ABSTRACT:  INTRODUCTION:  STUDY AREA:  METHODS:  RESULTS:   DISCUSSION:   ACKNOWLEDGMENTS:  REFERENCES:

Coastal ecotone between US 1 and Card Sound Road in 1940 showing white zone between coast and fresh water marsh.

We examined the vegetation of the Southeast Saline Everglades (SESE), where water management and sea level rise have been important ecological forces during the last half century.  Marshes within the SESE were arranged in well-defined compositional zones parallel to the coast, with mangrove-dominated shrub communities near the coast giving way to graminoid-mangrove mixtures, and then sawgrass marsh. The compositional gradient was accompanied by an interiorward decrease in total aboveground biomass, and increases in leaf area index and periphyton biomass.  Since the mid 1940’s, the boundary of the mixed graminoid-mangrove and sawgrass communities shifted inland by as much as 3.3 km. Moreover, the interior boundary of a low-productivity zone appearing white on both B&W and CIR photos moved inland by an average of 1.5 km.  A smaller shift in this "white zone" was observed in an area receiving fresh water overflow through gaps in one of the SESE canals, while greater change occurred in areas cut off from upstream water sources by roads or levees. These large-scale vegetation dynamics are apparently the combined result of sea level rise - approximately 10 cm since 1940 - and  water management practices in the SESE.

Keywords: Sea level rise; water management; salinity index; periphyton; aboveground biomass; leaf area index; correspondence analysis; Cladium jamaicense; Eleocharis cellulosa; Rhizophora mangle.

Nomenclature:  Long & Lakela (1971).

    Coastal wetlands reflect a dynamic hydrologic balance between the marine and upstream or upslope terrestrial ecosystems which bound them on all sides. In these transitional settings, ecological responses which arise principally as a result of changes in the marine system may be modified by physiography, management, or land use patterns in the terrestrial environment, or vice versa. One example of these interactions is found in the response of mangrove ecosystems to sea level rise, which varies in rate or even direction in different physical settings or under alternative water management scenarios (Meeder et al. 1993). At a landscape scale, the anticipated response of coastal wetlands to change in sea level is even more complex, especially if the dramatic vegetation zonation typical of many coastal wetlands is incorporated.

    Nearly a half-century ago, Dr. Frank Egler described the vegetation of the area south and east of the Atlantic Coastal Ridge in southernmost peninsular Florida, noting a conspicuous coastal zonation within the area he called "the Southeast Saline Everglades" (Egler 1952). Egler's description was based on 1938 and 1940 aerial photographs, and on field work undertaken in 1940 through 1948. He described the vegetation pattern in the coastal Everglades at the time as "fossil", responding slowly to a rapidly changing environment that included a rising sea level, a decline in the level of the surface freshwater aquifer, a reduction in the frequency of fire, and a range of anthropogenic modifications to natural drainage patterns. Egler documented several examples of local vegetation change over the period of his study, including the invasion of the halophytic red mangrove (Rhizophora mangle) into freshwater wetlands far from the coast. At the same time, he anticipated a continued interiorward shift in the coastal vegetation gradient.

    Egler's work preceded the connection of the South Dade Conveyance system to the sea via the C-111 canal. Completed in the late 1960's, this project allowed more effective drainage of the agricultural and urban lands abutting the Southeast Saline Everglades (SESE).  In conjunction with roads and agricultural ditches, operation of the canal system altered fresh water delivery to SESE wetlands, starving some areas of water while augmenting the supply to others.

    Five decades after Egler's ecological studies, we reexamined the vegetation of the SESE. Our objectives were (a) to describe current vegetation patterns of the area in relation to known environmental or geographic variables, and (b) to document and interpret changes in the coastal wetland vegetation since our predecessor's studies. We hypothesized that changes in SESE marshes would reflect salinization effects associated with sea level rise, and that these effects would be more severe in areas cut off from upstream water sources by canals or roads. Documentation of temporal change in SESE vegetation was derived from two sources: floristic surveys and aerial photointerpretation.
 

STUDY AREA:

    As described by Egler (1952), the Southeast Saline Everglades includes the broad band of wetlands extending from the southwest-curving Atlantic Coastal Ridge to the coast (Figure 1). Our study concentrated on the coastal wetlands north of Florida Bay, but also included several sites within the Biscayne Bay drainage, east of U. S. Highway 1 (Figure 1).

    Climate.  At 25⁰N latitude, the SESE climate is transitional between temperate and tropical environments. Mean annual temperature is about 25⁰C, with the difference between the warmest month (July) and the coolest month (January) less than 10^oC.  A more significant ecological factor may be the frequency of freezing temperatures, which at the nearest longterm weather station in Homestead averaged one event per year between 1949-1987 (Duever et al. 1994).  Mean annual precipitation (1951-1980) at Homestead is 155 cm, with a distinct dry season between November and April.  During many years, a large proportion of rainfall is attributable to Atlantic tropical storms and hurricanes.

