REPORT ON THE FLORIDA BAY SEAGRASS MODEL WORKSHOP

KEY LARGO, FLORIDA
12-14 JANUARY 1998

Submitted to the

PROGRAM MANAGEMENT COMMITTEE
FLORIDA BAY RESEARCH PROGRAM

Prepared by the

AD HOC REVIEW PANEL

Dr. Susan L. Williams, Chair, San Diego State University
Dr. Kenneth H. Dunton
, University of Texas
Dr. Kenneth L. Heck, Jr.
, Dauphin Island Sea Lab, University of South Alabama
Dr. W. Michael Kemp
, University of Maryland

2 February 1998

SUMMARY RECOMMENDATIONS

The major recommendations of the review Panel for the Florida Bay Seagrass Modeling Workshop are given below.

1. Seagrasses should be the central focus of Florida Bay management and scientific efforts because of the considerable influence they have on hydrodynamics, sedimentary processes, nutrient cycling, and other organisms.

2. The development of a predictive model for management objectives will require a long term, well-funded and well-coordinated program. In the interim, modeling will be a useful tool for understanding seagrass ecology in Florida Bay and will help scientists progress toward the ultimate goals of the Strategic Plan for the Interagency Florida Bay Science Program

3. A modeling approach should include several different kinds of seagrass models to adequately address the complex issues surrounding Florida Bay seagrasses. Appropriate approaches include:

Unit physio-ecological process models
Demographic models
Statistical models
Landscape models

4. The conceptual scenario for the original Thalassia die-off and subsequent seagrass declines put forth in the Strategic Plan can be evaluated only after translation into testable hypotheses followed by empirical research.

5. Coordination among managers, modelers, seagrass biologists and other scientists working on Florida Bay problems is essential for a successful effort.

6. Goals and expectations for the models should be made explicit in order to plan the research program and to set reasonable timelines.

7. The unit process model for Chesapeake Bay eelgrass and other species is not appropriate for Florida Bay seagrasses without significant modifications. Evaluation of alternative existing models or new ones is recommended before COE/WES uses the Chesapeake Bay unit model as a component in the Water Quality Model they are developing.

8. Restoration goals need to be developed more explicitly to help focus modeling efforts. The goal of hydrologic restoration is too narrow to encompass all the ecological issues perceived to affect seagrasses in Florida Bay.

9. Historical data on seagrass distributions, abundance, and species composition is critical to obtain for setting restoration goals and general understanding of changes in Florida Bay.

10. Modeling cannot be done without research. The seagrass research and modeling program should be multi-disciplinary and multi-institutional, resulting in collaborative efforts among scientists with experience in Florida Bay and other ecosystems.

11. It is imperative that there is no further delay in initiating a coordinated seagrass research program. The Workshop provided the basis for an RFP. We recommend:

Immediate allocation of some of the Department of Interior's Critical Ecosystem Science Initiative funding for the monitoring of key parameters identified in conceptual models (e.g.,. light, biomass, hydrographic data) and statistical modeling of existing data. Proposals for other studies (e.g.,. experimental research, model development, etc.) must not be included under this partial funding allocation.

Rapid posting of the RFP for research proposals and timely peer-review.

Development of a plan to sustain funding for the long-term research effort required to meet management and restoration goals set forth in the Strategic Plan.

12. Recommendations made here and by previous review panels with respect to the organization of the program management need to be pursued.

INTRODUCTION

The Strategic Plan for the Interagency Florida Bay Science Program was prepared by the Florida Bay Program Management Committee (PMC) and formalized in March 1997 after extensive internal and external review (Florida Bay Program Management Committee 1997). The Strategic Plan was developed to "focus the resources of the member agencies on a research strategy to provide science information critical to the restoration of Florida Bay." The plan was designed around five questions that center on recent changes in the Florida Bay ecosystem that precipitated public concern over biological resources and indicators of ecosystem decline, particularly seagrass mortality and algal blooms. The Plan states that "the primary thrust of the Florida Bay program is to test the validity" of the hypotheses for ecosystem changes "so as to understand the effects of past human actions and to provide a scientific framework for testing new hypotheses if warranted." In addition, a major objective of the Florida Bay Science Program is to provide the scientific knowledge requisite to help define restoration goals, "to predict system response to management actions, and to establish success criteria." The Florida Bay Program is one scientific component of the large South Florida Ecosystem Restoration Initiative headed by a Task Force.

