2
The Restoration Plan’s Adaptive Management Strategy

ADAPTIVE ASSESSMENT

The adaptive framework for implementation of the Restoration Plan has been referred to in project documents prior to 2002 as “adaptive assessment”. Recently, there was a decision to replace this term with “adaptive management” (Appelbaum, 2002), a more commonly-used term. However, this change in terminology does not imply a change in strategy. The following discussion of the strategy includes quotations from project documents that used the “adaptive assessment” terminology. It also provides a brief comparison of “active adaptive management” to “passive adaptive management” or “adaptive assessment” to clarify the type of adaptive strategy that the Restoration Plan will attempt to use.

“Adaptive assessment is a process for evaluating how well the phases of the Comprehensive Plan achieve their expected objectives, and for using these evaluations as a basis for refining future phases of the program. To be successful, an adaptive assessment process requires that the Comprehensive Plan be implemented iteratively, that a pre-determined set of targets be appropriately monitored, that it be possible to make changes in the design and sequencing of the plan in response to information learned from the monitoring program and from new research and modeling, and that a specific protocol for conducting the adaptive assessment process be in place throughout the life of the program.” (USACE and SFWMD, 1999).

The adaptive assessment strategy and monitoring principles for the Restoration Plan are outlined in the Recommended Comprehensive Plan (USACE and SFWMD, 1999) and in a draft white paper dated November 14, 2001 (AAT, 2001). The draft Monitoring and Assessment Plan dated March 29, 2001 describes the performance measures1 and parameters that will

1  

These are measures specifically chosen by the Restoration Plan to assess ecosystem performance during and after restoration, as opposed to the more general term “indicators.”



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 19
2 The Restoration Plan’s Adaptive Management Strategy ADAPTIVE ASSESSMENT The adaptive framework for implementation of the Restoration Plan has been referred to in project documents prior to 2002 as “adaptive assessment”. Recently, there was a decision to replace this term with “adaptive management” (Appelbaum, 2002), a more commonly-used term. However, this change in terminology does not imply a change in strategy. The following discussion of the strategy includes quotations from project documents that used the “adaptive assessment” terminology. It also provides a brief comparison of “active adaptive management” to “passive adaptive management” or “adaptive assessment” to clarify the type of adaptive strategy that the Restoration Plan will attempt to use. “Adaptive assessment is a process for evaluating how well the phases of the Comprehensive Plan achieve their expected objectives, and for using these evaluations as a basis for refining future phases of the program. To be successful, an adaptive assessment process requires that the Comprehensive Plan be implemented iteratively, that a pre-determined set of targets be appropriately monitored, that it be possible to make changes in the design and sequencing of the plan in response to information learned from the monitoring program and from new research and modeling, and that a specific protocol for conducting the adaptive assessment process be in place throughout the life of the program.” (USACE and SFWMD, 1999). The adaptive assessment strategy and monitoring principles for the Restoration Plan are outlined in the Recommended Comprehensive Plan (USACE and SFWMD, 1999) and in a draft white paper dated November 14, 2001 (AAT, 2001). The draft Monitoring and Assessment Plan dated March 29, 2001 describes the performance measures1 and parameters that will 1   These are measures specifically chosen by the Restoration Plan to assess ecosystem performance during and after restoration, as opposed to the more general term “indicators.”

OCR for page 19
inform the Restoration Plan adaptive assessment process. These documents, together with miscellaneous reports and several meetings between the CROGEE and Restoration Plan personnel, form the basis for the discussion that follows. Adaptive management is a general concept that could refer to a broad range of approaches to achieving ecosystem restoration. However, the minimal elements of any truly adaptive management scheme include (1) clear restoration goals and expectations, (2) a sound conceptualization of the system, (3) an effective process for learning from future management actions, and (4) explicit feedback mechanisms for refining and improving management based on the learning process2. The extent to which the Restoration Plan will meet the restoration goals and expectations rests in large part on a well-designed framework for creating and supporting these four elements. After a brief comparison of active and passive adaptive management, the Restoration Plan adaptive assessment strategy is examined from these perspectives. Overall, the conceptual planning for the Restoration Plan and the Restoration Coordination, and Verification (RECOVER) process are well grounded in the theory and practice of adaptive management. Likewise, current scientific theory and information, for the most part, have been well applied in formulating a strategy for the Restoration Plan Monitoring and Assessment Plan. Nevertheless, in moving towards implementation, there are some specific actions that can be taken to strengthen the monitoring and assessment program with respect to all four elements of the adaptive assessment process, especially with respect to feedbacks between the monitoring information and decisions concerning the implementation of the Restoration Plan. TYPES OF ADAPTIVE MANAGEMENT Walters and Holling (1990) defined three general ways to structure adaptive management: (1) trial-and-error, (2) active adaptive management, and (3) passive adaptive management. According to these authors, the trial-and-error or evolutionary approach (also referred to as disjointed incrementalism by Linblom, 1968) involves haphazard choices early in system management while later choices are made from the subset of choices yielding more desirable results. Active adaptive management strategies use the available data and key interrelationships to construct a range of alternative response models (scenarios) that are used to predict short-term and long-term responses based on small- to large-scale “experiments.” The combined results of scenario development and experimentation are used by policymakers to choose among alternative management options to identify the best management strategies. Passive adaptive management is based on historical information that is used to construct a “best guess” conceptual model of the system. The management choices are based on the conceptual model with the assumption that this model is a reliable reflection of the way that the system will respond. Passive adaptive management is based on only one model of the system and monitors and adjusts, while in active adaptive management a variety of alternative hypotheses are proposed, examined experimentally, and the results applied to management decisions. The restoration strategy outlined in the Restudy abandons the idea of large-scale, management experiments with controls and replicates, opting instead for incremental implementation in which “each incremental step in the plan is viewed as an experiment 2   Successful application of an adaptive management framework requires more than these four elements (e.g., collaborative working relationships, trust, a champion). These elements assure that the basis for adaptive management has been established; they are thus necessary but not sufficient conditions (Holling, 1978; Walters and Holling, 1990).

OCR for page 19
accompanied by one or more hypotheses that predict how that step will improve the system” (USACE and SFWMD, 1999). Pre- and post-implementation monitoring will be used to evaluate the effectiveness of that step of the restoration. In environmental impact assessment this approach is referred to as “Intervention Analysis,” because no control sites, which are presumed to be unaffected by the manipulation, are identified and monitored. This contrasts with “Before-After, Control Impact Analysis” or “Impact versus Reference Sites” designs that include simultaneous monitoring of control or reference sites (Stewart-Oaten and Bence, 2001). The latter may be especially important when inadequate baseline (pre-project) data are available. A major challenge faced by the Adaptive Assessment Team as it continues to design the MAP will be to maximize the information derived from this type of passive-adaptive approach that builds incrementally on one initial model. It is important to establish experimental controls wherever possible. This includes designing interventions so that their results can be interpreted as if they were controlled experiments. One of the justifications used for taking the passive approach to the Everglades restoration is that an active approach may be too risky for rare species (e.g., Cape Sable seaside sparrow). A reality of the restoration is that the Endangered Species Act (ESA) prohibits any action that jeopardizes the continued survival of listed species. As a result, this must be a primary consideration in the Restoration Plan’s design, implementation, and operation. Still, perceived risk to rare species is one of the most frequent causes of failure of adaptive management programs (Walters, 1997), and the Restoration Plan will be challenged to prevent concerns for rare species from crippling its attempt to manage adaptively. Some mechanisms are available to reduce the severity of such conflicts, such as the multi-species habitat-conservation plans allowed under the 1982 amendments to Section 10 of the Endangered Species Act (NRC 1996a). Another justification for a passive approach is that it is simpler than an active one because it limits the number of choices available to managers if undesirable ecosystem responses occur at any point in the restoration. However, taking this simple approach at a critical juncture could complicate later management options and hamstring subsequent actions. While limiting management options in a system as complex as the Everglades may be desirable from the management point of view, this lack of flexibility is also a fundamental drawback of the passive adaptive management approach (Walters and Holling, 1990). Another drawback is the inability to attribute any given ecosystem response to a specific causal factor, particularly in large-scale projects like the Everglades where there are multiple stressors or drivers acting on the ecosystem. RESTORATION GOALS AND TARGETS “In its original meaning, and when used with reference to a natural system under anthropogenic stress, ‘restoration’ means a return to a system that is not under anthropogenic stress. When used in the context of the south Florida wetland system, ‘restoration’ has come to mean the recovery of sustainable wetland systems at some higher level of ecological health than characterizes the current impacted systems. The broad goal is to recover and sustain the major defining ecological characteristics of the pre-drainage south Florida wetland systems over as large an area of the remaining wetlands as possible” (Ogden et al., 1997).

