FLORIDA BAY 2001 SCIENCE CONFERENCE SYNTHESIS

Backround

Seagrass beds are an important component in many coastal marine environments, however, there are few locations in the world where seagrasses are as dominant in the hydroscape as south Florida (Fourqurean et al., in press). Seagrasses are the dominant biological community in Florida Bay, historically covering over 90% of the 180,000 ha of subtidal mudbanks and basins within the Bay (Zieman et al. 1989). By comparison, mangrove islands cover only about 7% of Florida Bay. Because of the shallow nature of Florida Bay, (mean depth < 2 m, Schomer and Drew 1982), seagrasses are also the dominant physical feature of the Bay, and their presence greatly affects physical, chemical, geological as well as biological processes in this system (Zieman, 1982).

The shallow distribution of seagrasses places them in close proximity to the land/sea interface, a region experiencing rapid growth of human populations. This coastal distribution also places seagrass communities at the end of the watershed pipe, thus, their status reflects not only direct coastal influences, but larger, landscape-to-regional scale influences as well. Because most seagrasses are benthic-perennial plants, they are continuously subject to stresses and disturbances that are associated with changes in water quality along the land/sea interface. Seagrasses act as integrators of net changes in water quality variables which tend to exhibit rapid and wide fluctuations when measured directly.

Seagrass communities are also important to the economy of south Florida because they provide food and shelter to numerous fish and invertebrate species, many of commercial importance within the region (Powell et al. 1989; Thayer and Chester 1989; Tilmant 1989; Chester and Thayer 1990). Seagrass abundance to a large extent determines public perception regarding the "health" of the coastal waters of Florida (Goerte 1994; Boesch et al. 1995). Thus, the recent changes in the distribution and abundance of seagrasses within Florida Bay have been perceived as a change in the health of the Bay. For these reasons, seagrasses may be the best indicators of changes in the Florida Bay ecosystem Fourqurean et al, 1992.).

A widespread die-off of seagrasses within Florida Bay began in 1987 (Robblee et al., 1991). This event was first observed by back-country fishing guides who reported the occurrence of "potholes" in the seagrass beds of the bights along the north-central part of Florida Bay. Extensive areas of Thalassia began dying rapidly in central and western basins, and by 1990, 4,000 ha were completely lost and 24,000 ha were affected by the die-off (Robblee et al., 1991). Factors such as elevated water temperature, prolonged hypersalinity, and excessive seagrass biomass (due to lack of recent disturbances, including hurricanes and reduced salinities), leading to increased respiratory demands, hypoxia and sulfide toxicity are some of the physiological stressors thought to have contributed to Thalassia die-off (Robblee et al., 1991; Carlson et al., 1994). Observations indicated the waters of Florida Bay were generally clear, warm and hypersaline in most areas exhibiting die-off. High water column light attenuation from suspended sediments or eutrophic phytoplankton growth was not present at the initiation of the dieoff or for the first several years (Boyer et al, 1999).

Blade lesions were frequently observed on Thalassia in areas affected by die-off, although they were not universally present during the earliest phases. An undescribed species of marine slime mold in the genus Labyrinthula was the most common microorganism isolated from die-off affected short-shoots (Porter and Muehlstein, 1989). Physiological stress and a negative carbon balance are associated with infection of Thalassia by Labyrinthula (Durako and Kuss, 1994). Based on early observations and data, two conceptual models were formulated (the Zieman Model and the Carlson and Durako Model) incorporating distal and proximal elements thought to have contributed to die-off. These models also included predictions of negative cascades following die-off.

The patterns of changes in seagrass abundance in Florida Bay have recently undergone four phases: 1) primary die-off, 2) secondary dieoff, with mortality primarily due to light attenuation, 3) seagrass recovery associated with improving water clarity, and 4) renewed primary die-off in areas where die-off has not been previously observed. The initial phase of primary seagrass dieoff occurred only in Florida Bay, which is the most continentally influenced and least oceanically flushed of all of the extensive south Florida seagrass beds, and it occurred only in the most densely developed beds within the bay. To this day, no occurrence of primary seagrass dieoff has been found outside of the densest seagrasses within Florida Bay, and no primary seagrass dieoff occurred in sparse or medium density beds within the bay.

