Interim Review of the Florida Keys Carrying Capacity Study (2001)

Each of the technical modules that comprise the CCAM is discussedin detail later in this section. In assembling these comments itbecame clear that similar issues were raised for many of the modules.By first understanding and addressing these cross-cutting issues,progress can be made in many areas.

1. Inputs and outputs — Immediate and clear decisions should be made about exactly what inputsand outputs are needed for each module and for the CCAM as a whole.This problem was discussed in greater detail in Section 2, but itbears repeating.

2. Setting priorities — Creation of an all-encompassing assessment tool such as the CCAMwould be a challenging undertaking under any circumstances: The existingsevere limitations on time and money make it an even greater challenge.Each module should, before proceeding any further, have clear prioritiesfor what processes and subsystems will be included and at what levelof detail. These should be driven by the desired outputs from theentire assessment tool. Technical experts (including the module designers,the outside experts, and this committee!) appreciate all the complexitiesof the human and natural systems being explored and are reluctantto oversimplify. For the purposes of the CCAM, however, it will benecessary to focus on a subset of primary, driving mechanisms andensure that these are accurately modeled.

3. Data quality — CCAM designers should be selective about the data sets used. Oncedecisions are made about essential inputs and outputs, and then aboutwhich mechanisms and processes will be included, it should be easierto determine which data sets are most critical. Appropriate qualityassurance/quality control (QA/QC) should be performed for every dataset used in module calculations. The contractor should ensure thatall data used are reviewed for reasonableness and their authenticityand range documented to the best degree possible.

4. Definition of terms — In Section 1, this report discussed the ambiguities inherent insuch terms as “carrying capacity” and “thresholds.” In the modules, similar difficulties occur with theuse of terms such as “keystone species” and “indicator species.” Project designers should clearly define, oravoid, potentially ambiguous concepts. The expert advisors mightbe helpful in this task.

5. Use of expert opinion — As discussed in Section 2, for many important endpoints (such asthe minimum habitat required to sustain a given species), quantitativemeasures have not been or cannot be defined. In these cases it willbe necessary to rely on the consensus of an appropriate group ofexperts or relevant stakeholders. This kind of consensus judgementis preferable to an unreliable, inexact, or otherwise inappropriate“objective” measure.

6. Spatial and temporal scales — Consideration of the spatial and temporal scales in and betweenmodules is critical. Time- and space-averaged quantities will notalways reveal critical information. For example, the nearly instantaneouseffect of a heavy rainfall event on water quality and public bathingin the halo zone may be a significant output. Conversely, the slowrate of sea-level rise should be examined as an output of the dynamicnature of coastline habitats that will undergo spatial shifts alongthe edges of the Keys, especially the lower-lying middle Keys. Spatially,the nearshore distribution of propeller scars, boat groundings, and

marina impacts may be of most importance because of the greater likelihoodof thresholds being achieved, although there are certainly many humanimpacts farther offshore.

7. Uncertainty and variability— Every data set and numerical output that is part of the CCAM shouldinclude some measure of its variability. Variability can often beindicated by customary measures such as standard deviations, coefficientsof variation, quantiles (e.g., 10 percent and 90 percent values),ranges, or frequency histograms. This kind of statistical variabilityis just one source of uncertainty in the output. Other sources ofuncertainty include lack of complete understanding of the systemsbeing modeled, modeling simplifications, scale issues (e.g., applicationof data collected at a small scale to a much large study area), non-linearitiesthat affect scaling, changes that occur between the time of originalmeasurements and the time of application, and incorporation of unquantifiablefactors such as public attitudes. End-users should be clearly informedabout the different levels of uncertainty in the outputs —even whenthese values can not be easily quantified—and be advised of the implicationsof these uncertainties for assessing alternative policies, regulations,or actions. Sensitivity analyses should be performed on all outputswhen the draft CCAM is complete. These analyses will provide guidanceabout which input variables most influence the ultimate outputs andtherefore warrant the greatest effort in ensuring their reliability.At the same time, sensitivity analyses will aid in selecting thecritical processes to be included in the modules.

