Multi-taxon analysis of the "white zone", a common ecotonal feature of
South Florida coastal wetlands.
In Porter, J. W.  & K. G. Porter (eds.) Linkages Between Ecosystems in the South Florida Hydroscape:  The River of Grass Continues.
by
Michael S. Ross, Evelyn E. Gaiser, John F. Meeder and Matthew T. Lewin
 
July 16, 1999
                 Southeast Environmental Research Program, Florida International University, Univeristy Park, OE-148, Miami, Fl., 33199
                     Florida International University, Southeast Environmental Research Program, Copyright © 1998. All rights reserved
E-mail contact: rossm@fiu.edu

 
Introduction 
 
Methods 
 
Results 
 
Discussion 
 
Summary 
 
References 
 
Appendix A
 
 
Introduction

 

    Coastal ecosystems of South Florida are arranged in predictable zonal patterns that have been described by numerous authors (e.g., Harshberger 1914; Davis 1940; Stephenson and Stephenson 1950; Egler 1952; Craighead 1971; Ross et al. 1992; Ross et al. in press). These studies illustrate that local variations in biotic zonation are influenced by a number of interacting physical factors, including bedrock and surface topography, sediment type and depth, coastal exposure to wave energy, salinity and tidal amplitude of the adjacent marine system, and magnitude of upstream sources of fresh water. Coastal gradients in south Florida generally begin in fringing mangrove swamps or rock barrens and terminate in upland forests or freshwater marshes. Gradient length is highly variable, ranging from tens of meters to more than ten kilometers.

    A common feature of many South Florida coastal areas is a zone of low plant cover, clearly recognizable as a white band on black and white or color infrared photos, sandwiched between more densely vegetated fringing mangrove and interior ecosystems. Egler (1952) described such a "white zone" midway within the Southeast Saline Everglades (SESE), a broad expanse of marine, brackish, and fresh water wetlands extending south and east from the Atlantic Coastal Ridge. Stephenson and Stephenson (1950) also distinguished a white zone along rocky shorelines in the Florida Keys, where it is the most interior and least frequently inundated of five coastal bands recognizable on the basis of substrate color. With more limited exposure to waves, mangroves colonize such coastlines, but the white zone frequently remains as a reflective band beyond the reach of most non-storm tides, surrounding the broadleaved hardwood hammocks of the interior (Ross et al. 1992).

    What is known of the nature and dynamics of the white zone? In his monograph, Egler (1952) did not speculate on the causes of the sparse plant cover of the white zone, focusing instead on the transitional character of its vegetation, which included both graminoid elements from the interior marshes and dwarf mangrove forms from the coastal swamps. He argued that the mixture of interior and coastal taxa in the white zone reflected a balance between several types of disturbance. For instance, the periodic occurrence of fire and freezing temperatures were seen to limit the invasion of the fire- and cold-sensitive mangrove species into the interior. In contrast, major tidal events associated with hurricanes and tropical storms dispersed mangrove propagules well into the interior, and raised soil salinities above the tolerance of some freshwater species. Ross and others followed on Egler's pioneering work, sampling vegetation, soils, and site characteristics within a diffuse network of sites throughout the SESE, then examining aerial photos of 1940 and 1994 for evidence of changes in the position of the white zone that might be associated with sea level rise and/or water management. They found that the inner boundary of the white zone had shifted toward the interior by an average of 1.5 km during the period, with maximum change in areas cut off from upstream water sources by canals or roads. Surprisingly, the plant species composition of the expanded white zone had also changed in the ca half-century since the earlier study, i.e., there was a general increase in the relative abundance of mangroves, at the expense of graminoids more commonly associated with brackish or fresh water conditions (Ross et al. in press).

    These results indicate that, while the position of the sparsely vegetated white zone seems to provide an effective indicator of the extent of marine influence, the gradient in plant species composition responds independently to the same coastal influences. To clarify the significance of the white zone within coastal wetlands, and to elucidate its relationship with the local biota, we undertook a detailed examination of a single transect in the easternmost portion of the SESE, bordering on southwestern Biscayne Bay (Fig. 1). Since the white zone demonstrates potential for indicating environmental change, it is important to determine if its physical appearance is linked to biotic pattern. The relatively fine detail of the data collected along this transect allowed us to offer some more refined inferences about the relationships of several important environmental variables with the distributions of species (and assemblages) within three taxonomic groupings, and with their apparent productivity. These groups include the two main primary producers, plants and algae (primarily diatoms), and one group of consumers, the benthic invertebrates (predominantly mollusks). We hoped that understanding the ecological sensitivities of these assemblages might aid in developing indices of changing environmental conditions. Our analyses of these data focus on the following questions:

     1. What structural, compositional, or physical features define the white zone within SESE coastal wetlands?

    2. How are individual species and species assemblages arranged along important environmental gradients? Do the three taxonomic groups
        differ in their apparent sensitivity to the measured environmental variables?

    3. Is it possible to define distinct compositional ecotones, or boundaries, within the coastal gradient for any of the three groups of organisms?
        If so, do the positions of these ecotones correspond among groups? Do they correspond to the structurally defined boundaries of the white
        zone?



Methods

    Study Area---The study area selected was south of Turkey Point Nuclear Power Plant in southeastern Miami-Dade County, Florida (Fig. 1). These swamps contain the most extensive remaining example of the coastal wetlands of southern Biscayne Bay; although they have been impacted by drainage and other hydrologic modifications, they are less disturbed than sites further north within the SESE. Previous investigations of Egler (1952) and Ross et al. (in press) suggested that the white zone is well defined in the area.

    Field Sampling---During a two-day period in April 1997, a transect was established roughly perpendicular to the Biscayne Bay shoreline. Circular plots of radius 5 m were established at 100-m intervals, beginning 50 m from the coast and extending to the L-31E levee, ~ 3900 m distant. Shoot cover of all macrophyte species was estimated in octave categories within each plot, and the height and species of the tallest plant was recorded. At the center of every second plot, a 10-cm diameter PVC tube was pushed into the sediment to wall off surface water from entering, and a soil core was extracted from within the walls of the tube with a 4-cm diameter x 30-cm deep bucket auger. The core hole was then pumped free of water and allowed to refill from the sides. A sample of approximately 30 ml of porewater was drawn, and its specific conductivity was determined in the field with a Hanna 8733 conductivity meter.

     A survey of surface topography along the transect was completed in May 1997. Elevations of the ground surface at the vegetation plots and intermediate points were determined by theodolite, beginning from a USGS benchmark of known elevation adjacent to the L-31E levee. Soil profiles were examined intermittently along the transect, and depth of sediment deposited by Hurricane Andrew was determined at 100-m intervals by trenching and by probing with a 1-cm diameter aluminum rod.

    In April 1998 the transect was visited for a third time to determine the coastal distributions of diatoms and benthic invertebrates. Plots were established at 100-m intervals, and porewater conductivity was determined by methods identical to those used in 1997. Two additional reference plots were established on the west side of the canal, ca 4020 and 4120 m from the coast. Benthic invertebrates and diatoms were sampled from the surface sediments. At each plot, the presence and appearance of a periphyton mat was noted. For diatoms, five 1-cm deep cores were taken through the mat or surface sediments with a 4.2 cm2 coring device and composited. For invertebrates, the upper 1 cm of soil was collected from a 750 cm2 area, immediately washed through a 2 mm mesh sieve, and frozen upon return to the lab. Previous work determined that sediment accretion rates at Turkey Point averaged approximately 2 mm per year (Meeder et al. 1996); on this basis, we estimated that our collection represented about five years of accumulation.

    Laboratory Procedures---In the laboratory, surface soil samples were thawed for invertebrate analysis, and live and dead individuals identified to species using regional literature (Pilsbry 1946; Tabb and Manning 1961; Moore 1964; Abbott 1974; Thompson 1984). For diatoms, surface sediment cores were thawed, homogenized with a blender and quantitatively subsampled. A portion was dried to constant weight in a 600 C oven (dry mass) and again in a muffle furnace at 500 ° C (ash-free dry mass). A second subsample was cleaned for diatom analysis using a combination of strong acids and chemical oxidizers to remove calcite and organic material. Permanent slides were prepared and 500 diatom valves per slide were counted and identified on random transects. Identifications were made using standard and regional literature (Schmidt 1874-1959; Hustedt 1927-1966; Hustedt 1930; Patrick and Reimer 1966, 1975; Navarro 1982; Foged 1984; Podzorski 1985; Round et al. 1990; Krammer and Lange-Bertalot 1986-1997; Lange-Bertalot 1993). Digital photographs were taken of each taxon and stored in an Image Pro® database linked to corresponding ecological information.

    Analytical Procedures---GPS coordinates of the coastal and interior ends of the transect were determined and superimposed on a USGS digital orthophoto quarter-quadrangle (1:12,000) of January 1994. The photo was the basis for identifying and digitizing the interior and coastal boundaries of the white zone in the vicinity of the transect.