The Strategic Plan outlines an approach that combines three elements: 1) research leading to the understanding of the causality and mechanisms underlying ecosystem changes, 2) continued and expanded surveys and monitoring, and 3) simulation modeling of ecological processes. As a step toward reaching Program objectives, the Strategic Plan proposes a series of Workshops addressing critical research issues and questions. The Workshop on Seagrass Modeling was convened by the PMC to help address Central Question #4:

What are the causes and mechanisms for the observed changes in the seagrass community of Florida Bay? What is the effect of changing salinity, light, and nutrient regimes on these communities?

The Strategic Plan outlines development of one or more ecological seagrass models that "should be capable of key processes including those connected with putative causes of the die-off." The Strategic Plan provides the expectation that the model(s) should help to address each of the seven hypothetical causes proposed for seagrass die-off in the Plan:

1) altered freshwater flows to the Bay, including relationships to hypersalinity;

2) overmaturity and susceptibility of Thalassia beds;

3) reduction in storm-mediated disturbance of the sediments and seagrass beds;

4) altered sediment chemistry (e.g., sulfide build-up, iron limitation);

5) disease spread;

6) unusually severe climatic conditions at die-off initiation;

7) altered nutrient regime.

An important function of the seagrass model, as stated in the Strategic Plan, is to provide the requisite information for a landscape model within which growth, survival, and recruitment of key benthic species may be simulated.

 Workshop Organization and Review Panel Charge

The PMC organized a Seagrass Team, consisting of the PMC delegate (Mike Robblee) and Seagrass Team co-leaders (Michael Durako and Jay Zieman). The Seagrass Team also includes modelers and researchers working in Florida Bay. The Seagrass Team prepared a Workshop agenda with the assistance of other team members (Drs. Penny Hall and Chris Madden). The Florida Bay Scientific Oversight Panel (FBSOP) selected a panel of independent scientists with expertise in seagrass ecology and management but not involved in Florida Bay research to provide broad technical and management review of the Workshop. Susan Williams, a member of the FBSOP , was appointed as Chair of the Review Panel for the Seagrass Model Workshop. Members include Ken Dunton, Ken Heck, and Michael Kemp. Workshop participants were selected from seagrass biologists working in Florida Bay and similar seagrass systems, modelers with relevant experience, and resource managers.

While charges made to the Review Panel were broad, the Workshop products were identified by the Seagrass Team as:

1) Required elements for a seagrass model for Florida Bay;
2) Prioritized listing of empirical data and research needs;
3) Detailed timeline.

In accordance with the Strategic Plan, the goal of the modeling strategy was identified as:

"To evaluate responses of seagrasses to natural and managed hydrologic regimes."

Therefore, this report is written in the context of these three objectives and the overarching goals of the Strategic Plan.

History of Modeling and Strategic Plan Progress Relevant to a Seagrass Model

In October 1996 the Model Evaluation Group (MEG) for the Workshop on the Design and Specifications for the Florida Bay Water Quality Model reviewed the modeling efforts for Florida Bay, incorporating recommendations from reviews of previous workshops on Florida Bay models (e.g., the hydrodynamic and nutrient models, D'Elia et al. 1996). The recommendations of the MEG most relevant for the Seagrass Model were:

1) A seagrass model must be an element of an integrated water quality-sediment-seagrass model;

2) The integrated model must include 'longer-term stable' and 'shorter-term opportunistic' seagrass species.

In arriving at these recommendations, the MEG considered information from previous Workshops on Florida Bay on hydrodynamic modeling and nutrient dynamics.

Panel Summary of the Workshop

The Workshop was organized into 15 to 20 minute presentations by 15 scientists. The presentations on the first day were devoted to background information on Florida Bay seagrasses and the seagrass decline. Presentations on the second day were devoted to modeling programs in Florida Bay or to descriptions of relevant models. The Workshop organizers allowed flexibility in the schedule for more effective discussion of research questions and needs, particularly on the second day. Major points of general agreement are summarized here.