8. Future updates— Section 6 discusses the benefits that could be realized by maintainingand updating the CCAM over the coming years. Although this decisionwill probably not be made until later, a small effort now can makea huge difference in facilitating later updates. Each module shouldidentify key data sets and other factors that are likely to changeover time and processes that are likely to become better understood.Documentation should then include specific recommendations and instructionsaimed at future maintenance and improvement.

The suggestions above are of high priority and apply to all the modules.They are not repeated in the analysis of each module that follows,although specific examples are pointed out in some sections. Thefollowing sections are intended primarily for those directly involvedin module development. Thus these comments include more terminologyand a greater level of detail than previous sections.

This module plays three critical roles in the CCAM.

1. It is the interface through which alternative plan amendments, landdevelopment regulations, and other scenarios are input into the assessmenttool.

2. It must produce appropriate outputs, such as land use changes andpopulation projections, to be used as inputs to the other modulesin evaluating impacts on natural systems.

3. It must provide a range of important performance-measure outputsof its own.

All of these functions will need to be accomplished successfullyif the overall carrying capacity study is to be successful. The module,as described and presented by the project team, is not now sufficientlydeveloped to accomplish these crucial tasks.

As with all the modules, many uncertainties remain concerning inputsand outputs. In addition, this module has not yet developed a detailedapproach for delivering useful socio-

economic, quality-of-life, community character performance measuresand other important issues. The project team also needs to establisha more realistic portrait of where future development is likely tooccur and what patterns that development may take. As it stands,the module has a narrow focus on development that excludes redevelopmentpotential and the influence of land development regulations. Thismodule will need to provide information about what development orredevelopment activities occur and where these activities occur,at a scale and in a format consistent with the approaches taken bythe terrestrial, marine, and water modules.

The current plan to provide output as acres of development in generalcategories (as population and dwellings for 29 area units) does notappear consistent with the input needs of the other modules. Forexample, fragmentation of habitat, which depends on the locationand not just the total acreage of development, is important to theterrestrial ecosystems and species module. In determining wastewatervolume and contaminants it is important to know the types of developmentthat occur—fish processing plants versus shopping centers, for example—rather than aggregate categories. It is also important to accountfor land development regulations that govern the quantity and qualityof stormwater generated, such as limits on total impervious surfaceor design and performance standards for stormwater detention, retention,or treatment.

The socio-economics, land use, and human infrastructure module mustbe able to link population growth estimates (both in number, type,and location) to resultant plausible land use scenarios in orderto produce GIS-based maps of appropriate quantities, both for useas final outputs and as input to the other modules. It is highlyunlikely that population and economic growth will be concentratedon undeveloped, privately owned, upland sites. With 70 percent ofthe Keys' land already in public ownership, and most of the residualprivate property developed in some fashion, pristine available sitesare apparently quite limited despite an inventory of unbuilt plattedlots. This is especially true in considering demand for non-residentialspace (i.e., commercial, service, or hotel). Construction of non-residentialfacilities has been greatly constrained, perhaps even more so thanfor residential units, during the Rate of Growth Ordinance period.Instead, the main focus of future development, whether limited bygrowth restrictions or simply in response to market demands, is likelyto be on currently developed sites already disturbed beyond theirpristine state.

Some existing sites may be candidates for redevelopment, in otherwords, the replacement of old or obsolete structures by new usesat contemporary standards of construction. Other sites may be underutilized,that is, occupied by commercial or residential structures built wellbelow current zoning potential, to which new uses at higher densitieswill be attracted. Given the physical configuration of the Keys andthe evolution of development there over more than 150 years, suchconclusions appear inevitable. The very fact that Route 1 is theonly major road and its adjacent lands form the only corridor providingeasy accessibility for commercial and other non-residential uses(including public facilities such as schools and libraries) drivesthese conclusions. The module design team should identify sites thatare the most realistic redevelopment targets and then design severalredevelopment scenarios to assess the impacts of build-out to maximumallowable densities and intensities on the human and natural environments.