OCR for page 19
Science-based restoration will occur only if science is strongly integrated into the decision-making processes that most critically impact the state of the ecosystem. As discussed by Harwell et al. (1999), linking science and decision-making depends on how restoration goals, targets, and measures are arrived at and related to one another. Harwell et al. point out that society must define the restoration goals although the goals need to be constrained by scientific knowledge. In addition, scientists should make clear to stakeholders the degree to which achieving restoration goals will require allocation of finite resources. It is then up to scientists and stakeholders to translate those goals into explicit restoration targets, that is, the set of observable ecological and societal attributes that characterize the restored system; yet, the uncertainties associated with these attributes must be recognized. Finally, the choice of restoration measures, the actual variables used to evaluate restoration progress, is essentially a scientific problem. Science has the potential to inform ongoing restoration policy and management decisions to the extent that restoration targets and measures actually capture and measure progress towards society’s goals and objectives. As the Restoration Plan was developed, a great deal of analysis and political and scientific judgment were invested in specifying restoration goals, targets, and measures. Perhaps inevitably, building broad stakeholder support for the program has been achieved in part by promoting goals and expectations that may not be entirely achievable or even internally consistent. In particular, the following issues have emerged during discussions with RECOVER personnel and continue to be discussed by the Task Force. Each of these is discussed in more detail in the following subsections. Some of the specific restoration goals could be construed to be at odds with the general goal of ecosystem restoration. Restoration targets have not been reconciled with overarching forces of change in south Florida, especially population growth, land-use change, and sea-level rise. Targets and measures have not yet been defined for the broad goal of achieving compatibility of built and natural systems. There appear to be competing visions of what success actually means, in part because there is no agreement on how to define a “healthy” Everglades. What “restoration” of the Greater Everglades ecosystem consists of is not entirely clear, which makes it difficult for scientists to establish explicit restoration targets and measures. That in turn makes it difficult to develop an effective monitoring and assessment plan and to apply adaptive management. In the absence of clear and agreed-on restoration goals, scientists must use short-term management objectives to set interim targets and measures. Even though the Everglades is a water-driven ecosystem, the goal of “getting the water right” is somewhat ambiguous, and as a strict and only goal, it might be incomplete. It is somewhat ambiguous because it is difficult to know what historical water levels, quality, timing, and distribution were in a system that has been so altered by local, regional, and global human activities. It has the potential to be incomplete because the Everglades ecosystem has a highly variable hydrologic regime that invites over-management to achieve the non-ecological goals of the restoration. Further, focus only on this one factor risks excluding other important factors from considering. Also, because irreversible changes have occurred in the system, historical water levels might not be optimal at all locations for restoration of the remnant Everglades.

OCR for page 19
RESTORING SPECIES AND HABITATS OR THE ECOSYSTEM? The South Florida Ecosystem Restoration Task Force (SFERTF) established three strategic restoration goals: (1) get the water right, including both hydrologic regime and water quality; (2) restore, preserve, and protect natural habitats and species, including control of invasive exotic plants; and (3) foster compatibility of the built and natural systems (SFERTF, 2000). The stated goals and objectives of the Central and South Florida Restudy were to: Enhance ecological values. Improve habitat and functional quality. Improve native plant and animal species abundance and diversity. Increase the total spatial extent of natural areas. Enhance economic values and social well being. Increase availability of fresh water (agricultural/municipal & industrial). Reduce flood damages (agricultural/urban). Provide recreational and navigation opportunities. Protect cultural and archeological resources and values. These goals and their components—four ecological and five societal—were crafted to achieve a broad consensus among disparate stakeholders and the committee does not challenge them. However, there is potential for different interpretations of the ecological goals and a tension between those goals, which refer to natural areas, habitats, and species, and the larger aim of “restoring the Everglades ecosystem.” Habitat is an organism-specific concept and refers to the set of resources and conditions that allow an organism to occupy an area (Pianka, 1978). Thus, use of the term habitat in the restoration goals potentially puts the emphasis on restoration of a place to its former condition as opposed to restoration of desirable ecosystem processes that may or may not produce a return to the historical conditions in that place (Bradshaw, 1996). Similarly, setting goals for communities or species places sole emphasis on biological composition rather than on biological and physico-chemical processes that are associated with restoration of ecological functioning. In this sense, the Restoration Plan goals contrast with restoration goals for the Kissimmee River Project, which were framed in terms of ecological functioning instead of discrete taxonomic components or conditions (Light and Blann, 2001 working draft). Focusing on habitat and species restoration could impede adaptive, large-scale ecosystem restoration, especially in a setting such as the Everglades where places deemed critical habitats for threatened and endangered species could become locally less suitable for those species during plan implementation. An alternative approach is to aim more broadly for restoration of ecosystem processes at a large spatial scale (Walters, 1997; Cairns, 1988; Bradshaw, 1996). “In a general sense, do not attempt to restore the system (Everglades) to what it supposedly ‘was’ where it ‘was’, but attempt to restore critical functions and structures” (Holling et al., 1994). In establishing restoration targets, the Central and South Florida Restudy Alternative Evaluation Team (AET) defined restoration of the Everglades in a broad conceptual sense that is consistent with Bradshaw’s ecosystem restoration viewpoint: (1) low nutrient levels in marshes; (2) healthy plant mosaics; (3) strong food chains at middle trophic levels; (4) viable populations of animals with large spatial requirements; (5) an abundance of certain upper trophic level animals; (6) recovery of endangered species; (7) extensive, low-salinity estuaries;