    We examined plant, benthic invertebrate and diatom compositional patterns with respect to the measured physical variables, e.g., distance to shore, porewater conductivity, and elevation. Species abundances for the three groups (percent cover for macrophytes; total density for living and dead invertebrates and diatoms) were analyzed individually and at the community level. For the community analysis, the abundance of species that occurred in two plots or more were relativized and then fourth-root transformed to reduce the influence of the most abundant taxa. The data were ordinated with non-metric multidimensional scaling (NMDS) in PC-ORD version 3.11 (McCune and Mefford 1995), using the Bray-Curtis dissimilarity metric (Clarke and Warwick 1994).

    The strength of association between the physical variables and community composition was assessed in several ways. Using the BIO-ENV procedure in the Primer program (Clarke and Warwick 1994), we calculated a weighted Spearman rank correlation between the site-by-site dissimilarity matrices (Bray Curtis index) based on the transformed species abundances and the individual dissimilarity matrices for distance to coast, porewater conductivity, and elevation.

    We applied a weighted-averaging (WA) approach to estimate the optimum (abundance-weighted mean) and tolerance (weighted standard deviation of optimum) of each species for each environmental parameter based on its abundance along the gradient, using WACALIB 3.3 (Line et al. 1994). In this procedure, the value of an environmental parameter for each site is estimated as the average of the optima of the species present, weighted by their abundances and possibly tolerances (ter Braak 1987). Environmental values calculated from species distributions are then compared to the observed values using regression. If the relationship is strong, the calibration model can then be applied to modern monitoring or fossil biotic data to infer environmental quality with a measured degree of confidence. Because taxa had unequal occurrences, we followed the recommendations of Birks et al. (1990) and used the number of occurrences to adjust the tolerance assigned to each taxon. By incorporating tolerances into the model, we weighted species with narrow distributions along the gradient more heavily than species with broadly dispersed or erratic distributions. We used classical regression (Birks et al. 1990) to eliminate shrinkage in the range of inferred values, because it resulted in the most evenly distributed residuals. We used bootstrapping to calculate the root mean squared error of prediction (RMSE; Birks et al. 1990). The resulting bootstrapped estimates of the environmental variable were then plotted against the observed values and a regression coefficient calculated. Residuals from the regression of measured vs. species-inferred distance values were examined for trends, then plotted against measured conductivity and elevation values to determine if distance estimates were biased.

    We used two methods to examine the continuity of the biotic assemblages along the transect, and to identify any breaks or ecotones within the gradient. For all three groups, we visually identified peaks in the sequence of dissimilarity scores betweeen adjacent sampling stations, using a five-plot (500 m) moving window to smooth the curves (modified from Whittaker 1960).

    We also assessed the clustering of species boundaries (lower boundary, most coastward occurrence; upper boundary, most interior occurrence) by comparing the observed distribution with a null expected distribution derived through a Monte Carlo simulation, and obtaining a test statistic (Auerbach and Shmida 1993). This analysis was only applied to diatoms, because the number of plant or mollusk species boundaries was too small to provide adequate statistical power.

Results

    Delineation of the white zone---Aerial photos from 1994 indicate that the white zone was a broad and conspicuous band throughout the coastal swamps south of the Turkey Point power plant (Fig. 1). The outer boundary was very distinct, intersecting our transect at about 170 m from shore. The inner boundary, however, was less clearly defined. Beginning at approximately 2760 m, a speckled transitional zone of approximately 350 m width separated the white zone from distinctly darker-signatured vegetation. This transitional band, which may signal an incipient expansion of the white zone, was variable throughout the Turkey Point region, and was entirely absent in some areas.

    Physical variables---Soil profiles of the transect were affected by marine sediments deposited by Hurricane Andrew (August 1992). Deposits thinned from 12-16 cm adjacent to the coast to 2 cm or less at 2500 m. Beneath the storm deposits, sediments were characterized as peaty marls within 500 m of the coast; throughout the interior, they were predominantly marls (calcitic silts and muds) precipitated under fresh water conditions. West of 3800 m the marls are organic-rich, approaching peats in areas where sawgrass was dominant. Elevation generally increased from coast to interior, ranging from 15 cm above sea level at the coast to 40 cm ASL at the base of the L-31E levee (Fig. 2). Local depressions (ephemeral drainages) were present between 1000 and 2000 m; however, the overall correlation of elevation with distance was very strong (r = 0.95). Porewater conductivity in 1998, however, was only weakly correlated with both distance to the coast and elevation (r= -0.18 and r = -0.29, respectively). Through 2000 m, conductivity was similar to nearshore surface water (35-45 mS/cm/sec; Fig. 2). Beyond 2000 m, conductivity gradually rose, reaching an elevated plateau of 50-65 mS/cm/sec at 2900 m, then dropped sharply to near-freshwater conditions (<5 mS/cm/sec) within a few hundred meters of the levee. Similar spatial patterns in 1997 (not shown) suggest that the conductivity profile in Fig. 2 is fairly representative of conditions during the late South Florida spring, when osmotic moisture stress is apt to be highest during the year.

    Vegetation pattern---Total macrophyte species cover decreased from about 50% adjacent to Biscayne Bay to <10% between 900-2000 m (Fig. 3). Over the same sequence, the height of the tallest plant (R. mangle in all plots ) decreased from 125 cm to about 50 cm, with the most abrupt decrease occurring near 1200 m. Cumulative species cover increased sharply beyond 2000 m, exceeding 100% at several points near the L-31E levee. Maximum plant height in the interior half of the transect was variable and difficult to interpret because the tallest species differed among plots. The white zone at Turkey Point is therefore characterized by vegetation cover < 50% and maximum vegetation height <1 m.

    Spatial patterns in plant community composition are summarized in the NMDS ordination (Fig. 4), with sites labeled according to their distance from the coast. The general increase in Axis 1 scores with distance from Biscayne Bay indicates strong coastal zonation in species composition. On the right side of the diagram, stations 2100-3300 were also distinguished from stations 3400-4100 by their positions on Axis 2. The relative isolation of the latter group suggests that the freshwater community was somewhat distinct from the rest of the gradient.

    The distribution of individual species along the transect provides further detail about the zonal nature of the vegetation (Fig. 5; Appendix A). Two species (R. mangle and A. germinans) were found throughout all portions of the transect east of the levee, but the other nine species had ranges of 2500 m or less (Fig. 6). Within the first kilometer or so of Biscayne Bay, R. mangle and A. germinans were the sole vascular plant species present, with the former much higher in abundance. Moving inland to about 1300 m, L. racemosa and D. spicata became present at low abundances, and the cover of A. germinans increased. Juncus romoerianus was first sampled at about 2000 m from shore, but became much more abundant after 2400 m, coincident with a decline in the importance of R. mangle. Little subsequent change in the species mix was evident until about 500 m from the L-31E levee, where Conocarpus erecta and Cladium jamaicense became common in the marsh, and D. spicata cover increased sharply. In general, plant species richness increased from coast to interior (Fig. 7), with new species entering the community throughout and only Ruppia maritima dropping out entirely east of the L-31E levee. Over the same gradient, shrubs were replaced by graminoids as the dominant growth form. The two disjunct sites in the marsh west of the L-31E levee were similar in species composition to much of the freshwater Everglades, with C. jamaicense the dominant species and Eleocharis cellulosa in a subordinate canopy position.

    The strong zonation in species composition illustrated in Fig. 4 and 5 is further supported by results of the BIO-ENV and weighted averaging regression procedure. Plant species composition showed a stronger correlation with distance than with elevation or soil conductivity in Spearman rank correlations (r=0.75 vs. 0.59 and 0.17, respectively). This was confirmed by the weighted-averaging procedure also detected a strong fit between observed distance and the distance inferred from species composition (R2=0.87; RMSE = 526; Fig. 8). Residuals from this model exhibited little pattern with distance beyond the first 1 km of the transect, where vegetation composition was very homogeneous. More importantly, the residuals were uncorrelated with either porewater conductivity or elevation.

    The sequence of dissimilarities in plant species composition among adjacent points indicates a pair of compositional breaks at 1200 and 1700 m from shore (Fig. 9). These represent a relatively broad ecotonal zone associated with the shift from a virtually pure R. mangle community to one in which dominance was shared more evenly with several other halophytic species (e.g., L. racemosa, D. spicata, A. germinans, and Juncus romoerianus). A second ecotone is indicated between 3200-3500 m. High dissimilarities in this zone mark the transition from the halophytic assemblage described above to a C. jamaicense-dominated marsh community usually associated with freshwater conditions. This inner ecotone is also associated with a shift from R. mangle/A. germinans to L. racemosa/C. erecta dominance among woody plants.