Conceptual Model of Seagrass Die-Off

The Zieman-Durako scenario (Fig. 2 in the Strategic Plan) for seagrass die-off was accepted by consensus as a good working model for directing research hypotheses. In brief, the scenario describes a doubling in Thalassia biomass over the decade before the die-off, during a period when hurricane frequency was low and the Bay was largely hypersaline, followed by a precipitous die-off of only Thalassia in the basins with the densest vegetation. Hypothetical triggers for the decline include a negative carbon balance due to stress, sulfide toxicity, and disease.

Details of how to put this conceptualization to use in a research program were discussed. Various researchers place different weight on these postulated causative factors based on their individual perspectives. It was concluded that without research, the relative importance of causal factors or their synergism cannot be ascertained. In addition, the relation of "post die-off" declines of seagrasses (Halodule and Thalassia) have not been fully reconciled with this conceptual model. These recent declines appear to be associated with phytoplankton blooms and wind-generated sediment resuspension, and are occurring despite remission of hypersalinity conditions. It thus appears that these recent trends represent responses that are different from ones that caused die-off of Thalassia.

The Central Role of Seagrasses in Florida Bay Ecology

Discussions at the Workshop re-emphasized the importance of seagrasses in the Florida Bay ecosystem for sedimentary and biogeochemical processes (water column and benthic) and living resources.

Major Ecological Regions in Florida Bay

It was agreed among Workshop participants that critical regional differences in physiography, seagrass distribution, and general ecology need to be recognized in a research program. Another consensus was that the species composition of seagrasses in Florida Bay differs over physiographic regions within the Bay, from the northeast where water management efforts are currently concentrated to the west where the Gulf of Mexico apparently has considerable influence, to more local bank-basin features. The time constant for models might vary for the margins versus the center of the Bay.

Identification of Research Questions

Research gaps and questions were listed after existing data on Florida Bay seagrasses were presented. Listing was done with a view toward future critical evaluation of the conceptual model and the current declines in seagrasses. Although time was too short for consideration of details, the basic needs for a RFP were addressed.

It was evident that developing a seagrass model was very ambitious given the research gaps and questions identified. The approach to model development needs to incorporate the building of a data base for model development and calibration. The approach would have to be flexible and include various types of models and components because of the diversity of questions and problems evident. These include an understanding of seagrass physiology and growth in relation to physical factors of light, temperature and salinity and their interaction, how stress affects susceptibility to disease, the effects of seagrasses on sediment resuspension, shear stress on the bottom, and diagenetic processes in carbonate sediments. Salinity, despite its critical importance for Florida Bay, provides an example of the critical dearth of data. Whether hypersalinity was present historically is still equivocal and it is clear that current salinity regimes cannot be represented by the simple sinusoidal relationships used in many models (presentation by E. Chipouros).

Appropriate Modeling Approaches

The relative merits of different modeling approaches were discussed, including: 1) unit seagrass physio-ecological process models; 2) demographic models of seagrass growth; 3) statistical (also termed "empirical" models by Workshop participants) models of seagrass distribution in relation to environmental variables; and 4) spatially explicit landscape models of seagrass dynamics.

The unit seagrass physio-ecological process models would be based on understanding how physical factors drive photosynthesis and carbon and nutrient allocation to biomass. The basic structure for this unit model could be borrowed from existing ones for eelgrass or other Submerged Aquatic Vegetation (SAV) species in the Chesapeake Bay or from the Halodule productivity model developed for the COE by Burd and Dunton for the Laguna Madre. These structures would need to be modified for Florida Bay application. Unit process models are particularly useful for understanding how environmental parameters affect seagrass photosynthesis and carbon and nutrient balances.

Demographic models would be a complementary approach to understanding seagrass production. Such models could be extended from unit models if allocation of photosynthate between meristem and short shoots were known and alternatively from a knowledge of short shoot recruitment, mortality, and rhizome elongation. An existing data set collected as a Florida DEP project is available to begin a modeling effort based on the latter approach. Demographic modeling can explain seagrass population expansion and contraction.

The use of multivariate statistical ("empirical") models could be developed immediately to help isolate variables for further study. There are water quality data sets collected over the past few years that could be matched with seagrass data (shoot #, biomass, etc.) to investigate patterns of association to explore whether there are strong relationships between water quality variables and seagrass variables that might suggest concentrating on a subset of the water quality variables in the future. In addition, the presence of strong relationships between water quality variables and seagrass indices may allow more focused efforts on subsets of the water quality variables in the future.

