Evolution of Florida Bay from Island Stratigraphy

Evolution of Florida Bay from island stratigraphy

PAUL ENOS Department of Geological Sciences, State University of New York at Binghamton, Binghamton, New York 13501 RONALD D. PERKINS Department of Geology, Duke University, Durham, North Carolina 27708

ABSTRACT Florida Bay with time. Progressively larger areas of the bay are building into the supratidal zone through both vertical and lateral The sedimentary record of most Florida Bay islands is an asym- accretion. If the present depositional trend continues without sig- metric cycle consisting of a transgressive sequence followed by a re- nificant increase in the rate of sea-level rise, the present area of gressive sequence, both formed during a continuous Holocene rise Florida Bay will eventually be accreted to the Florida mainland as a in sea level. The principal sedimentary environments of Florida Bay coastal mangrove swamp, a supratidal "marl prairie," and a series and the south Florida mainland are represented in the cycle by an of shallow lake basins. upward succession of (1) freshwater pond, (2) coastal mangrove swamp, (3) shallow bay ("lake"), (4) mud bank, and (5) island. Setting Some parts of the cycle may be missing, but the sequence is always the same. Supratidal carbonate sedimentation on islands may de- Sediments in Florida Bay are being deposited disconformably on velop from coastal mangrove swamps or by mangrove colonization an almost planar surface of Pleistocene pelletoidal lime packstone of mud banks. Islands have developed from mud banks at many and grainstone (Perkins, 1977). Extensive sediment probing and different times during the rise of sea level into Florida Bay, indicat- observations both from the air and underwater reveal no appreci- ing that mud banks must have existed throughout most of the his- able Pleistocene depositional relief. The Pleistocene surface is pitted tory of the bay. Present trends of island formation and growth by small solution holes several centimetres (less than 1 ft) deep, suggest that Florida Bay will evolve into a coastal carbonate plain producing a microkarst topography similar to that exposed farther with inland mangrove swamps and freshwater ponds, very similar north in the Everglades. The surface slopes southward and south- to the present southwest Florida mainland. westward from the Florida mainland at about 1:100,000 (0.5 ft/mi) until it rises abruptly along the Florida Keys, reflecting depositional INTRODUCTION morphology of coralline limestone. Florida Bay thus occupies a shallow, wedge-shaped basin that is deepest (3 m; 10 ft) im- Florida Bay occupies the inner part of the south Florida shelf mediately behind the Florida Keys and along its southwestern (Fig. 1) and covers an area of 1,550 km2 (600 mi2; Scholl, 1966, margin. Table 1). The boundaries of the triangular bay are defined by bar- Sea level in south Florida has risen at a slightly declining rate for riers that restrict circulation. On the southeast, a ridge of Pleis- at least the past 5,000 yr (Fig. 2; Scholl and others, 1969). Flood- tocene coralline limestone (Key Largo Formation) forms the nearly ing of Florida Bay began in the southwest, where the rock floor lies continuous barrier of the Florida Keys, interrupted only by a few 3 m (10 ft) below present sea level, about 4,500 yr ago and con- tidal passes. The southwest border of Florida Bay is delineated by a tinues to the present. Despite the continuous rise in sea level, parts series of shallow mud banks that extend from Cape Sable at the of the south Florida mainland have prograded seaward more than 8 southwestern tip of the Florida mainland southeastward to km (5 mi) through lateral accretion of marine lime mud and peat. Matecumbe Keys (Fig. 1) and comprise Sandy Key Bank near Cape The present bathymetry of Florida Bay is dominated by a series Sable, Ninemile Bank, and Peterson Key Bank. The triangular area of shallow sublinear mud banks, which divide the bay into slightly south of Ninemile Bank, extending from Lower Matecumbe Key to deeper basins, or "lakes" as they are locally known. Systematic dif- Schooner Bank to Bamboo Banks and Vaca Key, may be considered ferences exist in the width of mud banks, the relative area occupied either as part of Florida Bay or transitional to the Gulf of Mexico by basins, and the abundance of islands in different parts of the shelf to the west. This area has fewer banks and greater tidal ex- bay. The width of banks increases progressively from the north- change than Florida Bay proper. The south Florida mainland forms eastern part of the bay toward the west. The relative area occupied the northern boundary of Florida Bay, consisting of a supratidal by basins is concomitantly reduced. Islands are more abundant in carbonate flat ("marl prairie") from Cape Sable to Flamingo and a the northeastern and particularly the central part of the bay than in series of mangrove-fringed embayments and discontinuous low, the western part. In northeastern Florida Bay (north of lat 25°N shelly beach ridges east of Flamingo (Fig. 1). and east of long 80° 43'W), Stockman and others (1967, p. 644) Florida Bay is compartmentalized by a series of carbonate mud have estimated that basins occupy 90% of the area and are floored banks that are awash at low tide. Small, mangrove-fringed islands, by an average of 15 cm (6 in.) of Holocene sediment. Banks and some partially flooded at high tide, dot the mud banks. The dis- islands occupy 10% of the area and average 1 m (3 ft) in sediment tribution and stratigraphy of these islands with their mud-bank thickness. Scholl (1966, p. 283) has estimated the area of mud nuclei indicate an increase in the size and number of islands in banks for the entire bay to be 496 km2 (192 mi2) or 32% of the

Geological Society of America Bulletin, Parti, v. 90, p. 59-83, 19 figs., 1 table, January 1979, Doc. no. 90112.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 61