    Diatom pattern---Diatoms were incorporated in a coherent benthic periphyton mat in some sections of the Turkey Point transect, but these were for the most part interior to the white zone. No mat was present until 1800 m, where patches of green filamentous algae were first observed at the base of some of the larger mangrove shrubs. Between 2200-3000 m, the sediment intermittently included a surface (<1 cm) crust that was reddish-yellow in color. This surface algal mat became continuous after 3000 m, thickening to 2 cm or more by 3600 m. The thickest algal mats were found at the stations west of the L-31E levee. Ash-free dry mass (AFDM) estimates of periphyton biomass reflected these visible changes in mat thickness (Fig. 3). AFDM ranged from 10-20 g/m2 in interior sites (2200-4120 m) and dropped to a low of 2-4 g/m2 in the middle portion of the white zone (200-2200 m). An increased organic component in coastal sites (0 and 100 m) may reflect an increase in mangrove-derived fibrous peat in the surface sediments rather than an increase in periphyton biomass.

    The NMDS diatom ordination demonstrated a zonation with distance from the Biscayne Bay coast that was even stronger than that exhibited in the vegetation (Fig. 4). Sites were essentially ordered by distance along Axis 1. Dissimilarity was low between sites representing the inner two-thirds of the white zone (1200-3000 m), resulting in a clumped distribution in the center of Axis 1. Interior (3100-4120 m) and coastal (0-1100 m) sites were distinct from the rest of the transect, yet a high degree of variation in their assemblages was denoted in Axis 2.

    A total of 154 diatom taxa, comprising 40 genera, were identified from the 41 sampling locations at Turkey Point. Because the flora of tropical coasts has been poorly examined taxonomically, we had difficulty assigning definitive names to many taxa and resorted to a numbering system for species and varieties of questionable identity. Pictures of some of these taxa, assembled according to ecological affinities along the coastal gradient, are shown in Fig. 10.

    Distance optima and tolerances estimated for each taxon by weighted-averaging (Appendix A) can be used to discriminate cosmopolitan taxa from those with more definite range restrictions. In comparison to plants, relatively few diatom species were cosmopolitan in range, and progressively smaller ranges were represented by increasing numbers of species (Fig. 6). Most cosmopolitan taxa were also relatively common in individual samples. For instance, Navicula rhynchocephala var. 01, Encyonema evergladianum, Navicula sp. 02 and Nitzschia amphibia were common at most sites along the coastal gradient. However, most taxa had more restricted ranges, including Brachysira neoexilis, Nitzschia palea var. debilis, N. serpentiraphe, Mastogloia smithii var. lacustris, Encyonema vulgare and E. silesiacum which were restricted to freshwater sites adjacent to the L-31E canal. Moving coastward, several taxa including Caponea caribbea, Caloneis sp. 01, Amphora sp. 01, Rhopalodia acuminata and Mastogloia erythreae attained highest abundance in the center of the gradient, thereby characterizing the white zone. The coastal flora was most distinctive - 55 taxa were restricted to the intertidal environment. Of these, several occurred in high abundance, including Tryblionella granulata, Psammodictyon mediterraneum, Tryblionella debilis var. 01, Caloneis permanga, Caloneis sp 02, Campylodiscus sp. 01, Surirella fastuosa, Mastogloia decipiens, Cocconeis scutellum, and Achanthes submarina.

    In contrast to the pattern among plants, diatom species richness was highest near the coast, with a maximum of 68 species at 100 m. Diversity declined rapidly over the first 500 m of the transect, then more gradually. Species richness was lowest in the freshwater marsh west of the levee, where only 15-25 species were found in each plot (Fig. 7).

    The BIO-ENV procedure detected a strong correlation of diatom compositional dissimilarity with distance and elevation (r=0.83 and 0.74, respectively), but little relationship to soil salinity (r=0.23). The WA model also produced an excellent fit of diatom-inferred distance to observed distance (r2=0.98, RMSE=253; Fig. 8). Nevertheless, residuals from the WA model predictions exhibited conspicuous patterning (Fig. 8). Distances were generally overestimated between 400-1700 m and beyond 3000 m, and underestimated adjacent to the coast and between 1800-2900 m. Further analyses of these residuals revealed no correlation with elevation, but a significant correlation with porewater conductivity (r=-0.39; p<0.02). Underestimates of distance were associated with zones of high conductivity, and overestimates were associated with fresher water.

    Two pronounced discontinuities in species composition were evident in the dissimilarity sequence for diatoms (Fig. 9). The first was a relatively sharp peak at 300 m and the second was a broader area of change between 3000-3300 m. These two ecotones reflected the presence of sizable groups of exclusively coastal and fresh water species, respectively, that were unable to thrive in the interior of the coastal gradient. The diatom dissimilarity sequence thus resembled the macrophyte sequence, except that the coastal ecotone in the diatom gradient was positioned about 1 km closer to Biscayne Bay.

    The analysis of diatom species boundaries confirmed the presence of two ecotonal zones, the first within a few hundred meters of Biscayne Bay and a second about 3 km inland (Fig. 11). Table 1, below, shows the site locations having significantly more or fewer than expected first or final appearances (boundaries) of species than would be expected in a random model.

 

Table 1. Sites with significantly more or fewer upper and lower diatom range boundaries than expected.
  Distance from coast (m):
More Upper Boundaries 100, 200, 3000 
More Lower Boundaries 100, 3100, 3200, 3300, 3500, 3600 
Fewer Upper Boundaries No sites
Fewer Lower Boundaries No sites
 

    Benthic invertebrates---Mollusks dominated the benthic invertebrate assemblage. Only one species of foraminiferan, Penoroplis planatus, comprised more than 1% of any sample. The other non-molluscan invertebrates sampled were: crabs (Uca, only six specimens), annelids (two specimens), and a second estuarine foraminiferan, Ammonia beccarii, that was usually below our minimum size limit. Density of invertebrates decreased precipitously with distance from the coast. Maximum density of live invertebrates of all taxa was 4600 m-2 at 500 m, and invertebrates were entirely absent at several sites beyond 3400 m. More than 99 % of the invertebrates sampled were dead specimens, which also exhibited an identical trend of declining density with distance from shore. Live invertebrates were predominantly immature forms. The most frequently found live specimens, in decreasing order, were Batillaria minima, Cyrenoida floridana and Certhidea beattyi near the coast, and Littoridinops monroensis closer to the interior (Fig. 12).

    The NMDS invertebrate ordination resembled the diatom ordination in general pattern (Fig. 4). The samples generally arc from sites closest to the coast to sites furthest from it, with a diffuse coastal arm (sites that include marine transported gastropods) and a more compressed interior arm (sites associated with freshwater pulmonates). The tightly grouped core of sites near the middle of the transect were characterized by Battilaria minima, Cyrenoida floridana and Cerithidea beattyi. However, the placement of sites with respect to distance was not as orderly as in the diatom analysis, partly because the coastal invertebrate assemblage is not well-developed at this location, and also because of influx and mixing of offshore marine invertebrates with the resident species.

    Fourteen of the 17 invertebrates were present within 100 m of the coast, but only three taxa extended beyond 3500 m (Fig. 12). Of these, the foraminiferan, Peneroplis planatus, and the mollusks Acteocina canalicolata, Triphora cf. nigrocincta and Rissoina catesbyana were residents of seagrass beds in the nearshore marine area. The result is a rather constant decline in species richness with distance (Fig. 7). This pattern resembles the diatom gradient, but invertebrates differ in the absence of species distributed exclusively in the freshwater portions of the transect. In general, invertebrate species are more broadly distributed along the transect than diatoms or plants, with only a single taxon (Triphora cf. nigrocincta) whose range was less than 500 m (Fig. 6).

    The correlation between distance and invertebrate species composition measured by BIO-ENV was somewhat lower than for plants and diatoms (r=0.68); little correlation was found between invertebrate assemblages and conductivity or elevation (r=0.27 and 0.53, respectively). However, based on the weighted-averaging regression, the invertebrates were intermediate between plants and diatoms in the strength of their relationship to distance. Unlike the other two groups, the fit of the unweighted WA model is superior to that of the tolerance-weighted model (R2=0.94, RMSE=457 and R2=0.90, RMSE=546 respectively, Fig. 8). Between 1800 and 3000 m, there is some evidence of patterning in the residuals from the unweighted model, but this pattern is uncorrelated with conductivity or elevation.

    Dissimilarities between neighboring invertebrate assemblages are low and relatively constant within the coastal two-thirds of the transect, though there is a plateau of slightly more rapid change between 1200-1500 m (Fig. 9). Dissimilarity increases sharply between 2800-3200 m, and remains high thereafter. This is the region where the freshwater pulmonates and the hydrobiid, Littordinops monroensis, become abundant. West of the levee, pulmonate density and diversity increases to typical Everglades marsh levels (Meeder et al 1996). The increase in point-to-point variability in the inner one-third of the transect is not associated with the addition of new species that are absent from coastal locations. Rather, the high variation is associated with a less speciose mollusk community, with minor species dropping out entirely, and dominant taxa becoming much less abundant than in coastal sites.
 