To do so this module will need to accept input parameters that reflectpotential amendments to land development regulations or changes tocomprehensive plans and future land use maps. These sorts of changeswill alter the impacts of development and redevelopment on such outputmeasures as environmental quality, socio-economic conditions, andquality of life.

Based on the January workshop presentations, it appears that thestudy team is treating potential development as an “on/off” switch. Theteam seemed to include only new growth (i.e., new dwelling units,square feet of retail, acres of commercial) and the extent to whichit “disturbs” land, thereby presenting an overly simplistic view of ecosystemimpacts and effectively ignoring socio-economic and quality-of-lifeimpacts.

Different comprehensive plan policies and land development regulationscan produce very different levels and kinds of disturbance. For example,a clustered, zero-lot-line project, with stormwater source controlsand substantial open space, may be more environmentally benign thana subdivision on conventional quarter acre lots with engineered stormwaterretention or detention systems. Developments with the same numbersof dwelling units can have significantly different occupancy ratesand can attract substantially different types of residents whoseimpacts on the natural environment and on public facilities and infrastructuremay also be significantly different. Thus the module should be ableto account for an array of land development regulation alternativesand the variable impacts that result.

The project team should consider the use of a regional economicsmodel capable of providing information on different sectors of theeconomy. The current aggregate approach overlooks differences inenvironmental and human impacts that arise from different types ofeconomic activity. Looking in more detail at various sectors willallow the project team a more accurate picture of the types of developmentpressure likely to face Monroe county and the likely environmentalimpacts that may result. The use of a more detailed regional economicsmodel also allows projection of employment and income patterns, importanteconomic performance measures in their own right. Given the timeand budget constraints, the team should adopt an existing regionaleconomic model, supplied with Monroe county data and adapted to fitlocal conditions.

Tourism is the principal economic activity in Monroe county and shouldbe an integral part of the CCAM. According to the U.S. Census's County Business Patterns (1998), more than 55 percent of the employment in the county wasin two tourism-oriented sectors: (1) accommodations and food serviceand (2) retail. It is appropriate that projections of future touristlevels are being included in estimates of functional population,but this is not enough: Land use, facility character, water accessrequirements, and traffic impacts of tourism should also be examined.Tourism experts should be enlisted to assist in estimating the effectsof alternative land development scenarios on the number and typeof tourists likely to be attracted to Monroe county. Attention shouldalso be given to identifying affordable housing options for tourismindustry support staff. Based on the geography of the Keys, long-distancecommuting or labor shortages are not viable options in this sector.

The existing market analysis framework assumes that housing demandis fixed and exogenous, instead of modeling the demand as a functionof housing prices. Housing prices are determined by considerationsof supply and demand, with prices rising and falling to equate supplyand demand. Ignoring price effects in the housing market is a seriousdeficiency. Monroe county currently has the highest median houseprice of any county in Florida. Because of continuing pressure onthe housing market and concerns over affordable housing, projectionsof land and housing prices should be an important performance measure.In the context of the Keys, where physical and regulatory constraintslargely limit the supply of developable land, changes in demand essentiallytranslate into changes in price.

One essential Keys Study goal is to assess the impacts of alternativeland development scenarios on the local economy and community. Todate, factors other than fiscal impact

analysis, land use, and population projections have been largelyignored, although the carrying capacity study was charged with consideringsocio-economic, quality-of-life, and community character issues.The range of expertise on the project team (including the expertadvisors) should be expanded to include individuals conversant withaesthetic, socio-economic, quality-of-life, and community characterissues and those knowledgeable about relevant design and performancestandards that may be incorporated in future land development regulations.