total area.1 He defined mud banks as areas less than 60 cm (2 ft) spite the limited range of the semidiurnal tides within the bay, the deep, thus excluding much of the gently sloping margins of the mud banks are awash at low tide. banks. The contrast in these two estimates primarily reflects the Prolonged winds may move larger volumes of water than the much greater area occupied by banks in the western part of the bay. lunar tides. Northeasterly winter storm winds tend to push water Mud banks of the western bay (west of long 80°50'W, north of out of the bay through the tidal passes, slightly reducing the bay lat 24°55'N) are broad, amoeboid-shaped shoals. Associated basins level. Prevailing wind direction and wind velocities are therefore ("lakes") of western Florida Bay are poorly defined and generally nontidal factors that significantly affect the water levels in Florida floored by as much as 1 m of sediment (Fig. 3). The basins are inter- Bay. connected by a few sluggish, poorly defined channels in the broad Currents within the open "lakes" of Florida Bay are less than banks. In contrast, basins of the eastern bay are interconnected by 1 cm/s (Gorsline, 1963, p. 131) but may reach 125 cm/s in tidal narrow, well-defined, active channels through most of the mud passes into the bay (Vaughan, 1935). banks. Salinities in Florida Bay range from 6°/oo to 70°/oo, depending Islands ranging in size from a few tens of square metres to about upon seasonal variations in rainfall and net evaporation within the 2 km2 (500 acres) are scattered along the mud banks, particularly in bay and in amount of mainland runoff (Ginsburg, 1956; McCal- central Florida Bay. Scholl (1966) estimated the total area of ap- lum and Stockman, 1961; Gorsline, 1963; Lloyd, 1964; Scholl, proximately 105 islands2 or keys at 67 km2 or about 4% of the bay 1966). Fluctuations are greatest in northeastern Florida Bay, the area. In contrast to the linear Florida Keys, which are underlain by area most isolated from normal oceanic water. Extremes diminish Pleistocene rock, the keys of Florida Bay are composed of muddy toward the southwest, so that salinities in most of Florida Bay are carbonate sediments extending a few centimetres to nearly 1 m between 35°/oo and 40°/no. Water temperatures within Florida Bay above the adjacent bank. Many of the islands are covered or range from 19 to 38 °C (Ginsburg, 1956). Even more extreme fluc- fringed by mangroves and are flooded at high tide. Other islands, tuations in both salinity and temperature occur in water a few cen- particularly those near the Florida mainland in western Florida timetres deep ponded on the islands. Bay, stand several decimetres above normal high tide and support Turbidity of Florida Bay water is generally high, imparting a salt grasses or palms and hardwoods. characteristic translucent milky-green color to the water. The floors Tidal deltas in the form of irregular or triangular-shaped banks of "lake" basins about 2 m (5 to 7 ft) deep are visible only after have accumulated along tidal passes through the rocky Florida calm periods lasting a day or more. Turbidity varies locally accord- Keys. Generally, the surfaces of the deltas are slightly submerged ing to the bottom conditions; turbidity is reduced in areas of thick except for natural levees along the tidal channels and for a few is- grass cover and shelly bottoms containing little interstitial mud. lands that rise slightly above the low-tide level. Circulation of open oceanic water into Florida Bay is restricted SEDIMENTARY ENVIRONMENTS by the rocky Florida Keys, which are broken by only five tidal passes into the bay proper, and by the broad mud banks of western The depositional record of essentially all modern sedimentary Florida Bay. Within the bay ,circulation is further restricted by the environments of Florida Bay and the adjacent mainland is repre- numerous mud banks. Tidal range is about 60 cm (2 ft) on the At- sented in cores from Florida Bay islands. A summary of modern lantic side of the Florida Keys and at Cape Sable; it is abruptly sedimentary environments and their resultant sedimentary prod- damped by the narrow tidal passes and mud banks to less than 15 ucts is given in Table 1. cm (0.5 ft) behind the first line of mud banks in Florida Bay. De- Freshwater Ponds

1 Percentage figures differ from those given by Scholl (1966) because he • Lime mud is being deposited in shallow freshwater ponds in the based calculations on long 81°05'W as a convenient western boundary of Florida Bay, rather than the smaller alternate boundary, defined by mud Everglades, beyond the normal limits of salt-water encroachment. banks, used here. This increased the area from 1,546 km2 to 2,179 km2. Most of the broad expanse of the southern Everglades is "saw-grass prairie" that is covered by water during the rainy season but stands dry during prolonged droughts. The saw grass grows on very sparse 2 More than 170 islands are shown on U.S. Coast and Geodetic Survey 1:80,000 charts (1250), but many are insignificant in size. soil overlying bedrock. Sediment accumulation is negligible and is largely confined to organic debris. o; 0 Depressions in the rock floor form semipermanent ponds that are t- populated by spike rush (Eleocharis) and algal mats (see Spackman z UJ I and others, 1964). In this environment, lime mud or "marl," as it is 2 ac precipitation is still debated. Freshwater lime mud is composed of nearly 100% calcite; marine mud of Florida Bay averages 58 % 3 < £ aragonite, 27% magnesium calcite, and only 15% calcite (Scholl, 1966). UJ The freshwater lime is typically dense white or cream lime 3 UJ 4 mudstone or wackestone (Fig. 4, A). Vague wavy laminae are found locally in the pond deposits; the fine-scale, regular lamina- 150 IOOO 2000 3000 4000 5000 tion of algal stromatolites has not been observed. In the present YEARS BEFORE PRESENT (RADIOCARBON) setting, freshwater ponds form directly on Pleistocene limestone. Figure 2. Sea level in south Florida to 5500 B.P. (radiocarbon 3 Depositional textures defined by Dunham (1962) are used here to de- years). Redrawn from Scholl and others (1969). scribe sediment, ignoring that the "stone" endings imply lithified rock.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 Figure 3. Thickness of Holocene sediments in Florida Bay and adjacent shoreline. Probe data from Shell Development Company files were sup- areas. Isopachs omitted locally on small islands where coincident with plemented by notations on U.S. Coast and Geodetic Survey charts.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 63 Consequently, basal pond deposits commonly contain lithoclasts grove swamps and peat deposits in south Florida has been the sub- derived from the underlying limestone. Concentrations of plant ject of numerous papers, notably those of Davis (1940), Scholl material formed in the ponds or by encroachment of saw grass pro- (1963, 1964), Spackman and others (1964), and Craighead (1964). duce dark-gray organic-rich layers or lenses. Sediment deposits in mangrove swamps are dark, fibrous to plastic peat (Fig. 4, A). Lime mud and shells may be widely dispersed or MANGROVE SWAMPS concentrated in irregular lenses and burrow fillings. Accumulations of peat in depressions on the Pleistocene surface Mangrove swamps are a prominent shoreline feature throughout have been cited as evidence that the processes of peat production southern Florida. They are also common on Florida Bay islands. dissolve limestone to initiate the depressions (Spackman and Mangroves may form dense forests extending for kilometres inland others, 1964). However, as with the chicken and the egg, it may from the actual shoreline, forests developed behind a narrow beach equally well be argued that the depressions retained the water or a broad supratidal mud flat, or scrubby growth on a rocky necessary for mangrove development and therefore preceded peat shoreline as along northeastern Florida Bay and Biscayne Bay. accumulation (F. C. Craighead, 1965, personal commun.). Climax mangrove forests develop on a substrate of thick sediment The peat of south Florida is frequently referred to as "basic peat" or peat that is not too compact for root penetration (F. C. because interstitial water typically has a pH of 7.5 to 8 (H. K. Craighead, 1965, personal commun.). The development of man- Brooks, 1965, personal commun.). This is interpreted to mean that carbonate-bicarbonate buffering controls pH in the interstitial water of the peat. If the formation of peat tends to produce acid conditions, some dissolution of the carbonate is necessary to produce the high pH values observed. The real question, at present un- answered, is whether the volume is significant. Lack of significant solution is suggested by good preservation of thin calcitic shells of tree snails commonly found in the peat. Shell layers (Scholl, 1964) and calcitic lime mud are commonly well preserved beneath thick peat deposits (Fig. 4, A) but have not been systematically studied. Craighead (1964; Pray, 1966) observed that layers of lime mud deposited by storms in the Everglades or on mangrove islands may disappear within a few years in locations where erosion by currents or by rain may be largely discounted. He concluded that quantities of this aragonitic lime mud are dissolved by metabolic processes or decay of mangroves. The potential of these processes, however, would appear to be severely restricted on small islands surrounded by marine water that provides a virtually infinite' reservoir of buf- fered water supersaturated with calcium carbonate.