Discussion

The nature of the white zone

    At Turkey Point, the white zone begins abruptly at about 200 m from the coast, but its interior boundary at 2.5-3 km distance is less easily defined. This pattern parallels that observed by Egler (1952; p. 256), who described the interior boundary as "... distinct, but not knife-edge, definite enough so that the changeover usually occurs within half a kilometer...". Our analyses suggest that its white appearance on aerial photographs is attributable to sparse cover of low-growing plants, in conjunction with the reflective quality of the exposed marl soils or the fresh storm deposits that cover them. The absence of a continuous periphyton mat in the white zone at Turkey Point may also be a contributing factor, because the chlorophyll present in algae is expected to produce a red color on color-infrared photos. However, the spectral signature of periphyton is complicated by the depth and quality of surface water, if present, and perhaps by the structure, composition, and calcite content of the mat itself (e.g., Rutchey and Vilchek in press). Moreover, we have observed a well-developed periphyton mat in other South Florida locations within the white zone, though that is not generally the rule.

    The sparse cover and dwarfed morphological forms that characterize the vegetation of the white zone suggest that these are very unproductive wetland ecosystems. Ross et al. (unpublished manuscript) recently estimated total aboveground plant production in a nearby dwarf mangrove basin in Biscayne National Park (BNP), where maximum plant height and total plant cover were considerably higher than midway in the Turkey Point white zone. Even in the relatively favorable BNP location, production was only 8600 kg ha-1 yr-1, which was 29% that of an adjacent coastal mangrove forest. At Turkey Point, productivity was undoubtedly much less. In the current study, we were not able to directly address the causes of the white zone's low productivity, and can only suggest a few factors that may combine to limit biological activity:
 

  1. Wide seasonal fluctuations in salinity and moisture content. As in the relatively steep, compressed Florida Keys gradients described by Stephenson and Stephenson (1950), the white zone of the more gradual mainland coast occupies predominantly a supratidal position, i.e., salt water reaches the interior only during spring tides and storms, and even its coastal portions are not wetted regularly by the semi-diurnal tides. Once tidal waters do enter the white zone, they drain slowly and unevenly, tending to pool in local depressions, where they evaporate over days or weeks. By April of a dry year, the process may result in both surface desiccation and saline or hypersaline porewater. These environmental conditions contrast sharply with those encountered during the fall months, when the combination of high tides and heavy rains often cause persistent flooding with fresh or brackish water. While certain plants may grow reasonably well in one or the other of these conditions, adaptations which allow them to persist in both may require physiological tradeoffs which result in reduced productivity (Ball 1988). Environmental variability may be moderated at the interior boundary of the white zone by freshwater influx, and at the coastal end by recurrent tidal flushing.
  2.  

  3. Absence of fresh water input from upstream sources. In our recent examination of historical changes in SESE coastal wetlands over the past five decades (Ross et al. in press), we found minimal change in the position of the white zone in Taylor Slough and north of Long Sound, but much more dramatic landward shifts north of Joe Bay, in the triangular area between Card Sound Road and US 1, and south of Turkey Point. We attributed this pattern to differences in access to upstream fresh water sources, because the latter three basins were cut off or deprived of water as a result of water management activities, while water supply to the first two basins remained high during the period. The mechanism for this effect is not entirely clear, but two possibilities surface: (1) by supplementing the nutrient supply to coastal marshes, upstream water sources may enhance productivity, thereby narrowing the width of the white zone, and (2) maintenance of connectivity between freshwater and marine water sources may increase production by buffering ecotonal areas from the seasonal fluctuations discussed above.
  4.  

  5. Heavy marl soils that may contribute to phosphorus limitation. White zone soils in much of the SESE are marls of the Perrine series, which are silt loam in texture (USDA 1996). As such, these soils present both physical and chemical impediments to plant growth. Because of their heavy texture, they drain slowly, and therefore may undergo long periods of anaerobiosis. Moreover, owing to their calcareous nature and high pH (typically 7.4 - 7.8), phosphorus availability may be limited. In the Biscayne National Park dwarf mangrove ecosystem described above, where soils were similar to those of the Turkey Point white zone, Jacobson et al. (unpublished manuscript) found porewater with very high alkaline phosphatase exoenzyme activity, indicating phosphorus limitation. Inasmuch as red mangrove leaf phosphorus content was also reduced in these locations in comparison to the coastal fringe, there is strong support for the notion that phosphorus availability limits white zone productivity.
  6.  

  7. Sporadic occurrence of natural disturbances. Egler (1952) believed that fire was an important factor in maintaining a mixture of graminoids and mangroves in the white zone. Today the large graminoids capable of carrying a fire (e.g., C. jamaicense and J. romoerianus) are absent or of low abundance in most of the white zone, and it is difficult to imagine these areas burning. Periodic freezes, however, may be more important in controlling the expansion of sensitive mangrove species such as R. mangle and L. racemosa. While recovery of the pre-disturbance canopy is probably achieved in less than five years in relatively productive mangrove communities, freeze events may occur frequently enough to limit mangrove cover in the white zone, where recovery is likely to be slower. South Florida coastal areas have been affected by killing freezes at least four times in the last two decades (Olmsted et al. 1993, Ross pers. obs.).
 
Species and community responses to coastal gradients

    Species of each taxonomic group sort predictably into bands that parallel the SESE coastline. Using abundance-driven optima and tolerances for distance from the coast, spatially explicit niches can be defined for most species. Detection of a strong assemblage response to the coastal distance gradient is to be expected, since sites are by definition spatially autocorrelated. However, the strong linear relationship observed for all three groups is striking, and suggests that distance effectively incorporates trends in physiologically significant environmental variables. Our sampling strategy was to measure several environmental factors (i.e., salinity and elevation, which is related to water depth) during the dry season, when we expected osmotic stress to be maximal; however, we found little or no correlation of these variables with the assemblage data. This suggests that species distributions within these tidal areas reflect the entire annual range of variation in such variables, rather than a level measured during any single time of year. By monitoring a suite of parameters along a coastal gradient regularly, one could characterize sites by parameter ranges instead of averages, and perhaps define ecological forcing factors more directly.

    Strong and analogous responses to distance are apparent among the three groups of species, indicating an overriding effect of this composite variable on assemblage composition. Nonetheless, the groups vary in the spatial scale at which they respond to variation in the coastal environment. By exploring these differences, we can illuminate ecological constraints that are imposed on these groups by their disparate natural histories.

    Vascular plants---Although several vegetation analyses show that plant community composition changes in a fairly regular manner with distance from the shore at Turkey Point (Fig.s 4, 5, 8), this clinal sequence is also marked by prominent discontinuities at 1200-1700 m and at about 3200 m from shore (Fig. 9). The more interior of these coincides well with the inland boundary of the white zone, but the coastal discontinuity is located well within the white zone borders. These ecotonal zones do not reflect wholesale shifts in species composition, but instead signify changes in the relative abundance of taxa, in conjunction with the addition of new species to the assemblage with increasing distance from the coast. The observed increase in plant diversity along this coastal transect represents the lower end of a longer gradient which ends in upland forests, especially pine rocklands, where species richness reaches its South Florida maximum (Ross et al. 1992; USFWS 1999).

    Not surprisingly, the increase in plant species diversity along the transect is accompanied by an increase in structural complexity, as graminoids of both rhizomatous and bunch-forming morphologies occupy the openings among the mangroves. Because the changeover to a mixed assemblage of graminoids and shrubs occurs midway within the white zone, it may be divided into inner and outer subzones on the basis of vegetation. While these subzones appear to share a low production potential, the contrast in their structural organization and physical setting suggests differences in disturbance regime (fires and freezes more frequent in inner zone, storm-associated sedimentation more common in outer zone), as well as in other ecosystem properties, including nutrient cycling characteristics, and faunal and algal associations. Thus, the diatom and mollusk assemblages we studied may be expected to respond to vegetation differences that affect light availability, protection from desiccation, and the types of substrate accessible for colonization and growth.

    All of the wetland plants sampled at Turkey Point are rooted, emergent, woody or herbaceous perennials. Most or all are characterized by seeds that are readily dispersed by water. Once they have germinated, however, they must survive through several to many annual cycles at a fixed location, though many are able to spread laterally through vegetative means. As a result of their life span and size, plants and plant assemblages may display a relatively coarse-scale response to environmental variation, i.e., integrating over relatively long time periods and wide spaces. This may explain, in part, the lack of association between plant species composition and dry-season salinity observed in this study.

    Diatoms---According to several measures, including weighted-averaging regression/calibration (Fig. 8), BIO-ENV correlation, and range analysis (Fig. 6), diatoms are more sensitive to the coastal distance gradient than either plants or mollusks. This may be attributed to their considerable species richness - an order of magnitude larger than the other two groups. The quality of WA regression models has been shown to be positively correlated with the number of species entering the model (ter Braak et al. 1993). Besides their diversity, however, diatoms are also several orders of magnitude smaller than the other organisms examined here; if their perceived microenvironments are scaled correspondingly, the spatial range at which they integrate environmental variation is also expected to be much smaller. In addition, because diatoms are single-celled organisms with short generation times (~1-2 divisions/day), they respond quickly to environmental changes that directly affect them physiologically. Over the course of the millions of years of competition, diversification, and specialization, these qualities have facilitated the evolution of diatoms into what we now consider to be an ideal indicator organism.