The original scope of work required the Keys Study to “describe site specific interactions between geology, surface water,coastal water, land use, nutrients, pollutants, runoff, erosion,and vessels.” This is a tall order, and central to achieving it willbe understanding the nature of water flows throughout the Keys. Inthe general area of water systems modeling, the CCAM identifies threeprimary subareas: stormwater, wastewater, and receiving water. Theseare further subdivided in the contractor's water-modeling flow chart,but it is these three subareas that are the focus of the discussionthat follows. The stormwater sub-module is farthest along towardcompletion, followed by the wastewater sub-module. Tracking water-qualityconstituents from their generation on land to their fate in near-shorereceiving waters is a very complicated process. In particular, thereceiving-water modeling effort has the potential for massive complexityand this element remained purely conceptual at the time of the Januaryworkshop. Given the pressing time constraints, the contractor shouldmake every effort to identify essential components that cannot beomitted or simplified and focus resources on those components. Establishmentof clear aquatic ecological indicators and the performance of sensitivityanalyses on important parameters and forcing functions will helpguide the effort.

The following comments are to a significant degree based as muchon what is planned as on what has been accomplished to date. In additionto participating in the January workshop, the committee has drawnsome inferences based on verbal evaluations of the water module madeby the expert advisors who attended the January workshop.

Overall, the committee finds that the stormwater modeling effortis on track. The contractor is using a simplified runoff predictionmethod based on daily rainfall records and runoff coefficients (a“spreadsheet” approach). This approach can be justified by the lack of a needfor surface flow routing (which would require shorter time step computations)and the current goal of evaluating only those receiving-water-qualityindicators that respond over a long time period. However, the committeenotes that hourly or even 15-minute runoff estimates (consistentwith available rainfall data) could be performed by the same spreadsheetmethod if necessary and could be helpful in analyzing certain kindsof scenarios.

The linkage between this module and land-use descriptions is consistentwith the need for data on imperviousness and other parameters thatinfluence runoff. Use of event mean concentrations is a reasonableway to develop water-quality loads (WEF and ASCE, 1998; NRC, 2000).The committee understands that event mean concentration data arenot currently available for the Keys, but the contractor should investigatethe transferability of urban data for South Florida, especially forimpervious surfaces. The limerock and sand stratigraphy of the Keysmake it harder to evaluate the impact of some stormwater best managementpractices, since

studies elsewhere reflect the mitigating effects of greater depthsand different types of soil. (The level of uncertainty introducedwith any such extrapolation should also be assessed.)

Wastewater loadings can reasonably be based on documented effluentquality of the treatment devices currently in use in the Keys (NRC,2000), including cesspools, septic tanks, aerobic septic systems,on-site wastewater nutrient reduction systems, and wastewater treatmentplants, as proposed by the contractor, with attention to the qualityand variability of the different data sets. Although gaps exist,a good effort has been made at defining the extent of each type ofdisposal system in the Keys and at assessing long-term plans forupgrading inferior systems. Nevertheless, the quality of the effluentthat reaches the coast following seepage from residential systemsor discharge from shallow (60 - 90 ft) boreholes for small wastewatertreatment plants is unknown. The ultimate fate of wastewater injectedinto shallow or deep wells is also unknown. If long-term solutionsfor the disposal of treated wastewater focus on injection wells,it will be important to know whether upward migration of the effluentplumes will occur. Finally, wastewater loadings should account forthe transient nature of population in the Keys, including the relativelyrecent appearance of large cruise ships in Key West. Initial estimatesshould be made of wastewater loadings caused by the influx of personsfrom such ships to determine whether these will be significant andto determine the impact of such short-term loading peaks on the performanceof the Key West wastewater treatment plant.

1. It is unclear what final parameters will be generated to evaluatethe Florida Keys carrying capacity based on aquatic water quality.The focus area seems to be the halo zone, 50-100 m of salt wateradjacent to the islands' beaches (defined at the January workshopas water approximately 1 m deep or less). Specific aquatic end pointsmentioned were seagrass coverage in the halo zone, the extent andlocation of benthic communities, coral reef status, the species associatedwith the previous three communities, and water clarity, perhaps includingwater clarity in finger canals.