Because landscape models include a spatial element, this approach would be useful in understanding: 1) the importance of the western edge of the Keys as a filter for the reef tract; 2) the importance of Syringodium beds in the west as a control on the magnitude of turbidity plumes; 3) disease transmission; and 4) organizing principles for seagrasses over the entire Florida Bay system.

It was the consensus of this Workshop that the Florida Bay seagrass modeling program should utilize a suite of complementary modeling approaches rather than one model.

Seagrass in the Florida Bay Water Quality Model

The appropriateness of the seagrass component of the Chesapeake Bay water quality model for Florida Bay was discussed because it is being considered for use in the WES simulation. The Chesapeake Bay model is a very well-developed, working predictive model. The consensus was that the Chesapeake Bay seagrass model, per se, is not appropriate for Florida Bay for a number of reasons.

1) The size of the seagrass variable in the model will be huge relative to other components.

2) The allocation of biomass in Thalassia is primarily belowground. This pattern is very different from SAV species in the Chesapeake Bay.

3) Extensive clonality in Thalassia results in growth, biomass, and survival responses lagging up to 6 months beyond short-term responses of photosynthesis and leaf elongation.

4) The term for photosynthate (or nutrient) transfer between leaves and below ground tissues and along the rhizome has not been established for morphologically simpler plants as eelgrass, and will require extensive research for Thalassia.

5) The physiography of Florida Bay is very different that of the Chesapeake Bay. The mean water depth is 1.5 m and the Bay currently is divided into basins isolated by banks. Hypersalinity and carbonate geochemistry are major features that are not included in the Chesapeake Bay model.

6) Rhizophytic benthic seaweeds and several species of seagrasses (Thalassia, Syringodium, Halodule) and more estuarine submersed angiosperms form complex communities that vary in composition over broad geographical regions but also locally due to unknown physical and biological factors.

7) Seagrass epiphytes are primarily faunal or calcareous, and few data exist on important controlling factors. This contrasts greatly with the dominant diatom epiphyte community of the Chesapeake Bay.

8) Light in Florida Bay is attenuated by chlorophyll but also by suspended carbonate particles. Attenuation coefficients for carbonate-dominated particulate matter have not been described.

9) The role of seagrasses in nutrient cycling in the water column and carbonate sediments is poorly understood, as emphasized by the MEG (D'Elia et al. 1996).

It was suggested that it would be easier to develop a Florida Bay seagrass model de novo than to try to modify the code for the Chesapeake Bay model. The existing COE/WES model for Laguna Madre Halodule in clastic sediments and another model for seagrasses in southern Florida (Fong et al. 1997) could be evaluated for their appropriateness for Florida Bay. The Laguna Madre has important similarities to Florida Bay; Thalassia and Halodule are the dominant competitive species that cover 90% of a shallow (1 m) subtropical lagoon characterized by high salinities and limited nutrient inputs. One difference is that the substratum is not carbonate as in Florida Bay.

Restoration Goals

Restoration goals will need to be better developed to tie modeling efforts to restoration. The discussion of restoration goals, and thus the ultimate objective for understanding seagrass ecology and modeling, was difficult for the participants because resource managers have not articulated specific restoration goals other than re-establishing historic freshwater discharge into Florida Bay. Because the effect of this discharge on the entire Bay is unknown but might have limited effect in the western and southern regions, seagrass restoration goals will probably be more complex. For example, if sediment turbidity plumes prevent re-establishment of seagrass beds in the west, then an engineering approach to mitigation of turbidity, or a biological approach using extensive seagrass transplantation, could be required. In addition, considerations of the value of seagrass beds for human utilization include several different restoration scenarios.

Panel Review of the Seagrass Workshop & Recommendations

Despite the very short time between announcing and convening the Workshop, the Panel was pleased that several important objectives were achieved. A consensus was reached on the conceptual model for Florida Bay seagrass ecology and on global research questions. In addition, specific research hypotheses relevant to testing the conceptual model were identified. Unfortunately, there was insufficient time to prioritize these or discuss specific details and data needs. Nonetheless, the elements for a RFP for seagrass research were outlined. The seagrass research team is commended for keeping the schedule flexible and for responding to the suggestions of the Panel. Although there was no time to discuss the planning agendas and questions developed for the Workshop canceled in June 1997, these documents will be very useful for RFP development. Because of the short notice, some key participants were unable to attend, and most of the Panel had very little time to prepare. In order to achieve maximum benefit from a variety of experts in the future, better and more timely organization of the seagrass program is encouraged.

