I. OVERVIEW OF THE DEVELOPMENT AND APPLICATION OF THE CONCEPTUAL MODELS

John C. Ogden, SFWMD

A. INTRODUCTION

The Central and South Florida “Restudy” Project team and the Science Coordination Team of the South Florida Ecosystem Restoration Working Group have adopted the “Applied Science Strategy” as a process for linking the sciences and management during the planning and evaluation of the south Florida ecosystem restoration programs (SCT 1997, Ogden et al. 1997; Figure 1). The creation of a science application strategy was motivated by the need for a broadly accepted process for organizing and converting large amounts of existing technical information into planning and evaluation tools that would directly support the restoration programs. Some sort of organizing process is clearly required where restoration planning occurs on such large scales, where information from many disciplines is widely scattered in time and place, where focused efforts are needed in order to identify and fill gaps in our information base with “best professional opinion,” and where a large degree of consensus regarding the major cause and effect relationships is necessary. To be successful, it was felt that this strategy must, (1) lay out a scientifically reviewed sequence of steps and tasks for converting research and modeling results into planning objectives, performance measures and evaluation protocols, (2) serve as a strong catalyst for promoting consensus among scientists and managers regarding the nature of the principle resource issues, and the probable routes for resolving these issues, and (3) be a process that can contribute to the objectives and needs of both the scientific and management communities in the regional restoration programs. The Applied Science Strategy is a total systems and multi-disciplinary process for determining the most appropriate restoration targets, and the best measures for each of these targets, during and following the implementation of the restoration programs. These tasks are prerequisite to the successful application of “adaptive assessment” during the implementation of the restoration programs.

An essential step in the Applied Science Strategy is the creation of a set of conceptual ecological models of the major wetland physiographic regions in south Florida. These simple, non-quantitative models are an effective means for developing a consensus regarding a set of causal hypotheses, which explain the affects that the major anthropogenic stressors have on the wetland ecosystems. Each model identifies the attributes in the natural systems that are the best indicators of the changes which have occurred as a result of the stressors. Each model also delineates the ecological linkages between the stressors and the attributes and the most appropriate measures for each of the attributes. The development of a consensus regarding the components and linkages in the conceptual models is the first step in the process of reaching agreement on specific hydrological, ecological, and biological measures of restoration success, and for designing a regional, performance-based ecological monitoring program. Conceptual models have been widely used for similar purposes in other regions of North America (e.g., pp. 31-38 in Gentile 1996; also see Rosen et al. 1995).

This report presents the schematic and narrative descriptions for eight conceptual ecological models that have been used by the C&SF Restudy team. These eight models provide a basis for developing performance measures for evaluating alternative restoration plans, for designing a performance-based monitoring program, and for adaptive assessment of the south Florida ecosystem restoration programs. The eight models are for Lake Okeechobee, the Caloosahatchee and St. Lucie estuaries (northern estuaries), the ridge and slough Everglades, the Big Cypress basin, the southern Everglades marl prairies, the southern Shark Slough and Florida Bay mangrove estuaries (southern estuaries), Biscayne Bay, and Florida Bay.

B. METHODS

The set of conceptual models was initially developed and reviewed by over 100 scientists and resource managers who participated in a concentrated program of workshops between October 1996 and July 1997. The list of participating scientists, and a listing of the models that each participant helped to develop, are presented in Table 1. These workshops were open to all interested participants. Special efforts were made to invite the field scientists in south Florida who have had considerable “hands-on” research experience in these landscapes, and therefore were well qualified to bring knowledgeable professional opinions for these systems to the modeling discussions. The models were developed at landscape scales, shown in Figure 2.

The initial steps in the development of each model were to use the informal format of the workshops to identify and discuss the causal hypotheses which best explain the major anthropogenically driven alterations in each landscape. From these discussions, the participants created lists of the appropriate stressors, ecological effects, and attributes (indicators) in each landscape. The objective was to identify the physical and biological components and linkages in each landscape which best characterized the changes explained by the hypotheses. Preparers for each model used the hypotheses and lists of components to lay out an initial draft of the model, and prepared a supporting narrative document to explain the organization of the model and the supporting science for the hypotheses. The drafts and narratives were reviewed in subsequent workshops, resulting in revisions to the models.

A schematic diagram and a narrative description are provided for each conceptual model. The diagrams follow a top-to-bottom hierarchy of information, which identifies the societal drivers (external sources), the specific stressors on the natural systems, the ecological effects resulting from the stressors, and the recommended ecological attributes (indicators) and measures for each attribute. The symbols used in the models to indicate each of these model components are as follows.