   The method for evaluating receiving-water quality in the halo zoneis still under development. The outline as presented at the workshopwas to drive a steady-state receiving-water-quality model using residualnear-shore currents based on the Corps' Florida Bay model or somealternative Florida Keys circulation model yet to be defined. Steady-statemodeling was justified by the response of seagrass—assumed to bethe primary aquatic endpoint—which occurs over a period of many months,making short-term variations in loadings and currents unimportant.This assumption may be correct for seagrass, but concerns remainregarding the necessary timescales. Some questions include the following:

   • To evaluate seagrass growth, what water-quality parameters are needed,both in the receiving water and from stormwater and wastewater runoffestimates? If dissolved inorganic nitrogen is to be used in a receiving-water-qualitymodel, can the necessary loadings be provided to drive the model?The contractors should consider the successful seagrass modelingperformed in Tampa Bay.

   • What parameters are required to estimate water clarity?

   • Can receiving-water quality be estimated solely on the basis of long-termloadings, allowing for use of a steady-state model, or is there alsoa need for short-term modeling for bacteria, pathogens, or biologicaloxygen demand (BOD)? For instance, water-quality standards

for coliform bacteria exist for Class III marine waters. These standards might serve as appropriate performance measures in the CCAM (§62-302.503 Florida Administrative Code). Other standards for Class III marine waters, such as those for biological integrity, dissolved oxygen, BOD, pH, transparency, and turbidity should also be considered possible performance measures and may be appropriate inputs to the marine ecosystem module. It may be appropriate to incorporate the prohibitions that result from designation of most of the waters surrounding the Key as outstanding Florida waters (§62-4, Florida Administrative Code).

• Is the water quality in finger canals an important CCAM output? If so, a different level of receiving-water model will be needed.

• Although the health of coral reefs was identified as an issue of concern in the scope of work, the modeling of reef response to anthropogenic forcing functions is very complex. Will it be possible to do this quantitatively within current project constraints? If not, how will this endpoint be evaluated?

• What are the implications for the Keys of possible long-term changes in the water quality of Florida Bay, and will the CCAM be able to incorporate such changes?

• Consideration of these questions is important to guide further development of this module and allocate resources wisely. Appropriate land-side loadings and temporal and spatial definitions will need to be resolved quickly to meet the needs of receiving-water modeling. Farfield forcing functions (such as regional current models) should also be better linked to the nearshore modeling efforts.

  2. The water systems experts must know what water-quality parameters(such as receiving-water concentrations) will be needed to evaluatethe status of various important aquatic species.

  3. There is an urgent need to provide quality assurance/quality controlon the data used in the water modules. For example, in the wastewatermodule the effluent quality predictions for BOD, Total SuspendedSolids, Total Nitrogen, and Total Phosphorous presented by the contractorwere probably overly optimistic, based on the immediate feedbackfrom the experts.

  4. Related to the issue of quality assurance/quality control are thoseof variability and uncertainty. The contractor should continue toseek out the best available data sources for items such as stormwaterevent mean concentrations or wastewater treatment performance measures.The use of these data, however, must be qualified by their inherentvariability. Variability can be assessed by using several sourcesof stormwater quality data in South Florida and variations over timeat individual sites. A realistic range of effluent quality shouldalso be employed for evaluation of wastewater discharges.

  5. Because the water calculations are driven by weather, the natureof the long-term weather patterns input to the CCAM will stronglyinfluence the water-quality endpoints. For instance, a wet year generallyprovides higher loadings than a dry year. With this in mind, thebasis of the weather scenarios should be clarified. Use of a numberof representative conditions (e.g., dry, average, or wet for stormwatermodeling and calm, average, or windy for circulation modeling) isone option for evaluating the impact of climatic variations. Continuoushydrological modeling (i.e., a rainfall and meteorological time serieslasting over a period of years) is another, more complex option thatcould be used to ensure realistic variability. However it is achieved,the water module should incorporate variations in climatic conditionsand should quantify these variations and their impacts on moduleoutputs.