OCR for page 19
(8) large spatial extent; and (9) dynamic water storage and sheet flow. These restoration expectations recognize that targets are a temporary set of expectations that will be found wanting, and replaced in time as new understanding of the ecosystem emerges through the process of adaptive assessment. However, the Task Force and RECOVER have also identified and promoted a large number of quite specific targets, for example: A 90 percent recovery of the acreage and number of tree islands existing in 1940, and a health index of 0.90 (where 0 = death is imminent, 1 = completely stress free). (Interim target: A 20 percent improvement in the general health index of the tree islands, and no further loss in the total number of tree islands by 2020.) Four thousand nesting pairs of wood storks in the Everglades and Big Cypress basins. (Interim target: Fifteen hundred nesting pairs by 2010.) Nesting Roseate Spoonbills in the coastal zone of the southwestern Gulf Coast between Lostman’s River and the Caloosahatchee River; and 1,000 nesting pairs in Florida Bay, including 250 nesting pairs in northeast Florida Bay. A 65-75 percent coverage of Florida Bay with high-quality seagrass beds. A long-term commercial harvest of pink shrimp on the Dry Tortugas fishing grounds that equals or exceeds the rate that occurred during the years 1961-1962 to 1982-1983; and an amount of large shrimp in the long-term average catch exceeding 500 pounds per vessel-day (McLean and Ogden, 1999). By producing a long list of specific targets, RECOVER has attempted to provide both specificity and accountability to the broader restoration goals and also to prominently identify criteria that are meaningful to various stakeholders and the public at large. Restoration Plan scientists recognize that many of the specific targets, which have been set using historical evidence, conceptual models, and dynamic hydrologic and ecological simulation models, are little more than “best guesses” at where, when and how populations and communities will respond to restored hydrologic conditions (Restudy AET, 1998). There is the danger that nonscientists will take these targets too literally and challenge the credibility of the restoration if those specific targets are not met. Furthermore, the way the targets are currently defined does not recognize possible ecological tradeoffs as restoration activities operate to benefit one species to the detriment of another, even though there is evidence that no one management strategy is best for all species (e.g., Curnutt et al., 2000). Finally, it is important that targets not be chosen if it is known in advance that they are unrealistic. Further refinement of the MAP targets will be limited by the lack of broad consensus on what Everglades restoration means, let alone how to achieve those goals. RECONCILING TARGETS WITH EXTERNAL FORCES OF CHANGE The Everglades ecosystem is not a closed system. It was an open system in the past and remains one. Its dynamics are driven in part by global processes that affect local weather and climate, introductions of exotic species through foreign trade, and periodic extreme weather events that are at least regional if not hemispheric in origin. Its surrounding environments also are affected by global and regional economic, political, and societal forces. Incorrect assumptions can slowly changing drivers and attributes like topography, climate, atmospheric pollutants, human population dynamics, land use, economic trends (including sugar supports

OCR for page 19
and trade policies), politics, etc., that create problems with monitoring and assessment projects. It is these slowly changing aspects of the system that can be expected to constrain the more rapidly changing elements of the system—the ones most frequently identified as restoration targets. Most of the performance measures identified in the MAP are variables that change on relatively short time scales (e.g., wading bird and alligator population size). Reducing uncertainty associated with the Restoration Plan depends on understanding the constraints imposed by these overarching forces of change on rapidly changing variables. ACHIEVING COMPATIBILITY OF THE BUILT AND NATURAL SYSTEMS Until targets and measures are set for defining compatibility of the built and natural systems, the Restoration Plan will not have explicitly and fully addressed possible tradeoffs and conflicts between ecological restoration and other policies, statutes, and social demands. Establishing these targets and measures depends on the ability to conceptualize and make credible forecasts (scenario analysis in adaptive-management terminology) of socioeconomic change in south Florida (e.g., population growth and distribution, land use change, changes in water demand, transportation networks). To date restoration planning has been driven by sophisticated hydrologic and ecological models dedicated to describing the internal dynamics of the Everglades ecosystem, either original or remnant. These models treat the social drivers that have produced the current diminished and ecologically degraded state of the Everglades as exogenous variables that will maintain constant or linear trends over the next half-century. For example, in the Restudy, the population of the lower east coast of Florida is assumed to increase by 72 percent by 2050 and the number of residents in the 16-county study area to increase from 6.3 to 11 million people. Leaving aside for the moment that this estimate might not be appropriate over the long run, or even now, it appears that the Restudy used a “future without plan” scenario for the human system, assuming that the Restoration Plan would not appreciably influence the future number or distribution of people and anthropogenic ecological stressors in south Florida. This approach is not scientifically credible. Experience in south Florida demonstrates the importance of scenario analysis. The Central and South Florida project required major modifications as conditions and assumptions changed3 based on population dynamics, climate change, etc. (Light and Dineen, 1994; Light et al., 1995). Conceptually separating the future dynamics of natural and human systems decouples the biophysical driving forces of the Everglades and its restoration from the dynamics of human habitability. To date, the difficult task of analyzing to what degree the built and natural systems are compatible has not been undertaken. Ultimately, demographic and economic dynamics will need to be brought into restoration planning and implementation to better anticipate, monitor and respond to the effects of the restoration on social and political systems, and vice versa. 3   For example, at first the Central and South Florida system was not constructed to deliver water to the Everglades National Park. Then the assumption was made that monthly allocations were sufficient. Finally, in 1980 the drought of record exceeded design specifications and expectations.

OCR for page 19
DEFINING “ECOLOGICAL HEALTH” FOR THE EVERGLADES ECOSYSTEM The Restoration Plan’s Adaptive Assessment Team (AAT) is developing a monitoring program based on a standardized set of monitoring and data-management protocols over spatial and temporal scales that are relevant to the Restoration Plan implementation schedule and to ecosystem responses. A challenge of defining those protocols is that neither current ecosystem conditions nor those that predated the Restoration Plan are well understood. The restoration is currently defined by a list of more than 100 hydrologic and ecological interim targets, as there is general agreement by scientists that the Everglades restoration should be thought of as an open-ended process (e.g., Davis and Ogden, 1994). It is within this context that scientists are striving to develop the MAP. What is lacking is agreement about which of these ecosystem features should receive the highest priority and the extent to which these features can be restored. The most immediate hurdle facing the AAT in development of the monitoring program is defining “ecological health” of the Everglades. What are the attributes of a “healthy” Everglades? The “health” of an ecosystem is not definable scientifically (e.g., NRC, 2000), and so the choices are to identify biological parameters as goals or to use societal values. Whatever the approach, the goals must be measurable and identifiable so that it is possible to determine whether they have been achieved or not. A focus on improving specific aspects of ecological functioning seems likely to be useful and also to help integrating biological parameters with societal values. Without definable goals or targets, it will be extremely difficult to select appropriate indicators and the performance measures needed to monitor progress towards those targets. That is, it is impossible to measure ecological recovery if ecological recovery is not clearly defined. Simply selecting variables to measure because they are likely to respond to altered hydroperiod is not a good basis for choosing a performance measure unless the objective is to demonstrate that the system will change as a result of altered hydroperiod and thus management is a “success”. Even the language of the South Florida Ecosystem Restoration Task Force’s documents shows a predisposition to reporting success, as for example, “success in the making” and “coordinating success.” An emphasis on “success” should not influence the selection of one type of performance measure over others. RESTORATION REFERENCE STATE “Baseline information provides the benchmark against which the progress of the restoration plan can be measured, and to understand the ranges of natural variability necessary to confirm when change has actually occurred. While some regions of the Everglades ecosystem have well-established monitoring programs, other areas have little or no baseline data. Plugging the gaps in baseline conditions is one of the critical components of the monitoring and assessment plan.” (USACE and SFWMD, 2001a) The committee defines the restoration “baseline” for the Restoration Plan broadly to include not only the record of observational data, but also the current state of understanding. That understanding includes conceptual and simulation models that have been used to define two “pre-Restoration Plan” reference states: the “pre-drainage” and “post-drainage, pre-Restoration Plan” Greater Everglades Ecosystem. The reference states provide the basis for assessing the magnitude and desirability of system responses to the Restoration Plan (NRC,