The first phase of primary seagrass dieoff occurred during the relatively dry and clear period of 1987 to early 1991. During this period standing crop and areal productivity of Thalassia were high, but then declined as die-off progressed (Zieman et al. 1999). Some stations had shown large increases in seagrass standing crop in the period prior to the dieoff (Zieman et al. 1999). In addition, there was rapid spread of Batophora oerstedi and Halodule into die-off patches (Thayer et al., 1994). At this time, Florida Bay as a whole remained as clear as in the years just prior to the dieoff, with limited turbidity from suspended sediments in the immediate vicinity of the denuded dieoff patches.

Several years after the initiation of the seagrass die-off, Florida Bay began exhibiting widespread and chronic turbidity with a concomitant decline in Thalassia and Halodule in Johnson Key Basin and Rankin Lake (Stumpf et al., 1999). The increase in turbidity, which began during the fall of 1991, was principally due to cyanobacteria-dominated microalgal blooms and resuspended sediments associated with the loss of seagrasses on the western banks, and was most severe in the western and central Bay (Phlips and Badylak, 1996). The blooms may have been initiated by the nutrients liberated from the die-off of seagrasses (Butler et al. 1995). Loss of seagrass cover was the major factor in the increases in sediment resuspension in the Bay (Prager, 1998). This resulted in a negative feedback loop in which loss of seagrass cover from die-off leads to exposed, easily resuspended sediments, and more, widespread loss of seagrasses due to turbidity. Sponge mortality, changes in juvenile lobster population dynamics (Butler et al. 1995) and indications of cascading effects on plant and animal communities in adjacent systems (e.g., sea urchin population explosions and unbalanced growth of Syringodium filiforme in the waters of the Florida Keys National Marine Sanctuary southwest of Florida Bay, Rose at al. 2000; Kenworthy et al. 1998) were also observed. From 1992 to 1995, salinities showed a progressive decline throughout the bay (Boyer et al. 1999).

During this time, there was much speculation regarding how much seagrass had been lost, with estimates as high as 100,000 acres (40,000 ha) of loss being stated with no data support. With the decline in water clarity, aerial photography became useless in determining seagrass distribution over much of Florida Bay. Because of the continuing concern regarding the extent of seagrass changes within Florida Bay, and the need to monitor the effects on seagrass communities of proposed water management alterations for the restoration of the Everglades/Florida Bay ecosystem, the Fisheries Habitat Assessment Program (FHAP) was initiated during spring 1995. The turbid conditions in western Florida Bay from 1995-1997 complicated measurements and interpretations of seagrass losses and changes in species’ distributions due to primary die-off versus secondary effects attributable to light limitation. Early FHAP data (!995-1997) indicated that seagrass decline in the western basins was primarily due to deteriorated water quality because mortality was evident as a general shoot thinning, rather than the occurrence of distinct die-off patches surrounded by dense beds. A comparison of seagrass distributions in Florida Bay between 1984 and 1994 (Hall et al., 1999) and between 1995 and 1998 (Durako et al., in press) also indicated that the chronically turbid regions had exhibited the most significant losses of T. testudinum.

Since 1998, there has been an upturn in abundance, productivity, standing crop, and flowering of Thalassia and Halodule which seems to be associated with improving water clarity (Durako et al., in press; Durako et al., Zieman et al.). Some of this improvement in water clarity may be due to a decrease in unvegetated bottom (Durako et al., in press). However as recovery begins to occur at the original dieoff sites, a new instance of primary seagrass die-off has been observed north of Barnes Key beginning in January 1999. This new primary die-off is occurring in an area where die-off has not been previously observed, but it has some characteristics very similar to those observed during the initial seagrass die-off in 1987. The area has excessively dense beds of T. testudinum with die-off patches interspersed among the dense beds, and the same unusual seagrass morphologies (‘twinning’) have been observed in surviving shoots (Zieman et al). Seagrasses affected by the new die-off exhibit symptoms like those of the 1987 event, i.e. the short-shoot meristem tissue appeared to be the tissue most immediately affected. Meristems seemed mushy and smelled like "mustard" while the rest of the blade looked green and healthy (Carlson).

Seagrass loss has also been recently observed in the northern portion of Madeira Bay from the mouth of the Taylor River to the south and west (Bacon et al.). This area has shown recolonization by Halodule and Ruppia. These observations of renewed seagrass die-off led to a plan for a series of observations and experiments to test several of the remaining hypotheses that might explain seagrass die-off.