    Physical environment.  The SESE is a flat coastal plain which ascends from sea level to approximately one meter above sea level at the base of the uplands, more than 10 km distant throughout most of the region. Soil substrates are mostly marls produced under fresh water conditions, or peats, or some combination of the two (Leighty et al. 1965). The lower half of the SESE plain is dissected by ephemeral creeklets which range in depth from several inches in the upper reaches to several feet near the coast. Much of the variation in vegetation structure, including a profusion of tree islands, appears to be associated with these and smaller local undulations in topography.

    Water levels at sites in the lower SESE respond primarily to fluctuations in the adjacent marine waters, while levels in the interior marshes are highly correlated with artificially maintained stages in the C-111 Canal. The contrast in hydrologic controls on interior and coastal sites may produce conditions where water levels are higher at either end of the SESE coastal gradient than at intermediate locations. Egler (1952) emphasized that periods of elevated salinity in the SESE are episodic, short-lived events, while background porewater salinities in most areas are low.

    Vegetation. Several general treatments of South Florida vegetation include characterizations of marsh communities present in the SESE (e.g., Harshberger 1914; Harper 1927; Davis 1943; Robertson 1955; Craighead 1971).  However, only Egler (1952) described the marshes of the SESE as a group.  Egler's classification divided SESE vegetation into seven concentric belts roughly parallel to the coast. The two most interior belts, the "Pine Forest" and the "Aristida Grassland", are upland or transitional communities described in passing. In Belt 3, the "Upper Saline Everglades", the dominant plant was sawgrass (Cladium jamaicense), and red mangrove (Rhizophora mangle) and spikerushes (Eleocharis spp) were notably absent.  Egler recognized two variants of this marsh type: a dry, interior, species-rich phase and a wet, more coastal, species-poor phase.

    In contrast, red mangrove was an important element in Egler's Belts 4-7. He described Belt 4 (the "Lower Saline Everglades") as a narrow zone characterized by "... the general paucity of vegetation..." and "... the 'dwarfness' of many of the mangroves..." (Egler, 1952; p. 256). As presented by Egler, Belt 4 marshes were a mixture of sawgrass, red mangrove, and spikerush, with the former the dominant species on the basis of cover. Egler considered fire regime and propagule availability to be the primary factors affecting the occurrence of red mangrove in Belts 3 and 4.

    Egler described Belt 5 ("Mangrove-sawgrass vegetation") as a mosaic of sawgrass patches, mixed mangrove forest (including black mangrove, Avicennia germinans), and open water. According to Egler, the interior border of Belt 5 represented the coastward limit of fire. In Belt 6 ("Mangrove - Tidal Marsh Vegetation"), sawgrass was replaced by salt-tolerant plants such as Batis maritima, Borrichia frutescens, and Juncus romoerianus, and red, black, or white mangrove (Laguncularia racemosa) were locally dominant. Finally, in the floristically simple Belt 7 ("Rhizophora Border"), the salt marsh species were less extensive, and R. mangle was the sole tree species of significance.
 

METHODS:

    Sampling design. Vegetation and soils were sampled at 54 individual locations during the winters of 1994-95 and 1995-96 (Figure 1). Twenty-six sites had previously been sampled by Tabb et al. (unpublished 1968 report), and another twenty-one were adjacent to hydrological stations in a network jointly maintained by Everglades National Park and the U. S. Geological Survey. Nineteen of the hydrologic stations consisted of wells sunk to bedrock, outfitted with water level recorders that had been in operation since at least 1992.  Seven sites along the western and coastal periphery of the study area were chosen to fill in gaps in sampling coverage.

    Vegetation sampling.  To minimize observer bias among the four people involved in vegetation sampling, we determined a rank abundance for each plant species present at our 54 individual sampling locations, based upon their cover. The vegetation sampler walked approximately 50 meters north from the plot center, tossed a 1.12 m diameter hoop (1 m²  area enclosed) to his left, and ranked the vascular plant species rooted within the hoop in order of their shoot cover. By pacing chords of 12^o arcs clockwise from the initial point, we located and sampled thirty subplots in a broad circle around the plot center. Species rank abundance at each site was equal to 100 times the sum of species abundance in the 30 subplots divided by the total abundance of all species, where the species ranked first in each plot was assigned an abundance of 10, the species ranked second an abundance of 5, the species ranked third an abundance of 2, and species ranked fourth or higher an abundance of 1. Species present in the marsh but not located in any subplot were assigned a cumulative abundance of 1.