OCR for page 19
2000). The Restoration Plan design is supported by a large body of scientific research, but the general question “What is the ecological reference state for the Restoration Plan?” remains unresolved. There is no simple answer to this question. The Everglades System has changed and will continue to change due to long-term variations in climate and sea level even with no further direct human impacts or restoration attempts. These forces and factors such as regional land-use change, pollution from remote sources, and invasive exotic species, will exert continuing effects on the system. Furthermore, the current system is still adjusting to recent historical changes in drainage and water quality (e.g., nutrient cycling, extent and distribution of tree islands) and one would expect significant lags in the system’s response to the proposed modifications in hydrologic regime and water quality. Given these dynamics, what constitutes the restoration reference condition trends and variability for the Restoration Plan? What period of record is appropriate for specific targets and measures? What quality standards will be applied to select the data used to quantify the restoration reference state? Scientists associated with the restoration have sought to define reference conditions using a mix of modeling and empirical studies. Despite considerable progress there is a great need for continued research to better conceptualize and describe the reference state, research that goes well beyond performance monitoring. This is not to say that the Restoration Plan should not proceed without a better-defined conceptualization of the restoration reference state, but only to point out that the adaptive assessment strategy should include monitoring in support of improved “baseline” data and model outputs as well as hypothesis-driven research to validate the underlying cause–effect relationships identified in the MAP conceptual models. The reference state for the pre-drainage Everglades has been reconstructed from paleoecological information and historical observations along with extensive use of the Natural Systems Model (NSM). The uncertainties associated with the use of the NSM are relatively well known, notably those related to specification of pre-drainage topography, patterns of precipitation and evapotranspiration, and surface roughness as well as the artifacts of the discrete space and time in the numerical model (Bales et al., 1997). Reviewers of the model have cautioned that in its current form the NSM can only indicate broad regional patterns of inundation over time as opposed to local discharges and flows. Despite these limitations, spatial hydrologic patterns predicted by the NSM have played a significant role in setting spatially explicit reference conditions for the Restoration Plan. The Restoration Plan has used both NSM and complex simulation models like the Across Trophic Level System Simulation (ATLSS) model (http://atlss.org) and the Everglades Landscape Model (ELM; http://www.sfwmd.gov/org/wrp/elm) to represent current understanding of the relationship between the hydrologic and ecologic variables of the historical Everglades system, to describe pre-restoration conditions of most ecological targets, and (in conjunction with the South Florida Water Management Model) to compare the ecological benefits of restoration alternatives. Although they are advanced and important restoration tools, ATLSS and ELM are still unreliable for establishing reference conditions and for forecasting outcomes of the restoration because of the complexity of the ecosystem. (For example, hydrologic data alone predict patterns of fish abundance at least as well as ATLSS does (J. Trexler, Florida International University, personal communication, 2001). Given the incomplete historical evidence, imperfections of the NSM, and the severe limitations of current ecological data and models, there is a healthy scientific debate over definition of the pre-drainage hydrologic and ecological reference conditions for the Restoration Plan. Some reference conditions, such as the historical vegetation, water quality and ecosystem

OCR for page 19
processes in what is now the Everglades Agricultural Area and the relationship of that area to the rest of the system, must remain somewhat speculative. Some responses of the biota in the current highly modified system might not accurately represent the way they would have responded to changes in the pre-drainage system. However, other scientific issues of critical importance to understanding ecosystem response to the Restoration Plan are more tractable. They include the following: Improving the resolution and accuracy of hydrologic model input data (e.g., surface topography and evapotranspiration parameters). Improving ecological model specifications for species and community ecological requirements and functional relationships to hydrologic regimes and water quality. Conducting empirical and modeling analyses to evaluate the importance of extreme events and variability as opposed to the mean and range of annual conditions in climate, hydrologic regimes and associated disturbance processes such as flooding and fire in maintaining the ecological characteristics of the Everglades. Investigations of hydrologic linkages between surface water storage and flows and near-surface groundwater flows and seepage processes. Research to determine the relative importance of water velocity versus hydroperiod in controlling plant and animal communities. Analysis of trends in key hydrologic and ecological variables due to recent climate change and modern sea level rise. Understanding the role of modern fragmentation of wetland and surrounding upland habitats on the ability to achieve desired conditions. Investigation of contaminant accumulation and transformation in soils/sediments and mobilization, and especially their effects on the biota. Research to better integrate social system dynamics into the conceptual and simulation models used for adaptive assessment (discussed in more detail in the next section). CONCEPTUALIZING THE HUMAN DYNAMICS OF THE EVERGLADES RESTORATION As discussed in the section on the Restoration Plan’s Goals and Targets, human dynamics were treated relatively simplistically in the Restudy as exogenous to and uncoupled from the Restoration Plan. Despite recommendations from social scientists for alternative approaches (e.g., Harwell et al., 1999), this view has also pervaded the reference state conceptualization of the Everglades Ecosystem as represented in both simulation models and conceptual models that have been used to set targets and choose performance measures. Two questions are raised by this approach. The first is how robust the Restoration Plan is to alternative scenarios of human dynamics in and around the Greater Everglades Ecosystem. The second is whether there is any possibility for strong feedbacks between implementation of the Restoration Plan and social dynamics that could significantly impact the ability to achieve the restoration goals. Future population growth and its distribution in the Restudy are based on current conditions and on the recent past, with assumptions that recent trends will continue. Much of the information is derived from local comprehensive planning documents. However, it is clear from the demographic and sociological analyses that the major social drivers such as population

OCR for page 19
change, urban growth, agriculture, long-term change in economic activity, and tourism have changed through time in ways that defy simple extrapolation (Solecki et al., 1999). How robust is the Restoration Plan to alternative scenarios of these exogenous drivers? For example, one could make a general case that land uses and other human activities are likely to change a great deal between now and the year 2050, if not sooner, based on very reasonable assumptions about physical factors (e.g., soil subsidence, peat loss and climate change) and socioeconomic factors (e.g., changes in the price and costs of production of sugar, citrus, winter vegetables, and cattle; the relative prices of land for other uses; competing hydrologic demands from both the natural system and the built environment; changes in political systems; or changes in societal values), all of which could profoundly influence the way humans interact with the Everglades, with or without the Restoration Plan. Social dynamics will likely interact with other exogenous drivers, like an accelerated rate of sea-level rise or increasing frequency of extreme weather events, that have been predicted to occur with ongoing global climate change. Such changes could create differential pressures on flood control, human fresh water demand, and water requirements for the coastal estuaries. These kinds of considerations lead us to conclude that the human dynamics of Everglades restoration require greater research, modeling and monitoring. This need has been recognized by social scientists involved in the restoration and is specifically addressed in the Restoration Plan Environmental and Economic Equity Program Management Plan (USACE and SFWMD, 2001b), which contains specific objectives for establishing socioeconomic and environmental justice baseline data (including alternative scenarios). That plan also calls for improved social-science research. In addressing the Restoration Plan’s objectives, then, those responsible for designing a monitoring and assessment program should consider the degree to which the following general issues are important. What are the critical human forces driving or affecting restoration? For example, how might changes in the size and distribution of human populations in south Florida affect restoration? To what extent can science and modeling examine the alternative restoration scenarios proposed in the Environmental and Economic Equity Program Management Plan? What social dynamics of importance are outside the management boundary? How should these be analyzed and modeled? For example, how should the Restoration Plan model and analyze such externally influenced factors as the price of agricultural commodities and energy, even though they cannot currently be predicted accurately? LEARNING THROUGH INTEGRATED MONITORING, MODELING, AND EXPERIMENTAL RESEARCH Learning is a critical part of adaptive management; if the agencies and scientists involved in the Restoration Plan emphasize adaptive management and assessment as a learning vehicle for the restoration efforts, the restoration is likely to be more effective than it would otherwise be. The learning process that will guide the “adaptive implementation” of the Restoration Plan will depend on a research strategy that effectively combines monitoring, modeling, and experimental research, with a high level of attention to information management, data synthesis and periodic re-synthesis of information throughout the implementation and