Basins ("Lakes") of Florida Bay

The largest entities in Florida Bay are the broad shallow basins, or "lakes," as they are known locally. Basins of eastern Florida Bay occupy about 90% of the total area (Stockman and others, 1967) and differ systematically from those of the western bay that occupy somewhat less than half of the total area. A typical "lake" of the eastern bay is roughly polygonal in outline, reflecting the linearity of the enclosing banks. It is 3 to 13 km (2 to 8 mi) in maximum dimension, 1 to 2 m (3 to 7 ft) deep, and floored by bare rock or a veneer of shelly sediment. Stockman and others (1967) cited an average sediment thickness of 15 cm (6 in.), which includes areas marginal to mud banks. Grass and green algal cover is sparse or lacking in the basins because of the lack of an adequate substrate to anchor roots and holdfasts. The depth and fetch of the basins are Figure 4. Cores of Florida Bay sediments. A. Lime mud from sufficient for wave action to winnow the fines produced in or trans- freshwater pond deposition overlain by peat from coastal man- ported into the basins. Most of the winnowed material probably grove swamp. Differential shrinkage and distortion of peat results accumulates on the leeward side of the mud banks where wave ac- from drying of original 90% (by volume) water content. Lime mud tion is reduced, or it is flushed out of the bay entirely by wind- is 94% low-magnesian calcite and 6% aragonite, by X-ray diffrac- driven currents and violent storms (Ball and others, 1967). Thus, tion analysis. Bottom of core F1.6, Calusa Key; see Figure 12 for the basin-floor sediment in eastern Florida represents a lag deposit location. B. Mud-bank sediment. Vertical tubes are formed by that spans 3,500 yr (Fig. 2; sea level at -2 m). More than 70% of roots of turtle grass; horizontal cracks result from shrinkage. Note the sand-size fraction (greater than 62 fiin) consists of molluscan shrinkage of geopetal internal fill in large articulated pelecypod skeletal fragments (Ginsburg, 1956). (Chione cancellata). Core F3.1, Upper Cross Bank east of Bottle The "lake" biota in general is impoverished (Ginsburg, 1956) Key (see Fig. 5 for location), 140 cm (4.6 ft) from sediment surface. because rock and coarse shells suuport neither a large infauna nor

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 64 ENOS AND PERKINS

TABLE 1. CHARACTERISTICS OF SEDIMENTARY ENVIRONMENTS OF FLORIDA BAY

Environments Lithified equivalents of Color Particles Grain Size Fines principal sediment types (% < 62 (im)*

Freshwater pond Mudstone; gastropod wackestone White to dark gray Lithoclasts, carbonaceous Fine High fragments, shells Coastal swamp and Peat Black Plant fibers, shells mangrove islands

Open basin or "lake" Coarse mollusk fragment pack- Gray, mottled Shells, lithoclasts, "gray Coarse 10-60 stone and wackestone grains"

Mud bank Mollusk wackestone; locally fine Medium gray Pellets, shells 0.025 mm; 40-90 skeletal fragment packstone poorly sorted

Windward fringe Mollusk fragment packstone shells Coarse 5-40?

Bare mud patches Very fine skeletal fragment wacke- pellets, grass blades Very fine 85 stone; pellet mudstone?

Channel Skeletal grainstone and packstone; Coarse to fine fine skeletal fragment mudstone Fine and wackestone Tidal delta Very similar to mud-bank deposits Island Supratidal flat Foram, gastropod, and/or pellet Tan, cream, Pellets, small shells, Fine (silt) High wackestone and mudstone light gray intraclasts

Mangrove fringe Peat and lime mudstone Black, dark gray Plant fibers

Beach Mollusk grainstone; foram/fine Shells, grass blades 5—high skeletal fragment packstone * From Ginsburg (1956), Taft and Harbaugh (1964) and Paul Enos and L. H. Sawatsky (unpub. data). From Paul Enos and L. H. Sawatsky (unpub. data).

an extensive grass community with its attendant epifauna. The sed- that would serve as a substrate for grass. The carpet of turtle grass, iment is thoroughly burrowed by worms, crabs, and shrimp once established, would serve both to trap additional sediment and (Ginsburg, 1957; Shinn, 1968b), and the mud-size fraction is to prevent winnowing (Ginsburg and Lowenstam, 1958). largely pelleted by these and other organisms. Many of the mollus- Coarse shelly units found near the base of many mud banks and can and foraminiferal grains are discolored gray or black, a process island cores are interpreted as basal "lake" deposits. They may, that is not well understood but is associated with areas of slow sed- however, antedate the development of the entire bank system. If so, iment accumulation. these shelly units would be more analogous to sediments of the The basins of western Florida Bay are 1 to 2 m (4 to 7 ft) deep, shallow, relatively open shelf bordering the Gulf of Mexico west of rather irregular in outline, and 4 to 9 km (2.5 to 5 mi) in maximum Florida Bay than to the sediments now accumulating in the shal- dimension. In contrast to the basins of the eastern bay, they typi- low, enclosed basins of Florida Bay. Sediment flooring the shallow cally contain as much as 1 m of muddy sediment (Fig. 3) that sup- Gulf of Mexico shelf is a lag deposit similar to that of eastern ports a dense grass population and a large infauna. Sediment in Florida Bay, but it contains more open-marine elements, especially these basins is similar to the sediment in the mud banks (see below), plates of the green alga Halimeda. The distribution of Halimeda in but it contains an appreciably greater percentage of shells. Less in- present-day Florida Bay is essentially restricted to those areas adja- tense winnowing of the western bay basins may result from (1) cent to tidal passes that transect the Florida Keys. Halimeda plates slightly smaller basin size with less effective fetch; (2) much broader are found near the base of many cores in Florida Bay, particularly enclosing mud banks, or (3) greater depth to rock floor, which may those in the western bay (P. Fickman, unpub. data), indicating dep- ultimately have placed the floor of enclosed basins below local ef- osition under more open conditions than have prevailed in Florida fective wave base. Thus, fine-grained sediment could accumulate Bay since the extensive development of mud banks.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 65

TABLE 1. (Continued)

Structures Geometry Thickness: Maximum Porosity1 Permeability* Remarks width thickness (%) (millidarcys) (m)

Wavy laminae Circular or irregular lens 1:100 1? 90% to 100% calcite; lithoclasts from (locally) underlying Pleistocene limestone Mangrove roots and Irregular ribbon or 1:10,000 4 90 1500 Generally contains lime-mud admixture; tubules, burrows seaward-thickening to 1:12 thickness: width = 1:12 for eroded (filled with lime wedge islands in Whitewater Bay mud) Burrows Sheet 1:50,000 0.3 (east), Blackened, corroded shells fairly 1 (west) common; conditions in western lakes with sediment and grass carpet approach those of banks Burrows; grass roots Elongate prisms (eastern 1:50 2 66-78 50-100 and rhizomes bay); triangular blanket (western bay) 1:5,000 4 Irregular lenses or ribbon 1:50 0.3 Winnowed shelly lags accumulate on windward (northeast) side of banks Millimetre lamination Lenses, circular 1:10 1 66-73 0.6-1 Includes negative-relief ("blowouts") and positive-relief ("mud dunes") features; most abundant on leeward of banks Cross-lamination Slip-off slopes are muddy Lamination Elongate prism 1:100 4 Abandoned channel fill