    Sampling at the 21 hydrologic stations also included an estimation of biomass and leaf area in the aboveground plant community. A 100-meter transect was established, beginning within twenty meters of the monitoring well and oriented perpendicular to the apparent slope direction. Within each ten-meter segment of the transect, a single 0.25 m² subplot was randomly selected, and all aboveground material harvested and separated into the following components: live vascular plant tissue by species, dead standing vascular plant tissue, litter, mat periphyton, and epiphytic periphyton. Leaf and structural tissues of woody plants were also separated. Harvested material was dried to constant weight at 65⁰C, and weighed. Inorganic and organic fractions of periphyton components were separated by combustion at 500⁰C, and biomass expressed on the basis of the latter. Leaf area index was calculated by multiplying live leaf biomass (woody plants) or live aboveground biomass (herbaceous plants) by specific leaf area for individual species, then summing for the plot as a whole. Estimates of specific leaf area for the three most abundant plants (C. jamaicense: 53.58 cm²/g, s = 5.03; R. mangle: 45.22 cm²/g, s = 8.39; and E. cellulosa: 121.79 cm²/g, s = 8.69) were based on samples of 4-6 leaves per plant from ten individuals at several locations within the study area.  A mean specific leaf area of 74 cm²/g was used for other species, which generally comprised less than 25% of total leaf biomass.

    Environmental sampling. At the 21 hydrologic sites, average elevation was calculated on the basis of a theodolite survey from the water level recorder table (whose elevation was known) to ground surface at five-meter intervals along the transect, as well as at the ten subplot centers. At all vegetation sites, soil depth was determined as the average of 4-6 probings to bedrock with a 1-cm diameter aluminum rod. Bulk density and organic matter content were determined from surface samples (upper 3 cm) at 42 sites, according to methods outlined in  DeLaune et al. (1987) and Dean (1974), respectively. From the fossil mollusk assemblage present in these surface soils, we calculated a salinity index equal to the density-weighted average of the salinity preferences of the sampled taxa. The salinity preferences of the mollusk species encountered, as well as citations for the literature on which those ratings were based, are listed in Table 1. It is recognized that this index provides an imperfect approximation of true salinity, since mollusks are also sensitive to other physical factors.

    Current patterns in marsh communities.  We used a combination of classification and ordination techniques to describe the current gradients in marsh composition in SESE, taking steps to ensure that our definitions of SESE plant communities could be directly compared to Egler's (1952) study.  Egler included quantitative data from two transects representative of Belt 3 marsh vegetation and one transect representative of Belt 4 species composition. His tables listed the frequency of occurrence of each species in fifty 10-m²  plots, and the proportion of occurrences in which it was "rare", "occasional", and "abundant". We assigned an abundance of 1, 2 and 10, respectively, to these descriptors, and calculated relative species abundance in his transects in a manner analogous to what we had done with our own data. We then applied the TWINSPAN classification procedure (Hill 1979) and Detrended Correspondence Analysis (DCA) (ter Braak 1987) to a 57-site marsh data set (our 54 plots and Egler's three). After eliminating species that occurred in fewer than three sites, we transformed the relativized data into octave categories for the DCA analysis, and chose pseudospecies cut levels of 0, 2, 8, 33, and 66 for TWINSPAN.

    Comprehensive, spatially explicit, appropriately scaled environmental information was extremely limited for the study area. We therefore considered only a few variables in our examination of vegetation-environment interactions: the three soil variables described above (depth, organic matter percentage, bulk density), salinity index, hydroperiod, and distance to the coast. The latter was a spatial variable intended to represent the composite effects of many unmeasured environmental factors. At the nineteen sites for which continuous surface hydrologic data was available for the 1992-94 period, we calculated hydroperiod as the mean annual number of days that average water level was more than five cm above the mean soil surface. Interactions between the environmental measures and SESE marsh vegetation were analyzed via Canonical Correspondence Analysis (CCA), using PC-ORD version 3.11 (McCune & Mefford 1997), which incorporates the stricter convergence criteria suggested by Oksanen and Minchin (1997).  The CCA analysis was applied twice: (1) to a 42-site data set in which distance to coast, the three soil variables, and salinity index were the environmental variables, and (2) to a 19-site data set in which distance to coast and hydroperiod were sole environmental variables examined. To determine the best model, we applied Monte Carlo permutation tests to variable combinations arrived at through a forward stepwise selection process.

    Temporal change in marsh vegetation patterns - 1940 to 1994. Our analysis of changes in the position and species composition of SESE marsh communities were based on comparisons with Egler (1952). These comparisons were intended to document the distance and direction of movement, if any, in the landward border of a  "... conspicuous white band, forming a smooth arc which parallels, and is inland from, the ocean shores." (Egler 1952; p. 256). Egler's manuscript included copies of several 1940 aerial photos (Soil Conservation Service, CJF series), whose captions clearly illustrated the zone to which he referred. In Egler's classification, the interior border of this white band demarcated the transition between Belts 3 and 4, and represented " ... the upper limit of invasion by Rhizophora propagules sufficient to form a complete coverage..." (Egler 1952, p.231). The white band remains remarkably prominent on current photos, with a landward boundary that remains "... distinct, but not knife-edge, definite enough so that the changeover usually occurs within half a kilometer..." (Egler 1952; p. 256). We therefore compared the 1940 and 1994 positions of this boundary, hereafter referred to as the "white zone", using vegetation data from both studies as an aid in interpreting its ecological significance.