Each stresssor is linked to one or more attributes. Measures of responses by the stressors and attribute in each model, and in all models combined, are recommended as the minimal set of components in any regionally comprehensive monitoring program for the purpose of determining the success of the restoration programs.

The major components of the models are defined as follows.

Drivers: The major external driving forces that have large scale influences on natural systems. Drivers can be natural forces (e.g., sea level rise) or anthrogenic (e.g., regional land-use programs).

Stressors: The physical or chemical changes that occur within natural systems that are brought about by the drivers, and which cause significant changes in the biological components, patterns and relationships in natural systems.

Ecological effects: The biological responses caused by the stressors.

Attributes: Also known as indicators or endpoints. A parsimonious subset of all potential biological elements or components of natural systems, which are representative of the overall ecological conditions of the system. Attributes typically are populations, species, guilds, communities or processes. Attributes are selected to represent the known or hypothesized effects of the stressors (e.g., numbers of nesting wading birds), and the elements of the systems that have important human values (e.g., endangered species, sports fishing).

Measures: The specific feature(s) of each attribute to be monitored to determine how well that attribute is responding to projects designed to correct the adverse effects of the stressors (i.e., to determine the success of the project).

Critical pathways: The known or hypothesized linkages among stressors, ecological effects and attributes, which explain a high percentage of the adverse ecological responses shown in the conceptual model (e.g., the ecologically most significant pathways leading from a stressor).

Each model narrative includes, (1) a brief introduction to the dynamics and problems of the landscape, (2) descriptions of specific stressors, ecological attributes and measures, (3) descriptions of the major ecological linkages affected by the stressors, and (4) an initial set of recommended performance measures and restoration targets for the stressors and attributes.

Earlier drafts for the four southern Everglades models (ridge and slough, marl prairies, mangrove transition, and Florida Bay) were previously reported in Ogden et al. 1997. These models should be viewed as “working tools” in support of the restoration programs. The models will undergo regular review and revision as new information becomes available.

The organization and content for each model, and the supporting text, were prepared by an editor (or editors), based on the information collected from the workshops. Each model report in this document is a product of the editor(s).

C. DISCUSSION

The Applied Science Strategy that is being used to link the sciences and management in the south Florida ecosystem restoration programs is derived from the “Ecological Risk Assessment” process of the U.S. Environmental Protection Agency (EPA 1992, Gentile 1996). Risk assessments are used by the EPA as a guideline for conducting ecological evaluations of proposed management actions in both North America and Europe. A similar approach was used by the Man and the Biosphere Human Dominated Systems program to define sustainability goals and identify ecological endpoints for a series of restoration scenarios for south Florida (Harwell and Long 1992, Harwell et al. 1996).

Gentile (1996) suggested that risk assessment and recovery of ecological systems “…can be viewed as opposite sides of the same coin.” Risk assessment “… is the process of determining the probability (with associated uncertainty) of a particular event occurring as a result of the action of a specific agent or stressor…” Recovery of ecological systems “…can be viewed as the process for determining the probability (with associated uncertainty) of a particular event occurring (e.g., recovery to a…ecologically desired sustainable state) as the result of mitigating the action of a specific agent (e.g., canals, berms) or stressor (e.g., phosphorus). Gentile (1996) listed three principal functions of the risk assessment process: 1) identification of potential causal relationships between stressors and effects; 2) selection of endpoints (attributes), indicators, and success criteria; and 3) development of a scale-dependent conceptual model that describes the inter-relationships between multiple stressor pathways and multiple ecological receptors.

The development and application of conceptual models provide benefits to the scientists who create the models, and to the managers and public who use them to guide and implement resource policy (Gentile 1996, Ogden et al. 1997, SCT 1997). The process of creating and reviewing conceptual models aids scientific endeavors by, 1) establishing a forum for open, multi-disciplinary exchanges of ideas and information pertaining to complex ecological issues; 2) developing scientific consensus regarding current understandings of ecosystems; 3) creating working hypotheses which serve to guide both research and management; 4) better defining (and reducing) areas of scientific uncertainty; and 5) providing a framework for continuing discussions and revisions as new information becomes available. For managers and the public, the models serve to, 1) de-mystify the science; 2) provide a means for converting broad policy level goals and objectives into specific, measurable targets; 3) provide a visual description of the rationale for the prevailing hypotheses and management priorities; 4) reduce the complexity and dimensionality of the problems; 5) separate essential from non-essential information; 6) provide a tool for improved communication.

