Abstract Introduction Acknowledgments Methods Results Discussion Conclusions References

DISCUSSION

Modern Environmental Data

Our temperature and salinity measurements are in agreement with climatologic and oceanographic data gathered during the same time period in Florida Bay (Halley, http://www.sofia.usgs.gov/projects/circulation; Johns, et al., 1999; NOAA NCDC website http://www.ncdc.noaa.gov/). In the northern and eastern portions of Florida Bay, salinity varies more than temperature at each site on both a seasonal and an annual scale. Typically at each site, there is less fluctuation in salinity from year to year in February (the "dry" season), than during July (the "wet" season). The exception is the 1998 pattern; during 1997-1998, a very strong El Niño affected regional rainfall throughout southern Florida, bringing heavy wintertime rain. This was followed by a strong La Niña causing a hot dry summer in southern Florida. Text-figure 3 illustrates the effect of the anomalous regional wintertime rainfall on the salinity of northern and eastern Florida Bay. February 1998 resembles a typical summer pattern, whereas July 1998 resembles a winter pattern.

Text-Figure 13. - Map showing range of distribution of presence-absence clusters over course of sampling period from February 1996 through July 1999. This represents a compilation of data shown on individual seasonal plots in Text-figure 3. Clusters B and E are not shown; the distributions are limited and non-contiguous. [larger image]

Modern Analogues

Five distinct molluscan associations were identified from the analysis of the modern presence-absence database. Assemblages identified by clusters B and E (Text-fig. 4) are minor components of larger communities within Florida Bay, and are not considered significant for this study. Text-figure 13 shows the geographic range of five communities in Florida Bay, and Text-figure 3 shows the changing distribution from season to season over the period of observation.

Batillaria/Melongena Assemblage

The Batillaria/Melongena assemblage defined by cluster A represents a unique set of environmental conditions. This assemblage frequently co-occurs with the Brachidontes assemblage (cluster C), but occupies a unique subenvironment at each site. Subenvironments were not analyzed separately, therefore, the occurrence of the Batillaria/Melongena assemblage is not always reflected on the distribution maps (Text-fig. 3). This assemblage appears to be restricted to eastern and central Florida Bay, in very shallow areas with sparse subaquatic vegetation and in close proximity to islands. Dense Batillaria minima consistently have been present at site 14 since the beginning of our observation period (Text-fig. 3). Salinity does not appear to be a strong controlling factor; Batillaria minima and Melongena corona have been found from mesohaline to hypersaline conditions (Table 6). These observations are in general agreement with Turney and Perkins (1972).⁵ They identify Melongena corona as a common species in the northern subenvironment and Batillaria minima as an intertidal species found in sediments near land; we have commonly found Melongena corona in the northern transitional and eastern divisions of Florida Bay.

Brachidontes Assemblage

The Brachidontes assemblage defined by cluster C is the most predominant molluscan community found in eastern and central Florida Bay. Brachidontes exustus is commonly found alive at most sites in this region of the bay, regardless of salinity, water depth, substrate, or clarity. The low fidelity values for Brachidontes exustus in cluster C (Table 4) are due to the ubiquitous nature of the species. Although Brachidontes exustus is euryhaline, ranging from 10.2-41.3 ppt, the average value is 27 ppt .⁶ Brachidontes exustus have been found in the deepest parts of the basins in eastern Florida Bay, bysally attached to exposed limestone, and on the tops of mud banks attached to subaquatic vegetation. Brachidontes exustus are found attached to almost all common types of vegetation, including Thalassia, Halodule, Sargassum, Chondria, and Laurencia, and to exposed limestone and shell debris. When Brachidontes exustus are very abundant they commonly form "nests" in clumps of Chondria, Laurencia, and dead Thalassia, and they were observed forming a dense "Brachidontes-mat" by attaching to each other across the top of an expansive shoaling area (2/97, site 12). The most abundant Brachidontes exustus concentrations we have observed are typically found on the sides of the mud banks in clumps of Chondria and/or Laurencia in relatively murky water. In the process of collecting and processing samples, we have determined that Brachidontes exustus are very tolerant of low water quality; they have been found alive several days after collection in sealed plastic bags, filled with decaying organic material. Given the wide range of environmental conditions under which Brachidontes exustus have been found, it is difficult to determine what factors control their distribution. Given these tolerances it is noteworthy that Brachidontes exustus are primarily restricted to the northern transitional, eastern, and central portions of Florida Bay. A possible explanation is that their distribution is a function of decreased competition or fewer predators in this region of the bay.