    We first delineated the interface described above, based on both 1940 and 1994 aerial photos. The former were same images (1:24,000, black and white) used by Egler, and the latter were NAPP 18" x 18" 1:29,000 color infrared photos. The images were rectified utilizing natural or anthopomorphic landmarks as control points. Using ATLAS-GIS software (Strategic Mapping Inc.), the historical and current white zone interfaces were digitized and superimposed on boundary files generated from U.S. Geological Survey 1:100,000 digital line graphs. Distances between the 1940 and 1994 interfaces perpendicular to the general direction of the coastline were calculated at ca 520-meter intervals along the coast, and examined graphically. Because the coastal boundary of the white zone was not clearly defined on either set of photos, we did not attempt to delineate its position.
 

RESULTS :

    Current marsh communities.  Axis 1 of DCA explained 21% of the variation in marsh species composition, and Axis 2 accounted for an additional 9% (Figure 2). The first and strongest division in the TWINSPAN analysis (eigenvalue=0.43) divided Group 1 from Groups 2- 4, largely on the absence or presence, respectively, of a significant component of R. mangle. Subsequent divisions in Group 1 were weak, and the removal of individual sites commonly caused significant rearrangements in the subgroup assignments. Both of Egler's Belt 3 samples fell within Group 1. The Level 2 division distinguished Group 2 from the remaining sites (eigenvalue=0.24), based on the presence of C. jamaicense and the absence of Ruppia maritima. Further divisions within the group (eigenvalues < 0.2) were not recognized, leaving Egler's Belt 4 site located centrally within Group 2.  Level 3 division of the remaining sites distinguished Groups 3 and 4 (eigenvalue=0.31) on the left side of Figure 2; R. maritima and Utricularia foliosa were indicators of the former, and Conocarpus erectus and Laguncularia racemosa indicators of the latter group. Other divisions beyond Level 3 were not considered to be significant at the scale of the study area.

    Based on the analyses described above, we recognized four vegetation units among SESE marshes and non-forested swamps (Table 2).  Sawgrass Marsh (Group 1 in Figure 2) was compositionally equivalent to Egler's Belt 3. It was characterized by the overwhelming dominance of C. jamaicense and the common occurrence of E. cellulosa, in mixture with a suite of graminoids (e.g., Schoenus nigricans, Rhynchospora spp.) and forbs characteristic of freshwater marshes further north. Group 2, Sawgrass - Spikerush - Mangrove Marsh, was equivalent to Egler's Belt 4. The two named graminoids and R. mangle were of roughly equal abundance in this transitional community, while other plants characteristic of Sawgrass Marsh were conspicuously absent. Groups 3 and 4 included mangrove-dominated species mixtures that Egler divided among Belts 5 through 7. In both groups, sawgrass (C. jamaicense) was restricted to small patches in elevated areas adjacent to tree islands. Mangrove Scrub (Group 3) was a monotonous assemblage characterized by red mangrove shrubs, with E. cellulosa and/or several salt-tolerant aquatic herbs (Ruppia maritima, Utricularia foliosa, U. purpurea) in the large gaps between shrub clumps. Coastal Prairie (Group 4) was characterized by the presence of white mangrove (L. racemosa) and buttonwood (C. erectus) shrubs in mixture with R. mangle, and the scattered occurrence of several halophytic herbs (Aster tenuifolius, Fimbristylis castanea, Distichlis spicata).

    The 21 biomass sampling stations included no examples of Coastal Prairie vegetation (Table 3). Mean aboveground biomass (including dead standing material, but not litter) averaged 619 g/m² in the Sawgrass Marsh, 800 g/m² in Sawgrass - Spikerush - Red Mangrove Marsh, and 1072 g/m² in Mangrove Scrub.  Differences in standing crops among the three community types were largely attributable to the relative importance of the woody plant (e.g., red mangrove) component. However, leaf area index in the predominantly woody Mangrove Scrub (0.54 m²/m²) was little more than half that in Sawgrass Marsh (0.96 m²/m²) and Sawgrass - Spikerush - Mangrove Marsh (0.99 m²/m²) . The importance of the algal component also decreased with total standing crop; total periphyton constituted 29% and 23% of total aboveground biomass in the Sawgrass and Sawgrass - Spikerush - Mangrove Marsh, respectively, decreasing to 9% in the Mangrove Scrub. Finally, the high proportion of dead standing material in all three community types was in part the result of the winter sampling period and a recent (December 1989) freeze.

    The canonical correspondence analyses (Figure 3) indicated that SESE marsh vegetation was arranged primarily along a gradient of distance to the coast. For the 42-site data set, linear combinations of the five environmental variables explained 29% of the variation in species composition. Distance to coast alone explained 16% of total compositional variation (p < 0.001), and, according to the Monte Carlo permutation test, the addition of only depth to bedrock (p = 0.002) and salinity index (p = 0.08) improved the single-variable model further. Both depth to bedrock and salinity index were negatively correlated with distance to coast (r = -0.62 and r = - 0.52, respectively). However, depth to bedrock also exhibited a secondary trend of increase toward the east, and salinity index varied considerably in its rate of decrease with distance among different SESE drainage basins. Neither bulk density nor organic matter content showed a strong spatial trend across the area as a whole, and neither was significantly associated with marsh species composition.  Finally, for the 19-site data set, hydroperiod was uncorrelated with coastal distance (r = -0.20), and its inclusion did not significantly improve a model based on coastal distance alone (CCA diagram not shown).