Seagrasses continue to be the dominant biological community in Florida Bay. Of the more than 14,000 Braun-Blanquet samples (0.25 m²) taken in the Bay from 1995 to 2000 by the Fish Habitat Assessment Program (FHAP, Durako et al.), approximately 97% contained seagrass. In northeastern Florida Bay, Thalassia was present at 75.9% and Halodule was present at 69.0% of the 762 randomly selected stations sampled from May 1999 to May 2000 (Bacon et al.). The entire South Florida coastal zone, including the areas west of Florida Bay and within the Florida Keys National Marine Sanctuary, is dominated by seagrass habitats. Fourqurean et al. (in press) assessed seagrass species composition and density at 1207 sites distributed across 19,402 km² of nearshore marine and estuarine environments in south Florida. At these sites, a total of 8434 quadrats (0.25 m² ) were sampled from 1996 to 1998, covering an area of 2108.5 m². At least one species of seagrass was observed at 1056 of the 1207 sites, or 87.5 percent of all sampling sites. Thalassia testudinum (Turtle Grass) was the most commonly encountered species, being found at 898 sites. Halodule wrightii (Shoal Grass) was the second most commonly encountered species, occurring at 459 sites; followed by Syringodium filiforme (Manatee Grass, 239 sites), Halophila decipiens (Paddle Grass, 96 sites), Ruppia maritima (Widgeon Grass, 41 sites) and Halophila engelmannii (Star Grass, 28 sites).

Following the initial die-off and period of widespread turbidity, seagrass abundance has shown three distinct phases (Zieman et al. 1999; Zieman et al). At stations associated with primary seagrass dieoff, standing crop declined from 1989 to 1995. It remained stable from 1995 through 1997, and has increased during the past several years (1998-2001). Thalassia has shown little net change in abundance (" 8% of the mean) at the Bay scale from 1995-1999, although at the basin scale abundance has varied by an average of " 30% (losses then gains in the west, contrasted with stability or gains in the middle and east basins). Most of the increase in seagrass abundance has resulted from expanding coverage by Halodule (+ 200% Baywide, + 450% in Johnson Key and Rabbit Key Basins, Durako and Hall 2000). Thus, the dominance of Thalassia is declining and mixed turtlegrass and shoalgrass beds are becoming more common. In the past 5 years, relative Thalassia abundance has dropped from being over 5 times that of Halodule to being less than three times more abundant; in spring 1997 Halodule replaced Thalassia as the most abundant seagrass in Johnson Key Basin (JKB, Durako et al.). During fall 1996, the small-bodied, low-light adapted seagrass Halophila engelmannii was observed at one station in JKB. By spring 1998 this species was present at 15 of the 32 stations in this basin.

The passage of hurricane Georges west of the Bay in Fall 1998 uprooted much of the Halophila in JKB (it was only observed at 5 stations nine days after the storm), reduced the cover of Thalassia, especially in areas where it had been sparse, and it removed much of the litter layer on the bottom. However, by spring 1999 Halophila cover increased in JKB and this species was observed in Rankin Lake (RAN), Whipray Basin (WHP) and Twin Key Basin (TWN). This rapid increase in spatial distribution suggests the hurricane may have played a role in distributing propagules (Durako et al.). Sediments were also resuspended by Georges (longer internodes of Barnes Key SS suggests recent sediment deposition- meristems still relatively deep). Hurricane Irene reduced cover in areas of sparse Thalassia and reduced Halophila distribution as determined by before and after sampling by FHAP.

Another recent dramatic change in the ecology of seagrasses in Florida Bay is the widespread occurrence of flowering, in both Thalassia and Halodule. In spring 1999, reproductive short-shoots of Thalassia were present at 19 sites in 7 basins across the Bay; reproductive short-shoots of Halodule were present at 24 sites in 4 basins in the western part of the Bay (Durako et al). In spring 2000, flowering Thalassia was observed at 19 sites in 6 basins. The recent increases in seagrass cover and the dramatic increase in flowering may reflect improvements in water quality, although shifts from turtlegrass to shoalgrass often are associated with declining light availability, as is the appearance and spread of Halophila engelmannii.

To assess the relationship between water clarity and seagrass recovery in Florida Bay, Carlson et al. have continuously measured subsurface and bottom PAR at seven stations in Florida Bay since fall 1998. In addition to continuous light data, discrete water samples for analysis of turbidity, color, chlorophyll and total suspended solids and plant samples for epiphyte light attenuation measurement have been collected monthly. Diffuse attenuation coefficients vary seasonally and among basins within Florida Bay: attenuation is higher in winter than in summer and generally higher in the basins which lost large amounts of seagrass in die-off episodes between 1987 and 1991. Persistent phytoplankton blooms in the north-central region of the Bay are associated with high (>3) Kd values, but water clarity at most sites was higher in 2000 than in 1999.