This module, like most, was in a very early stage of developmentat the time of the workshop. The module designers were unable toanswer many questions, because they had not yet considered all theimportant issues. The largest efforts so far seem to have been madein reducing the list of species to be considered in evaluating theimpacts of development. As of January 2001 the list had been reducedfrom an initial 128 to 17 and was reduced further over the courseof the workshop. The species chosen include federally listed endangeredspecies (both threatened and endangered), a rare-plant communitycomplex, a suite of fruit-producing species (incorrectly describedas keystone species), some species that are dependent on fire andpineland, and several species that are good indicators of hardwoodhammock habitat.

This module cannot and was perhaps not intended to function as anecological population dynamics model that might typically be usedto quantify biological carrying capacity. The module was also notintended to be a dynamic habitat or ecosystem function model. Instead,it provides a very simplified measure of future habitat losses andfragmentation. The processes, database, and methods described atthe January workshop will produce an assessment tool that providesa basic illustration of probable degrees of impact from future development.

Other modules may require maps of upland and wetland habitat distribution.Such maps are the basis for all impact evaluations planned for thismodule, however it is not clear that up-to-date and professionallyagreed coverage maps for these habitats currently exist. It alsoappeared that the actual geographic distributions for some of thespecies being examined are in dispute (e.g., the lower Keys marshrabbit). For some species (e.g., the white-crowned pigeon) informationfor the lower Florida Keys will be needed in addition to the informationbeing relied upon for the upper Keys. Such information will haveto be obtained and integrated immediately if these species are tobe included.

To date, no habitat- or ecosystem-level outputs based on physicalfactors or processes have been developed. The design team agreedthat any additional losses of hammock or pineland would be unacceptablefor many of the chosen indicator species or groups (e.g., Schaus's butterfly, wooly croton, rare plants, tree snails, wood rat, Keydeer, tree cactus, forest birds). The team stopped short of concludingthat mapping additional habitat fragmentation or loss of acreagein pinelands or hammocks would in itself serve as an effective outputto identify development impacts. This is a simple and obvious additionto this module's outputs, but it requires input from the socio-economicmodule that is spatially explicit about where land use changes willoccur.

As mentioned above with respect to pinelands and hammocks, the contractorexpressed an intention to look more closely at habitats or ecosystems,but no specific proposal was presented for review. Of particularconcern at this time is the lack of any clear plan for reviewingthe distribution, status, or impacts of development on wetlands.Some indirect evaluation would be derived from looking at singlespecies that depend on freshwater wetlands, such as the mud turtle.Some level of wetland response may also be derived from the assessmentof red mangrove impacts. However, no habitat level assessment oftidal wetlands is currently included in the module. Wetlands composenearly 60 percent of the study area and have a history of directand secondary cumulative adverse impacts from development. The wetlandsareas deserve a much higher profile and level of evaluation thanallocated to date. Different parts of this evaluation may involvethe water, terrestrial, and marine modules. Some level of gradientanalysis of water quality in tidal wetlands should be included asan output from the CCAM.

In the case of the Key deer the opinions of the module design teamappear to be at odds with accepted scientific information and opinion.For example, workshop participants were told that the informationnecessary to conduct a population viability analysis was not availablefor any of the team's chosen species, however a thorough populationviability analysis was performed for the Key deer in 1990 (Seal etal., 1990). The module design team should take advantage of the extensiveinformation available on this well-studied endangered species. Theoutside expert advisors could be very helpful with this task. Doinga top-notch job with this high-visibility species would add a gooddeal of credibility to the overall effort.