Funding for Seagrass Research

Despite the recognized importance of seagrasses in Florida Bay, comprehensive and coordinated seagrass research, current or retrospective, has not proceeded to any appreciable degree. Beyond the allocations made by independent agencies to supplement existing programs, very few funds have been available for seagrass research. As a result, many of the hypotheses developed and reviewed by the FBSOP over the past 4 years remain unsupported by data. The new data that are available resulted from small projects and the persistence of investigators concerned with Florida Bay. Until a coordinated new seagrass research program is underway, there is little chance that the questions regarding seagrasses in Florida Bay will be addressed or that existing data sets on seagrass distribution and abundance and physical parameters can be analyzed. Developing a coordinated well-funded program for seagrass research is absolutely necessary for the success of the Strategic Plan.

Articulation of Testable Hypotheses

A major step forward occurred in translating the broad conceptual model for seagrass die-off and subsequent declines into mechanisms and testable hypotheses with identifiable response variables. The Panel encourages the seagrass team to continue this effort.

The Panel echoes the concern of the FSBSOP on the assumption that salinity stress initiated the die-off and that reduced freshwater flow was the major cause of hypersalinity throughout the Bay. Focus on freshwater management has the potential to preempt other serious problems, e.g., turbidity (phytoplankton and carbonate) that affect seagrasses.

Because the seagrass issue in Florida Bay is very complex, the objectives of the seagrass program need to be prioritized in a hierarchy based on needs for management and on reasonable time estimates for accomplishment. There is little scientific understanding of the most basic principles that control the productivity, species composition, and distribution of seagrasses in Florida Bay. These include photosynthesis and its relationship to salinity and temperature, minimum light requirements, nutrient uptake rates, carbonate geochemistry, interactions among species, partitioning of carbon, and recruitment/colonization processes.

Modeling cannot be done in the absence of research. In addition, a comprehensive monitoring program needs to proceed to support research needs and model development and calibration. The Chesapeake Bay modeling program provides a valuable example of the effort required to develop a useful model for management. An initial research program lasting a decade was funded before serious modeling efforts were successful. The recent (1998) development of a new model for Halodule in the Laguna Madre of Texas (funded by the COE) was possible in two years with $4.5 million largely because of the existence of nearly continuous measurements of situ light and seagrass biomass from a previous decade of research.

Critical Areas for Seagrass Research

The Panel endorses the critical areas for seagrass research in Florida Bay that were identified during this Workshop. Among the most important research areas agreed on by Workshop participants include:

1) effects of salinity, temperature, light, nutrients, and their interactions on photosynthesis, respiration, carbon balance, growth and distribution of Florida Bay seagrasses;

2) effects of stress (reduction in carbon or nutrient acquisition) on plant survival and susceptibility to disease;

3) carbon and nutrient allocation within and among seagrass short shoots;

4) seagrass nutrient requirements and uptake kinetics for leaves and roots and their interaction;

5) limitation of growth and photosynthesis by inorganic carbon;

6) diagenetic processes in carbonate sediments;

7) controls of seagrass species composition and distribution;

8) evaluation of the importance of grazers, particularly mesograzers, as epiphyte controls that could mask responses to nutrient loadings;

9) light (PAR, PUR) attenuation and partitioning in the water column and epiphytic matrix;

10) rates and controls of evaporation in Florida Bay waters.

The workshop participants identified many scientists that either have expertise relevant to ecology and biology of seagrasses or in the design of plant demography and production models. It is imperative that these individuals are given the opportunity to use their expertise in addressing various hypotheses raised by the conceptual models or in developing an integrated model for Florida Bay. Effective communication among modelers, scientists, and managers is critical.