There are spatial and temporal gradients in species composition of epiphytes with coralline reds being more dominant in western high-flow areas of the Bay (Frankovich 1999). Fleshy epiphytes have restricted distributions only occurring near bird islands and near the Keys. The cyanobacteria Lynbya is very common in the Syringodium filiforme beds in the west (Frankovich and Zieman) . Epiphyte attenuation is higher in winter and spring (30-50%) than in summer and fall (15-30%), and values are higher in the eastern region of the Bay (>40%) than in the west (ca. 20%). Calcium carbonate derived from calcareous algae and resuspended sediment comprise more than half of the epiphyte load. Epiphytic light attenuation has also been measured using Mylar strips and a light attenuation measurement apparatus that has been successfully employed by investigators in Australia and Chesapeake Bay (Frankovich and Zieman). The Mylar strips have been set out within various seagrass meadows across the Bay for a sufficient time period to allow for the accumulation of epiphytic organisms. Distinct epiphytic communities consisting of benthic diatoms, coralline red algae (Melobesia membranacea, Hydrolithon farinosum), and filamentous red and brown algae occur seasonally in various regions across Florida Bay. These various epiphyte functional forms, and combinations thereof, result in differing levels of light attenuation at the leaf surface relative to the amount of epiphyte loading. Epiphyte loads at Barnes Key are at lower 5% of baywide averages (Frankovich and Zieman). Highest epiphyte loads in areas of highest seagrass productivity, so epiphytes not reducing production (Frankovich and Zieman).

Grazers may have an important role in controlling epiphyte abundance. Snails (50/m2) and hermit crabs are conspicuous at Rabbit Key Basin (Frankovich). Baywide, Bittium occurs at densities of 3/m2, versus in Indian River Lagoon were densities of 20/Halodule short-shoot or 1000's/m2 have been observed.

Macroalgae such as Laurencia are patchy in abundance. It is unknown at present whether it’s abundance has changed in response to the changes in seagrass abundance (although these data are in the FHAP dataset). Jeff Holmquist found that Laurencia accumulation did not kill underlying Thalassia.

Light penetration to the leaf blades, which sets the limits for seagrass photosynthesis, varies as a function of depth, turbidity, and epiphyte cover. The occurrence of the marine slime mold Labyrinthula has also been shown to affect photosynthetic characteristics of T. testudinum (Durako and Kuss 1994). Each of these factors plays an important role in survival and the potential for regrowth into previously inhabited areas. Seagrasses in Florida Bay are meristem dependent and depend on rhizome growth and branching to maintain or increase their populations. Excess carbon from photosynthesis is needed for production and growth of apical meristems to allow lateral growth and spread of populations. Thus, apical meristem density and branch frequency (rhizome apical density normalized to short-shoot density) may act as "ecoindicators" of the potential for loss or gain of seagrass density. Core samples (15cm diameter) obtained during spring 1998 and 1999 FHAP sampling indicated that the mean branching frequency was 0.24 during this period, or one rhizome apical branch produced for every four short-shoots (Paxson and Durako). This is very similar to the branch frequency of 0.28 observed during the early die-off from 1989-1990 (Durako, 1994). The relationship between rhizome branching in 1998 and shoot density changes from 1998 to 1999 was significant (p< .0001), however, the r² value of .12 was low. Labyrinthula occurrence was found to have significant negative influences on branching frequency (p= .03), shoot density (p= .004), and apical density (p= .033) suggesting this micro-organism has a negative influence on seagrass growth and carbon balance (sensu Durako and Kuss, 1994).

Photosynthetic yields of Thalassia testudinum leaf material measured in situ at Sunset Cove and Cross Bank, using a submersible pulse amplitude modulated fluorometer (Diving-PAM), were significantly lower for regions of Thalassia leaves that had visible lesions. These patterns agree with those previously reported by Durako and Kuss (1994). However, close interval PAM fluorescence measurements along an individual leaf with several visible lesions indicated the reductions in photosynthesis were restricted to the immediate area of the lesion. Quantum yields of lesion-free leaf regions of short-shoots also declined along transects from dense, apparently-healthy beds to recent die-off patches at both Barnes Key and Cross Bank. The photosynthetic characteristics of solitary short shoots within the die-off patches were significantly lower than those of shoots along the ecotones and shoots 1 m inside the bed. This indicates that photsynthesis may be reduced even in the absence of visible lesions. The PAM data also reveal that the severity of stress imposed by the leaf lesions will be a function of the proportion of total leaf surface that is necrotic and it remains to be determined what is the lethal threshold for lesion coverage.

