The panel concludes that the draft BDCP is missing critical elements, including an effects analysis, a description of how and where scientific information was used in the draft BDCP, and a description of the BDCP’s relationship to other ongoing efforts. In addition, the draft has several structural or systematic problems, including lack of clarity as to the purpose of the BDCP; an unclear linkage of various parts of the BDCP to the effects analysis4 and among its other components; and lack of detail about analyses of various future scenarios, including a lack of analyses of tradeoffs among the BDCP’s goals in various scenarios. The panel offers some guidance on how these systematic problems might be addressed and how the draft BDCP might be completed more usefully.
At the outset of its review the panel identified a problem with the geographical and hydrologic scope of the draft BDCP. The BDCP aims to address management and restoration of the San Francisco Bay Delta Estuary, an estuary that extends from the Central Valley to the mouth of San Francisco Bay. Thus, given that the BDCP purports to describe a Bay Delta Conservation Plan, the omission of analyses of the effects of the BDCP efforts on San Francisco Bay (aside from Suisun Bay) is notable. This omission should be of concern to all BDCP parties because the Bay-Delta system is an estuary, and there are significant physical, biogeochemical, and ecological connections between the various sub-embayments as well as between the Bay-Delta and the Pacific Ocean (e.g., Cloern et al., 2010). In particular, changes in outflows and in the tidal prism associated with changing water-project operations and restoration actions would be expected to cause changes in San Francisco Bay, and not only in the Delta. A plan intended to be comprehensive should incorporate these fundamental features of the system. Although the statutory basis of the BDCP may argue against consideration of the effects outside the statutory Delta, the BDCP’s failure to address issues related to San Francisco Bay is a significant flaw that should be corrected in subsequent versions of the plan.
4 Even though the effects analysis is not yet complete, the BDCP’s authors should at least be able to describe how the completed parts of the BDCP will be linked to the effects analysis.
The draft BDCP describes an effects analysis as:
“the principal component of a habitat conservation plan [HCP]. . . . The analysis includes the effects of the proposed project on covered species, including federally and state listed species, and other sensitive species potentially affected by the proposed project. The effects analysis is a systematic, scientific look at the potential impacts of a proposed project on these species and how these species would benefit from conservation actions” (draft BDCP p. 5‐2).
Clearly, such an effects analysis, which is in preparation, is intended to be the basis for the choice and details of those conservation actions. Its absence in the BDCP, therefore, is critical gap in the scope of the science and the conservation actions. Nevertheless, the panel presents its vision of the structure and content of a useful effects analysis.
The above description of the effects analysis to be included in the BDCP is rather narrowly cast, because it focuses on the BDCP as a habitat conservation plan (HCP), that is, as an application for an incidental take permit. It thus presupposes the choice of the project to be permitted. By contrast, a broadly focused conservation strategy, which the draft BDCP also says it is5, requires a similarly broadly focused, comprehensive effects analysis. Such an effects analysis would include a systematic analysis of the factors affecting species and ecosystems of concern and the likely contribution of human-caused changes in the system. Such an analysis would then lead to the informed choice of options for reversing the decline of the ecosystem and its components, rather than only analyzing a pre-chosen option. What would such an effects analysis look like?
Effects analyses are used in a range of disciplines to understand complex systems. As noted in the quote above, their main attribute is that they are systematic scientific analysis. Their precise form is not critical. For example, failure mode and effects analysis (FMEA) is commonly applied in the automotive, aerospace, and software industries to understand whether and how the failure of individual components impact the reliability of the overall system (Gilchrest, 1993; McDermott et al., 2009). In the environmental field, effects analyses are used to understand and compare likely responses to alternative management schemes (e.g., Marcot et al., 2001). The National Research Council has reviewed the application of effects analysis within the environmental arena (NRC, 2009). In addition, several NRC reports have discussed or applied the techniques of effects analysis even though they were not necessarily called “effect
5 The following statement appears on p. 1‐1 of the draft BDCP: “The [BDCP] sets out a comprehensive conservation strategy for the Delta designed to advance the co‐equal planning goals of restoring ecological functions of the Delta and improving water supply reliability to large portions of the state of California.”