OCR for page 19
hypothesis-based hydrologic and ecological variables at several scales. The water-management problem of actually achieving the desired distribution and timing of water to benefit wood storks in the park and Big Cypress Basin may require yet another kind of learning to tune the operation of the water-delivery system (e.g., a denser network of water-level and precipitation gauges). The March 29, 2001 version of the MAP describes the basis for selection of the variables or performance measures that will be required to assess hydrologic conditions in the ecosystem and the ecosystem response to hydrologic conditions. Two aspects of the monitoring plan will need to be addressed in the future: 1) spatial and temporal distribution of the performance measure sampling effort and 2) integration of monitoring supported by the Restoration Plan with ongoing long-term monitoring conducted by groups like South Florida Water Management District, U.S. Environmental Protection Agency, U.S. Geological Survey, National Park Service, university projects (e.g., Florida Coastal Long Term Ecological Research (LTER) project), the Tribes, and private organizations (e.g., The Audubon Society). The March 29 version of the MAP list of 156 performance measures—including hydrologic, soil, water-quality, and ecological performance measures—have not yet been integrated into a coherent monitoring plan. Rather, the list of performance measures is held together only very loosely by regional conceptual models. True ecological indicators of Greater Everglades ecosystem functioning (see NRC 2000) have not been developed. However, efforts to integrate the performance measures into five categories called functional groups by the Restoration Plan (wetland trophic relationships, wetland landscape patterns, estuarine epibenthic communities and habitats, Lake Okeechobee pelagic and littoral zones, and biota of special concern) were ongoing when this report was written. Comments specific to the hydrologic and ecological performance measures are discussed in following sections. Hydrologic Performance Measures The basic premise common to the wide variety of published restoration goals is that water management to mimic pre-drainage hydrologic conditions can provide sustainability of the human system and improvement in the ecological “health” of the natural system. This hypothesis led to development of the South Florida Water Management Model and Natural System Model used to design the Restoration Plan. Many of the 78 hydrologic performance measures included in the MAP are essential variables to understanding the spatial and temporal distribution of water in the natural system and water supply to the human system. These measures are being updated; for the latest version of them, please see the RECOVER web site for a draft report: http://www.sfwmd.gov/org/ema/everglades/consolidated_03/ecr2003draft/ (as of February 2003). Other hydrologic performance measures included in the MAP are important water quality measures. Originally, the list of hydrologic performance measures included approximately 900 hydrologic and water-quality performance measures. These measures were developed by the Restudy Alternative Evaluation Team to evaluate alternative plans for achieving the water management targets set out in 1998/1999 during the Central and South Florida Restudy process. This number was reduced to 24 hydrologic and 54 water-quality performance measures by the time the MAP workshop was held. These 78 measures will be used to monitor water management as restoration proceeds and will provide information about water levels, water flow, duration of flooding, and water quality (especially P and less frequently N). The

OCR for page 19
hydrologic performance measures are not response variables in the same sense as the ecological performance measures and “functional groups.” Rather, they are measures of factors identified as stressors in the regional conceptual models. Although socioeconomic drivers are prominent in all of the conceptual models, only the hydrologic stressors are included in the current list of performance measures. The hydrologic performance measures were largely developed for the purpose of designing the Restoration Plan. In this context they were very effective. However, their use in future applications, such as refinement of the Restoration Plan’s design and adaptive assessment, would benefit from some adaptations. One limitation of the current hydrologic performance measures derives from the fact that their use in designing the Restoration Plan is a very different exercise than their use in adaptive assessment. In the former, the measures are applied to the results of multi-year model simulations of the performance of alternative formulations. Each simulation is based on a single sample of the relevant hydrologic variables under the assumption that a particular formulation has been completely executed. Adaptive assessment is a much more difficult problem. To be effective for adaptive assessment, measures must provide information early enough to allow for corrective action. But the present measures will be of limited value when applied in the first decades of the Restoration Plan, largely because many critical features of the CERP will not have been implemented. Even after the Restoration Plan has been completed, observed measures will be confounded by temporal variability in water flows and levels due to climatic variability. For example, it may take many years before there are climatic conditions that provide opportunities to test measures involving maximum and minimum limits, such as limits on water levels in Lake Okeechobee. Furthermore, the value of the “test” will depend on the severity of the conditions. For measures depending on mean values of hydrologic variables, such as the duration of flood conditions, their variability will limit their usefulness for many years. Another limitation of the current hydrologic performance measures is that they cannot be aggregated to provide an overall measure of system performance. This is because they do not quantify the “damage” associated with failure to meet the targets on which they are based. An aggregate measure was not required in the initial Restoration Plan design, apparently because the design was based on the assumption that all objectives would be met. However, future decisions on the design may require compromises in the face of budget limitations and ecological realities; an aggregate measure would enable such compromises to be effected in a consistent and efficient manner. Fortunately, these limitations can be readily addressed. The hydrologic model can be used in such a way that observed climate and the status of the Restoration Plan implementation are used as input for it. This would be most useful for adaptive assessment. Furthermore, hydrologic model outputs can be the attributes of ecological habitat suitability functions for selected ecological indicators. Composite values of these time series of habitat-suitability-index values can serve as surrogate indicators of “damage” functions, indicating the relative benefit or loss associated with water management outcomes. These can be compared to previous values to enable the development of an aggregate performance measure. Such a measure would enable refinement of the Restoration Plan design based on multi-year simulations (Tarboton et al., 2003). Hydrologic modeling can be used readily to account for the dynamic elements of the restoration, such as climatic variability and implementation. As various features of the Restoration Plan are implemented, the model could be modified accordingly. On an annual

OCR for page 19
basis the evolving model could be used to simulate the system based on the observed meteorological conditions. Hydrologic performance measures could be computed from the simulation results, and compared to performance measures computed from actual field data. For each measure, error analysis could be used to evaluate the significance of the observed difference between the two values. Significant differences could trigger studies to determine their causes, such as errors in model input, parameters, or structure. These errors could involve the natural system, such as estimation of flow resistance, or the engineered system, such as components of the Restoration Plan. In any case, periodic comparisons of predicted and actual hydrologic variables would enable continual improvement of the hydrologic model. ECOLOGICAL PERFORMANCE MEASURES Conceptual Models for the Restoration Plan Conceptual models of each of the Greater Everglades Ecosystem’s nine major physiographic regions in south Florida (e.g., ridge and slough, marl prairies) are the basis of the Restoration Plan ecological monitoring program. The March 29, 2001 Monitoring and Assessment Plan lacks system-wide performance measures (see Appendix E for lists of the performance measures as of that time).4 Thus, these broad-scale measures are not evaluated in this report. However, suggestions are provided about a useful approach for developing such a set of measures. Since the November 2001 monitoring and assessment workshop, some progress has been made towards development of a system-wide conceptual model that will be the basis for incorporating system-level performance measures into the Restoration Plan’s Monitoring and Assessment Plan. The conceptual models were intended to offer a non-quantitative conceptualization of the causal linkages between ecosystem drivers and attributes of the physiographic regions. These models differentiate variables hierarchically into “drivers”, “stressors”, “ecological effects”, “attributes”, and “performance measures” (USACE and SFWMD, 2001a). The drivers represent the major external forces that have large-scale (spatial and temporal) impacts on the natural system, such as climate change or sea-level rise. Stressors are physical or chemical changes to the natural system that are caused by the drivers, which ultimately lead to biological and ecological effects. Attributes and performance measures represent known effects of the stressors and are features that can be monitored to determine progress towards restoration goals and objectives (e.g., number of nesting wading birds). It is the attributes and measures of the conceptual models that form the basis of the monitoring program and are described as ecological performance measures. Some of the attributes and measures have characteristics of ecological indicators--measures of ecological condition, ecosystem functioning, or ecological capital—as described by the NRC (2000). A specific example would be extent of plant cover. Other attributes and measures do not measure ecological condition, such as phosphorus concentration. Development of the conceptual models began in 1996 (Ogden et al., 1997) and the models were refined in the summer of 1999 during meetings among experts working in each of the Everglades subsystems (i.e., the conceptual model teams) during workshops held in summer 1999 (see RECOVER AAT homepage, http://www.evergladesplan.org/??/recover/aat/cfm). 4   A total system conceptual model was completed in 2003, and will be used as a basis for developing system-wide performance measures.