The mesocosm studies (Chesnes et al., Montague et al., 1999) have produced data on how fluctuating salinities may affect turtlegrass, shoalgrass and widgeon grass. Unfortunately, operational difficulties plagued the mesocosm facility, which slowed progress and resulted in problems with maintaining proper controls.

The role of hypersalinity in seagrass dieoff still remains unclear. Salinities during the initial dieoff episodes in 1987-89 were accompanied by hypersalinities ranging from 45 to 70 ppt., in addition to the other environmental stresses. The Barnes Key dieoff clearly did not experience this level of salinity but experienced many of the other stresses that accompanied the original primary dieoff. While the role of hypersalinity may vary relative to other stresses, what is clear is that low salinity does provide a refugia from Labyrinthula, as infection does not occur at < 15 ppt (Blakesley et al., 1999).

The potential links between elevated sediment sulfide concentrations and seagrass mortality (an hypothesis long proposed to account for seagrass die-off) have been explored by Carlson et al. and Erskine and Koch. Hydrogen sulfide is a known plant toxin. High sulfide levels in sediments have been observed during, and after, die-off episodes (Carlson et al. 1994). Sulfide concentrations > 2-4 mM have been measured in die-off sites, but it is not known if this is a cause or an effect of die-off. If sulfide concentrations are < 2 mM, no die-off has been observed. Carlson has observed that high porosity sediments correlate with high hydrogen sulfide because of low permeability. Barnes Key surface sediments very high porosity (lots of water and fine sediments). Durako and Kuss (1994) suggested that Labyrinthula infection reduces oxygen production and may increase susceptibility to sulfide. However, photosynthetic rates in Thalassia increase as a function of increasing sulfide up to 6 mM, and sulfide levels up to 10 mM have failed to produce visual signs of acute sulfide toxicity, although high sulfide levels have been shown to result in reduced leaf elongation rates (Erskine and Koch, 2000). A series of field experiments has been undertaken to help understand how sulfide concentrations may influence seagrass growth and abundance and clear up uncertainties between the two conflicting sets of results. A submersible pulse amplitude modulated fluorometer (Diving-PAM) was used to investigate photosynthetic yields of Thalassia testudinum leaf material in situ within bucket experiments established by Carlson. Photosynthetic yields were significantly lower in buckets with added glucose and acetate, both treatments that should have increased sediment sulfide levels (Durako). These results indicate that reduced photosynthetic capacity, may be caused by Labyrinthula-induced lesions or elevated sulfide. This may make Thalassia more susceptible to sulfide toxicity, hypoxia, or disease by imposing a negative carbon drain on belowground tissues.

Seagrass leaves grow by the proliferation of cells from a basal meristem. This is the most metabolically active tissue in the plants. Recent high-resolution, in situ oxygen measurements show the short-shoot meristems of Thalassia at the Barnes Key dieoff site go anoxic during the night in November and remained anoxic for up to 5 hours (Borum et al.). In contrast, leaf meristems at Rabbit Key Basin did not go hypoxic. The Barnes Key Samples showed a more rapid decline in oxygen concentration after sunset, and slower internal oxygen concentration increases in the morning than samples from Rabbit key Basin. Thalassia at Barnes Key is very dense with over 1200-1500 short-shoots/m2, has a very high biomass (SC 300 g/m2), and low turnover rate (1.2%/ day) with a very thick litter layer. In Rabbit Key Basin densities are also high (1200-1300 short-shoots/m2) but biomass (SC 109 g/m2) and turnover rates (1.7% /day) are more moderate (Zieman et al). The long short-shoot stems at Barnes Key lead to increases in diffusion distance for oxygen. Long internodes may be a response to hypoxia or rapid sedimentation. Carlson et al. have reported that alcohol dehydrogenase (ADH) increases in Thalassia under hypoxic conditions, exacerbating the carbon drain on belowground tissues.

Die-off has only been observed in biogenic carbonate sediments. This sediment type is usually low in iron. However, Florida Bay carbonate sediments have relatively high iron levels for carbonate sediments (Carlson et al). Spatially, iron is high near mainland and high toward west. Atmospheric deposition, as proposed by Shinn et al. is still not known. Addition of iron to sediments decreases the flux of sulfide to Thalassia (Carlson bucket experiments) and results in a significant. but small increase in growth. Iron distribution may also have a role in controlling phytoplankton blooms in the Bay.