    SESE marshes were zonally arranged, with the four community types falling neatly into parallel bands. Coastal Prairie was closest to the shore of the SESE interior bays, followed by Mangrove Scrub, Sawgrass - Spikerush - Red Mangrove Marsh, and Sawgrass Marsh toward the interior (Figure 4). Among the sampled sites, mean distance to the coast was 1.0, 2.1, 4.5, and 8.1 km for the four community types listed above, with some overlap between the latter two groups (Table 4). Site salinity index, highest in the two coastward community types, decreased sharply in the two types interiorward of the Mangrove Scrub. Soil depth in the Coastal Prairie type (mean = 1.3 meters) was more than twice that in Sawgrass Marsh (mean = 0.5 meters), with intermediate values in the two intervening zones. The other three environmental variables did not exhibit substantive differences among community types.

    Historical changes in SESE vegetation. The white band discussed by Egler (1952) was a narrow coastal feature in 1940 (Figure 5); the mean distance from its inner edge to the northern coastlines of Joe Bay, Long Sound, Barnes Sound, and Card Sound at that time was 1.13 km (range 0.56-2.22 km) (Figure 4). By 1994, the interior boundary of this zone was, on average, 1.46 km further from the ocean (Figure 5). In general, this trend was less pronounced west of U.S. 1 (mean = 0.82 km) than east of it (mean = 2.24 km). However, west of U.S. 1, the interiorward shift increased with distance from the highway, while the opposite was true east of U.S. 1. Finally, while Egler's observations indicated that in 1940 the interior boundary of the white zone separated pure Sawgrass Marsh to the north from the mixed Sawgrass/Spikerush/Mangrove community, today the white zone border is near (in places south of) the boundary between Sawgrass - Spikerush - Red Mangrove Marsh and Mangrove Scrub.  Along a transect north from Joe Bay, the current interface between Sawgrass Marsh and Sawgrass-Spikerush-Mangrove Marsh is about 3.3 km further inland than in 1940, while the interior border of the white zone has moved north by only 1.2 km (Figure 4).
 

DISCUSSION:

Environmental underpinnings of coastal zonation:

    Like our predecessor (Egler 1952), we found the matrix of low vegetation in SESE to be arranged in a long gradient perpendicular to the South Florida coastline.  Compositional units along this gradient were aligned in distinct zones, with little geographic overlap between one group and the next. However, the low eigenvalues for TWINSPAN divisions beyond Level 1, and the overlap in species composition among groups, suggest that these are not so much discrete units as nodal regions along an ecocline from mangrove-dominated coastal shrublands to sawgrass-dominated interior grasslands.

    The clinal nature of the SESE vegetation is further demonstrated by the strong coupling of species composition with distance to the coast in the CCA analysis, i.e., the position of a site with respect to the coast was a much better predictor of its plant species composition than the measured physical variables.  This pattern suggests that the positional variable "coastal distance" integrates a complex of physical factors which are strongly intercorrelated, but among which no single variable type is dominant.
 
    Spatial variation in edaphic, hydrologic, climatic, and disturbance variables is not well known within the SESE study area, but many of these variables are believed to change in a predictable way with distance to the coast.  Soils become thinner, lower in organic matter content, and higher in bulk density with increasing distance from the coast,  with attendant feedbacks on the availability of oxygen, water, and nutrients for plant growth.  For instance, along a more abbreviated South Florida coastal transect that also included a transition from peat to marl, sediment phosphatase activity --- a good indicator of phosphorus limitation --- increased inland by nearly an order of magnitude over 500 meters (Jacobsen et al. in prep.).  Hydrologically, the coastal gradient is one in which marine waters and tidal processes dominate at one end, and the highly managed runoff of terrestrial water sources prevail at the other.  A wide range of water quality parameters vary predictably based on site location within this mixing zone; salinity generally decreases with distance from the coast, though hypersaline conditions may develop in interior basins.  Some ecologically significant climatic variables change sharply at the coast and immediately inland.  Based on temperature isoclines of Thomas (1974), freeze events, which have been identified by a number of authors as important influences on vegetation patterns in South Florida (Craighead 1971; Egler 1952; Olmsted et al. 1993), may be less frequent and severe immediately adjacent to the coast than in the interior portions of the SESE.  Finally, the principal natural disturbances impacting SESE ecosystems are wildfires and hurricanes. Fire frequency is presumed to parallel the distribution of fine fuels (e.g., graminoids) in increasing with distance from the coast (Egler 1952).   As for hurricanes, while there is little reason to expect their occurrence within SESE to exhibit a strong spatial pattern, their characteristics may differ predictably within the study area in several ways.  For instance, hurricanes accompanied by large storm surges may deposit several inches or more of mud in areas closest to the coast, causing extensive mortality in the mangrove forests (Craighead 1964).