Seagrass Modeling Approaches

The Panel endorses a flexible, multifaceted approach to modeling seagrasses. We recommend that the Department of Interior seagrass research program consider support for four complementary modeling approaches: 1) unit physio-ecological process models; 2) demographic plant population models; 3) statistical models of seagrass distribution and environmental conditions; 4) spatially explicit landscape models. Not only are these four approaches complementary, but their development can (and should) be coordinated toward a long-range objective of generating a spatial model of abundance, distribution, growth, physiology and ecology. Each of these modeling approaches has merit for particular suites of hypotheses and research questions. As with empirical research supported in this program, modeling studies must incorporate the physiographic, hydrologic and oceanographic differences among major regions of Florida Bay. It is important to realize that one approach will not address all relevant scientific questions. An effort to develop mechanistic models should be combined with statistical models that use existing data sets from monitoring programs. Additional exploration of modeling approaches will be required.

There was confusion over whether the COE/WES currently was pursuing use of the Chesapeake Bay seagrass model in a Florida Bay water quality model application. In part, this confusion resulted from a last-minute substitution for Mark Dortsch's participation. If the Chesapeake Bay seagrass model is being considered, the Panel concurs with the consensus that this model should not be 'plopped' down on Florida Bay. As it stands now, the Chesapeake Bay seagrass model is not appropriate for the reasons listed above in the Workshop summary. A major impediment to using the model in Florida Bay is that the data base does not exist to support it and is not likely to accumulate in the 2 years specified in the COE contract.

In addition, there is continuing concern over model issues that have been previously expressed by the FBSOP. These include running time and whether a reduced grid of 500-1000 cells for the RMA10 is still too fine (why not 50?). The collapsed grid modification for the model apparently has not been tested. There are probably more time and cost-effective alternative approaches to using even a modified Chesapeake Bay model.

Alternative approaches to the Chesapeake Bay seagrass model need to be investigated for use in simulating seagrass dynamics in the COE Water Quality model for Florida Bay. As indicated earlier in this report, we recommend that the WES modelers examine the Fong et al. model (1997) developed for Biscayne Bay and the Burd and Dunton model in Laguna Madre as options for an initial seagrass model structure to be used (and further developed) in the Florida Bay Water Quality Model. While the development of research models in this program must be closely coordinated with empirical studies, their development should also be integrated into the continuing evolution of an improved seagrass component in the Florida Bay Water Quality Model. The COE/WES personnel involved in Florida Bay should communicate with the WES group that is developing the hydrodynamic model for the Laguna Madre. The Laguna Madre has important similarities to Florida Bay; Thalassia and Halodule are the dominant competitive species that cover 90% of a shallow (1 m) subtropical lagoon characterized by high salinities and limited nutrient inputs. A critical difference is the carbonate substratum in Florida Bay.

The COE and the PMC should work together closely to ensure that there is not duplication of effort in developing a unit seagrass model component for the Florida Bay Water Quality Model.

Although it is unreasonable to expect to develop a seagrass model component for the Water Quality Model in 2 to 3 years, it is not unreasonable to develop small, simpler models that will guide understanding and research. There are undoubtedly pieces of a seagrass model that the COE could develop productively with collaboration within the 2-year contract time. One example would be the effect of seagrasses on sediment resuspension, and the consequent effect of sediment resuspension on seagrasses. Simple models could address the goal of understanding the distribution, physiology, and growth of seagrasses, light attenuation, and carbonate geochemistry. These models can help refine hypotheses and identify research needs, providing the foundation for more elaborate models with predictive capacities for management purposes. Interim progress could be demonstrated by development of such simpler models as well as by statistical models of existing data sets.

It is also recommended that the COE/WES modelers develop linkage structure for a seagrass model component given that a seagrass model cannot be developed rapidly. Recommendations for linking the hydrodynamic-water quality models with sediment resuspension submodels cannot be made at this time. However, work can be proposed under an RFP that links sediment resuspension with light attenuation characteristics of the water column (and hence seagrass photosynthetic production). Linkage of models is critical to the overall science program but this issue needs to be addressed in the future.

Timeline for Seagrass Model Development

Recognition that a carefully planned and coordinated seagrass research program has not been funded since the seagrass die-off also forces a re-evaluation of the timeline for a seagrass module in the Water Quality Model. The goal of the Seagrass Workshop was to develop a strategy to meet the needs for a seagrass module in the Water Quality Model and to develop a RFP for Florida Bay seagrass research relevant to management needs. The focus on re-establishing natural salinity regimes as the restoration goal in the Strategic Plan and to assess seagrass response to changes in hydrologic management (Workshop goal) is inadequate to encompass the goal of understanding the role and responses of seagrasses in Florida Bay as a whole. This view echoes the concern of the FBOSP (Boesch 1997).