Establishing the relative contribution of dieoff versus light-stress induced mortality to the recent losses of Thalassia in western Florida Bay is problematic. There is a high spatial coincidence among the distribution patterns of seagrass loss, Labyrinthula abundance (Blakesley et al. 1998), high sediment sulfide levels (Carlson et al., 1994), and turbidity (Phlips et al. 1995; Stumpf et al. 1999). Increases in Halodule in the Bay may reflect its lower light requirements (Williams and McRoy 1982; Dunton and Tomasko 1994), ability to rapidly spread into areas where the Thalassia canopy has been removed (Thayer et al. 1994), or resistance to disease. There has been little change in Thalassia and Halodule abundances in basins that are periodically subjected to low salinities. This may reflect the effects of intermediate disturbance (sensu Connell’s Intermediate Disturbance Hypothesis) in maintaining a mixed-species subclimax (Zieman 1982, fig. 13), as well as offering a low-salinity refugia from disease since Labyrinthula has never been found in Florida Bay at salinities below 15 ppt (Blakesley et al. 1998).

The spatial patterns of abundance changes from 1995-2000 suggest that, presently, the most perturbed environment in Florida Bay with respect to seagrasses, is along the western and southern Bay margins. Much of the focus of management and restoration efforts in South Florida have been directed toward landscape-scale modifications to an extensive flood-control system to increase the quantity of freshwater delivered to northeast Florida Bay, and more recently, Shark River slough. Although the initial die-off initiated in the interior basins of the Bay (Robblee et al. 1991), the greatest changes in seagrass abundance in the present system are occurring far from the Everglades/Florida Bay land/sea interface. In addition, the waters of western Florida Bay form a hydrodynamic link between the Everglades and the coastal waters of the southwestern Florida peninsula/eastern Gulf of Mexico, to the north, and the Florida Keys reef tract and the Atlantic Ocean to the south (Schomer and Drew 1982). The seagrass communities of this region form an important buffer by intercepting the flow of water along this region and reducing nutrient and particulate loads in the waters reaching the reef tract (Kenworthy et al. 1998). Continued losses of seagrasses along this margin, coupled with the proposed increase in water flows out of the Everglades could result in greater fluxes of material out of Florida Bay and onto the reef tract (Kenworthy et al. 1998). Resource managers will need to consider actions that might aid in the re-establishment of continuous seagrass cover along western Florida Bay. This would be an important step in reducing sediment-resuspension induced turbidity along this boundary that could reverse the cascading declines that characterize the present system.

Early in the dieoff studies conceptual models were developed of hypothesized dieoff mechanisms and the conditions and processes leading up to the dieoff. Initially the model developed by Zieman, Fourqurean, and Robblee was a more process-oriented model, and placed strong emphasis on the historic conditions leading up to the dieoff. The other conceptual model was developed by Carlson and Durako and was a more mechanistic conceptual model with more emphasis on the dieoff process. These conceptual models are shown in the accompanying figures.

The Zieman et al model has three major phases. A Developmental Phase (A-C), where a combination of natural and anthropogenic processes contributed to an extensively developed (overdeveloped actually) Thalassia ecosystem. An Initiation Phase ((D-G) where the heavily developed system interacted with a suite of environmental stresses to produce the initial dieoff episodes, and a Maintenance Phase (G-L-G and repeat) where the process became self sustaining. Here interactions of the dense Thalassia and the environmental stresses formed the primary trigger to the initial dieoff episodes.

The Carlson and Durako model included the over-developed Thalassia as a component, but focused more on the role of physiological stress, especially hypoxia and sulfide toxicity as major drivers. In addition, this model gives a much larger role to the effects of the slime mold, Labyrinthula as a causative element. Although both of these models have matured as research has progressed, the process has been one of small refinements, as both are little changed in the past decade.

As research progressed and the conceptual models matured, seagrass researchers found much common ground on which to agree.

Among these points of agreement are:

• 1. Primary seagrass dieoff is species specific - Thalassia.