An examination of the associated species that separate the sub-clusters within C indicates some of the environmental variables that control the distribution of the Brachidontes assemblages. Sub-cluster C1 is defined by the co-occurrence of Brachidontes exustus with Prunum apicinum; these species are most commonly found together on clumps of Laurencia or Chondria. The co-occurrence of Brachidontes exustus and Pteria longisquamosa with Argopecten irradians or Chione cancellata delimits sub-cluster C2. This group of species is indicative of relatively clear water and abundant healthy Thalassia; macro-benthic algae may be present. Sites with Brachidontes exustus alone or with Batillaria minima form sub-cluster C3. Batillaria minima are present when the water is very shallow, the substrate is generally soft mud, and the vegetation is sparse. Due to Brachidontes exustus tolerance of varying salinity and substrate conditions and relatively poor water quality, the presence of B. exustus alone may indicate the environment is not suitable for other epifaunal molluscs. Sub-clusters C4 and C5 have relatively high values for faunal richness. Pteria longisquamosa, Crepidula spp., Bittiolum varium, and Prunum apicinum may be present with Brachidontes exustus, and a number of other species in R-mode cluster Ia. Species from R-mode cluster Ib are included in sub-cluster C5. These diverse faunas are typically present when the substrate consists of mixtures of Thalassia and macro-benthic algae including Polysiphonia, Chondria, and Laurencia. Sub-cluster C6 is defined by the presence of Brachidontes exustus and Crepidula spp.; Crepidula are generally found on Thalassia, and like Brachidontes exustus, appear to be relatively tolerant of reduced water quality.

Count-data from the modern vegetation samples and the push cores are in agreement with presence-absence observation data, and provide further evidence of the predominance of Brachidontes exustus in the molluscan fauna of Florida Bay. Out of 11,597 total individuals collected on vegetation samples in July 1998, we found 5713 Brachidontes exustus (49.26%). These numbers are comparable to the findings of Lyons (1998). He found Brachidontes exustus comprised 78% of all the molluscs he collected in an extensive sampling program throughout Florida Bay in the summer of 1994 that gathered nearly 14,000 living specimens greater than 3mm. In the pristine fraction of the push core samples, Bittiolum varium comprises 30.3% of the specimens, and Brachidontes exustus 13.0%, followed closely by Transennella sp. at 11.43%. When all the fragments and worn specimens are included from the push core data, Brachidontes exustus comprises 20.64% of the total molluscan specimens, and Bittiolum varium 19.74%. The lower values for Brachidontes in the pristine sub-set are probably due to the fragility of the shell.

Pteria Assemblage

Pteria longisquamosa [= Pinctada radiata Turney and Perkins] and its associated species Modulus modulus, and Argopecten irradians define cluster D. This association of species is typically found on the sides of mudbanks (40-150 cm of water) in dense Thalassia beds, relatively clear water, and salinities between 20 and 40 ppt ⁷ (Table 6), in eastern, central, western and Atlantic transition zones of Florida Bay. Although Pteria longisquamosa has been found at its densest concentrations attached to Thalassia, it has also been found on macro-benthic algae, particularly Chondria and Laurencia, and it has the ability to camouflage itself to match the color of the vegetation to which it attaches. The occurrence of Pteria longisquamosa in relatively clear water raises an interesting question of whether P. longisquamosa occurs there because the water is clear, or whether the water is clear due to the filtering activity of P. longisquamosa. The latter especially may be true where Pteria longisquamosa are extremely abundant. Pteria longisquamosa, however, seem unable to survive in water of diminished quality; they generally die within hours of collection. Typically, the Pteria assemblage occurs outside the 20-ppt contour and to the east, south, or west of the Brachidontes assemblage (Text-fig. 3). Based on our observations, we believe the distribution of the Pteria assemblage is controlled by a combination of salinity, substrate, water depth, and water clarity.⁸ The dominance of this combination of species in a death assemblage would be indicative of polyhaline to euhaline relatively clear water, deposition on the side of a mud bank in 40-150 cm of water, and the presence of a relatively dense Thalassia bed.