Wedge 1:5,000 5 67 40

Lamination, algal Irregular to elliptical 1:300 4 61-66 2,000- Islands most abundant on narrow banks stromatilites, mud lenses to 1:1,000 25,000 in central bay cracks, "birds eyes," burrows, crumbly texture, root perforations 1:1,000 2? Best developed on leeward sides of islands

Narrow prism 1:10 0.3 On windward sides of islands In intertidal zone

Banks they are not produced by wave erosion of the grass cover but rather that the grass is killed locally by other causes exposing the underly- Most of the sediment of Florida Bay is contained in shallow mud ing sediment to erosion. Erosion of the sediment below the sur- banks, particularly the very broad irregular banks of the western rounding surface produces a bowl-shaped depression with a basal bay (Fig. 3). The tops of most mud banks are flat and generally lag that subsequently becomes a sediment trap during calmer awash at low tide. The flanks slope away from bank crests at angles periods. Sediment settling out of suspension fills the depression, that do not exceed 1°, even in the narrow banks of the eastern bay producing well-developed laminae, which are accentuated by detri- (Fig. 5). Most of the bank surface is covered by a thick carpet of tal grass blades. Ultimately, green algae and grass may repopulate turtle grass (Thalassia testudinum). the surface. Thus, a lens of laminated mud is incorporated into the Narrow banks that intercept the effective winter storm winds typical shellier sediment of the bank (Fig. 5). from the northeast quadrant have distinct windward and leeward Bare areas with positive relief are traction deposits or "dunes" of sides. The windward side is slightly steeper (Fig. 5) and generally is pelleted mud ("soft sand" of R. N. Ginsburg, 1964, personal com- mantled by a skeletal lag. The leeward parts of mud banks slope mun.), as indicated by a distinct surface asymmetry and ornamen- more gently, have a thicker carpet of grass associated with muddier tation by low-relief (oscillation?) ripples. Laminae are apparently sediment, and are characterized by patches of bare muddy sedi- produced by dune migration, but they are so nearly horizontal that ment (Fig. 6). The bare patches have either negative or positive re- it is doubtful that dunes deposits can readily be distinguished from lief of a few decimetres or centimetres relative to the surrounding fill of the depressions in the stratigraphic record. Mud dunes appear grass-covered surface. Patches with negative relief are commonly to be most common on "spits" that project from the leeward side of referred to as "blow-outs" because plumes of sediment-laden water mud banks such as the two immediately south of Bottle Key (Figs. 1 drift from them during prolonged winds (Fig. 6). Their greater and 6, A). abundance on the leeward side of banks suggests, however, that Channels that cut through mud banks (Fig. 6, B) typically are

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 CROSS BANK SEA LEVEL

PLEISTOCENE

O 100 200 300 0 PACKSTONE r WACKESTONE

2T( SCALE, FEET VERTICAL IOOX MUDSTONE

Figure 5. Cross section of Florida Bay mud bank. Cross Bank is typical of narrow mud banks of eastern Florida Bay. Note slight asymmetry; steeper (north) side faces dominant northeast winter storm winds. Apparent bedrock relief in section is fortuitous. Numerous prob- ings from many banks show no consistent bedrock relief associated with mud banks. Cores by E. A. Shinn and P. R. Rose.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 67

floored by a lag of shells, but deposits of the channel banks are very Islands in the early stage of development, which constitute the muddy and locally contain well-laminated slip-off—slope deposits. majority, are characterized by a low beach ridge on the windward The many subenvironments of the bank combine to produce a side, a periphery of red mangrove (best developed on the leeward complex mosaic of sediment types. Most of the sediments, how- side), an inner zone of black mangrove, and an interior open tidal ever, consist of skeletal wackestones, thoroughly burrowed and flat or shallow pond (Fig. 7). penetrated by roots and rhizomes of turtle grass (Fig. 4, B). The windward sides of low islands have distina erosional nips at the shoreline. Sediment in the intertidal part of a nip is locally fine- Islands grained skeletal sand. The beach ridge, generally less than half a metre high, is constructed immediately behind the nip (Fig. 8). The Most islands of Florida Bay are partially or entirely flooded at ridge consists of marine shells, intraclasts of soft muddy sediment spring high tide, but on some, mostly in northwestern Florida Bay, eroded from the nip, and lodged pieces of marine grass and island sediment accretion has reached half a metre or more above normal vegetation (Fig. 8, B). Beach-ridge sediments give way to laminated high-tide level. Craighead (1964) distinguished "early," "middle," muds with a barely perceptible slope toward the island interior. and "late" stages of island formation on the basis of the transition Red mangroves (Rbizophora mangle) line most of the island from (1) algae and mangroves to (2) less salt-tolerant plants to (3) shores. On the windward side, clumps of red mangroves with tropical hardwoods that require fresh ground water. The vegeta- myriad prop roots form "buttresses" anchoring erosional cusps. tion primarily reflects the elevation of the island above sea level. The mangrove fringe is generally wider, denser, and taller on the

A A

Figure 6. Aerial views of Florida Bay mud banks. A. Low Figure 7. Aerial views of Florida Bay islands. A. Crane Key, oblique north-north east across Ramshorn Shoal toward Bottle Key looking northwest. Windward side is at top. Profile northwest- (center). Bank in foreground is a dead-end bank or "spit." Note southeast in Figure 13 runs through prominent elongate open tidal turbid water drifting from grass-free patches. B. High oblique of flat in center of island. Note density of mangrove forest on leeward Cross Bank looking north, near location of cross section in Figure sides of island, particularly adjacent to shallow mud banks. 5. Windward side (top) is indicated by small breakers and linear B. Stake Key, looking southeast (see Fig. 5 for location). Windward patterns, which reflect in part accumulations of shelly sediment. sides are at left (northeast side) and in foreground. Shallow mud Leeward side has prominent patches bare of grass. Light linear banks with local mangrove colonization are prominent at top and scars are furrows left by wayward motor boats at bank edge. right. Concentric tree lines apparently reflect former positions of Natural tidal channel across bank exposes Pleistocene bedrock lo- leeward margin of island; generally no visible relief is associated. cally. Dark neighboring feature (right) may be filled and abandoned Relief is very slight at present leeward margins. Windward margins channel. are generally undergoing erosion.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 68 ENOS AND PERKINS