    Compared to other regions of North America, the diatom flora of the southeastern United States has been poorly explored. In South Florida, the flora of the freshwater Everglades has recently received attention related to management and restoration of this fragile system (Swift and Nicholas 1987, Slate 1998, McCormick and O'Dell 1996), yet very little of this research has had a taxonomic focus, and virtually none has emphasized the coastal saline flora. As a result, the flora presented here includes many species that are new to the North American literature and several taxa for which we have been unable to find descriptions.

    Like other biota in South Florida, the coastal diatom flora contains a mixture of temperate and tropical taxa (Long and Lakela 1971). Many of the taxa have not been previously reported in standard North American or European literature developed for predominantly temperate floras, but can be found in tropical floras, particularly of the Caribbean basin (DeFelice 1978, Montgomery 1978, Sullivan 1981, Navarro 1982, Reimer 1996). The flora of the SESE mangroves is genus-rich, typical of tropical coastlines that experience large fluctuations in salinity (Østrup 1913, Mann 1925, Hagelstein 1938, Sullivan 1981, Foged 1984). The SESE mangroves share with these studies the same dominant genera -Mastogloia, Amphora, Nitzschia, and Caloneis. There is substantial overlap between the Turkey Point flora and the diatom floras described for Cuba by Foged (1984 - 55 taxa), coral reefs in the Florida Keys (Montgomery - 41 taxa), mangroves in the Bahamas (Sullivan 1981 - 28 taxa; Reimer 1996 - 15 taxa) and Southwest Florida (Navarro 1982 - 30 taxa) and Jamaica (Podzorski 1985 - 15 taxa).

    Based on their distance response, diatoms of the SESE can be sorted into four categories: (1) coastal, (2) transitional, (3) interior or (4) cosmopolitan (indifferent). The sharpness of the boundaries between the zones varies; the coastal-transitional is graded, with species terminating in clumps spanning several hundred meters, while the transitional-interior is more abrupt with more range terminations than expected over a short distance. The exterior limit to the coastal zone is fixed at 0 m; the interior boundary appears to fall between 1100 and 1600 m, which includes the upper and lower boundaries of 12 and 13 taxa, respectively (Fig. 11). Another boundary within the coastal zone occurs between 100-300 m (see Table 1) that defines the end of the range of many (18) marine taxa. At the other end of the transect, the interior zone has a diffuse lower limit that falls between 2600-3300 m, where 15 and 29 taxa appear and disappear, respectively. Over 70% of these taxa have short ranges (less than 300 m) suggesting that this is a zone of transition along the gradient (also shown in the dissimilarity diagram, Fig. 9b). The rest of the taxa of the interior zone continue into the freshwater marsh west of the canal. The transitional zone can be defined by default as the upper and lower boundaries of the coastal and interior zones, respectively, and includes the visually defined white zone. The fourth group contains taxa that are cosmopolitan, which we arbitrarily defined as having ranges greater than 60 % of the total transect length (2500 m). A final group may be considered that contains remaining taxa with short ranges that extend into more than one zone (only 6 taxa).

    The coastal zone is well defined in terms of diatom composition, containing many (55) species that are not found elsewhere along the gradient. Diatoms inhabiting the coastal zone must be able to withstand tidal fluctuations in water depth and salinity, and associated changes in dissolved gasses and nutrient availability. This zone is also characterized by thick and unconsolidated sediments that were deposited during Hurricane Andrew. These sediments and associated diatom communities are prone to redistribution and scouring during storms and high tides. Because of this, diatom communities of intertidal zones usually comprise a varied assemblage of marine planktonic taxa and loosely attached epiphytes that have been transported into the shallow sediments and benthic taxa that reside within the sediments themselves (Vos and de Wolf 1993). The dominant genera in the coastal portion of the Turkey Point transect (ie., Mastogloia, Amphora, Nitzschia, Tryblionella) are common constituents of soft-intertidal sediments in the tropics (Foged 1984, Navarro 1982). Many of the taxa were also recorded from coral reef communities in the Florida Keys (Montgomery 1978). Several notably marine taxa were probably transported from the ocean to coastal sediments during storm tides (ie., Cyclotella striata, Catacombas gaillonii, Biddulphia spp., Terpsinoë musica).

    Whereas several species extend into the transitional zone from adjacent zones, only a few (19) are restricted to it and these are relatively rare. This decline in species richness into the transitional zone parallels an apparent decrease in algal productivity in the white zone. Few taxa thrive in this portion of the gradient. The diatom assemblage of the transitional zone (~1100-3300 m) is dominated by Navicula rhynchocephala var. 01, Nitzschia amphibia, Amphora holsatica, Navicula scopuloroides, and Synedra filiformis var. exilis. These taxa are wide-ranging and are found in abundance throughout the transect. Only a few taxa are restricted to and therefore diagnostic of the transitional zone. These include Caponea caribbea, Mastogloia fallax var. 01, Amphora sp. 01, Rhopalodia acuminata, Mastogloia erythreae and Fragilaria exigua var. 01. These taxa must be capable of surviving large fluctuations in salinity and frequent desiccation. Of these, complementary ecological data could only be found for Caponea caribbea, which was recently described from collections by Podzorski (1984) from algal mats in hypersaline pools along the Broad River in Western Jamaica.

    The interior zone is characterized by 44 taxa, most of which are common elsewhere in the freshwater Everglades system. This flora has been described in Swift and Nicholas (1987), Slate (1998), McCormick and O'Dell (1996) and is typically dominated by Mastogloia smithii var. lacustris, Nitzschia serpentiraphe, Fragilaria synegrotesca, and Brachysira neoexilis. These diatoms are most frequently found in association with the thick cyanobacterial mats that typify Everglades marshes (Browder et al. 1994). Their decrease in abundance toward the coast reflects, in part, the disorganization of this complex mat community in addition to the coastward increase in osmotic stress. In addition to the expected dominants, we also collected Plagiotropis lepidoptera var. proboscidea, several Mastogloia spp., Navicula scopuloroides from interior sites. These taxa are classically considered halophilic (Patrick and Reimer 1975, Hustedt 1927) and, in our transect, increase in abundance toward the coast. Their presence in interior sites suggests that salinity in interior regions is also frequently elevated or that there is substantial physical mixing during high tides that transports diatoms between zones.

    The final group contains 26 cosmopolitan taxa that are found throughout the gradient, including Encyonema evergladianum, Navicula rhynchocephala var. 01, Navicula trivialis, Navicula digitoradiata, Amphora veneta, Nitzschia amphibia and Mastogloia cf. quinquecostata. These were also among the most abundant diatom species in each sample, suggesting superior dispersal capabilities and an ability to thrive in a variety of environmental conditions. Most of these taxa have been frequently encountered in other areas of the Everglades (ie., E. evergladianum, A. veneta; see Slate 1998) and in the Caribbean basin (N. rhynchocephala, N. trivialis, N. digitoradiata; see Foged 1985, Podzorski 1984), again indicating that they are ecological generalists.

    The assortment of most diatoms into one of three zones along the coastal gradient suggests taxon-specific differences in physiological tolerances to salinity, nutrient availability, pH, and water depth or hydroperiod, all of which have linear or complex relationships to the distance gradient. Of the variables that we were able to measure, salinity significantly biased the results of the weighted-averaging distance model, indicating a diatom response to "pockets" of high or low salinity that are unrelated to the distance gradient. The WA distance model predicted more interior locations for assemblages between 1000-2000 m, related to low salinity values recorded here; conversely, the model predicted more coastal locations for assemblages between 2500-3100 m, where hypersaline pockets occurred within a region of lower salinities (Fig. 8e). In general, however, coastal regions contained mostly halophilic taxa that have been described from benthic collections elsewhere in the tropics, whereas interior regions contained mostly taxa with freshwater affinities. Within the broader area of the white zone, however, taxa range from halophilic to halophobic and there exist an abundance of generalists with an apparent ability to withstand a broad range of osmotic stress. Thus, the boundaries of the white zone are best defined by the taxa comprising well-defined zones adjacent to it, rather than by the presence of diagnostic forms within the zone itself.

    Invertebrates---Like other Everglades consumer assemblages, mollusks numerically dominate the large-bodied invertebrate fauna of the shallow coastal SESE (Rader 1994). The abundance of periphyton available for grazing taxa and particulate organics for filter feeding species contribute to the production and pattern of composition of primary consumers in the Everglades and other coastal environments. In addition, mollusks are highly sensitive to salinity and typically sort predictably along salinity gradients (Moore 1964). While our study found a strong correlation of distance and mollusk assemblage composition (Fig 8b), the effect of salinity was negligible. However, only a short distance of the gradient, 3300-4100 m, was characterized by very low salinity (Fig 2). Freshwater invertebrates thrived in this zone while salt-tolerant species occurred throughout the rest of the transect (0-3300 m). Occasional appearance of larvae or dead adults of marine taxa in freshwater zones is most likely a result of intermittent high or storm tides that transport these individuals landward. In general, mollusk species appear to respond in one of three ways to the coastal gradient: (1) abundant only in coastal regions of the transect, (2) present in low abundance in the center of the transect (white zone), or (3) present in low abundance throughout the transect. The plotted change in compositional dissimilarity between adjacent samples (Fig 9c) shows a major transition between 3000-3300 m that marks the interior extent of coastal and white zone taxa. Samples from this part of the transect also appear as outgroups in the NMDS ordination (Fig. 4c). Looking at species ranges (Fig 12), another less obvious transition occurs between 1100-1700 m that reflects the exterior extent of transitional zone taxa and interior extent of several coastal species.