Gentile (1996) summarized the importance and appropriateness of the risk assessment and conceptual model process for setting ecological targets and conducting ecological evaluations in regional ecosystem restoration and management programs. The points emphasized by Gentile were, 1) that there is a large body of peer-reviewed literature that describes these methods and processes for conducting ecological assessments; 2) this literature represents an internationally accepted framework for structuring assessments; and 3) considerable effort has been devoted during recent years to formalize the process of selecting and classifying the endpoints, indicators, and metrics used in assessments.

D. CONCEPTUAL MODELS IN SOUTH FLORIDA

The overall Restudy and SERA strategies are to use the conceptual models as a basis for developing performance measures and targets for the stressors and attributes in each model. These targets, collectively, describe the physical and biological conditions that will be used to define successfully restored natural systems. The rationale for having performance measures and targets for each stressor is that the stressors are known or hypothesized to be the immediate causes of the ecological problems in each landscape. A successful restoration program must remove the adverse affects created by each stressor. A performance measure identifies which elements of each stressor must be corrected, how those elements should be measured, and how those elements must change (i.e., the restoration target) in order to eliminate or reduce their adverse effects.

Performance measures are also developed for each attribute in the conceptual models. The attributes have been identified as the biological or ecological elements that are the best indicators of responses in the natural systems to the adverse effects of the stressors. The hypotheses used to construct the conceptual models show how each attribute is linked to the stressor(s) which are most responsible for change in that attribute. If the hypotheses are correct, reducing the adverse affects of the stressor will result in a predictable positive response by the attribute. The performance measure developed for each attribute identifies the element of that attribute which should respond, how that element should be measured, and how that element should respond (i.e., improve, thus setting the restoration target for that attribute) once the affects of the stressor are removed.

The attributes that are selected must provide measures of responses at appropriate temporal and spatial scales, and from different taxonomic and hierarchial levels in ecological systems. They must also be measurable, and their historical patterns, relationships and functions well enough understood, so that responses can be correctly determined and interpreted. The model teams also considered the need for a mix of attributes that can serve either as indicators of ecological conditions (e.g., diversity and production) or of societal priorities (e.g., endangered species and high quality sports fishing). Finally, the model teams considered the most appropriate number of attributes for each model. The desirable trade-off was to have a sufficient number to adequately reflect the major system responses, while not having more than would be necessary for this purpose or that would contribute to an unmanageably large monitoring program.

The modeling teams also identified critical linkages in the conceptual models. Critical linkages were defined as those ecological links between one or more stressors and attributes, which seem to explain most of the ecological and biological changes in the systems. The assumption was that the stressors that are part of critical linkages should have a priority during restoration planning over other, less influencial stressors. The Restudy alternative evaluation team (AET) generally placed higher priority on meeting performance targets for stressors with strong critical linkages to the attributes than for other stressors in the models. For example, the Everglades ridge and slough conceptual model suggests that stresses caused by reductions in the duration of uninterrupted surface water hydroperiods and by compartmentalization may explain much of the adverse changes that have occurred in the long-hydroperiod slough systems. Correction of these stressors became the highest priority in evaluating the predicted affects of alternative restoration plans in the sloughs of the Everglades.

The performance measures and targets from both the stressors and attributes are used for two primary purposes. Performance measures function during the planning phases of the restoration projects as assessment tools for alternative plan design. What combination of structural and operational components is most likely to achieve the desired objectives of the project, as determined by how well a plan is predicted (by modeling) to meet each of the targets set by the performance measures? In this planning role, the measures and targets not only are used to determine which plan is most likely to be successful in achieving its objectives, but also are used to influence the design of the project as efforts are made to evaluate the combination of features which can best moderate or eliminate the adverse affects of the stressors.

The second primary use of the measures and targets is in the design of a system-wide, ecological monitoring program. The purpose of system-wide monitoring is to measure how elements in the natural system actually respond following the implementation of each iteration of a restoration program. Because the design of a restoration program is determined by the features that are predicted to best meet a suite of regional performance measure targets, the monitoring program must measure the responses by the same set of performance targets. Comparisons between the predicted responses among stressors and attributes and the actual responses among the same stressors and attributes, provide a basis for making revisions to the causal hypotheses and conceptual models, and a means for structuring an adaptive assessment strategy throughout the implementation of the restoration program. The overall focus of the monitoring program must be performance-based, i.e., the elements of the natural systems to be monitored must be those that provide actual measures of the stressor and attribute targets.