analysis” (e.g., NRC, 1995, 2002, 2004a, 2004b, 2005; Appendix E of this report provides an example of an effects analysis from NRC 2004b). Effects analyses are commonly used because they integrate empirical data and expert opinion to guide management decisions (e.g., NRC, 2004b). The analytical approaches used in the different types of effects analyses vary from classical risk priority numbers, to simulation modeling (e.g., Legault, 2005), to complex Bayesian network models (Ellison, 1996; Uusitalo, 2007). However, certain important elements are common to all of these analyses, including the need to describe how individual components in the system are connected. It is an effects analysis of this scope that the panel envisions for the BDCP. Here, the panel provides guidance regarding the structure and essential elements that it would expect to see in the completed effects analysis for the draft BDCP. The panel draws on a recent paper by Murphy and Weiland (2011) for a description of a useful effects analysis, itself based to some degree on NRC (2009), because it sets forth specifics for an effects analysis that would be appropriate for the Delta. The panel agrees with Murphy’s and Weiland’s general approach.
An effects analysis is an essential element of the final BDCP, because it will help meet the legal requirement for a habitat conservation plan to evaluate whether the preferred action aids in the recovery of the species (state requirement) and does not appreciably reduce the likelihood or the survival and recovery of the listed species in the wild (federal requirement). These requirements are initially triggered because as an HCP/NCCP (natural communities conservation plan), the BDCP deals with listed species. However, even if this were not the case, an effects analysis provides the framework within which the impacts of alternative management options can be compared and thus could be justified from a purely logical point of view. An effects analysis is further justified because it also may inform the adaptive management process by identifying which components or processes are the most sensitive indicators of the status and structure of the ecosystem (McCann et al., 2006).
Once the goal of the effects analysis has been defined, the first element of any effects analysis must be an integrated description of the components of the system and how they relate to one another. This description should include a clear statement of the alternative management actions proposed, including that of no action. The activities in this first section naturally lead to a clear definition of the management goal and the temporal and spatial domain of the impacted area. At this introductory level, it is not necessary to quantify the relationships. One needs to mainly indicate the connections. Such a description is essential for several reasons. Most important, it formalizes the understanding of the connections among processes and components in the system. It defines which processes and components are expected to respond to any perturbation and which ones will not. Secondarily, in formulating the problem, a conceptual diagram can serve to identify and rank in importance data on different processes and components within the system. Finally, the system description provides a broader context into which information on the status and trends of species covered by state and federal statues can be placed―such that the dependencies of
these listed species on processes and components of the system are identified.
The second stage of the effects analysis should be the collection, review and critical assessment of the best relevant scientific information available. The determination of which data need to be assembled is guided largely by the conceptual framework identified in the first stage. It is neither necessary nor helpful for the assembled data to be encyclopedic in coverage. However, it is essential that data on those processes and components identified in the first stage are compiled, assessed and summarized. This information may be in the form of empirical data or in instances where data are unavailable, in the form of expert opinion. Expert and stakeholder opinion has been successfully used in several management questions involving water use or fish stocks (Borsuk et al., 2001; Miller et al., 2010). The objectives of the data assembly phase are to clearly describe the baseline or reference condition6 and to quantify the expected relationships among system processes and components. An important feature of this stage is the need to include information on the uncertainties around estimates of processes or component levels. Additionally, the spatial and temporal scale of processes and components under consideration are a vital concern. Different processes and components likely respond at characteristic spatial and temporal scales. For example, the response of many chemical or physical variables might scale with the residence time of water in the system, whereas the response of biological variables might scale with the generation time of the organisms involved. Similarly, salinity gradients affect much of the central and western Delta, while some organisms like salmon, which spend a portion of their life cycles in sea water, occupy much of the North Pacific as well as the Delta and its tributaries. Within the biological realm, rates of primary production, nutrient and oxygen cycling, as well as microbial growth may respond rapidly to ecosystem conditions whereas the abundance of long-lived animals such as sturgeon is expected to integrate ecosystem dynamics over extended periods. The Comprehensive Everglades Restoration Plan (CERP) provides a good example of the use of measurable outcomes for these purposes (NRC, 2008, 2010c).