Development of a data base for understanding the fundamental biology of Florida Bay seagrasses and for modeling exercises will require a well-planned and sustained research program organized in tiers based on prioritization of research questions and data gaps. Although important progress can be expected after a RFP has been issued, and proposals have been reviewed and studies have been initiated, it is unreasonable to expect a level of understanding Florida Bay seagrass ecology equivalent to that in the Chesapeake Bay without a similar extensive, well-funded and long-term research program. Construction of seagrass models in the Chesapeake Bay is still ongoing after two decades of research. This effort was based in part on eelgrass, for which a considerable data base and biological understanding already existed. The situation is very different for seagrasses in Florida Bay.

The Need for Historical Ecological Data

The implicit premise of the Florida Bay seagrass studies and modeling effort is that the abundant changes in seagrass abundance and species composition observed during the past decade are significantly different from changes that have taken place historically. While there are many reasons to believe that this is so, the Panel believes that an exhaustive effort to search for information on pre-1980s seagrass abundance would be extremely beneficial. There is a very simple reason for this: it is impossible to establish seagrass restoration goals, or to determine whether restoration is necessary, without knowing what the previous state of seagrass resources in Florida Bay has been across geological as well as ecological time scales.

The Panel believes that aerial photographs that may be found in Soil Conservation Service, Defense Mapping Agency, or Department of Transportation archives, for example, could help in understanding whether there have been previous fluctuations in seagrass abundance during the past half century that approach those recently recorded. Because it is usually not possible to determine species composition from aerial photography, such information, if it exists, will not address the important issue of changes in seagrass species composition. Nevertheless, the Panel feels that any new information discovered about previous seagrass presence/absence will be quite valuable. It is also possible that reports of earlier inwater seagrass mapping efforts that are unknown to most researchers still exist, which might also help in addressing the issue of historical seagrass abundance.

Combined with ongoing paleoecological research which may provide evidence for long-term seagrass presence, a search of archival data could be critical in reconstructing the historical record of seagrass presence and abundance in the Bay during the latter half of this century. Just such a reconstruction, using newly discovered aerial photography dating to 1937 and fossil evidence of submerged aquatic angiosperm presence and species composition (Brush and Davis 1984), clearly established that unprecedented losses in submerged aquatic vegetation had taken place in Chesapeake Bay during the 1970's (Orth and Moore 1983). More importantly, the reconstruction demonstrated that such changes were well beyond the bounds of "natural variability". This reconstructed record of prior submerged aquatic vegetation presence also served to guide the development of goals for restoration of the vegetation.

Therefore, we recommend that an effort to recover archival evidence of seagrass presence and abundance be given high priority when developing an RFP to address the issues involved with Florida Bay seagrass loss during the past decade.

Panel Recommendations for Management of the Seagrass Program

1) Timing. It is imperative that initiation of the seagrass research program is implemented immediately, particularly given that the $2 million appropriation is in its second of 2-3 years. We strongly suggest that the RFP be posted without delay and certainly by the May Science Conference. Valuable time has already been lost due to the postponement of the Seagrass Workshop for nearly a year.

We perceive that key PMC personnel are overburdened with administrative and clerical details and might require more administrative and/or secretarial assistance. If so, the investment would be worthwhile because it is imperative to initiate the RFP.

2) The Funding Process. Moving ahead quickly on a seagrass research program is imperative. Important issues are: 1) to allocate the currently available Department of Interior funds promptly and efficiently without compromising the quality of the research; 2) to develop a long-term plan for sustained funding; and 3) to demonstrate progress in addressing the goals of the Strategic Plan. The first step will be rapid posting of a RFP for the seagrass research described above.

The RFP process must be an open one; it should be announced broadly and with sufficient time for fair response. Resulting proposals must be externally peer-reviewed. As clearly articulated by workshop participants, the success of the research effort will depend on the collective experience of seasoned Florida Bay scientists and the participation of experts that represent a variety of disciplines outside the existing program. The RFP should therefore encourage interdisciplinary proposals that involve the collaboration of investigators from outside Florida Bay with those having extensive knowledge of the system to produce the most competitive proposals.