E. ADAPTIVE ASSESSMENT

The key to successful implementation of a regional ecosystem restoration program, e.g., maximizing the effectiveness of the program and reducing uncertainty during its implementation, is an adaptive assessment strategy. The conceptual ecological models are an integral part of this overall process. The conceptual models are revised based on interpretations of actual system responses following the implementation of each restoration project. These revisions in the conceptual models influence predictions of system responses, and the choice of attributes and measures, for future iterations of the restoration program. As the accuracy of the hypotheses in the conceptual models improve, the opportunities for making pre-construction improvements in project design are also enhanced. Greater accuracy in the conceptual models leads to improvements in the choice of measures and targets used to refine the design and judge the predicted performance of each project design.

F. FUTURE TASKS

The conceptual ecological models, and the planning and evaluation tools coming from the models, still require several priority improvements. These are: 1) completion of a conceptual model technical report, and peer review of this report; 2) new workshops to complete the development of performance measures for the biological attributes identified by the models; 3) workshops to design a performance-based monitoring program for the south Florida ecosystem restoration program; and 4) the development of a consensus among the participating agencies for a system-wide adaptive assessment strategy for the ecosystem restoration program.

Table 1. List of participants and model workshops

Name	AFFILIATION	Models
G. Abbott	SFWMD
K. Alvarez	FDNR	BC
C. Anastasiou	Univ. FL	FB
T. Armentano	NPS/EVER	R/S, MP, SE, FB
A. Arrington	SFWMD
T. Bancroft	Arch. Biol. Stat.	R/S, MP, SE, BC
I. Barnett	FDEP
O. Bass	NPS/EVER	R/S, MP, SE
S. Bellmund	NPS/EVER	R/S, MP, SE, BB
R. Best	USGS/BRD	FB
R. Brock	NPS/EVER	FB
J. Browder	NOAA/NMFS	R/S, MP, SE, FB, BB
C. Buckingham	FWS	BC
G. Burzycki	Dade Co. DERM	R/S, MP, SE, BB
D. Busch	NPS/EVER	R/S, MP, SE
K. Cairns	FWS	NE
E. Carlson	Natl. Audubon Soc.	BC
P. Carlson	FDEP/FMRI	FB
B. Chamberlain	SFWMD	NE
E. Chipouras	Univ. FL	FB
J. Colvocoresses	FDEP/FMRI	FB
W. Cropper	U. Miami/RSMAS	FB, BB, SC
K. Cummins	SFWMD	SC
S. Davis	SFWMD	R/S, MP, SE, BC, FB, SC
R. Day	Ind. R. Lagoon NEP	NE
D. DeAngelis	USGS/BRD	SC
P. Doering	SFWMD	NE
M. Duever	The Nature Conserv.	BC
M. Durako	FDEP/FMRI	FB
J. Englehardt	U. Miami	SC
E. Estevez	Mote Marine Lab	NE
M. Gaines	U. Miami/Biology
D. Gawlik	SFWMD	R/S, SE
J. Gentile	U. Miami/RSMAS	SC
C. Goodyear	NOAA/NMFS	LO
G. Graves	FDEP	NE
S. Gray	SFWMD	NE, FB
L. Gunderson	U.Florida/Zoology	BC, FB
D. Hanisak	Harbor Branch Inst.	NE
C. Hanlon	SFWMD	LO
M. Harwell	U.Miami/RSMAS	SC
D. Haunert	SFWMD	NE
K. Havens	SFWMD	LO
L. Heisler	FGFWFC	LO, R/S, SE
L. Hornung	USACOE	SC
A. Houser	USACOE	BC
Soon-Jin Hwang	SFWMD	LO
B. Irlandi	U.Miami/RSMAS	FB, BB
T. James	SFWMD	LO
T. Janicki	Coastal Environ.	NE
D. Jansen	NPS/BICY	BC
Kang-Ren Jin	SFWMD	LO
K. Kibby	Lee County	NE
W. Kitchens	USGS/BRD	SC
M. Koch	FL Atlantic Univ.	SE, SC
J. Koebel	SFWMD
B. Kruczynski	EPA	FB
J. Layne	Arch. Biol. Station	BC
D. Lirman	U.Miami/RSMAS	BB
W. Loftus	USGS/BRD	R/S, MP, SE
J. Lorenz	Natl. Audubon Soc.	SE, FB
L. McCarthy	FDACS	LO, R/S, MP, SE
C. Madden	SFWMD	NE, R/S, MP, SE, FB
L. Manners	USACOE	LO
S. Markley	Dade Co. DERM	BB
F. Mazzotti	U.Florida/IFAS	R/S, MP, SE, BC, FB
J. Meeder	FL.Int.Univ./SERP	MP, SE
S. Melvin	SFWMD
R. Montgomery	Coastal Environ.	NE
D. Morrison	Natl. Audubon Soc.	R/S, MP, SE, FB
J. Moulding	USACOE	R/S, MP, SE
A. Nath	SFWMD
S. Newman	SFWMD	SC
J. Obeysekera	SFWMD	SC
J. Ogden	SFWMD	R/S, MP, SE, BC, FB, SC
S. Olson	SFWMD	FB
P. Ortner	NOAA/AOML	FB
R. Pace	FWS	LO, NE, R/S, MP, SE
T. Pernas	NPS/BICY	BC
S. Perry	NPS/EVER	R/S, MP, SE
M. Poole	FGFWFC	R/S, MP, SE
R. Punnett	USACOE	SC
G. Redfield	SFWMD	LO
C. Regan	NPS/EVER	FB
L. Richardson	FWS	BC
J. Rippe	SFWMD
M. Robblee	USGS/BRD	R/S, MP, SE, FB
B. Robertson	USGS/BRD	R/S, MP, SE, SC
B. Rosen	SFWMD	LO, NE
D. Rudnick	SFWMD	NE, MP, SE, FB
P. Sime	SFWMD	NE
F. Sklar	SFWMD	R/S, SE
T. Smith	USGS/BRD	SE
S. Snow	NPS/EVER	R/S, MP, SE
J. Snyder	USGS/BRD	BC
M. Steinkamp	FWS	R/S,MP,SE
A. Steinman	SFWMD	LO
P. Strayer	SFWMD
D. Strom	FDEP	NE
T. Tisdale	SFWMD	LO
S. Tosini	U.Miami/RSMAS
L. Toth	SFWMD
T. Towles	FGFWFC	R/S,MP,SE
S. Traxler	USACOE	NE, SE, FB
J. Van Arman	SFWMD	NE, BC
G. Warren	FGFWFC	LO
D. Weeks	NPS/BICY	BC
D. Welcher	NOAA/AOML	FB
H. Zebuth	FDEP	LO, NE, R/S, MP
M. Ziminske	USACOE	LO