    Invertebrate diversity increases toward the coast to a maximum of 13 taxa. Like the diatom assemblages, the coastal mollusk fauna includes several species that live within the marsh and others that have been transported into the shallow intertidal by storms. The latter group is characterized by species that are normally found as epifauna on sea grass blades (Abbott 1972) including Acteocina caralicolata, Triphora cf. nigrocincta and Rissoina catesbyana as well as the foraminiferan species Peneroplis planatus. These taxa are marine (tolerating lower salinities for short episodes but not desiccation) and were likely transported into the area as adults by high and storm tides (Wilbur and Yonge 1966). The furthest inland that P. planatus was transported probably represents the maximum distance of marine transport because their small size, light weight, and discoid shape are more conducive to transport than heavier mollusk shells. Other larger taxa drop out between 1600-2000 m, marking the eastern extent of the supratidal mollusk zone. In addition to allochthonous marine species, several taxa were found living in the surface sediments of the intertidal zone. These include the only filter feeder observed, Cyrenoida floridana, and the algal browsing gastropods, Cerithidea beattyi and Batillaria minima. These three species have broad salinity tolerances, thriving in brackish to saline conditions, and are able withstand periods of desiccation. During episodes of drying, these species aestivate by closing their shells tightly and secreting a muscous-like substance that prevents water loss from the shell. Of these species, Batillaria minima is the most tolerant of salinity and dessication (Abbott 1972). In coastal samples (0-1100 m) it was the most abundant taxon in the adult size range (15 mm). Several common coastal invertebrates were notably rare or absent from this zone, including Littorina, Melampus, oysters, and the fiddler crab Uca; this may be due to the lack of a coastal berm -- associated larger trees and peat soils that characterize other coastal areas where these taxa proliferate.

    The transitional zone is characterized by a composite of marine and more euryhaline "freshwater" species. Only two species have ranges that do not extend to either end of the transect, intersecting the white zone, Planorbella scalaris and Pomaceae palundosa. These are freshwater, salt-intolerant gastropods that may colonize the white zone during the rainy season. All except four of the coastal-zone taxa were also found in the white zone, but in low abundance. Based on the observed lack of adults and living specimens, most of these marine taxa are probably transported into the white zone during high and storm tides. The exception is Batillaria minima which was the most abundant mollusk found with viable living populations in the white zone.

    Only three taxa were collected from the freshwater marsh between 3200-4120 m, on either side of the L-31E canal. These included the pulmonate gastropod Physella cubensis, the hydrobiid Littordinops monroensis and, in very low abundances, the clam Cyrenoida floridana. Physella cubensis is common in the freshwater Everglades, but L. monroensis and C. floridana are brackish-water species that are typical of the lower saline Everglades (Thompson 1984). They are capable of tolerating low salinity for short periods of time and may disperse into the interior marsh as larvae from adult populations further coastward. The coastward boundary of this zone (~2800-3300 m) marks the western edge of the range of many of the transitional and intertidal taxa. This compositional ecotone coincides with a precipitous decrease in salinity that may account for the transition in biota.

    Summary to Species Responses---The remarkable zonation of plants, diatoms and mollusks within the 4 km transect at Turkey Point clearly demonstrates the overwhelming effect of coastal stresses on species composition. Within each group, a well-defined coastal ecocline is further differentiated into relatively homogeneous coastal, transitional, and interior units. Our physical measurements suggest that these compositional units differ fundamentally in hydrologic regime, though the hydrology of the Turkey Point area was not directly examined. In general, the coastal unit may be characterized as intertidal, experiencing regular tides during much of the annual cycle; the middle unit is supratidal, receiving tidal waters infrequently; and the interior unit is usually dominated by local rainfall and upstream sources (in this case, seepage from the L-31E Canal). Whereas the boundary between coastal and transitional units differs among plants, diatoms, and invertebrates, the ecotone between transitional and interior units is roughly coincident for all three groups at about 3000 m from shore. This position also coincides with the interior of the white zone, a broad zone characterized by low plant cover and a highly reflective surface substrate, that occupies a largely supratidal position.

    Species to the west of the interface between transitional and interior compositional units, or the inner boundary of the white zone, have mostly freshwater distributions, and indeed the area marks the beginning of a precipitous landward decrease in porewater salinity. Few freshwater taxa are able to tolerate the fluctuating salinities characteristic of the white zone and areas further coastward. The white zone itself is poorly defined in terms of species composition; the relationship of species assemblages to salinity breaks down within this zone. It appears to be a mixing zone where conditions fluctuate widely; halophilic and aerophilic diatoms and mollusks suggest at least occasional hypersaline episodes in addition to the stress of periodic desiccation. Marine taxa are also present in diatom and mollusk assemblages, implying substantial transport of small detached epiphytes and plankton from the ocean into the interior during storm tides. Although the white zone is not the customary habitat of these species, their remains are present in surface sediments and must be taken into account in any paleoecological application in the area. Plants respond less favorably to the extreme conditions of the white zone; taxa thriving at either end of the gradient are present in low abundance or absent in the white zone.

    The boundary between coastal and transitional biotic units is reached at about 1100-1700 meters from shore, though diatom assemblages continue to change somewhat closer to the coast. For all groups, this boundary is significantly interior to the white zone boundary as it is defined visually (~ 200 m). Low plant species diversity in this zone is typical of coastal mangrove forests (Janzen 1985), but has never been adequately explained. In contrast, mollusk and diatom diversity is high in this zone. This is most likely the result of shoreward transport of marine species from the ocean during high tides. Most of the diatoms and mollusks found living in this zone are characteristic of saltmarsh, estuarine or off-shore benthic environments, and thus are able to withstand significant changes in salinity.

    In conclusion, the interior edge of the white zone is also an ecotonal boundary in terms of diatom, mollusk and plant species composition. Although the visual appearance of the boundary is more a refection of transition in productivity rather than composition, both are caused by a dramatic coastward increase in salinity (and probably salinity variation). The coastal edge of the white zone (100-200 m) also marks a productivity boundary, while compositional transitions occur more smoothly between 1100-1700 m. This boundary is less a function of salinity than of decreased fluctuation in water availability and tidal flushing.

    As a group, diatoms are the most reliable quantitative indicator of coastal distance because of their immense species richness and narrow ecological tolerances. However, when viewed individually, the distribution of plant and mollusk species are clearly defined and, because they are easily and quickly identified, offer an accurate first assessment of the environmental signals related to coastal distance.
 

Planning and evaluating coastal wetland restoration

    The coastal wetland gradient we studied at Turkey Point was vastly different from its condition of only a century ago, due to the combined effects of sea level rise, drainage, and interference with freshwater sheet flow. A visible sign of change is evident from a comparison of aerial photos of 1940 and 1994, which indicate a landward movement in the interior boundary of the white zone of about 1800 m. Further evidence is contained in dated sediment cores along the transect, which show a consistent upcore increase in the salt-tolerance of the fossil invertebrate assemblage (Meeder et al. 1996). These signals of ecological change are not unique to our study area; they represent a widespread problem of saltwater intrusion and mangrove encroachment throughout South Florida. Restoration of coastal wetlands is an essential component of the effort to restore the greater Everglades ecosystem, especially since the condition of these areas is strongly linked to the health of adjacent water bodies, as well as the viability of populations of higher trophic organisms, such as fish and wading birds (Lorenz in press, Ogden 1994). The relationships described above may aid in setting goals for coastal wetland restoration, as well as in assessing the success of restoration efforts once they are undertaken.

    Despite their altered state in comparison to the pre-development condition, Turkey Point wetlands offer well-organized coastal gradients in plant, diatom and invertebrate assemblages. These gradients are characterized for the most part by clinal change in species composition with distance from the shore, but may be further differentiated into coastal, transitional, and interior assemblages. Knowledge of these distributions may, for a given coastal basin, provide four possible indicators of wetland status. These all reflect the hydrologic balance between rising sea level and changes in the magnitude of terrestrial water sources (i.e., freshwater sheet flow and groundwater discharge), and only indirectly reflect on wetland health per se. They are: (1) the position of the white zone, especially its leading edge, and (2-4) the position or range of coverage of the coastal, transitional, and interior compositional units for plants, diatoms, and invertebrates. All four geographic variables may be used to develop rational restoration objectives based on retrospective (i.e., paleoecologic) analyses, and in tracing the trajectory of coastal change following restoration through direct sampling. The potentials of each are discussed briefly below.

