. However, some comments and suggestions are in order.
First, as mentioned above, an adaptive management program requires clear goals. This point often is overlooked. If the project’s management goals are not clear, then it will not be evident how to adapt management in the face of new information. The BDCP does not explain how its multiple goals are to be integrated, but the problem goes deeper: some agreed-on goals, such as sustainability of the ecosystem or having a healthy ecosystem, may no longer be acceptable to all parties when they become more specific or when it becomes clear that not all aspects of the ecosystem can be rehabilitated simultaneously. This problem is not unique to the Delta: it affects other large restoration efforts as well, for example, the Everglades (e.g., Milon et al., 1998; NRC, 2010).
Second, adaptive management requires a monitoring program to be in place. The draft BDCP describes its monitoring plan in considerable detail: Table 3-20, which describes the monitoring for effectiveness of conservation actions, runs more than 80 pages, implying a large amount of monitoring activity. However, because there is no effects analysis, it is difficult to evaluate the scientific basis or to justify the appropriateness of individual elements of the monitoring program, elements which clearly should be tied to the results of the effects analysis. In addition, the panel questions the availability of resources necessary to accomplish the all monitoring described in Table 3-20, especially because additional baseline, compliance, and other monitoring also are described in the BDCP as being necessary.
Third, although all of the elements of an adaptive management program are present in the draft BDCP, some of them are not described in detail and some do not appear to be incorporated into the framework in Figure 3-63 (shown in Appendix E of this report). The panel emphasizes again how important it is for a meaningful adaptive management program to be tied to the results of the effects analysis, or at least related to the same issues being addressed by the effects analysis. If it is not, then it is difficult to see how the monitoring and adaptive management program can inform the implementation of the plan and inform decision makers.
The draft BDCP appropriately cites the Independent Science Advisors’ Report on Adaptive Management conclusion that:
“the weakest aspect of most adaptive management plans is in the sequence of steps required to link the knowledge gained from implementation monitoring and research and other sources to decisions about whether to continue, modify, or stop actions, refine objectives, or alter monitoring” (draft BDCP, p. 3-577).
This issue has been addressed by NRC reports on the Everglades restoration (e.g., NRC, 2006, 2008), and it is taken seriously by the Comprehensive Everglades Restoration Program. The panel recognizes the difficulty of understanding from the outside how decisions actually are made, and those elements of the BDCP’s adaptive management program that require publication of scientific results and provision of the resulting scientific advice to program managers are a good step in that direction. However, a clearer description of the mechanisms that will enable the scientific results to inform management decisions would be helpful.
Details of two other aspects of adaptive management, stakeholder engagement and interagency coordination, are vague. The way that agencies coordinate their activities and that stakeholders participate in the process can have significant consequences. For example, Linkov and his colleagues (Linkov et al., 2006a,b) have described the use of multicriteria decision analysis to enhance adaptive management, and the NRC (2004b) has provided worked examples of such an approach applied to restoring Atlantic salmon in Maine. Those approaches all depend on input from stakeholders. The concepts of a stakeholder committee to receive public input and a “Decision Body” to adjust water operations are too vague and their functions appear to be too limited to provide guidance. The panel recommends that the BDCP take advantage of the literature on this topic—beginning, but not ending, with the material cited above—to inform its processes.
Finally, the importance of action-related triggers related to environmental conditions or the status of covered species is briefly mentioned in the draft BDCP (draft BDCP, Section 3.7.4, pp. 3-586-3-587), but there is no discussion of their importance and role in the adaptive management program and their relation to the effects analysis.
The essence of adaptive management is to identify major uncertainties about the efficacy of policy actions, then to design field tests or management experiments to directly measure efficacy. Such tests can include field evaluation of alternative feedback decision rules that do or do not include thresholds or triggers for action. Initial adaptive management modeling exercises may screen out policies that require triggers by illustrating the challenges associated with uncertainty about the best triggering conditions. In some cases, however, triggers for action can and have been used, often in conjunction with multiobjective structured decision analysis that includes the values and alternatives
preferences of the various stakeholders involved (e.g., Karl et al., 2007; Kiker et al., 2008; Miller et al., 2010).
One such example is a recent effort on the Colorado River, where managers are seeking to establish flow releases to control non-native fish below Glen Canyon Dam9 (Runge et al., 2011). Through the decision-analysis process, objectives were identified (e.g., manage resources to protect tribal sacred sites and spiritual values, maintain and promote local economies and public services, operate within the authority, capabilities, and legal responsibility of the Bureau of Reclamation). In addition, management strategies were evaluated against the objectives, and tradeoffs between strategies were considered. The process identifies specific triggers (e.g. following High-Flow experimental floods, abundance of native or introduced fish species, flow and sediment load) for management actions (e.g., removal of non-native species, fine sediment slurry, release of stranding flows), while other actions (e.g., mechanical or chemical disruption of fish spawning areas, augmentation of fine sediment) are recommended without triggers. The value of triggers is in the efficiency of managing the system, minimizing expensive actions to when and where they are thought to be necessary for and beneficial to species recovery. Such triggers also would help to design a more-focused monitoring program. However, the challenge of using triggers is in the uncertainty in establishing thresholds for triggering actions. Thus, (Runge et al., 2011) caution that their results do not provide the final decision but instead provide guidance for further consultation by the decision makers. That consultation is likely to require experimentation, modeling, and continued adaptive management.
In summary, the BDCP’s adaptive management program is not fully developed. In addition, there remain significant scientific, policy, and management uncertainties about the BDCP’s purpose and organization. The panel concludes that the BDCP’s developers can benefit significantly from experiences in adaptive management attempted in other large-scale ecosystem restoration efforts. One such example is the CERP, where adaptive management has been a key component since its inception in 1999 (USACE & SFWMD, 1999). As recognized by the NRC (2006), the CERP adaptive management strategy provides a sound organizational model for the execution of a passive approach. More recent activities also include examples of active approaches where field tests have played a major role in the early phases of selected projects (RECOVER, 2010). Key components of the CERP adaptive management program are:
• CERP Adaptive Management Strategy (RECOVER, 2006a);
• Monitoring and Assessment Plan and an Assessment Strategy designed to monitor system-wide responses to determine how well CERP is achieving its goals (RECOVER, 2004; 2006a,b; 2009); and
9 The panel provides this example as a good use of action‐related triggers. The success of adaptive management in Glen Canyon in general has been questioned (Susskind et al., 2010).
• CERP Adaptive Management Integration Guide (available in draft form) (RECOVER, 2010).
The above documents detail more than five years of progress in implementing adaptive management in the CERP. The CERP’s program includes nine activities, which have been effectively integrated into the standard practice of project planning and life-cycle analysis (NRC, 2006). The integration guide describes how to apply adaptive management concepts to the CERP program and related projects through the identification of key uncertainties and the incorporation of activities into the existing CERP planning and implementation process. Even a soundly implemented adaptive management program is not a guarantee of a successful restoration effort, however. As described in several NRC reports and other documents, several factors outside the purview of the adaptive-management teams and even the program managers have hindered restoration progress in the Everglades. They include financial, political, bureaucratic, legal and other obstacles (e.g., NRC 2006, 2008, 2010), factors certain to influence the implementation of the BDCP as well. But a well-designed and implemented program should improve the likelihood of success in implementing the BDCP.