A theoretical model for die-off as a disease (Blakesley et al.1999) has suggested 3 different roles that Labyrinthula might play in Florida Bay under different environmental conditions. These included: (1) a nonpathogenic parasite; (2) an opportunistic secondary pathogen; and (3) a primary pathogen. Five different factors were discussed as critical elements in determining the role(s) of Labyrinthula in seagrass health at a particular site in Florida Bay (Blakesley et al., 1999b). Salinity controls infection (infection does not occur at < 15 ppt). Seagrass density determines the extent to which Labyrinthula infection spreads because the slime mold transmission is thought to depend on blade-to-blade contact (Muehlstein, 1992). Pathogenicity of a particular strain of Labyrinthula will determine severity of infection. Environmental stressors (abiotic factors) such as low light or high temperatures may weaken Thalassia and, in combination with the infection by pathogenic Labyrinthula, cause seagrass die-off. Resistance to disease due to genetic factors or production of phenolic compounds may be important in determining the health of Thalassia in Florida Bay. The model predicted that in areas with high seagrass density, high salinity, "suboptimal seagrass conditions (environmental stress)", and presence of pathogenic Labyrinthula, the slime mold could contribute to either chronic or acute die-off acting as an opportunistic secondary pathogen. With the same conditions, but without environmental stress, it was suggested that Labyrinthula could still cause "thinning" or patchy die-off acting as a primary pathogen (Blakesley et al., 1999).

The recurrence of an acute die-off in Barnes Key presented an opportunity to test a portion of the theoretical model by comparing the symptoms and progression of an acute event in the Barnes Key mudbank area with the symptoms and progression of what we believed to be chronic die-off in Sunset Cove. The hypothesis was that in Sunset Cove Labyrinthula acted as a primary pathogen in an environmentally unstressed site whereas in Barnes Key Labyrinthula more likely played the role of a secondary pathogen in an environmentally stressed site.

Comparisons of results for the Sunset Cove and Barnes Key sites revealed that although active die-off was occurring in both places, the sites were very different. At both sites the pattern of die-off was patchy, suggesting disease processes rather than a physical process as the primary cause. Both sites had high salinities (> 15 ppt) and dense Thalassia beds - necessary elements for Labyrinthula infection and transmission. However the data from the two sites revealed important differences summarized below:

These differences strongly suggest that the mechanisms for the die-offs in Barnes Key and Sunset Cove are not the same. Blakesley et al propose that the acute die-off in Barnes Key results from a series of events beginning with heat stress and an initial infection/disease (not Labyrinthula-induced) which rapidly kills the infected seagrass. The resultant large amount of decaying below-ground biomass from the rapidly dying Thalassia roots and rhizomes promotes microbial activity that in turn elevates the sediment sulfide levels selectively in those vegetative zones where the die-off is occurring or has recently occurred. The high sediment sulfide levels do not kill seagrass outright, but instead further stress the other seagrass in the immediate area. Finally, Labyrinthula, acting as an opportunistic secondary pathogen, infects the already weakened remaining seagrass. In contrast, the chronic die-off in Sunset Cove is hypothesized to be directly caused by the presence of the slime mold Labyrinthula acting as a primary pathogen. Sediment sulfide levels may remain relatively low in all the vegetative zones tested simply because the slowly dying Thalassia roots and rhizomes result in a smaller decaying below-ground biomass so that the sediment sulfide levels may remain relatively low in contrast to the Barnes Key scenario. Such chronic seagrass die-off is still ongoing in many parts of Florida Bay where the Thalassia beds are dense enough for transmission of the disease and the salinity is high enough for infection to occur.

BIG QUESTION REMAINS - WHY FLORIDA BAY?

Unique circulation- central basins isolated rainfall and evaporation most important

Continental influence

Other areas have densities as high

Other areas have carbonate sediments

Lakes region has banks and basins but has open boundaries

The production of distribution and abundance maps of the seagrasses have proven valuable and they provide a quantitative record of distribution and abundance at both basin and Bay scales. The change maps provide clear visualization of where and how much the distribution and abundance changes, thus can point to "hot spots" that may be related to specific forcing events or conditions. The use of cover has become an assessment standard in region (FKNMS, DERM, FHAP) and cover may be a more sensitive performance measure to short-term changes than density since density changes require mortality and recruitment of short-shoots while cover can change also in response to just changes in the leafiness of short-shoots. Short-shoot density may be a better longer-term performance measure and it is a quantitative measurement. Density is the net effect of mortality, recruitment, and life span (demographics).