    Within the environmental complex discussed above, hydrologic variables are most directly sensitive to small changes in sea level.  One may reasonably expect sea level rise to result in surface and pore waters of more marine character throughout the SESE wetlands, along with greater water depths and longer hydroperiods.  Vegetation response to elevated salinity ought to parallel the salinity tolerances of the three dominant plant species, which, though broadly overlapping, are in the order R. mangle > E. cellulosa > C. jamaicense (e.g. Koch 1996; Rejmankova et al. 1996; Lin & Sternberg 1993; Ross et al. unpublished manuscript).  Hydroperiod and water depth are considered to be major determinants of vegetation patterns in the exclusively fresh water portions of the Everglades (Busch et al. 1998; Newman et al. 1996; Craighead 1971; Loveless 1959), where E. cellulosa is generally found at more persistently and deeply inundated sites than C. jamaicense (McPherson 1973;  Herndon et al. unpublished manuscript).  Periodicity or depth of flooding may also distinguish mangrove species composition within the relatively narrow range of salinities that characterize these coastal forests (Davis 1940).  However, perhaps because of the relatively small number of sites and short period for which hydrologic data were available, interactions of hydroperiod with plant species composition were undetectable by our methods in the SESE coastal zone.  We expect that more extensive sampling would show that hydroperiod effects are pertinent here as well, though they may be secondary in importance to salinity and associated water quality measures.
 

Historical changes in SESE ecosystems:

    Our analyses documented two significant vegetation-related changes in the SESE over the last half century: (1) a movement toward the interior of a band appearing white on both color infrared and black and white aerial photos, and (2) a similar interiorward shift in the composition of plant communities comprising the coastal gradient.  In a parallel investigation of the paleoecology of the study area, we also traced a century-long increase in the proportion of marine-associated mollusk species preserved in SESE soil cores (Meeder et al. unpublished manuscript).  Several of these trends varied dramatically among drainage basins of contrasting character and history, with areas cut off from upstream water sources becoming altered most rapidly.  Such changes have taken place in the context of a global rise in sea level that has been 9.9 cm since 1940 at Key West (Maul and Martin 1993). We therefore believe that these are different manifestations of the same forces, i.e., the shifting balance between sea level and freshwater discharge in SESE.

    The white zone.  Since Egler's time, the white zone has changed from a band of mixed vegetation including approximately equal proportions of C. jamaicense, E. cellulosa, and R. mangle, to a monodominant community of R. mangle, with C. jamaicense virtually absent.  Because the white zone community of a half century ago differed from that present today, we surmise that the characteristic appearance of this interface on aerial photos results from factors other than species composition.

    Our data instead suggest that the appearance of the white zone on aerial photographs results from the combined effects of a reflective substrate and low plant cover. South and southwest of the C-111 Canal, the current white zone lies entirely within Mangrove Scrub vegetation, where leaf area index is less than 60% of that in the adjacent Sawgrass - Spikerush - Mangrove type. This contrast probably becomes accentuated during late spring or summer when new herbaceous shoots emerge in the latter. Given low macrophyte cover, the white appearance on CIR photos may be attributable to equal reflection in green, red, and near infrared bands by the surface substrate, which is a light-colored peaty marl, or, in some areas a mixture of marl with and recent storm deposits (e.g., Hurricane Andrew) (Meeder et al. unpublished manuscript).  Field observations indicate that the white zone usually does not extend into areas in which a coherent periphyton mat is at least seasonally present. Other factors which may affect the CIR appearance of SESE coastal zonation are dead biomass in the plant canopy, salt accumulation on leaf surfaces, the wetness of the soil surface, and leaf moisture content (Hardisky et al. 1986), or the species composition of the periphyton mat (Rutchey and Vilchek in press).

    Observed spatial and temporal variation in the position of the SESE white zone are consistent with the hypothesis that hydrological variables associated with the balance of marine and fresh water sources are somehow implicated in the presence of this unproductive band of vegetation.  Comparison of the 1940 and 1994 photo series indicates that the white zone has extended furthest inland in areas most deprived of access to fresh water. The areas east of U. S. 1 have long been entirely cut off from upstream sources of water by canals or roads. When these areas are flooded, it is for the most part by saline tidal waters of Barnes or Card Sound. In contrast, net outflow of fresh water through gaps in the C-111 Canal into marshes to the south (mostly Long Sound drainage) averaged about 175,000 acre-feet of water per annum during 1990-95 (Everglades National Park unpublished manuscript).  Here, the unculverted U.S. 1 serves to intercept fresh water flow and retain it within the basin. When flooded by wind-driven tides, both Long Sound and Joe Bay drainages are inundated by estuarine Florida Bay waters whose salinity is usually less than that of open sea water. Finally, the Joe Bay drainage is open at its upper end, but the effect of water management has been to shunt water either to the east or the west.