* Models: Lake Okeechobee (LO); northern estuaries, Caloosahatchee & St. Lucie (NE); Everglades ridge & sloughs (R/S); southern marl prairies (MP); southern Florida Bay/Shark Slough estuaries (SE); Big Cypress basin (BC); Florida Bay (FB); Biscayne Bay (BB); model steering committee and overall review (SC).

Figure 1: APPLIED SCIENCE STRATEGY FOR ECOSYSTEM RESTORATION

Figure 2: LOCATION MAP FOR CONCEPTUAL MODELS

G. REFERENCES

Environmental Protection Agency. 1992. Framework for ecological risk assessment. EPA/630/R-92/001. Washington, D.C.

Gentile, J.H. 1996. Workshop on “south Florida ecological sustainability criteria.” Final Report. University Miami, Center for Marine and Environmental Analysis, Rosenstiel School of Marine and Atmospheric Science, Miami, FL. 54 pp.

Harwell, M.A. and J.F. Long. 1992. U.S. M.A.B. Human-Dominated Systems Directorate workshop on ecological endpoints and sustainability goals. Univ. Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, FL. 115 pp.

Harwell, M., J.F. Long, A. Bartuska, J.H. Gentile, C.C. Harwell, V. Myers and J.C. Ogden. 1996. Ecosystem management to achieve ecological sustainability: the case of south south Florida. Environ. Mgmt. 20: 497-521.

Ogden, J.C., S.M. Davis, D. Rudnick and L. Gulick. 1997. Natural Systems Team report to the Southern Everglades Restoration Alliance. Final draft. July 1997. South Florida Water Management District, West Palm Beach, FL. 43 pp.

Rosen, B.H., P. Adamus and H. Lal. 1995. A conceptual model for the assessment of depressional wetlands in the prairie pothole region. Wetlands Ecology and Management 3: 195-208.

Science Coordination Team. 1997. Integrated science plan. Report to the South Florida Ecosystem Restoration Task Force and Working Group. Office of the Executive Director, SFERTF, Florida International University, Miami, FL.