The next stage of the effects analysis is the most challenging–that of representing the dynamic response of the system. For simple systems, this may be in the form of a simple model. For example, decisions regarding quota levels in fisheries management are often made with guidance from a single assessment model, albeit one with hundreds of parameters (Miller et al., 2010). However, even in simple systems, the level of uncertainty present in individual processes and components of the system may be of such magnitude that state-variable models are unreliable. In these cases probabilistic models have been developed (Legault, 2005). More recently, Bayesian approaches have been used to guide management in the face of uncertainty for complex environmental questions
6 Large restoration programs usually include methods for assessing their effects so that adaptive management can occur. The basic prerequisite for such assessments is the establishment and characterization of a reference condition against which future conditions and proposed alternatives can be compared.
(Borsuk et al., 2004; McCann et al., 2006; Rieman et al., 2001). For an example of incorporating uncertainty into management options, see Box 1.
In the case of the BDCP, it is unlikely that a single analytical framework, even one as flexible as Bayesian network analysis, will be adequate. Thus, it is likely that multiple models will be used to assess the response of different system components to each management alternative. Ultimately a range of integrated scenarios should be developed that link the models’ outputs to an integrated response. It is particularly important that each set of the models and analyses be clearly related back to the original conceptual framework generated in the first stage of the effects analysis. Analysts should be explicit about the model inputs and assumptions for each stage of the process. One of the risks of this approach is error propagation, that is, that uncertainty inherent in the forecasts made for one component are not fully carried forward to models of other components.
It would be highly advantageous if outcomes in the effects analysis were quantifiable empirically and could thus become components of the BDCP’s Monitoring and Evaluation Program (e.g., NRC, 2000, 2008; Orians and Policansky 2009). As noted above, the CERP has considered and described these issues in considerable detail (NRC, 2008and references therein). This information, when gathered in the BDCP’s Monitoring and Evaluation Program, could then be used to conduct statistical analyses and calibrate models and the modeling framework to inform the adaptive management phase over the decades following implementation of the BDCP actions.
The 2008 Federal Columbia River Power System Biological Opinion
A suitable example of an attempt to incorporate uncertainty is evidenced in the 2008 Federal Columbia River Power System (FCRPS) Biological Opinion (NOAA, 2008) and in the 2010 Supplemental FCRPS Biological Opinion (NOAA, 2010) prepared after the 2008 opinion was voluntarily remanded. The com- prehensive analysis in this biological opinion focused on determining the effects of different dam operation alternatives, on key ESA-listed anadromous salmonid populations in the Columbia River Basin. In that analysis, water delivery and dam operation models create conditions that route juvenile salmon through different routes at eight dams in the FCRPS, resulting in net smolt survival downstream of the last dam (Bonneville). Changes in smolt system survival associated with different operation-alternatives are then linked to a broader life-cycle analysis to assess the potential for population level responses to selected management actions.
During the meeting on December 8, 2010, in San Francisco, presenters indicated that the effects analysis that will be included in the BDCP will be only a first step, that is, that it would be iteratively updated as empirical data from the operation of the approved alternatives become available. This approach is certainly compatible with the use of the effects analysis framework as the foundation of the adaptive management framework. If this is indeed how the BDCP developers intend to use the effects analysis, the panel recommends that the final version of the plan articulate a clear vision of how the effects analysis will be updated and how these results will be used to generate the ranges that will be the foundation for subsequent adaptive management.
As an example, much of the recent discussion of changes in the Delta ecosystem has focused on declining planktonic primary production in the Delta and Northern San Francisco Bay (Jassby et al., 2002) as driving food-web changes, notably declines in planktonic grazers (secondary producers), that may underlie to some extent the decline of pelagic fish species like delta and longfin smelt (Baxter et al., 2008). Accordingly, significant elements of the BDCP involve efforts to enhance primary and secondary production through creation of additional tidal wetlands mostly around the edges of the Delta, a plan that strongly echoes CALFED’s earlier focus on the creation of shallow water habitat (c.f. Brown, 2003). The bases for this strategy are twofold: (1) in the face of light limitations, shallow water habitats for which the photic zone is a greater fraction of the water column should have higher rates of primary production than deeper waters, e.g., channels (Cloern, 2007); and (2) empirically it is observed that the periodically flooded shallow waters of the Yolo Bypass can support high rates of export of phytoplankton biomass (Schemel et al.,2004).