Turney and Perkins (1972) found Brachidontes exustus, Pteria longisquamosa, Cerithium muscarum and Bittiolum varium to be the characteristic species of their interior subenvironment, which corresponds approximately to the eastern and central divisions of our current usage. Their results agree with our analyses. The distribution of the Brachidontes and Pteria assemblages overlap in central and eastern Florida Bay (Text-fig. 13). Turney and Perkins (1972, p. 10) state that the molluscs in the interior subenvironment are able to "survive large salinity fluctuations and other effects of poor circulation." They point out the ability of the characteristic species to survive in other subenvironments, but suggest that one reason they flourish in the interior may be a lack of competition. Our data support the idea of wide salinity tolerances for these assemblages, but our data indicate Pteria longisquamosa is sensitive to water quality. This apparent discrepancy can be explained by the fact that Turney and Perkins were analyzing death assemblages.

The "Western" Assemblages

The Turbo, Tegula, and Columbella association that forms cluster F1 has been found at three sites (17, 18, 23) in the western, Atlantic and Gulf transition zones; these sites have a mixture of various types of sub-aquatic vegetation including Thalassia, Syringodium and Halodule and calcareous green algae. These sites also have varying substrates and water depths, because each site has a channel bordered by a mudbank or sandbar. Currents flowing through the channels can be strong at times, and the water is typically clear; salinities have ranged from 25-38ppt. Based on our observations, we consider this assemblage to be indicative of near normal marine conditions (upper polyhaline to euhaline), relatively clear water, and the presence of sub-aquatic vegetation. Currents and varying substrates also may be determining factors in the distribution of the F1 assemblage, but we have not observed this assemblage frequently enough to reach a conclusion.

Cluster F2 is a weak association of species including Cerithium spp., Chione cancellata, Carditimera floridana, and Columbella spp. and other faunal groups from R-mode cluster Ia. The cluster is formed of samples from sites 13, 17, 21, 24, and 26 in the central, western and Gulf transition zones of Florida Bay. Salinities for the sites in cluster F2 range from 30-39. Our data suggest this assemblage represents euhaline conditions, but the limited observations and weak association of fauna make this a speculative conclusion.

Pinnidae and Pleuroploca gigantea are the critical fauna in cluster G. These taxa are generally found in the western most sites (23 and 24) in the Gulf transition zone and they are representative of open marine euhaline conditions. Pinnidae are found in shallow shoaling areas on top of banks, partially submerged in the substrate. The substrates have varied from muddy calcareous sand to firm calcareous mud. These fauna have been rarely observed, but seem indicative of euhaline conditions.

Our observations of the more euhaline marine assemblages differ from Turney and Perkins (1972). Lithopoma americanum [=Astraea americana Turney and Perkins, 1972] and Tegula fasciata were listed as characteristic species of Turney and Perkin's (1972, p. 13) Atlantic subenvironment, and they stated that "the Atlantic group is composed of species which are almost entirely restricted to that part of the bay which is influenced by the Atlantic Ocean." Although our observations of living Lithopoma americanum and Tegula fasciata are limited, we have observed both species in the central (site 17), Atlantic (site 18), and Gulf transition (site 23) zones.

Comparison of presence-absence data with other data

Through quantitative and qualitative comparisons of the death assemblages represented by the push core data with other data sets (Lyons, 1995, 1998; Turney and Perkins, 1972; and our modern vegetation and observation data discussed above) we have determined that the push core data are a valid representation of the fauna that have lived at each site during the 0-10 years preceding collection. The prominence of Brachidontes exustus and Bittiolum varium in the push cores from sites 8, 12, 13, and 20 agrees with our modern vegetation data from those sites. Transennella sp. and other infaunal organisms (absent in the presence-absence observation data and the vegetation sample data) are a significant component of the push-core assemblages. Their absence from the observation and vegetation data sets does not indicate a lack of agreement with the push core data but is simply an artifact of our method of observing and sampling (see discussion under methods). Turney and Perkins (1972) did a detailed analysis of the death assemblages at over 75 stations in Florida Bay in the 1950's and they concluded that "transportation of shells larger than silt size is not significant in Florida Bay" (1972, p. 32). Only local migration of shells takes place and "faunal dislocations" are rare. Diagenetic and depositional processes have little effect on the distribution of molluscan assemblages (Turney and Perkins, 1972, p. 35). Thus, their work supports our conclusion that push core data provide a valuable addition for down-core comparison to the historical data set.