Permanent sites for doing leaf punch productivity measurements should continue and more sites should be included in areas that may be initially affected by changes in water management practices. Leaf productivity is a dynamic measurement that relates to carbon balance and should be sensitive to the physiological condition of Thalassia. Leaf productivity and turnover are also important components for modeling carbon flow to higher trophic levels. With the development of a reliable submersible field instrument, pulse amplitude modulated fluorescence measurements provide a nondestructive rapid assessment of photosynthetic performance which are very sensitive to stress at short time scales and small spatial scales. Future work should be directed to calibrating PAM data against leaf punch data and determining sources of variability in quantum yields in seagrasses. The recent observations of widespread flowering in Thalassia and Halodule should be examined in the context of resource allocation, recruitment potential, life history strategies, and how water quality may affect the roles of vegetative versus sexual reproduction in the establishment and maintenance of seagrass populations. Flower production, sex ratios, and fruit production indicate sexual recruitment potential, however, successful seedling recruitment is what is may be most important to recovery and expansion. Sexual reproduction may be a good performance measure; sex is costly so reflects system health.

Statistical modeling was commissioned by the Restudy consistent with recommendations from the 1998 Seagrass Modeling Workshop (Fourqurean et al.). The goal of this work is to seek relationships between water quality variables and seagrass species composition and abundance, which if sufficiently strong, can be used to predict the effects of various alterations in Florida Bay salinity regimes. The statistical models developed in this project will be used in conjunction with output from other models to predict the effects of Restudy scenarios on the benthic habitats of Florida Bay. The statistical model will not address mechanisms or degree of change in water quality that result from Restudy scenarios; other models (like the NSM, FATHOM, and the Florida Bay salinity transfer function models currently employed by the Restudy) must simulate water quality changes across Florida Bay that will provide the input to the new models developed in this project. As a consequence, it is anticipated that the benthic habitat change predictions of the statistical models will be the most reliable in the regions most closely coupled with water management practices, i.e. in the enclosed, mangrove-lined estuaries on the fringe of Florida Bay. As the fidelity of the physical water quality models to the behavior of the system declines, the benthic changes predicted by our model will also decline. But, since the statistical relationships will be based on data from more marine areas as well as upper estuaries, the basic relationships between actual water quality (not modeled) and benthic habitats will be robust.

In addition, other seagrass modeling efforts have recently been initiated. These include the development of both seagrass unit models for Thalassia and Halodule and a landscape model that will take output from the unit models.

Madden and Burd have initiated the development of process-based unit models for Thalassia and Halodule. Smith et al. are developing a hierarchical approach to modeling the interaction between plant and physical processes in the Florida Bay that involves two distinct spatial scales – the demographic unit (ca. 10 m²⁾ and the landscape unit (ca km²). Plant processes will be modeled at a spatial scale ca. 10 m², the spatial scale that we refer to as the demographic unit. A preliminary landscape model of Thalassia that explicitly relates patterns of photosynthesis, respiration and carbon allocation to environmental conditions that include salinity, temperature, PAR and nutrient availability ahs been developed. Physical processes such as sedimentation, decomposition and nutrient cycling will be modeled on a spatial scale of km², defined as the landscape unit. The approach provides a hierarchical framework where the demographic units used to simulate plant processes exist in the context of the landscape units, which will define the underlying physical environment. The plant characteristics that are relevant to the feedbacks with the physical environment, such as primary productivity, inputs of dead organic matter, etc. are described statistically for the demographic units and used to define the biological environment for each landscape unit. In this manner, the framework functions as a dynamic, interactive GIS where each parameter and process is described and simulated at the appropriate time and space scale.

While not part of the seagrass research program, the paleoecological investigations add relevant information for the understanding of how recent seagrass changes fit into the matrix of historical expansion and contractions of seagrass cover. At present it appears that seagrass coverage, as estimated both by the abundance of seagrass-associated microfossils and by chemical signatures in the sediments, has shown repeated cycles of expansion and contraction. Thus, the recent changes in seagrass cover apparently are not unprecedented. However, the paleoecological data come from only a few selected locations, limiting the confidence that can be placed in their general applicability. Therefore, it would be extremely valuable to have additional paleoecological information from more sites throughout the Bay.

REFERENCES

PHILLIPS, R. C. AND E. G. MEÑEZ. 1988. Seagrasses. Smithsonian Contr. Mar. Sci. No. 14, 104 pp.

THAYER, G. W. AND A. J. CHESTER. 1989. Distribution and abundance of fishes among basin and channel habitats in Florida Bay. Bull. Mar. Sci. 44: 200-219.