    The presence of a band of low apparent plant production midway along a coastal transect has been reported elsewhere.  Montague and Wiegert (1990) described several Florida coastal sequences in which a zone of low statured vegetation, sandwiched between taller marshes or swamps, occupied a belt coastward of the mean high tide line. Carter (1988) presented general marsh profiles for Britain and eastern North America in which the zone of sparsest vegetation occupied a region near the intersection of salt and fresh water bodies. In this position, one might expect maximum variance in salinity and in other water quality parameters which differ between marine and terrestrial sources. Depending on topography, frequent wetting and drying are also likely. A plant in such an environment must function under a wide range of conditions, and the ability to do so may require physiological adaptations that result in slow growth. In developing this concept, Ball (1988) pointed out that mechanisms of salt tolerance such as salt exclusion, conservative water use, and the associated need for control of leaf temperature through leaf display adaptations all incur carbon costs. Her studies showed that mangrove species with the widest range of salinity tolerance have the slowest growth at optimal salinity. Ball's arguments were applied to interspecific variation, but similar reasoning may be relevant for variation within populations, or for variation in other environmental variables.

    Changes in species composition and community structure.  Based on our analyses of current and historical SESE vegetation, marsh species composition - as represented by the interface of the Sawgrass Marsh and the Sawgrass - Spikerush - Mangrove Marsh - has been even more sensitive to sea level rise than has the position of the white zone.  Like any comparison of mapping results, this one depends on the accuracy of both modern and historical efforts, and on the equivalence of the mapping units utilized in each study.   In this case we were able to address the latter point by referencing Egler's original survey data in our classification procedure.

    The change in plant species composition documented in this study --- an advancing wave of replacement of C. jamaicense-dominated marsh by low R. mangle-dominated mangrove swamp --- signifies an extension of marine and brackish water conditions into formerly fresh water wetlands, as well as a structural shift from herbaceous to woody vegetation.  The structural changes carry ecosystem consequences that may be overlooked in analyses of coastal responses to sea level rise. In shrubland ecosystems, significant proportions of fixed carbon and absorbed nutrients become tied up in stem and branch pools with relatively slow turnover rates. To a considerable extent, turnover of woody material results from periodic disturbances rather than predictable seasonal cycles.  Moreover, from the perspective of associated faunal groups, low swamp communities provide a very different vertical and horizontal structure than herbaceous marshes.  Algal communities are also likely to respond to structural differences that alter the light or nutrient environment at the soil surface, or that interrupt the cohesiveness of the periphyton mat.  Structural impacts on hydrologic processes such as evapotranspiration or resistance to surface water flow are unknown for these ecosystems, but may be substantial.  Finally, sediment accretion processes and rates differ between SESE mangrove and graminoid ecosystems (Meeder et al. unpublished manuscript).  Because the process of mangrove encroachment has been so extensive in SESE, these changes may have important impacts at the level of the south Florida ecosystem as a whole.

 ACKNOWLEDGMENTS:

The project was funded by the South Florida Water Management District (Contract # C- 4244).  We especially thank Janet Ley, David Rudnick, and Rick Alleman of the District for their advice and cooperation throughout the project. Tom Armentano, Robert Fennema, Jerry Lorenz, Dewitt Smith, Alexander Sprunt, Bob Zep all provided information and discussion.  The following people provided field assistance and plant species identification: Jennifer Alvord, Zachary Atlas, Allen Herndon, Luz Romero, Carmen Navedo-Cordiva,  Joesph O’Brien, Kathy Rodriguez,  Ana Maria Rodriguez-Urquijo, and Carl Weekley.  This is contribution number ** from the Southeast Environmental Research Center at FIU.

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

Busch, D.E.,  Loftus, W.E. & Bass, O.L., Jr.  1998.  Long-term hydrologic effects on marsh plant  community structure in the southern Everglades.  Wetlands 18:230-241.

Carter, R.W.G. 1988. Coastal environments. Academic, London, U.K. 617 p.

Craighead, F.C., Sr. 1964. Land, mangroves, and hurricanes. Fairchild Tropical Gardens Bull.  19:5-32.

Craighead, F.C., Sr. 1971. The trees of south Florida, Vol. I: the natural environments and their  succession. Univ. of Miami, Coral Gables, FL. 212 p.

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

Davis, J.H., Jr. 1943. The natural features of southern Florida, especially the vegetation, and the  Everglades. FL Dept. Conserv. Geol. Bull. 25:1-311.

Dean, W.E., Jr. 1974. Determination of carbonate and organic matter in calcareous sediments and  sedimentary rocks by loss on ignition: comparison with other methods. J. Sed. Petr. 44(1):  242-248.

DeLaune, R.D., Pezeshki, S.R., & Patrick, W.H., Jr. 1987. Response of coastal plants to increase  in submergence and salinity. J. Coast. Res. 3(4):535-546.