Historical Piston Core Data

Three primary clusters are formed by the analysis of the compiled data set (modern vegetation samples, push cores, and piston cores) (Text-fig. 11). Only the Brachidontes assemblage that defines cluster C in the modern analysis is represented in the compiled data set as cluster K. The other modern assemblages identified in the cluster analysis of the presence-absence data set are not represented in the piston core data.

The absence of the modern assemblages (with the exception of the Brachidontes assemblage) from the piston cores can be explained by a number of factors. The Batillaria/Melongena assemblage typically is limited to shallow-shoaling areas in relatively close proximity to land. The four piston cores were not taken under these conditions in the modern environment, and the nearly complete absence of Batillaria minima and Melongena corona from the piston cores implies these conditions did not exist in the past at the localities where piston cores were taken. Clusters B and E, the Ostrea/Arcopsis assemblage and the Laevicardium assemblage, are limited to only a few sites in the modern sampling, are loosely associated groups, and are generally affiliated with Brachidontes. When these species are present within the piston core, the samples join with other clusters. Pteria longisquamosa are present in the piston cores (primarily RB19B and PK37), but do not form the discrete Pteria assemblage seen in cluster D of the modern analysis. Instead, Pteria longisquamosa are a component of the Brachidontes assemblage in cluster K1a. The fragility and low preservation potential of Pteria longisquamosa shells most likely affect their relative abundance in the piston core data, although they are easily recognizable from small fragments. This observation agrees with the results of Turney and Perkins (1972); they did not distinguish separate Pteria longisquamosa = [Pinctada radiata Turney and Perkins 1972] and Brachidontes exustus groups in the death assemblages from the interior subenvironment where they found both species to be abundant.

The key species in the two western assemblages (clusters F and G) are completely missing from the piston cores. Turbo castanea, Tegula fasciata and Columbella spp., which comprise cluster F1, are not present in any piston core samples. Cluster F2 is a more loosely associated group, defined by the presence of Columbella spp., Cerithium spp., Chione cancellata, and Carditimera floridana; Carditimera and Columbella are absent from all piston core samples. Cerithium and Chione are present, but typically in association with Brachidontes or Transennella dominated assemblages. Pinnidae and Pleuroploca gigantea (cluster G) are absent from the piston cores, but Lithopoma americanum is a rare component in the upper 50cm of piston core BA6A. The absence of the Turbo/Tegula/Columbella and Pinnidae/Pleuroploca assemblages from the piston cores indicates that during historical times conditions in northern, eastern and central Florida Bay did not reach the more open marine conditions currently seen in the western, Gulf, and Atlantic transition zones.

Compiled Data

Text-Figure 14. - Relationship between Q-mode cluster assignment (compiled data set, Text-fig. 11) and sample position. Specific push-core and vegetation samples are identified. [larger image]

Samples from piston cores RB19B and PK37, and the upper portion of BA6A, are primarily grouped in cluster K (Text-figs. 11 and 14). This is equivalent to the Brachidontes assemblage identified in the modern presence-absence data analysis. Modern push cores and vegetation samples from the corresponding sites (12, 13, and 20) also fall into cluster K in the combined analysis. These sites occur within eastern and east-central Florida Bay, and they can be characterized as shallow mud banks with sub-aquatic vegetation and macro-benthic algae present. Salinities have ranged from 17-33 ppt at site 20, 19-37 ppt at site 12 and 27-41 ppt at site 13 during our periods of observation. Typically, either the Brachidontes or Pteria longisquamosa assemblages have dominated at these sites in the modern presence-absence data (Text-fig. 3). The prevalence of the Brachidontes assemblage down-core for PK37 and RB19B, and down to approximately 54cm in BA6A, indicates that the general environmental conditions were the same in the past. The common species in cluster K correspond to the characteristic species identified by Turney and Perkins (1972) for their interior subenvironment.