OCR for page 19
These conceptual models and the 150+ ecological performance measures that were derived from the models were the focus of the July 6-7, 2000 CROGEE meeting. Further refinement of the ecological performance measures occurred during the summer of 2000, when each conceptual model team reviewed the model it had produced during the previous summer. As part of the reevaluation, the teams were asked to refine and rank the performance measures to (1) produce a relevant, practical, and parsimonious set of performance measures that would indicate the ecological health and recovery of their region, and (2) specify parameters and locations for monitoring the performance measures. They were also asked to identify, for each performance measure, experts to write a 1-2 page document describing uncertainties behind the performance measure and their impact on understanding the results of the restoration monitoring program. Additionally, each team was asked to identify research that would be essential for interpretation of the changes in the performance measures. Based on the summer 2000 workshops and the 1-2 page performance measure documents, the AAT reevaluated the ecological performance measures and produced the draft (March 29, 2001) of the MAP reviewed at the MAP workshop. This draft includes 61 performance measures distributed among the five “functional groups” (i.e., wetland trophic relationships, wetland landscape patterns, estuarine epibenthic communities and habitats, Lake Okeechobee pelagic and littoral zones, and biota of special concern). The conceptual models are not simulation models of the subsystems. Rather, they were an exercise designed as a first step in designing the monitoring program. These models are based on a “best guess” about how the major Greater Everglades Ecosystem physiographic regions function. They are crucial to identifying gaps in knowledge that might impact design of a monitoring program and are important tools for synthesis and integration of scientific information into the adaptive management process. The ecological performance measures included those attributes of the physiographic regions that the conceptual model teams judge to be the most likely to respond to the proposed water management scheme described in the Restoration Plan. Assessment of Conceptual Models and Ecological Performance Measures Given the incomplete characterization of pre-drainage conditions of the Greater Everglades Ecosystem, the Adaptive Assessment Team has no alternative but to develop a monitoring plan based on indicators that will provide an assessment of the current status of the ecosystem. As the Restoration Plan is implemented, this reference condition will be used to assess the trajectories of populations (e.g., organisms of special concern including threatened, endangered, and invasive species), communities (e.g., tree islands), and ecosystem processes (e.g., net primary productivity, formation of soil organic matter). This type of hierarchical approach will provide a relatively comprehensive evaluation of the system’s ecological response to restoration projects and changing environmental conditions (e.g., sea level rise, climate change). The difficulty facing the Adaptive Assessment Team is selecting appropriate, practical, and informative performance measures. If they are well chosen, they should also provide a context or framework for refining the series of conceptual models that are currently the basis of the monitoring plan. In turn, as the conceptual models are refined they should provide a context or framework for choosing and interpreting more specific performance measures.

OCR for page 19
The process of documentation of the uncertainties associated with each of the performance measures is ongoing and will be essential to decisions about the usefulness of each measure. Even with this documentation, the choice of performance measures may be difficult because many of the regions covered by the conceptual models currently are not well understood. While the conceptual models are useful tools for identification of ecological performance measures, these models do not provide insight into the temporal or spatial sensitivity of the measure’s response to altered hydroperiod. How can data on performance measures and the associated variation be used in combination with mechanistic models to insure that performance measures are monitored at appropriate temporal and spatial scales? Most of the ecological performance measures currently under consideration are area and species- or community-specific. A few system-wide measures are currently included in the MAP’s list of potential ecological performance measures, but these focus primarily on endangered or threatened species to the exclusion of ecosystem functioning. The current monitoring plan would benefit from a few ecosystem-level, system-wide indicators (as opposed to performance measures). The NRC (2000) proposed a suite of such indicators that were intended for national-scale assessment but also are applicable to regions like the Everglades. That report recommended indicators that would help assess the extent and status of an ecosystem type, ecological capital, and ecological functioning or performance. Indicators such as land cover and land use that have powerful influences on landscapes and adjacent ecosystems can be used to define the extent and status of an ecosystem (or subsystem type). Indicators of ecological capital include total species diversity, native species diversity, nutrient runoff, and soil organic matter. Such whole-system indicators could also be usefully applied to the Greater Everglades Ecosystem. The MAP should consider monitoring spatiotemporal patterns of total species diversity as well as of native species diversity. Additionally, given that the Restoration Plan is focused on a well-defined ecosystem an Index of Biotic Integrity-like measure (Karr and Chu, 1999; NRC, 1994; NRC, 2000) would be appropriate and useful. The purpose of such a measure would be to provide a “multimetric” that would integrate several key indicators to represent the changing status of the Greater Everglades Ecosystem. Indicators also are needed to provide information about ecosystem functioning in a broad sense and to provide information about the ecosystem’s capacity to respond to changes. They should include indicators of production capacity, such as total chlorophyll per unit area (or in aquatic regions, chlorophyll per unit volume). Carbon storage, particularly in wetlands, as indicated by sediment organic matter, is also informative about ecosystem functioning. Indicators of nutrient balance would provide information about environmental loading of nutrients to the ecosystem. Remote sensing using today’s operational systems provides relatively inexpensive and consistent estimates of several key ecosystem parameters that are common to all subsystems. Digital aerial photography and Landsat Thematic Mapper (TM) imagery have already proven useful in mapping and monitoring plant communities, land use and land cover (Doren et al., 1999). Vegetation classification can, however, be complicated by factors such as water depth or color, effects of fire, periphylon species composition, and growth morphology within a single species (Rutchey and Vilcheck, 1999). NASA’s suite of Earth Observing System sensors offers greatly expanded abilities for synoptic monitoring. For example, the MODIS (the Moderate Resolution Imaging Spectrometer) sensor provides an excellent opportunity for synoptic monitoring of land use/land cover, surface reflectance, spectral vegetation indices such as the Normalized Difference Vegetation Index (NDVI), the absorbed fraction of photosynthetically active radiation (FPAR),