leeward sides of the islands (Fig. 7). Although mangroves form an Sediment deposited on the open supratidal flat is fine-grained, abrupt wall on the leeward side, there is little surface relief at: the pelleted lime mud and skeletons (mainly miliolid foraminifers) transition from the island to the soft, muddy bank accumulating in transported in suspension by flooding tidal currents. Sediment the shelter of the island. layers as much as 5 cm (2 in.) thick may be deposited by a single The inner edge of the red-mangrove fringe merges with scrub storm tide (Ball and others, 1967). Algae subsequently recolonize black mangroves. Black mangroves (Avicennia nitida) tolerate a the fresh sediment surface and form a new mat. The resultant salinity range from at least 50°/oo to nearly freshwater (Davis, sedimentary record consists of dense mud layers separated by 1940), but they do not form the pioneer stand at the shoreline (or at organic-rich concentrations of algal mats with scattered skeletons the limit of salt-water influx in the Everglades) as do the red man- of organisms that live in and on the algal mats, notably cerithid gas- groves, perhaps because the black mangrove lacks the prop root tropods. buttresses needed for colonization into open water. A distinctive Floating blades of turtle grass may lodge in thick windrows on characteristic of the black mangrove is its network of protruding the islands during storms. In this oxidizing environment, the only breathing roots (pneumatophores) that radiate outward several sedimentary record of grass deposition is a distinctive concentra- metres from the main trunk. These roots disrupt sedimentary struc- tion of flat-sided, grass-encrusting organisms such as the spirorbid tures in the surrounding sediment. worms, foraminifera, and red algae that are oriented horizontally, The interior of most low or "early" islands is an open supratidal reflecting the final resting position of flat grass blades. flat populated by mats of blue-green algae. Storm tides and spring Island-interior sediments (Fig. 9, A) are characterized by (1) high tides periodically flood the flats. Between floodings the sea laminations formed by intermittent deposition and algal-mat de- water may be diluted by rainfall or concentrated by evaporation to velopment, (2) desiccation cracks, (3) light colors produced by oxi- salinities of more than 80°/oo. Infiltration and evaporation may lead dation, (4) small fenestral voids or "birds-eyes" formed by concen- to surficial desiccation, cracking, and curling of the algal mats. The trations of gas bubbles (Shinn, 1968a) and by plant rootlets, and water temperature, as well as the salinity, fluctuates widely. The (5) a distinctive crumbly or flaky texture apparently developed by thin film of water cools rapidly with the lowering of air tempera- repeated wetting and drying. Shells deposited on islands commonly tures; absorption of the sun's rays by the dark algal mats sometimes are filled with flakes or clots of sediment from the desiccated island warms the water so that it becomes too hot to stand in. surface rather than the fine pelleted mud that typically filters into shells deposited beneath water. A larger percentage of unfilled shells also characterizes island sediments. Although burrowing is less intense than in the shallow-marine environment, very large A burrows (as much as several centimetres across and half a metre deep) formed by land crabs are fairly common (Fig. 9, A). Sediments deposited in island interiors populated by mangroves are dominantly unlaminated carbonate with organic-rich lenses and root remnants ranging from minute hairlets to the tough black sheaths of larger roots. Peat, containing admixtures of lime mud and shells, is formed only where mangrove growth is dense and well established, generally on the leeward sides of islands. If sediment accumulation on the island exceeds the rate of sea- level rise, the island surface will accrete progressively higher above the normal tide range. As the frequency of flooding is thereby reduced, salt-tolerant shrubs, sedges, and trees, such as buttonwood (Conocarpus erectus), may establish themselves. These contribute organic matter and act as baffles that trap additional sediment during storm tides so that accretion of the island continues. Eventually land grasses may be established. As island accretion reaches several decimetres above sea level, population by palms and tropical hardwoods begins. Although the conditions on mud banks adjcaent to the islands differ little from those found elsewhere on the banks, the proximity of an island may be detected in the molluscan and foraminiferal as- semblages (Turney and Perkins, 1972; P. R. Rose, unpub. data; Ginsburg, 1964) or by concentrations of the black fibrous roots of = , 1 1—— mangroves. Many islands are partly surrounded by a distinct TOv P 0I I iINCHE S I SI "moat" of deeper water which may extend nearly to bedrock. The moat is perpetuated, and likely was formed, by tidal and wind- Figure 8. Beach ridges. A. Shelly beach ridge, approximately 30 driven currents deflected around the island. cm high, with storm-generated spillover lobe at north point of Bot- Distinctive deposits formed by concentrations of floating turtle- tle Key. B. Core of beach ridge at northeast side of Calusa Key. Top grass blades have been found in embayments of several islands. Re- is at left. Inclined beds are dominantly shells, mud, or organic mat- ducing conditions are established in the shallow water, preserving ter. Below shrinkage crack at right is burrowed mud-bank sedi- distinctively laminated, organic-rich deposits containing the as- ment. Core F1.5; see Figure 12. sociated grass-encrusting biota (Fig. 9, B).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 69

ISLAND STRATIGRAPHY evolution of Florida Bay. Individual cores are shown graphically in Figure 10. Isolated sediment cores from several Florida Bay islands had been interpreted to indicate that islands have occupied their present Bottle Key positions throughout the flooding of Florida Bay (Fleece, 1962, p. 70-71, 85; R. N. Ginsburg and E. A. Shinn, 1964, personal Bottle Key is a low-relief, early-stage island in eastern Florida commun.). Formation of islands late in the history of Florida Bay, Bay which projects northward from a wish-bone-shaped mud bank subsequent to mud-bank development, had been suggested by (Fig. 1). The surface of Bottle Key is ornamented by a series of tree Davis (1940), Gorsline (1963), and Craighead (1964) and as an al- lines which suggest scars of lateral accretion toward the south (Fig. ternative by Fleece (1962, p. 84). Systematic coring of Bottle Key by 11). No noticeable relief is associated with these lines, except for R. L. Walpole, N.D. Jeffries, and R. M. Lloyd in 1964 showed a the present beach ridge at the north end. The surface of the entire history of very recent island development. This led to a restudy of island is soft and marshy. available island cores, to refinement of criteria for recognition of Cores show a thin, patchy layer of freshwater lime mud and peat island, bank, and basin sedimentation, and to coring of additional overlying the Pleistocene rock (Fig. 11). This is overlain by a thin islands. In this section, the stratigraphy and history of the islands but persistent shelly layer that apparently represents the basal studied are discussed as a basis for generalized interpretation of the marine deposit as the sea transgressed into Florida Bay. This sedi-

A A*. Urlili- ' -,

J3* '">1 'V-Vv " t I Jïjsv

i { v : y- -

. 4

CM 0 10 15

INCHES

Figure 9. Island sediments. A. Sediment from open, interior algal flat of Crane Key. At top is slightly disturbed surficial algal mat, living when sample was taken. Dense, light-colored layers immediately below are mud deposited by storm tides; thickest layer was perhaps deposited during Hurricane Donna in 1960. Other dense, light layers are faintly visible below. Crack at center was widened by sample preparation, but algae growing into cleft show that it was open mud crack prior to sampling. Dark laminae beneath dense mud layers are compressed algal mats. Large disturbed area in center is burrow, probably of land crab. Vertical cut through plastic- impregnated box-core sample. B. Laminated sediment formed by mat of grass blades floated into small embayment adjacent to an island. Top highly distorted. Wide gaps between layers were formed on drying and are now impregnated with dark-blue plastic. Underlying sediment is mud-bank deposit. Core F1.7, Calusa Key; see Figure 12. Both photos at same scale.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 SEDIMENTARY STRUCTURES SKELETAL PARTICLES

"Birdseye" structure

Burrow FORAMINIFER

Lamination miliolid

77 - Lens ¡) peneroplid

—v~ Mud crack & coiled, general

RABBIT KEY =)c= Mud cracked layer encrusting

DI25.I DI25.2 DI25.3

PARTICLES MOLLUSK TOP 0.55 M SCALE MISSING I -- MD ê Intraclasts * 1 MD fa gastropod „ Pellets oyster co cc •-: • Quartz sand S pecten