However, if an effects analysis is indeed “the principal component of a habitat conservation plan” (draft BDCP p. 5-2), then it is difficult to see how these and other conservation strategies described in the BDCP can be scientifically justified before the effects analysis is completed.
The legal framework underlying the BDCP is extraordinarily complex. In attempting to comply with all relevant laws and regulations, the BDCP’s authors have undertaken to develop a habitat conservation plan of great importance, scope, and difficulty. The panel recognizes that the authors face significant challenges and that the BDCP is a work in progress. With these caveats in mind, the panel observes that it would be helpful for the draft BDCP to clarify and place into context a number of legal issues, because their nature and interpretation are closely tied to the BDCP’s scientific elements. Any lack of legal clarity makes it difficult for the panel and the public to properly understand, interpret, and review the science of the BDCP.
Ambiguous Role of Co-Equal Goals and Their Relationship to the BDCP
According to the draft BDCP (p. 1-8), it:
“has been prepared as a joint [habitat conservation plan] HCP/ [Natural Communities Conservation Plan] NCCP, which will support the issuance of incidental take authorizations from the US [Fish and Wildlife Service] FWS and [National Marine Fisheries Service] NMFS pursuant to Section 10 of the [federal Endangered Species Act] ESA and take authorizations from the California Department of Fish and Game (DFG) under Section 2835 of the [Natural Communities Conservation Planning Act] NCCPA to the non‐federal applicants. The BDCP has also been designed to meet the standards of Section 2081 of the California Endangered Species Act (CESA). The BDCP will further provide the basis for biological assessments (BA) to support the issuance of incidental take authorizations from USFWS and NMFS to [the Bureau of] Reclamation pursuant to Section 7 of the ESA, for its actions in the Delta.”
Thus, the BDCP is clearly and specifically an application for the incidental take of listed species as set forth in federal and state statutes.
To apply for an exemption from the § 9 “take”7 prohibition of the federal Endangered Species Act (ESA), the water users must submit a habitat conservation plan (here, the BDCP) that will minimize and mitigate the harmful impacts of their water usage. HCPs prepared as part of an application for an incidental take permit under federal law are not required to help listed species recover, but they must demonstrate that“the taking will not appreciably reduce the likelihood of the survival and recovery of the species in the wild” (ESA § 10).8 Under state law, the water users must submit a Natural Community Conservation Plan (NCCP) that, among other things, “aids in the recovery of the species.” (Natural Communities Conservation Planning Act [NCCPA], Cal. Fish and Game Code §§ 2800-2835). Neither the ESA nor the NCCPA specifically requires applicants to advance the “co-equal goals.”
Despite this, the first paragraph of the draft BDCP (p. 1-1) states that it “sets out a comprehensive conservation strategy for the Delta designed to ad-
7Take means “to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect, or to attempt to engage in any such conduct.” ESA, Section 3, 16 U.S.C.1532. Harm, within the statutory definition of “take” has been further defined by regulation: “Harm in the definition of take in the Act means an act which actually kills or injures wildlife. Such act may include significant habitat modification or degradation where it actually kills or injures wildlife by significantly impairing essential behavioral patterns, including breeding, feeding, or sheltering.” 50 C.F.R.17.3.
8 ESA § 10 also requires successful applicants to demonstrate that (1) “ the taking will be incidental [to an otherwise lawful activity],” (2) “the applicant will, to the maximum extent practicable, minimize and mitigate the impacts of such taking,”(3) “the applicant will ensure that adequate funding for the plan will be provided,”and (4) “[such other measures that the Secretary may require as being necessary or appropriate for purposes of the plan] will be met.” 16 USC § 1539(a)(2)(B).
vance the co-equal planning goals of restoring ecological functions of the Delta and improving water supply reliability to large portions of the state of California.” This and similar statements throughout the plan make it difficult to understand and evaluate the purposes of HCPs and NCCPs, and the methods of implementing them. Moreover, the methods of implementation are considerably different from the purposes and methods for achieving the two co-equal goals specified in California statutes. Indeed, California has begun to develop a broader “Delta Plan” in accordance with a recent state statute (Cal. Water Code §§ 85300-85309). Thus, the question arises as to the degree of importance to the BDCP of its purpose as an HCP/NCCP and of its purpose as a broader conservation plan designed to achieve California’s two co-equal goals. The BDCP and the Delta Plan address the same ecosystem and are somewhat overlapping, but their goals and legal requirements are not identical. Unless the BDCP’s relationship to the Delta Plan is clearly described, and its purposes clearly delineated, it will be difficult to assess the BDCP’s underlying scientific basis, because the purposes of a broad conservation plan like the Delta Plan are not necessarily the same as those of a habitat conservation plan.