OCR for page 19
leaf area index (LAI), net primary productivity (NPP), and land surface temperature. The MODIS instrument acquires image data in 36 spectral bands at spatial resolutions from 250 m to 1 km over the entire globe every two days. A series of standard land products are being produced from these data by the MODIS Land Discipline Group (MODLand). Several NSF Long Term Ecological Research (LTER) sites are involved in coordinated field research and monitoring efforts to validate standard MODIS products (e.g., Lefsky, 2001). The Restoration Plan, in collaboration with the newly established Florida Coastal Everglades LTER, could serve a similar role for the Greater Everglades Ecosystem. The higher-resolution data provided by IKONOS satellite imagery is presently being evaluated for use in mapping Everglades vegetation, especially invasive exotic plants, using the spectral reflectance characteristics of various vegetation species. These results are being compared to those using aerial photography. Information on this can be found on: http://www.sfwmd.gov/org/wrp/wrp_evg/projects/ikonos_satellite.html. Synthetic Aperture Radar (SAR) imagery and radar altimetry may provide a means for system wide monitoring of water level and inundation patterns. For example, Alsdorf et al. (2000) have demonstrated that interferometric processing of SAR phase data can be used to infer stage changes of less than 0.1 m in Amazon floodplains, and a similar approach may be feasible over much of the Everglades. (This technique has been tested in the Everglades by Kasischke and Bourgeau-Chavez, 1997). Although remote sensing is a valuable tool for monitoring, it does not eliminate the need for ground-based monitoring (NRC, 2000). Rather, these two approaches are complementary. A well-designed ground-based monitoring and experimental plan should provide the process information necessary to interpret remotely sensed instantaneous information about ecosystem condition and to provide the resolution needed to detect and characterize changes in ecosystem heterogeneity. Indicators of populations and ecosystem functioning sensitive to the restoration efforts need to be identified. Based on the workshop discussions, it is clear that the Adaptive Assessment Team recognizes the importance of selecting meaningful ecological indicators of populations, communities, and ecosystem functioning. Additionally, it is clear that the restoration effort will be considered a success only if the effects of altering water hydroperiod result in measurable improvement of the “ecological health” of the Everglades. Setting Monitoring Priorities Strategies are necessary for prioritizing the importance of the ecological performance measures. Given that the resources available for monitoring are limited, are there measures that are more important to monitor than others? In the ideal monitoring program, all aspects of community structure and ecosystem functioning would be monitored. However, because it is unrealistic to measure everything everywhere—let alone interpret all the data that would result—ecological/biological performance measures are the only practical way to inform us about the status of the ecosystem and/or the response of the system to changing conditions. The working hypotheses associated with the conceptual models of ecosystem functioning have been ranked as low-moderate-high certainty based on whether or not the hypotheses are supported by published, peer-reviewed quantitative relationships or predictive models versus best professional judgment. This is useful in understanding the amount of science supporting the conceptual models, but it is less useful for understanding the relative importance/relevance of particular hypotheses and ecosystem dynamics. Additionally, a goal of

OCR for page 19
the monitoring program is to determine the progress of the Restoration Plan towards meeting the restoration goals and objectives. Therefore, it is critical that the Monitoring and Assessment Plan identify the guidelines for deciding how funding priorities will be established for the monitoring effort. It also is important to consider how often a measure needs to be taken and at what season or seasons might be most informative. Characteristics of the measures that should be considered in setting these priorities include the following: relevance to restoration goals, sensitivity to the Restoration Plan’s design and operation; “normal” variation is known, potential to help identify knowledge gaps that are most critical to the largest number of future projects, relevance to predictive models, and importance to stakeholders. EXPERIMENTAL RESEARCH AND MODELING Experimentation must be a critical component of monitoring and assessment to improve understanding of cause-effect relationships. Thus, experimentation is needed for adaptive assessment to be an effective management tool. Monitoring is not enough; there is a need to understand processes, mechanisms, and inventories and use this information to construct mechanistic models. The Monitoring and Assessment Plan should seek to create a proper balance between modeling, monitoring, and experimentation. Each effort should support the others. All three are cornerstones of adaptive assessment. The key to addressing the uncertainties associated with a best-guess model is to integrate experimentation and monitoring with modeling to provide a more mechanistic understanding of the ecosystem. An additional advantage of incorporating mechanistic models into the Restoration Plan is that they are flexible in that they can be efficiently adapted to new situations and new stressors (or even drivers). Thus, when a new ecosystem stressor arises or there is a significant change in a recognized driver, these mechanistic models provide a means for incorporating the effects of the stressor into our understanding of the system. Even so-called quasi-experimental design, in which there is not random assignment of experimental and control groups, can be useful (e.g., Gribbons and Herman, 1997), especially in a situation like the Everglades, where experiments are only partially controllable. The best examples of experimental programs in the current Restoration Plan are the pilot studies to assess aquifer storage and recovery and seepage management (Sidebar 2-1). However, care must be taken in the development of implementation plans to assure that the pilot studies provide opportunities to improve understanding of processes as well as simply demonstrating that the techniques are feasible.

OCR for page 19
MODEL REFINEMENT AND SENSITIVITY ANALYSES DeAngelis et al. (2003) described three reasons for close integration of models into a monitoring program. First, models may be required to relate restoration targets to indicators or measures that can be directly and practically monitored. Second, models formalize and make explicit assumptions and hypotheses about causal mechanisms that link restoration actions to ecological outcomes. Third, models are the only credible means of forecasting to evaluate the possible ramifications of a restoration or management plan. Models may also facilitate understanding and they can provide the basis for a shared vision among all stakeholders of what alternatives are best according to their economic and social costs and their ecosystem and political impacts. The Restoration Plan provides a case study in the implementation of computer-based decision-support systems combining mathematical models describing the natural phenomena with the human interface for effective communication among the models and humans. These interactive modeling systems are being used to explore, identify, and evaluate various aspects of multiple restoration alternatives and their impacts. Models range from regional to local scales and together include ground water, surface water quality and ecosystem variables and indicators. The models also vary in resolution and applicability to the questions that need to be resolved in order to develop detailed designs and an operation plan for the region. While hydrologic and water-quality modeling are relatively advanced and sophisticated, modeling for ecological outcomes is much less complete. The population dynamics of a few species can be adequately simulated (e.g., Cape Sable seaside sparrow and alligators), but most cannot be. If ecological outcomes are to be evaluated on the basis of models, these ecological models must be improved through additional research/experimentation and through monitoring designed to define mechanistic relationships. One useful approach is to monitor some ecosystem process or processes or changes that are inputs (including parameters, values, and drivers) to ecosystem models. One example of this is the work of Robert B. McKane and SIDEBAR 2-1 Role of ASR pilot projects in adaptive assessment. The Restoration Plan includes a set of “pilot projects” associated with proposed aquifer-storage-and-recovery (ASR) systems, in-ground reservoirs, and seepage control technologies. These projects are designed to serve a variety of purposes, including demonstration that particular technologies are feasible and to acquire information needed for detailed engineering designs. They may also serve as the first step of a phased implementation. For example, the wells that will be constructed during the ASR pilot studies are intended to be permitted as operational ASR wells as part of the final regional systems. The Restoration Plan pilot studies provide a variety of opportunities for learning through active experimentation, but these opportunities can only be realized through careful design of the projects to allow testing of hypotheses that can enhance understanding of critical processes. This committee reviewed initial plans for ASR pilot studies (NRC, 2001) and identified a number of improvements that could be made in their potential to contribute to improved design and implementation of regional scale ASR systems. An ASR Regional Study has also been added to the project to address questions raised with respect to predicting the regional scale changes in aquifer hydrodynamics. This regional study is also intended to reduce uncertainties related to water quality changes during subsurface storage and to effects of aquifer heterogeneity on recovery.