The module-design team had previously agreed upon using geographic,spatially explicit mapping of habitat loss and fragmentation as themeans for evaluating various development scenarios. To date, theimpact analyses have been made by selecting specific habitat patchsizes below which a given species will no longer inhabit that patch.The outputs produced are summary statistical tables and color-codedmaps indicating the number of patches from which a species is eliminateddue to a particular development scenario. The scenarios tested todate simulate losses of habitat caused by additional upland development.The rules that determine minimum patch sizes needed are generallywell thought out, easy to understand, and easy to measure, howeverthese rules have not been fully developed or accepted by a consensusof experts. Additional consultation and review will be needed toensure the credibility of the outputs. It is essential, as notedabove, that the output from the socio-economic and land use moduleprovide spatially explicit data about where land use change willoccur.

Beyond the basic rule on patch size described above there is someconcern that within-patch variability has been ignored. All patchesof a given size are not the same from a species point of view. Considerationof inter-patch variability for such features as time since last fire,presence of specific host species, or sources of drinking water couldsignificantly and easily improve the module output. By includinginformation on conditions immediately surrounding each patch (e.g.,type of development, availability of drinking water, prevailing winddirection) the outputs can be made more detailed with relativelylittle effort.

This module was also in a very preliminary stage of development atthe January workshop. The contractor presented some interesting concepts,such as the approach to assessing the impacts of boat scarring onseagrass beds, but many elements of the module were still undeveloped.As discussed in Section 2, lack of sufficient coordination betweenmodules has been a major barrier to progress. For example, the marineand water modules will need to reach an immediate decision aboutthe modeling of Florida Bay waters and their impacts on Keys waterquality. Much of the effort to date in the marine module has beenspent on gathering useable data sets, anticipating the data thatmay be generated by other modules, and exploring other potentiallyavailable data.

Three main metrics were identified for marine ecosystems: seagrasscover, water clarity, and contaminant loads. Because the data arerelatively sparse and the diversity of marine fauna is high, themodule design team decided to focus on a process-oriented approachrather than a species-specific approach as used in the terrestrialmodule.

Although seagrasses play a central role in determining the overallhealth of the marine ecosystem under the proposed approach, the mapsof species-specific seagrass distribution were not well developed.Because all seagrasses are not the same (in terms of growth patterns,

nutrient requirements, recovery time after prop scarring), furtherwork will be necessary to map the distribution of seagrasses aroundthe Keys.

Up until now almost no effort has been made to understand eitherthe impacts of fishing on marine ecosystems or the impacts on fishpopulations of other activities. There was some reference to gatheringdata on the increase in marinas and boats, but it was unclear howsuch data would be used. If such information is desired, it shouldbe produced as an output from the socio-economic and land use module.With little time remaining, many important issues remain to be investigated:collisions with manatees, fishing pressure per boat, size of boats(which relates probably non-linearly to the number of fishers perboat), and reef damage due to recreational diving, boat anchors,and removal of upper level predators. These constitute real impactsto the health of the Florida Keys.

Because the marine module design team's approach so far is spatially explicit and does not consider temporalaspects, the assessment tool will be able to provide only a basicvisualization of relative degrees of impact and fragmentation fromfuture development, and only for a limited marine community. Forexample, the current approach would produce a tool that could notexamine the impacts of such short-term events as sediment pulsesfrom storms on communities of concern at some vital time in organismor community development. It would also be unable to assess synergisticor cumulative issues. Furthermore, the spatial arrangement of impacts(mangroves to seagrass to inshore patch reefs to offshore reefs)should be examined, given the juxtaposition of these important habitattypes in space and the documented connections between these importanthabitats and proper nursery function throughout the Keys (e.g., forsnappers, grunts, and groupers).

Finally, the marine module does not currently look at species onthe Federal Endangered Species List or the Official Lists of Florida's Endangered Species and Species of Special Concern, such as themanatee, crocodile, mangrove rivulus, and numerous marine turtles.It also does not consider important intertidal marine habitats, likemangroves. The inclusion of these federal- and state-protected speciesin this module is vital and will provide important performance measuresfor the carrying-capacity analyses.