NK P/MO ÜJ 3S Shell fragments ÜJ S 1 WK 4 Carbonaceous flake WORM ( \ 1 i - A tube, general 1 < _ LITHOLOGIES* ^ encrusting P ' 1.

m ~ GR GR Grainstone "l, MD * i I ^ WK PK Packstone OSTRACOD

MD WK Wackestone general MANGROVE P SWAMP MD Mudstone PLANT PARTICLES I-,' P Peat

îiSïj Peat Icore log) ALGAE \yi Shrinkage Loss /N on drying MANGROVE "J iL. OT* Mats (blue green) SWAMP Probed depth to rock o Halimeda (coring lossl 0 Calcisphere

MANGROVE P SWAMP *Depositional texture terms (Dunham, 1962) MANGROVE ANGIOSPERMS SWAMP of lithified equivalents. Grass, largely Thalassia

1 root

—' rhizome

Figure 10. Stratigraphy blade

of selected islands of Mangrove Florida Bay. For locations of cores, see Fig- $ root ures 1, 13, 14, 15. ^ tubule

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 CRANE KEY

0127.1 DI27.2 DI27.3

WK/GR_ 55W WK/PK m

m - c

WK/PK H P WK

W f Vi» >-f ' aP j/J, y ES WK 7m

ili"" y M 04 • ®S H 'Jiié r-i* SCALE O k ir Y. f >

l4 Figure 10. (Continued).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 CALUSA KEY

FL.I

PK WK WK/PK WK WK

<3wo MARSH SWAMP FRESH it'' WAT tH X POND << V

SCALE

r O r o

Figure 10. (Continued).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 BOTTLE KEY

N S D4I.I D4I.2 D4I.3 D4I.4 D4I.6 D4I.5

CROSS SECTION SCALE Figure 11. Cross sec- SEDIMENT FACIES tion of Bottle Key, east- ISLAND ern Florida Bay. See Fig- (SUPRATIDAL) ure 1 for location. Dashed lines on index MUD • BANK map trace tree lines like those seen in Figure 7, B. H "LAKE" MARINE Cores by R. M. Lloyd, 100 200 300 M R. L. Walpole, and N. D. SWAMP Jeffries. (PEAT) 10 FRESHWATER POND (CALCAREOUS MUD)

PLEISTOCENE INDEX MAP H LIMESTONE

500 I000FT

CALUSA KEY WNW ESE Fl.3

SEDIMENT FACIES ISLAND (SUPRATIDAL)

MUD • BANK "LAKE" MARINE 200 M

f||i| SWAMP Hi (PEAT)

FRESHWATER POND iiSiil (CALCAREOUS MUD)

PLEISTOCENE LIMESTONE 500 FT

INDEX MAP NNE ssw Fl.5 Fl.6 Fl.4 Fl.3 I I I • I

Figure 12. Cross sections of Calusa Key, central Florida Bay.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 CRANE KEY

DI27.3 DI274 DI27.I DI276 DI27.5

RABBIT KEY

wsw ENE DI25.I DI25.2 D 125.3 DI25.4

SEDIMENT FACIES

ISLAND (SUPRATIDAL)

MUD • BANK [3=1 "LAKE" MARINE 100 M SEDIMENT FACIES 2 SWAMP ¿älä (PEAT) ISLAND (SUPRATIDAL) PLEISTOCENE LIMESTONE j j MUD Ea BANK

r^l LAKE 100 M CS=J MARINE 500 FT SWAMP NW SE (PEAT) 0127.2 DI27.I DI 27.7 DI 27.8 PLEISTOCENE I (PROJECTED) S LIMESTONE

"0 500 FT Figure 14. Cross section of Rabbit Key, western Florida Bay.

Figure 13. Cross sections of Crane Key, south-central Florida Bay. Dashed lines index map trace vegetative patterns; see Figure 7, A.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 75 ment is virtually identical to that now being deposited in open ba- position (Fig. 2) than did Crane Key and likewise shows asymmet- sins of the bay, but it may have preceded development of mud ric development with slow growth eastward (Fig. 14). A moat cuts banks. The next overlying unit, which forms the bulk of the sedi- the broad bank around the north and west sides of the island; this ment beneath Bottle Key, represents typical mud-bank deposition. may retard accretion in these directions. Supratidal lime mud and peat deposited in an island environment form only a superficial layer beneath the island. Slightly thicker is- Sandy Key land sediment is found offshore at the north point of the island (core D41.1, Fig. 11), where peat is being exposed to erosion over a Sandy Key is a high island at the extreme western edge of the considerable area in water less than 1 m deep. broad mud banks of western Florida Bay (Fig. 1). The exposed Either Bottle Key has rapidly colonized large areas of mud bank western side of the island is a shell beach with a recurved spit at from a small nucleus in the north end of the present island, or the either end (Fig. 15, A). The back beach is populated by mangroves entire island has migrated rapidly southward, with concomitant and palms. Cores at the edge of the island contain nearly 1 m of erosion at the north end. These interpretations are actually quite coarse shell debris overlying mud-bank sediment. Apparently the is- similar, but rapid migration would imply that a considerable vol- land is formed almost entirely of coarse skeletal sands deposited by ume of island sediment may have been eroded from the north end storm-driven waves and tides breaking over the mud bank that of the island. Colonization from a small nucleus is perhaps the forms a small slope break at the western edge of Florida Bay (Ball, more reasonable interpretation because the erosion of compact 1967, p. 577). The setting of the bank, exposed to the open shelf of peat, which forms most of the substrate at the north end of the island, is slow relative to the erosion of the shelly lime mud of the banks. The present configuration of Bottle Key and the adjacent bank (Fig. 1) suggests that the island acted as a buttress or "an- * chor" as the remainder of the mud bank migrated southward.

Calusa Key

Calusa Keys are four small islands in an early stage of development in central Florida Bay (Fig. 1). The history of the largest of the four is considerably different from that of Bottle Key (Fig. 12). The outer fringes of the island record a transgressive sequence (in ascending order) freshwater lime, coastal mangrove peat, and open- bay deposits, capped by a regressive sequence of shallow mud bank and supratidal island sediments. In core F1.4 (Figs. 10, 11), however, lime mud was deposited in a supratidal mud flat developed over the mangrove swamps at the shoreline. Vertical accretion in pace with the rise of sea level maintained the island while the surrounding swamps were inundated during the development of Florida Bay. Analogous but much larger supratidal flats are now found along the shoreline between Cape Sable and Flamingo. The embryonic island subsequently spread laterally to form a larger nucleus, as evidenced by cores Fl.l, F1.2, F1.3, and F1.6. Early growth westward to the edge of the mud bank was rapid, followed by essentially vertical accretion (core Fl.l), although this is a relatively protected side of the bank. The island has subsequently grown more slowly to the north and the southeast to attain its present size.