Text-Figure 15. - "Blow-out" areas at Bob Allen mudbank, February 1999. Sharp transition between lush Thalassia beds and areas of gelatinous calcareous mud are visible. Photo courtesy of Thomas Scott, Florida Geological Survey. [larger image]

Samples from the lower portion of piston core BA6A are dominated by the Transennella assemblage that defines cluster L. This assemblage also is seen in portions of RB19B (30cm and 118cm) and PK37 (6cm). This is an infaunal assemblage and so it was not identified in the modern presence-absence analysis. Transennella sp. are present in the modern push core samples, but they are components of diverse Brachidontes assemblages, and thus the samples join cluster K. Bob Allen mudbank today (site 13) is distinctive among the sites discussed here. Lush Thalassia beds adjoin "blow out" areas at site 13 (Text-fig. 15) and the transition from grass to gelatinous carbonate mud is very sharp. Low diversity Transennella assemblages seen in the piston core may be the down-core representation of these modern "blow-out" areas, but we have not sampled the modern analogue for this cluster and can only speculate on what it indicates about the environmental conditions.

Piston Core T24

Historical changes in molluscan fauna at the mouth of Taylor Creek (piston core T24) are shown in Text-figure 7. The gradual decrease of the R-mode XIc and the increase in Brachidontes exustus, Crepidula spp., and Cerithium muscarum up-core illustrate the gradual increase in salinity that has occurred at this site during the time of deposition. Brachidontes exustus increase significantly in the upper 20cm of the core. Modern analogue data (discussed herein; Turney and Perkins, 1972; Lyons, 1996, 1998, 1999) have demonstrated that Brachidontes exustus seem most abundant in areas of fluctuating salinity and in areas where macro-benthic algae are available for habitation. Fluctuations in salinity at this site could be attributed to changes in fresh-water flow due to climatic change, or changes in water management practices, or it could be due to long-term changes in sea level. Fluctuations in the relative abundance of Hydrobiidae may be directly indicative of changes in fresh water flow. Hydrobiidae are minute (generally less than 3mm) fresh-water gastropods that can not survive in estuarine waters. They arrive in Little Madeira Bay via terrestrial freshwater flow through Taylor Creek, and perhaps their abundance can be related to flow rates, but additional field-testing is necessary. Despite the obvious changes in the relative abundance of individual species, the entire piston core falls within cluster J and Anomalocardia auberiana, Turney and Perkins' (1972) indicator of the northern subenvironment, is present throughout the piston core. In addition, diversity, evenness, faunal richness, and number of individuals remain relatively constant compared to the other piston cores.

Piston Core PK37

Deposition at Pass Key (piston core PK37) (Text-fig. 8) has been relatively rapid, and the piston core represents only the latter half of this century. Brachidontes exustus has dominated the assemblage at this site during the period of deposition, and below 26cm, the relative abundance of B. exustus is fairly constant. Above 26cm, however, a great deal of fluctuation occurs. Seemingly dramatic changes in species dominance near the top of piston core PK37 are due to extremely low numbers of individual shells and low faunal richness; which is probably a result of the rapid sedimentation rate. The complete disappearance of all species except Brachidontes exustus and Transennella sp. in the upper 6cm may be indicative of rapid sedimentation at the site and/or conditions unfavorable to other species. The increases in Anomalocardia auberiana and R-mode XIc group may denote periods of decreased salinity; this assemblage is characteristic of Little Madeira Bay, and the 10cm sample joins cluster J.

Piston core RB19B

The piston core from Russell Bank (RB19B) (Text-fig. 9) records approximately 110 years of change at the site. The lower portion of the piston core (~1880-1922) shows fluctuations in faunal richness and numbers of individuals, but the relative abundance of the key species does not fluctuate dramatically. Transennella sp. is the most abundant species in this portion of the core, but the presence of Brachidontes exustus, Cerithium muscarum, Bittiolum varium and other species in most of the samples indicates the presence of sub-aquatic vegetation. Significantly, Anomalocardia auberiana and other species from the R-mode XIc assemblage typical of the northern transition zone, are present before 1922. The presence of these species may indicate lower salinities and possibly the influence of a lower salinity plume. In the modern Florida Bay, a less saline plume of water is sometimes present on the western edge of the basin south of Little Madeira Bay (Text-fig. 3, July 1996; Halley, http://www.sofia.usgs.gov/projects/circulation). After 1922, significant changes occur in the relative abundance of the key fauna. Transennella sp. decrease, and Brachidontes exustus, Bittiolum varium and other components of the Brachidontes assemblage increase somewhat. After 1942, fluctuations become even more pronounced, until approximately 1980 when Brachidontes exustus increase dramatically in relative abundance, and all other key species, with the exception of Pteria longisquamosa and Modulus modulus, decline. The increase in Brachidontes exustus in the upper 20cm of the piston core corresponds to and increase in numbers of specimens (this is consistent with the findings of Lyons 1998) and a decrease in faunal richness, evenness and diversity.