The body of the BDCP contains some elements of both purposes, but not in a coherent and consistent way. For example, despite the statement that achieving the two co-equal goals is one of its purposes, the BDCP focuses on one of the goals at the expense of the other. Additional sources of the confusion are multiple, but two stand out. First, the BDCP document lists some eight planning goals of which providing a “basis for permits necessary to lawfully take covered species” is only one of these eight goals (draft BDCP, p. 1-6). Yet, the remainder of the BDCP appears to focus disproportionately on this goal. As such, much of the BDCP appears to be a post-hoc rationalization of the water supply elements contained in the BDCP.
A consequence of the lack of clarity is related to this post-hoc rationalization. To the extent that the BDCP is simply a request for an incidental take permit then the water users would first identify their desired action (such as construction of a specifically configured “alternative conveyance”), and then analyze its impacts and to develop measures to minimize and mitigate adverse effects. However, to the extent that the BDCP seeks incorporation into the broader Delta Plan, then an effects analysis would precede the choice of all conservation and alternative-operation options, and only then would an effects analysis of those options be performed. That is, if the proposed conveyance system and other measures such as wetlands restoration have been developed as measures to further the restoration of the Delta ecosystem, then one would expect that the effects analysis would be completed before coming to a conclusion as to the preferred type of water delivery system. The absence of an effects analysis and of consideration of water supply alternatives (other than the 45 mile tunnel or possibly an open canal; see section below on alternatives) suggests that the BDCP’s major purpose is to provide the basis for an application for an incidental take permit. Yet, this is contrary to what is stated throughout the plan with respect to the attainment of co-equal goals.
Despite these ambiguities, the draft BDCP has concluded that an “isolated conveyance facility” should be constructed consisting of a 45-mile tunnel or pipeline, capable of conveying 15,000 cubic feet per second (cfs) of Sacramento River water around the Delta to the south Delta’s existing water export pumping plants, to allow for“dual operation” with the existing south Delta diversion facilities (draft BDCP, Chapter 18.104.22.168.1 and Table 4-1). (Again, the “note to reviewers” on p. 4-14 of the draft BDCP suggests that the conveyance system might be a canal, but there is no analysis of a canal in the draft BDCP or even a statement as to whether the findings from the analysis of a canal would differ from the analysis of a tunnel system.)
To support the issuance of an ESA § 10 take permit, the BDCP must specify “what alternative actions to such taking the applicants considered and the reasons why such alternatives are not being utilized” (ESA § 10, 16 U.S.C. § 1539(a)(2)(A)). Even if the proposed action has been decided on, an analysis of alternatives is still required. This analysis does not appear prominently in the draft BDCP. Not only is the analysis a legal requirement, but it also is important scientifically, because to the degree that the reasons for not utilizing the alternatives are scientific reasons, the absence of the analysis hinders the ability to evaluate the BDCP’s use of science. If the BDCP also seeks incorporation into the Delta Plan (and thereby qualifying for state funding of public benefits), then it should also include an analysis of “conveyance” alternatives. As a prerequisite to incorporation, the BDCP must undertake “a comprehensive review and analysis of . . . [a] reasonable range of Delta conveyance alternatives, including through-Delta, dual conveyance, and isolated conveyance alternatives and including further capacity and design options of a lined canal, an unlined canal, and pipelines” (Cal. Water Code, § 85320). Finally, the federal approval process also will require an environmental impact statement that considers alternatives to the “proposed action,” which includes construction of the alternative conveyance (National Environmental Policy Act, 42 U.S.C. § 4332(2)(C)(iii)). Once again, this legally required analysis of alternatives is scientifically important. Therefore, to permit a complete scientific evaluation of the BDCP, it should include an analysis of such alternatives to “take”and to the construction and design of the contemplated isolated conveyance.