Crane Key

The larger of the Crane Keys in south-central Florida Bay (Fig. 7, A) has a history similar in several respects to both Bottle Key and Calusa Key. The island, still in an early stage of development, has grown outward from a nucleus that developed on bank sediments Figure 15. Aerial views of Florida Bay islands. A. High oblique (Fig. 13, cores D127.1 and D127.2). The northwest-southeast looking north across Sandy Key, westernmost Florida Bay (Fig. 1). profile (Fig. 13) shows that subsequent growth has been slightly Island is formed of shell sand; note accretion scars at near end. asymmetrical. Slower accretion toward the northwest, the wind- Shallow channel through far end of island with ebb and flood sand ward side, is evident. lobes was formed by breaching of island during Hurricane Donna in 1960. Darker area in foreground is deeper water (3 m; 8 to 10 ft) Rabbit Key marking west edge of Florida Bay. B. Low oblique, looking west across Triplet Keys along north border of Florida Bay. Islands are North Rabbit Key, an early- or middle-stage island in western merging from three or four nuclei by mangrove colonization of in- Florida Bay, developed at a somewhat higher, hence later, sea-level tervening shallow mud bank.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 76 ENOS AND PERKINS

the Gulf of Mexico, and its prominent shell beach make it unique relative to rise in sea level. Accumulation of island sediment is de- among Florida Bay islands, or at least an isolated end member. The pendent primarily on exposure to sediment-laden storm tides and beach forms a protected area for accumulation of finer grained sed- the trapping of sediment by vegetation. iment and the growth of plants. ORIGIN OF ISLANDS Other Islands Several hypotheses have been advanced to explain the origin of Fragmentary information from other islands indicates histories the islands in Florida Bay. Davis (1940) outlined the colonization similar to those studied in detail. A single long core and several of shallow banks and coastal areas by red mangroves. This process short cores from Cluett Key, an early-stage key in western Florida of island nucleation has been widely accepted by subsequent work- Bay (Fig. 1), shows a supratidal section overlying about 1 m of ers. It appears that this is the most important and fundamental of mud-bank deposits. The base of the section contains some peat. De- several processes that can lead to the development of islands. tailed work in progress by R. Steinen (1976, personal comrnun.), appears to confirm and enhance these interpretations. Murray Key, Mangrove Colonization a middle-stage island in northwestern Florida Bay (Fig. 1), was cored near the south shore by F. C. Craighead (1965, personal The red mangrove is viviparous; the cigar-shaped pod that drops commun.), and the molluscan fauna was studied by Durbin Tabb. from the parent tree at various times during the year, particularly in At the base of the core is a few centimetres of peat, overlain by 60 July, contains a live seedling (Davis, 1940). The encased seedling cm (2 ft) of lime mud containing a brackish molluscan fauna, and that falls into the water floats horizontally and may be dispersed by 1.5 m (nearly 5 ft) of lime mud containing a normal bank fauna. over wide areas. As the floating seedling matures, the root end be- The top 60 cm (2 ft) of the core consists of peat deposited near the comes heavier so that the floating position becomes vertical. Leaves edge of the island. Palm Key, also in northwestern Florida Bay (Fig. and roots project from opposite ends of the sheath. After it has at- 1), is a late-stage island. Two probes made with a hand auger indi- tained this position, the seedling may take root any time it drifts cated more than 2 m (6 to 8 ft) of white, crumbly lime mud, con- into a favorable substrate shallow enough for it to touch bottom. taining a few pulmonate gastropods, overlying a basal peat. This This may include a rocky shoreline with scattered soil pockets as suggests that the entire Holocene record is that of supratidal well as a mud bank. The floating mangrove seedling gradually be- sedimentation. comes heavier than water and sinks if it does not find a suitable substrate, but it may float for more than a year (Davis, 1940). The Summary depth of water in which the seedling can root is determined by its submerged length. The initial length of the sheath is 15 to 20 cm (6 It is apparent that the sedimentary history of Florida Bay islands to 8 in.), but the seedling increases in size with the development of is varied. A few islands developed from coastal tidal flats and were roots and leaves. never flooded during the Holocene rise in sea level. The present The rooted seedling develops prop roots that help trap flotsam, sites of other islands were flooded during the transgression of including other mangrove seedlings. A cluster of mangrove seed- Florida Bay and islands developed at later times overlying a basal lings further contributes to its substrate by acting as a baffle to cur- transgressive unit and a mud-bank nucleus. rents, which promotes deposition of suspended sediment, and by The transition from mud bank to island indicates sediment contribution of plant debris and the skeletons of mangrove epi- accretion at the erstwhile sea-level position. Therefore, the onset of zoans. Epiphytic algae, such as Batophora, attach to the prop roots island deposition can be dated approximately from the sea-level and enhance the baffle effect. As with any pioneer stand, most of curve (Fig. 2). Comparison of Figures 11 through 14 shows that is- the mangrove seedlings that colonize mud banks do not survive. lands developed at many stages during the flooding of Florida Bay. The chief cause of mangrove mortality is probably the strong cur- Since islands developed from mud banks that stood near sea level rents and high tides of the major storms (Craighead, 1964, p. 8). and many different levels of bank-to-island transition may be seen, Flotsam and epiphytes, which contribute to substrate development even under a single island, at least some mud banks existed during slow accumulation, give storm currents added leverage to throughout the history of Florida Bay, excepting perhaps the very uproot seedlings. earliest. The different depths of bank-to-island transitions indicate that is- Other Concepts land inception and growth are not primarily dependent on the present relatively slow rate of sea-level rise (Fig. 2). Some islands de- Gorsline (1963, p. 137-139) briefly considered the origin of veloped during the more rapid rise of sea level which prevailed ear- Florida Bay islands. He observed that enclosed lagoons and linea- lier in Holocene time. tions in vegetation patterns indicated lateral accretion of the keys Islands at similar stages of development show marked contrasts and suggested that "growth of the keys was primarily vertical dur- in age of nucleation. Bottle Key, an early-stage island, is relatively ing the initial flooding of the bay floor, then later changed to mainly young, but Calusa Key, also an early-stage island, is as old as lateral accretion because the rate of sea-level rise markedly de- Florida Bay. Palm Key, a late-stage island, has likewise existed since creased about 3,000 Before Present. . . ." Profiles of the islands initial flooding of the bay, but Murray Key, a middle-stage island, (Figs. 11—14) show some periods of dominantly vertical accretion, may be as young as 2,000 yr. Clearly the stage of island formation indicated by steep contacts between bank and island units, and is not a function of absolute age. It is a relative sequence analogous other periods of more rapid lateral accretion, indicated by nearly to the stages of youth, maturity, and old age in landform develop- horizontal contacts of island units overlying bank deposits. How- ment. Stage is determined by rate of vertical sediment accretion ever, the sequences appear to be too variable between islands or