    The position of the white zone---Given the availability of appropriate aerial imagery, comparisons in position of the white zone over different time periods is an expeditious means of characterizing changes in the coastal gradient over large areas. Interpretation and comparison of photos using modern GIS techniques necessarily involves some subjectivity, and may be complicated by short term phenomena such as standing water, or a recent history of freeze, fire, or storm that may obscure the boundaries of the white zone. The retrospective use of aerial photos is limited by the inability to fieldtruth the photointerpretation, and because the first comprehensive photos of the South Florida coastline date only to the 1930's, while Everglades drainage began several decades earlier. Nevertheless, the value of the spatial perspective predicted by aerial photos cannot be overemphasized.

    Zonation of vascular plants---Delineation of the position of vegetation units along selected coastal transects is a relatively straightforward field and analytic task, and could form the heart of a network directed at tracking coastal change. Utilization of aerial photos to monitor plant species composition throughout the South Florida coastal zone at the necessary level of detail is currently problematic. However, this too may soon become feasible, as the availability of appropriate remotely sensed data improves, along with our ability to interpret them. Changes in rooted vegetation may track saltwater encroachment in a coarse-scale manner (perhaps associated with disturbances) because, once established, most plant species can tolerate more severe conditions than during colonization. Thus, changes in plant communities may reflect more long-term and permanent changes in wetland status than mollusks or diatoms, whose distributions may shift in response to finer-scale variation. In considering the use of plant zonation as an indicator of coastal change, one should recognize that these patterns differ regionally. For instance, several species that were well-represented in the supratidal zone at Turkey Point (e.g., Avicennia germinans, Distichlis spicata, and Juncus romoerianus) are relatively uncommon in SESE wetlands adjoining northeastern Florida Bay (Meeder et al. 1996).

    Diatom zonation---Of the three taxonomic groups we studied, diatom assemblages were most closely aligned with coastal distance, yet also exhibited anomalies that appeared to be associated with salt-tolerance (Fig. 8). We therefore expect their community composition to exhibit a relatively fine-scale sensitivity to saltwater intrusion, i.e., responding on an annual or even seasonal basis. They should indicate on a short-term basis the ecological effect of management efforts to increase freshwater flows to the coast. Diatoms have frequently been used for reconstructing coastal paleoenvironments and sea-level change (Gell and Gasse 1990, Juggins 1992, Denys and de Wolf 1999); their sensitivity to salinity variation has been documented in a variety of settings (Ehrlich 1975, McIntire and Moore 1977, Fritz 1990) ). Although the quality of diatoms preserved in coastal sediments is often compromised by severe physicochemical stresses (Sherrod et al. 1989), more thorough paleoecological reconnaissance surveys in the region are needed before ruling out the application of diatoms in reconstructing salt-water encroachment in the SESE. If vertical sequences of diatoms can be extracted from sediments, the WA distance-transfer function provided here could be applied to fossil assemblages to document past rates of coastal transgression. Sequences that pre-date modern landscape modifications (primarily canalization) would be helpful in more completely assessing anthropogenic influences on the rate of salt-water encroachment into the SESE. In addition, because there is a high degree of overlap between the SESE and Florida Bay diatom flora (DeFelice and Lynts 1978, Huvane et al. - this volume), species affinities defined in this study could be applied to paleosalinity reconstructions elsewhere in the South Florida marine environment.

    Invertebrate zonation---Invertebrates were intermediate between diatoms and plants in the strength of their association with coastal distance, suggesting that they can be good indicators of coastal wetland status under future management scenarios. They can also be extremely useful in paleoecologic reconstructions, because they remain well-preserved in both carbonate and peat sediments. In conjunction with 210Pb-based estimates of sediment accretion rates, Meeder et al. (1996) applied published mollusk habitat affinities to species profiles in soil cores, thereby documenting an increase in the rate of saltwater intrusion in SESE within the twentieth century. In several basins, including Turkey Point, the increase coincided with management actions that reduced freshwater supply.

    The four elements discussed above complement one another as monitoring/planning tools for restoration of coastal wetlands. As a group, they provide broad ranges in (1) expected speed of response to saltwater intrusion, (2) time periods for which precise information about the historical coastal gradient may be provided, and (3) cost of data acquisition. As discussed earlier, they are descriptors of the position of the gradient, not of the condition of ecosystems within it. Perhaps maintaining or restoring the spatial arrangement of coastal plant, algal, and invertebrate communities, based on knowledge of current species-environment relationships and historical species distributions, is a first step toward ensuring adequate ecosystem function.



Summary

    The white zone is an important ecological phenomenon of South Florida coastal ecosystems, yet its potential as an environmental monitoring tool remains untapped. By investigating factors that influence the position of its boundaries and its relationship with several multi-species groups, we have taken a first step in exploring the capacity of the white zone as an environmental indicator. Our results are summarized as follows:

  1. The white zone is a region of low productivity characterized by low vegetation cover and canopy height (< 50% and < 1 m, respectively). From remotely sensed images, it appears as a reflective white band, resulting mainly from marl, fresh storm deposits and, in some areas, periphyton. The majority of the white zone corresponds with the supratidal region of the coast, which is irregularly flooded by tidal waters. Some areas, however, are more regularly inundated by semi-diurnal tides.

  2.  
  3. Over the past 50 years, the interior boundary of the white zone has encroached inland by an average of 1.5 km. Maximum shifts occurred in areas cut off by canals from upstream fresh water input (1.8 km at Turkey Point).
  1. Plant, diatom and mollusk species assemblages correlate strongly with the coastal gradient, and may be separated into coastal, transitional and interior units. For all groups, the transitional-interior interface corresponds strongly with the inner boundary of the white zone, but no correlation exists between the coastal-transitional interface and the coastal boundary of the white zone.

  2.  
  3. The coastal gradient is characterized by a host of environmental variables (salinity, elevation, wave energy, wind speed etc...). Of the environmental variables we measured (salinity and elevation), diatoms are weakly correlated with elevation and show significant correlation with salinity after accounting for spatial autocorrelation. Neither plants nor mollusks show any correlation with salinity or elevation.

  4.  
  5. The position of the white zone and coverage of the coastal, transitional and interior compositional units of plants, diatoms and mollusks may provide an indirect assessment of wetland status. These variables reflect the balance between ambient sea level and freshwater discharge, though each is likely to respond in a unique fashion.

  6.  
  7. Future investigations should address other influential environmental variables (e.g., soil nutrients, hydrologic variation, disturbance) to further the understanding of the factors controlling species composition. Also, studies should emphasize monitoring change in the position of the white zone and species distributions over varied time scales. This could include long-term monitoring, or retrospective studies involving paleoecological methods.


References
 

Abbott, R. T. 1972. American Seashells. Van Nostrand.

Auerbach, M., AND A. Shmida. 1993. Vegetation changes along an altitudinal gradient on Mt. Hermon, Israel - no evidence for discrete                 communities. Journal of Ecology 81: 25-33.

Ball, M. 1988. Ecophysiology of mangroves. Trees 2: 129-142.

Birks, H. J. B., J. M. Line, S. Juggins, A. C. Stevenson, AND C. J. F. ter Braak. 1990. Diatoms and pH reconstruction. Phil. Trans. R. Soc. Lond. B 327: 263-278.

Browder, J. A., P. J. Gleason, and D. R. Swift. 1994. Periphyton in the Everglades: spatial variation, environmental correlates, and ecological implications. In S. M. Davis and J. C. Ogden [eds.], Everglades: The Ecosystem and its Restoration. St. Lucie.

Clarke, K. R., AND R. M. Warwick. 1994. Change in marine communities: an approach to statistical analysis and interpretation. Natural Environment Research Council, UK.

Craighead, F. C., Sr. 1971. The trees of south Florida. Vol. 1: The natural environments and their succession. University of Miami.

Davis, J. H., Jr. 1940. The ecology and geologic role of mangroves in Florida. Pap Tortugas Lab. 32: 304-412. Carnegie Institute, Wash. Publ. No. 517.

DeFelice, D. R. 1975. Model studies of epiphytic and epipelic diatoms of Upper Florida Bay and associated sounds. M.S. Thesis, Duke University.

-----------and G. W. Lynts. 1978. Benthic marine diatom associations: Upper Florida Bay (Florida) and associated sounds. J. Phycol. 14: 25-33.

Denys, L., and H. de Wolf 1999. Diatoms as indicators of coastal paleo-environments and relative sea-level change. In Stoermer, E. F. and J. P. Smol [eds.] The Diatoms. Applications for the Environmental and Earth Sciences.

Egler, F. E. 1952. Southeast saline Everglades vegetation, Florida, and its management. Veg. Acta. Geobot. 3: 213-265.

Ehrlich, A. 1975. The diatoms from the surface sediments of the Bardawil Lagoon (Northern Sinai) - Paleoecological significance. Beihefte zu Nova Hedwigia 53: 253-277.

Foged, N. 1984. Freshwater and littoral diatoms from Cuba. Bibliotheca Diatomologica. 5, 243 pp.