Piston core BA6A

Bob Allen mudbank (piston core BA6A) preserves a record of deposition since approximately 1784. No modern analogue has been identified for the Transennella sp. assemblage (cluster L) that dominates the lower portion of piston core BA6A, before 1922 (54cm). The lowest portion of the piston core (140-158 cm) indicates either a rapidly fluctuating environment, or deposition at the transition zone between a grass bed (indicated by cluster K assemblage) and a low diversity, sparse Transennella sp. habitat. Beginning around 1890 to 1901, Brachidontes exustus becomes a significant component of the molluscan assemblage in piston core BA6A, and around 1901 the relative abundance of Transennella sp. declines and never again constitutes greater than 51% of the assemblage. The Brachidontes assemblage (cluster K), indicative of the presence of sub-aquatic vegetation) dominates the upper portion of piston core BA6A. Beginning in the late-1920's, Lithopoma americanum, indicative of the more euhaline western environments, is occasionally present at Bob Allen mudbank. Around 1981, Brachidontes exustus becomes the most dominant species in the piston core. In general, the measures of diversity (faunal richness, evenness, and the Shannon's diversity index) all fluctuate significantly throughout the deposition of this core, and the values for numbers of individuals and faunal richness are lower in piston core BA6A, in comparison to the other piston cores examined.

Summary of Significant Historical Findings

1) Molluscan assemblages that exist at each piston core locality today, typically existed there throughout the period of deposition represented by the piston cores. Shifts have occurred in the patterns of dominance and diversity within each piston core, but the faunal composition has not changed significantly.

2) Piston core T24, at the mouth of Taylor Creek, records gradually increasing salinity, an increase in euryhaline molluscan fauna, and a decline in mesohaline molluscan fauna up core. Anomalocardia auberiana, typical of the northern transition zone, is present throughout the core.

3) Pass Key piston core (PK37), southeast of Little Madeira Bay, shows significant fluctuations in Anomalocardia auberiana and the other species indicative of the northern transition zone.

4) Between 1913 and 1933, distinctive changes occur in the molluscan assemblage at Russell Bank (piston core RB19B). Anomalocardia auberiana, consistently present in low percentages prior to 1922, completely disappears from the site after that time.

5) At Bob Allen mudbank (piston core BA6A) a dramatic change in faunal assemblages occurred between 1890 and 1901, from a Transennella sp. dominated assemblage to a Brachidontes exustus dominated assemblage.

6) From approximately 1930 to 1980, Lithopoma americanum, typical of euhaline western and Gulf transition environments, is occasionally present at Bob Allen mudbank (piston core (BA6A).

7) From approximately 1930-1980 dramatic shifts in percent abundance of different species within the Brachidontes assemblage occur at both Bob Allen mudbank and Russell Bank.

8) Beginning around 1980, Brachidontes exustus becomes the dominant mollusc at Russell Bank, and faunal diversity and evenness decline. Similar increases in Brachidontes exustus are seen in the upper portion of all four piston cores analyzed.

⁶ The mean and median salinity values for Brachidontes are 27.45 and 28.8 ppt, based on 85 observed occurrences of the species. These values do not take into account the abundance of the species at a site, and may in part reflect the frequency of sampling within higher salinity ranges (see Text-fig. 2).

⁷ The mean and median salinity values for Pteria longisquamosa are 29.3 and 29.7 ppt, based on 62 observed occurrences of the species at a site. These values do not take into account the abundance of the species at a site, and may in part reflect the frequency of sampling within higher salinity ranges (see Text-fig. 2).

⁸ Water depth, water clarity, and vegetation are not independent of each other. The tops of the mudbanks typically have less vegetation and more material in suspension than the sides of the banks.