OCR for page 19
colleagues at the Western Ecology Division of the U.S. EPA’s Ecology Laboratory on the effects of stressors, such as pollutants, climate change, or forest management practices, on the structure and quality of wildlife habitat, using the General Ecosystem Model (GEM). The value of this approach is that there is a cause-effect relationship established between what is being measured and the result (in this case, tree growth predicted for the next 20 years). Two other examples of this approval derive from the Everglades restoration program itself. The first is the ATLSS model (Curnutt et al., 2000), described earlier in this report. Some of the ATLSS submodels are relatively sophisticated, being spatially explicit. The second is work by Tarboton et al. (2003) linking hydrologic model outputs to habit-suitability-index models (used to predict species’ demographic or behavioral responses to various sets of environmental conditions) for selected ecosystem indicators. The values of these habitat-suitability functions depend on hydrologic attributes that can be managed. Thus different time-series of hydrologic attributes resulting from different water-management policy simulations can be converted to time-series of habitat suitability function values, each of which can then be combined and averaged in various ways to provide quantitative indications of the relative ecological impacts of alternative water-management policies. SCIENTIFIC FEEDBACK TO GUIDE AND REFINE IMPLEMENTATION OF THE RESTORATION PLAN The usefulness of an adaptive assessment program in improving the potential for the Restoration Plan to meet its hydrologic and ecological restoration goals depends on the opportunities available to modify design or operational features of components as understanding of drivers and system responses improves through monitoring, experimentation, and modeling. Some Restoration Plan components will be designed and put into operation early in the project life. These components may not benefit directly from the adaptive assessment program in the design phase, but any flexibility introduced in a later redesign phase might improve their performance later in the project. These projects can be used as experiments to maximize learning. The Canal-111 (C-111) project—a pre-Restoration Plan project authorized in the Water Resources Development Act of 1994—has served many such purposes. The C-111 project is designed to restore the hydrologic conditions in the Taylor Slough and Eastern Panhandle areas of Everglades National Park and eliminate damaging freshwater flows to Biscayne National Park, while maintaining flood protection. Components include the degrading of existing spoil mounds of the canal to promote sheet flow, and the construction of new pump stations, bridges, and detention areas. The project has served many such purposes by allowing National Park Service scientists to monitor ecosystem response and establish a “footprint” of impact on the natural system. This project is an example of how trade-offs between seepage, infiltration, and water quality are being addressed in the basin. The C-111 project also will provide an opportunity to look at the groundwater/surface water interface and impacts on woody vegetation and ecology of the park west of the project. Additionally, C-111 has served as a test case for learning how to work through scientific and institutional differences to reach tentative consensus that will allow implementation of the Restoration Plan to go forward (USACE, 2002). Other components of the Restoration Plan, which are not scheduled for construction until later years, could be significantly modified to take advantage of learning from the adaptive assessment process. Some components, such as the regional ASR systems and the seepage-

OCR for page 19
control strategies, are specifically planned to include pilot projects and a phased implementation in order to make use of the pilot study results in the final design (see Sidebar 2-1). Completion of some projects should be viewed as providing tests of important mechanisms in hydrologic and ecological models. At the very least, possible outcomes of a project should be articulated a priori, as well as the appropriate response to each outcome in terms of alterations in design of future projects. The less reversible a planned action is, the greater is the incentive to reduce uncertainty about its potential effectiveness and effects, because the consequences of being wrong are greater. However, sometimes bold, irreversible actions have greater potential for success than smaller ones. This is a challenging problem for a sequential restoration program like that planned for the Everglades. The committee has no clear solution to this problem except to suggest that it may be helpful to take an incremental approach and couple it with focused research and adaptive management to reduce uncertainty associated with major actions that are likely to be difficult to reverse. To the degree that it is possible and consistent with the program schedule, postponing major actions whose consequences are both uncertain and difficult to reverse would provide additional time to reduce the uncertainty associated with them. In addition, a careful evaluation of the degree of reversibility of major components of the Restoration Plan would be helpful, as well as consideration of whether more reversible options might be available for those components that appear to be least reversible. As described above, the ability to learn from these actions and associated monitoring and research should be a major consideration in planning them. There is likely to be considerable variability among components of the Restoration Plan in the potential for flexibility in design or operational features. Some components may present a range of design choices but only limited flexibility to modify operations once a specific design is selected and the component is put into operation. For example, decompartmentalization for some portions of the water conservation areas could be accomplished by strategies ranging from installation of a few opening and control structures through an existing levee to complete breach of a levee and filling of the adjacent canal. Removal of the levee would provide the maximum connection between formerly compartmentalized storage areas. However, this strategy would offer no options to control flow between the areas if it is later determined that such flow has undesirable consequences such as transfer of excess nutrients into a nutrient-poor region. As part of an overall evaluation of the likely success of the monitoring and adaptive assessment program in the Restoration Plan, it would be useful to conduct a preliminary, but systematic, inventory of opportunities for flexibility in design and operational features of the Restoration Plan components. A systematic review and listing of the opportunities for both design and operational flexibility for major Restoration Plan projects would be useful to identify which projects could be modified in an adaptive assessment process. Results of this inventory could be used in an evaluation of project scheduling and in prioritizing monitoring, experimental, and modeling activities to provide input to the components for which adaptive assessment is likely to have the greatest impact on decisions regarding design and operations. Much has been written about institutional barriers to adaptive ecosystem management, lamenting the difficulty of maintaining scientifically-based adaptive strategies in an environment of stakeholders, bureaucracies, and political processes (e.g., Holling, 1995; Gunderson, 1999). During the November 2001 workshop much discussion focused on systemic barriers to the Restoration Plan adaptive assessment strategy that need more attention. These include the concern that existing laws could override proposed modifications to water delivery systems and the need for a more thorough policy analysis. There was also concern that current

OCR for page 19
incentive systems in the Army Corps of Engineers and the South Florida Water Management District that reward personnel for completing large projects “on time and under budget” tend to promote inflexibility in decision making. This is not to say that delays and cost overruns are desirable, but to point out that there are no obvious incentives for taking a more precautionary approach to the restoration with more reliance on pilot projects, contingency planning, and non-structural solutions to achieve ecological goals. Others noted that the current organizational strategy does not provide a direct linkage between science and decision making related to water management. Finally, some workshop participants noted that there are no guidelines or “policies” for when to change the plan. This is especially difficult given multiple time lags between implementation of a restoration project and ecosystem responses at different spatial and temporal scales. Just as flexibility of Restoration Plan design and operation is crucial to adaptive assessment management of the Everglades, societal flexibility and acceptance of scientific uncertainty are essential to the adaptive assessment mix so that modifications of policy that require changes to ecosystem drivers to achieve restoration goals and objectives are understood and accepted. Education and outreach about the scientific issues is central to fostering societal flexibility and acceptance of uncertainty by the public, decision-makers, and legislators. As a result, education is central to the success of the project. Institutional arrangements for transferring science advice to policy-makers and education of the public must be clearly identified in the Monitoring and Assessment Plan. Specific arrangements need to be made to communicate scientific conclusions about the functioning of the ecosystem to the decision-makers in the executive and legislative branches of government. Linkages should be designed to connect the RECOVER Senior Management Team and the Science Coordination Team to the Restoration Plan decision-makers. The key is to have frequent conversations with the decision-makers to inform them of the changing state of knowledge, so that they can make decisions based on current scientific information. Another important linkage, which does not seem well developed thus far in the Everglades would be between the adaptive assessment process and citizen advisory groups. This linkage has been developed in other restoration efforts such as the Glen Canyon Dam project (NRC, 1996b). In summary, it is not clear if there is enough flexibility in the Restoration Plan design to provide opportunities to respond to ecosystem response “surprises” or indeed other operational and system changes that will probably arise during implementation of the plan. Most of the flexibility within the Restoration Plan appears to be related to operational features rather than primary construction. To maximize the potential to apply results of increased understanding of the ecosystem, project design should attempt to maximize the range of operational conditions. Monitoring and process studies should focus on hydrologic and ecological features for which improved prediction of response can lead to project modification that will lead to a more successful result.