Duever, M.J.,  Meeder, J.F., Meeder, L.C., & McCollom, J.M. 1994. The climate of South  Florida and its role in shaping the Everglades ecosystem.  In:  Davis, S.M. & Ogden, J.C.  (eds.)  The Everglades: the ecosystem and its restoration,  pp. 225-248. St. Lucie, Delray
Beach, FL.

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

Hardisky, M.A., Gross, M.F. & Klemas, V.. 1986. Remote sensing of coastal wetlands.    Bioscience 36:453-460.

Harper, R.M. 1927. Natural resources of southern Florida. Fla. Geol. Surv., 18th Annual Report.  pp. 27-206.

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

Hill, M.O. 1979. TWINSPAN: A FORTRAN program for arranging multivariate data in an  ordered two-way table by classification of the individuals and attributes. Microcomputer  Power, Ithaca, NY, USA.

Koch, M.S.  1996.  Resource availability and abiotic stress effects on Rhizophora mangle L. (red  mangrove) development in South Florida.  Ph. D. dissertation, University of Miami, Coral  Gables, FL.  161 pp.

Ladd, H.S. 1957. Paleoecological evidence. In: Ladd, H.S., (eds). A treatise on marine ecology  and paleoecology,  Geol. Soc. Amer. Mem. 67:599-640.

Leighty, R.G., Gallatin, M.H. Malcolm, J.L. & Smith, F.B. 1965. Soil associations of Dade  County, Florida. Inst. Food Agric. Sci. Exp. Stn. Circ. S-77A, Univ. of Florida,    Gainesville.

Long, R.W. & Lakela, O. 1971.  A flora of tropical Florida.  Univ. of Miami Press, Coral Gables,  Florida.

Loveless, C.M. 1959. A study of the vegetation in the Florida Everglades. Ecol. 40: 1-9.

Maul, G. and Martin, D.M.. 1993. Sea level rise at Key West, Florida, 1846-1992:  America's longest instrument record? Geophysical Res. Letters 20:1955-1958.

McCune, B. & Mefford, M.J.  1997.  Multivariate analysis of ecological data. Version 3.11.   MJM Software, Gleneden Beach, OR, USA.

McPherson, B.F.  1973.  Vegetation in relation to water depth in Conservation Area 3, Florida.   U. S. Geological Survey Open File Report 73025.  Tallahassee, FL, USA.

Meeder, J.F., Ross, M.S., & Ford, R.G.  1993.  Mangrove ecosystem expansion in south Florida  under conditions of accelerating rate of sea level rise: results of conceptual depositional  and spatial models.  In: Brunn, P. (eds).  Proceedings Hilton Head Island, S. C., USA,  International Coastal Symposium.  Volume 2, pp. 431-445.

Montague, C.L. & Wiegert, R.G. 1990.  Salt marshes.  In: Myers, R.L. & Ewel, J.J. (eds.)  Ecosystems of Florida.  University of Central Florida Press, Orlando, FL. 765 p.

Moore, D.R. 1961.  The marine and brackish water Mollusca of the State of Mississippi.  Gulf  Research Reports 1:1-58.
Newman, S., Grace, J.B. & Koebel, J.W. 1996. Effects of nutrients and hydroperiod on Typha,  Cladium, and Eleocharis: implications for Everglades restoration. Ecol. Appl. 6(3):774-  783.

Oksanen, J. & Minchin, P.R.  1997.  Instability of ordination results under changes in input data  order:  explanations and remedies.  J. Veg. Sci. 8:447-454.

Olmsted, I., Dunevitz, H. & Platt, W.J.  1993.  Effects of freezes on tropical trees in Everglades  National Park, Florida, USA.  Tropical Ecology 34(1):17-34.

Rejmankova, E., Pope, K.O., Post, R. & Maltby, E. 1996. Herbaceous wetlands of the Yucatan  peninsula: communities at extreme ends of environmental gradients. Int. Revue Ges.  Hydrobiol. 81(2):223-252.

Robertson, W.B., Jr. 1955. An analysis of the breeding-bird populations of tropical Florida in  relation to the vegetation. Ph.D. thesis, University of Illinois, Urbana.

Rutchey, K. & Vilcheck, L. 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.

Tabb, D.C. & Manning, R.B. 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(4):550-647.

ter Braak, C.J.F. 1987-1992. CANOCO - A FORTRAN program for canonical community  ordination. MicroComputer Power, Ithaca, NY, USA. 95 p. + addendum..

Thomas, T.M. 1974. A detailed analysis of climatological and hydrological records of south  Florida with reference to man's influence upon ecosystem evolution. In:  Gleason, P.J.  (eds). Environments of South Florida, present and past, pp. 82-122. Miami Geological  Society, Miami, FL

Thompson, F.G. 1984. The freshwater snails of Florida: a manual for identification. University of   Florida Press, Gainesville, FL. 94 p.

Turney, W.J. & Perkins, B.F. 1972. Molluscan distribution in Florida Bay. Sedimenta III. The  comparative sedimentology laboratory. Univ. of Miami, Miami, FL. 37 p.