It is also very important to support and expand on-going monitoring efforts and statistical analysis of existing monitoring data bases. From the Workshop discussions, it was clear that this effort is already being supported to some degree and that an additional allocation could be modest. We suggest that priority be given to additional funding for: 1) expanding the monitoring to parameters critical for seagrasses that currently are not being monitored adequately (e.g., PAR, PUR); and 2) statistical ("empirical") modeling of existing data sets. The PMC could solicit short proposals for this directly from investigators, which then would be reviewed rapidly by the PMC and selected external reviewers. This would allow immediate initiation of the seagrass research program and also ensure against loss of valuable monitoring data needed for statistical multivariate models and for calibration of models to be developed. In addition, development of empirical statistical models would be a short-term effort that would demonstrate progress toward modeling success as well as serving to refine hypotheses.

3) Structure of the Research Program. The seagrass research program should be multi-faceted and closely coordinated in its structure and long-range in its scope. The modeling approach should be pluralistic, including ecological process models, demographic models, statistical models, and landscape models. Both modeling and experimental research should be designed to address key hypotheses about factors regulating seagrass distribution, abundance and biomass. The monitoring program should be coordinated with experimental and modeling studies and priority should be given to collection of key information needed for seagrass and related models. Examples include measuring variability of light regimes and attenuation characteristics and observations on evaporation rates contributing to development of hypersalinity in Florida Bay. Scientific modelers should work closely with empirical scientists and with engineers developing water quality models for Florida Bay.

Workshops should be convened annually to assess the progress by investigators. Coordination of the science and modeling efforts should be a number one priority for consideration in future workshops (which need to be carefully planned in advance).

4) Seagrasses as an Integrative Element in Florida Bay Ecology. The Panel encourages the PMC to continue to places seagrasses as a central element of the resource management programs in Florida Bay. The seagrass research program should be coordinated not only within itself (i.e., integrating modeling, experimental, and monitoring efforts), but also with other studies in Florida Bay. Seagrasses are of central importance in terms of: 1) the hydrography and sedimentology of the Bay; 2) habitat for commercially important fishes and invertebrates and other species; and 3) sites of crucial biogeochemical processes.

5) Sustaining a Seagrass Research and Modeling Program. The Panel's view is that developing a robust seagrass model will require more funding support over a longer time frame than what will be available from the Department of the Interior's Critical Ecosystem Science Initiative. While the available resources will go a long way toward launching an important research program which expands and deepens the understanding of the Florida Bay ecosystem, a plan should developed to continue support of an integrated long-term program which continues for a decade or longer.

Conclusion of the Panel Review

This Workshop has helped lay the foundations for an exciting program which will combine experimental research, field monitoring, and ecological modeling to deepen scientific understanding of seagrass dynamics in the Florida Bay ecosystem. This Panel strongly endorses actions which will serve to create and sustain a coordinated multi-investigator, multi-institutional program of research, modeling, and monitoring. Such a program will serve to improve fundamental understanding of processes that control temporal and spatial patterns in seagrass abundance and distribution and of ecological consequences of changes in these patterns. Ultimately, this effort will strengthen the ability to predict seagrass responses to management actions, which alter hydrology and hydrography of Florida Bay and associated transport of water, nutrients, and sediments to and from the region.

 References

Boesch, D.F., N.E. Armstrong, J.E. Cloern, L.A. Deegan, S.C. McCutcheon, R.D. Perkins, and S.L. Williams. 1997. Annual report of the Florida Bay Science Oversight Panel: Perspectives from the 1996 Florida Bay Science Conference and Review of the Strategic Plan.

Brush, G.S. and F.W. Davis. 1984. Strategraphic evidence of human disturbance in an estuary. Quaternary Research 22:91-108.

D'Elia, C.F., E. Callender, Wu-Seng Lung, and S.C. McCutcheon. 1996. Workshop on the design and specifications for the Florida Bay Water Quality Model: Report of the Model Evaluation Group.

Florida Bay Program Management Committee. 1997. Strategic plan for the interagency Florida Bay Science Program.

Fong, P., M.E. Jacobson, M.C. Mescher, D. Lirman, and M.C. Harwell. 1997. Investigating the management potential of a seagrass model through sensitivity analysis and experiments. Ecological Applications 7:300-315.

Orth, R.J. and K.A. Moore. 1983. Chesapeake Bay: an unprecedented decline in submerged aquatic vegetation. Science 222:51-53.