Fritz, S. C. 1990. Twentieth-century salinity and water-level fluctuations in Devil's Lake, North Dakota: test of a diatom-based transfer function. Limnol. Oceanogr. 35: 1771-1781.

Gell, P. A., and F. Gasse. 1990. Relationships between salinity and diatom flora from some Australian saline lakes. Proceedings of the 11th International Diatom Symposium 631-647.

Hagelstein, R. 1938. The Diatomaceae of Porto Rico and the Virgin Islands. New York Acad. Sci. Scientific Survey of Porto Rico and the Virgin Islands 8: 313-450.

Harshberger, J. W. 1914. The vegetation of south Florida Trans. Wagner Free Inst. Sci. Philos. 3: 51-189.

Hustedt, F. 1927-1966. Die Kieselalgen Deutschlands, Österriechs und der Schweiz. In L. Rabenhorst [ed.], Kryptogamen-Flora von Deutschland, Österriech und der Schweiz. 7. Akademische Verlagsgesellschaft.

-----------. 1930. Bacillariophyta (Diatomaceae). In A. Pascher [ed.], Die Süßwasser-flora Mitteleuropas. Gustav Fischer Verlag.

Janzen, D. H. 1985. Mangroves: where's the understory? J. Tropical Ecology 1: 89-92.

Juggins, S. 1992. Diatoms in the Thames estuary, England: Ecology, paleoecology and salinity transfer function. Bibliotheca Diatomologica 25:1-26.

Krammer, K., and H. Lange-Bertalot. 1986-1997. Bacillariophyceae. In H. Ettl, J.Gerloff, H. Heynig and D. Molenhauer [eds.], Süßwasserflora von Mitteleuropa. 2/1-4. Gustav Fischer Verlag.

Lange-Bertalot, H. 1993. 85 Neue Taxa und über 100 weitere neu definierte Taxa ergänzend zur Süßwasserflora von Mitteleuropa 2/1-4. J. Cramer.

Line, J. M., C. J. F. ter Braak, and H. J. B. Birks. 1994. WACALIB version 3.3 - a computer program to reconstruct environmental variables from fossil assemblages by weighted averaging and to derive sample-specific errors of prediction. Journal of Paleolimnology 10: 147-152.

Long, R. W., and O. Lakela. 199x. A Flora of Tropical Florida. University of Miami.

Lorenz, J. J. In press. The response of fishes to physical-chemical changes in the mangroves of northeast Florida Bay. Estuaries X: XXX-XXX.

Mann, A. 1925. Marine diatoms of the Philippine Islands. U.S. Smithson. Inst. Bull. 100: 1-182.

McCormick, P.V., and M. B. O'Dell. 1996. Quantifying periphyton responses to phosphorus in the Florida Everglades: a synoptic-experimental approach. Journal of the North American Benthological Society 15: 450-468.

McCune, B., AND M. J. Mefford. 1995. PC-ORD: multivariate analysis of ecological data. Version 2.0. MJM Software Design.

McIntire, C. D., and W. W. Moore. 1977. Marine littoral diatoms: ecological considerations. In D. Werner [ed.], The Biology of Diatoms. University of California.

Meeder, J. F., M. S. Ross, G. Telesnicki, P. L. Ruiz, AND J. P. Sah. 1996. Vegetation analysis in the C-111/Taylor Slough basin. Final Report to the South Florida Water Management District, Contract C-4244. West Palm Beach, FL.

Montgomery, R. T. 1978. Environmental and ecological studies of the diatom communities associated with the coral reefs of the Florida Keys. Ph. D. Dissertation. Florida State University, Tallahassee, FL.

Moore, D. R. 1964. Mollusca of the Mississippi coastal waters. Gulf Coast Bull.1. Beloxi, MS.

Navarro, N. 1982. Marine diatoms associated with mangrove prop roots in the Indian River, Florida,USA. Bibliotheca Phycologica 61: 1- 154.

Ogden, J. C. 1994. A comparison of wading bird nesting colony dynamics (1931-1946 and 1974-1989) as an indication of ecosystem conditions in the southern Everglades. In S. M. Davis and J. C. Ogden [eds.], Everglades: the ecosystem and its restoration. St. Lucie.

Olmsted, I., H. Dunevitz, and W. J. Platt. 1993. Effects of freezes on tropical trees in Everglades National Park, Florida, USA. Tropical Ecology 34: 17-34.

Østrup, E. 1913. Diatomaceae ex Insulis Danicis Indiae Occidentalis imprimis a F. Børgesen lectae. Dansk Bot. Arkiv 1: 1-29.

Patrick, R., and Reimer, C. W. 1966. The diatoms of the United States I. Acad. Nat. Sci. Philad., Monogr. 13: 1-688.

------------ 1975. The diatoms of the United States II, part 1. Acad. Nat. Sci. Philad., Monogr. 13: 1-213.

Pilsbry, H. A. 1946. Land mollusca of North America (north of Mexico). Acad. Nat. Sci. Phila. Monogr. 3: 237-302.

Podzorski, A. C. 1984. Caponea caribbea Podzorski, a structurally unique new diatom from Jamaica. Nova Hedwigia 40: 1-8.

----------- 1985. An illustrated and annotated check-list of diatoms from the Black River Waterways, St. Elizabeth, Jamaica. Bibliotheca Diatomologica. 7: 1-177.

Rader, R. B. 1994. Macroinvertebrates of the northern Everglades: Species composition and trophic structure. Florida Scientist 57: 22-33.

Reimer, C. W. 1996. Diatoms from some surface waters on Great Abaco Island in the Bahamas (Little Bahama Bank). Beiheft zu Nova Hedwigia 112: 343-354.

Ross, M. S., J. J. O'Brien, and L. J. Flynn. 1992. Ecological site classification of Florida Keys terrestrial habitats. Biotropica 24: 488-502.

-----------, J. F. Meeder, J. P. Sah, P. L. Ruiz, and G. J. Telesnicki. in press. The Southeast Saline Everglades revisited: a half-century of coastal vegetation change. Journal of Vegetation Science XX: XXX-XXX.

Round, F. E., R. M. Crawford, and D. G. Mann. 1990. The Diatoms: Biology and Morphology of the Genera. Cambridge.

Rutchey, K., AND L. Vilchek. In press. Air photo-interpretation and satellite imagery analysis techniques for mapping cattail coverage in a northern Everglades impoundment. Journal of Photogrammetric Engineering and Remote Sensing. X: XXX-XXX.

Schmidt, A. (et al.) 1874-1959. Atlas der Diatomaceenkunde. R. Reisland, Ascherleben.

Sherrod, B. L., H. B. Rolins, and Kennedy. 1989. Subrecent intertidal diatoms from St. Catherines Island, Georgia: Taphonomic complications. Journal of Coastal Research 5: 665-677.

Slate, J. 1998. Inference of present and historical environmental conditions in the Everglades with diatoms and other siliceous microfossils. Ph.D. Dissertation. University of Louisville, Louisville, KY.

Stephenson, T. A., AND A. Stephenson. 1950. Life between tide-marks in North America: I. The Florida Keys. Journal of Ecology 38: 354-402.

Sullivan, M. J. 1981. Community structure of diatoms epiphytic on mangroves and Thalassia in Bimini Harbour, Bahamas. In R. Ross [ed.], Proceedings of the 6th International Diatom Symposium, Budapest. O. Koelty.

Swift, D. R., and R. B. Nicholas. 1987. Periphyton and water quality relationships in the Everglades Water Conservation Areas. Technical Publication 87-2. South Florida Water Management District, West Palm Beach, FL, USA.

Tabb, D. C. and R. B. Manning. 1961. A checklist of the flora and fauna of northern Florida Bay and adjacent brackish waters of the Florida mainland collected during the period July 1957 through September 1960. Bull. Mar. Sci. 1: 550-647.

ter Braak, C. J. F. 1987. Calibration, p. 78-90. In: R. H. G. Jongman, C. J. F. ter Braak, and O. F. R. van Tongeren [eds.], Data analysis in community and landscape ecology. Wageningen, Pudoc.

-----------, S. Juggins, H. J. B. Birks, and H. van der Voet. 1993. Weighted averaging partial least squares (WA-PLS): definition and comparison with other methods ofor species - environmental calibration. In G. P. Patil and C. R. Rao [eds.], Multivariate Environmental Statistics. North Holland, Amsterdam.

Thompson, F. G. 1984. The freshwater snails of Florida: a manual for identification. University of Florida.  U. S. Department of Agriculture. 1996. Soil survey of Dade County area, Florida.  U. S. Fish and Wildlife Service. 1999. South Florida multi-species recovery plan. Atlanta, GA.

Vos, P. C., and H. de Wolf. 1993. Diatoms as a tool for reconstructing sedimentary environments in coastal wetlands; methodological aspects. Hydrobiologia 269/270: 285 296.

Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30: 279-338.

Wilbur, K. M, and C. M. Yonge. 1966. Physiology of Molluska. Vol. 2. Academic.