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 77

even within the same island to permit correlation of periods of more favorable conditions are found. Trees near the center of the accretion. More recent sea-level curves (Fig. 2; Scholl and others, islands may be killed by water stagnation and sedimentation over 1969) indicate a gradually declining rate of sea-level rise during the the weakened roots (Craighead, 1964). Another mechanism for de- past 5,000 yr rather than a sharp change in rate at 3000 B.P. as veloping open areas in the island interior is by mangrove coloniza- proposed in earlier curves. tion across the mouth of a bight or embayment in the margin of an The implication by Gorsline (1963) that islands have persisted island. throughout the history of flooding of Florida Bay is strengthened by The stratigraphic record shows that many islands have traces of the statement that the linear trends of mud banks in the inner bay peat or organic-rich mud at the base of an island section; this could have originated from "the growth of spits and shoals from the reflect the migration of the mangrove periphery of the island or ini- mangrove island nuclei" (Gorsline, 1963, p. 139), implying that is- tial colonization of the entire area by mangroves that were later lands preceded mud-bank development. In discussing the history of killed by sediment accumulation. Other islands lack concentrations the keys — in particular their distribution between the eastern and of organic matter at the base of the island section. This suggests central parts of the bay — however, Gorsline (1963) considered at that at least the center of the island was never colonized by man- least three possible origins: (1) colonization of high parts of the groves. One might also argue that it shows that mangroves were shoal by mangroves; (2) development from an initial substrate not instrumental in the formation of the island. However, the prominence or patch of marsh; and (3) drowned topography (ap- windward fringe of the embryonic island is the part most vulner- parently bedrock highs to distinguish this from origin 2). He re- able to erosion; thus, the record of mangrove colonization in this jected the second and third hypotheses because radiocarbon dates critical part of the island might well be eroded. The center of the indicated that all the sediments originated in the present cycle (less island, although it may be composed of lime mud, which is more than 4000 B.P.) and because coring (Fleece, 1962) showed no un- easily eroded than peat, would be protected by its interior position. usual relief under the keys. Both of these observations are corrobo- Deposits on the leeward side of the island should also be perferen- rated by the more extensive coring and probing done for this study. tially preserved; peat deposits are more common in this position, Gorsline (1963) concluded that the best explanation is "the spread although not invariably present. of mangroves from the land areas of the Florida Peninsula out over Another process that may be important in the growth of islands the shoals to form the nuclei of the keys." This implies the possibil- is the merging of adjacent islands. This appears to be occurring in ity that islands may originate at some time after the initial flooding the Triplet Keys in northern Florida Bay, where mangroves are of the bay and the formation of mud banks, a central conclusion of bridging a narrow expanse of bank between adajcent islands (Fig. this study. 15, B). This type of development may account for some lines of Craighead (1964) observed the different stages of island de- vegetation on islands that do not conform to normal patterns of velopment in modern Florida Bay and commented, "Only when lateral accretion. this [mechanical] accumulation reaches a depth that would permit Peat and/or freshwater lime mud immediately overlying the Pleis- surface exposures at low tide would it be possible for the man- tocene rock are commonly preserved beneath islands. Peat is found groves to become established." This downgrades the role of man- in several locations beneath banks, but it is much less common. groves in island formation but would be consistent with develop- Preservation beneath islands does not imply that islands develop ment of islands at any or several stages during the flooding of only on depositional highs formed by peat, because basin and bank Florida Bay. Sandy Key is an example of an island nucleated by deposits commonly overlie the peat and precede island deposition. mechanical accumulation of sediment. Rather, it probably reflects the stability of the islands, which has protected the underlying deposits from erosion. Mangrove swamps Present Interpretation probably formed in most areas of the bay near the shoreline of the transgressing sea. Any resulting peat deposits have been eroded Cores of Florida Bay islands show great variations in strati- from most areas, either at the shoreline during the initial transgres- graphic history, or at least in the duration of island sedimentation. sion, on the floor of the lake basin, or along the windward side of a Supratidal sedimentation has continued throughout the Holocene bank where it may be exposed by slow bank migration. Most areas rise of sea level in the center of Calusa Key (Fig. 12); in contrast, of the bay floor probably have been subjected to wave erosion at most of Bottle Key probably developed within the past 1,000 yr some time during the Holocene transgression, except where forma- (Figs. 11, 2). Islands with a continuous history of supratidal tion of an island provided a continuous protective cover. sedimentation nucleated from a coastal tidal flat built on a peat de- An alternative explanation to account for the patchy distribution posit or directly from a mangrove swamp. Most islands developed of peat is that its original distribution was uneven, perhaps from mud banks, probably by mangrove colonization. confined to areas of temporary stabilization of the shoreline. The Mangrove colonization is possible on most Florida Bay mud island distribution could then reflect shoreline positions, but since banks at present (Fig. 1), but it is a slow, hazardous process. Col- many of the islands were preceded by mud banks, the banks should onization would be favored by slight elevations such as shell ac- be the direct descendants of stabilized shoreline positions. It is cumulations of the windward lag heaped up or strewn across a difficult, however, to interpret present locations of mud banks as bank top by the winter storm winds. Sediment piled up by burrow- former shoreline positions, because most banks trend oblique to the ing organisms might also be effective. Once initiated, trapping of slope of the rock floor that would initially determine the position of sediment and debris by the mangrove colonies can build a bank top the shoreline. Moreover, one must still postulate bank migration to into the supratidal zone. account for the general absence of peat beneath the banks. The characteristic open interior areas of algal mats, scrub man- In summary, islands may develop during the initial flooding from groves, or grass may develop in at least two ways. Mangrove de- coastal tidal flats or mangrove swamps. Most islands nucleated velopment is invariably more extensive at the island fringes, where later in the history of the bay on mud banks. Mangrove coloniza-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 78 ENOS AND PERKINS

tion is considered to be the most important process in inititating of the bay, reflected in the contrasting patterns of banks. He drew island formation, but purely mechanical accumulation of sediment an analogy between the broad shoals at the western edge of the bay can also generate islands. and bay-mouth bars. While the analogy may be geomorphically in- triguing, it doesn't suggest mechanisms applicable to the formation ORIGIN OF MUD BANKS of the mud banks. The sediment of typical bay-mouth bars is contributed largely by longshore drift. The west coast of Florida has a The origin of mud banks has not been adequately explained and southward longshore drift pattern, but the sediment in transit there would be outside the scope of this report except for the trouble- is almost entirely quartz sand, which is virtually lacking in Florida some fact that mud banks form the substrate for many islands. As Bay (Ginsburg, 1956). Most of the quartz is arrested in shoals and previously noted, mud banks are grass-covered shoals submerged bays north of the shell beaches of Cape Sable. Gorsline (1963, less than 60 cm. Banks in eastern Florida Bay are narrow, sub- p. 139) suggested that the narrow banks extending into the bay linear, and generally asymmetric; western banks are much broader from the rocky Florida Keys may originate where tidal jets from ad- and less regular in outline. Approximately 50 cores from mud jacent channels through the keys meet at a null point. Sediment de- banks reveal totally marine sequences except for local swamp deposited in such a favored location would scarcely constitute a bank, posits immediately overlying the Pleistocene rock. The coarse sedi- but it could support grass and green algae, which promote bank ment fraction is biogenic and shows little evidence of transport development by trapping sediment and by greatly increasing sedi- (Ginsburg, 1956, 1957). Island stratigraphy shows that banks have ment production. existed throughout most of the present depositional cycle (see Pres- Regarding the more difficult question of the origin of interior ent Interpretation) and have apparently kept pace with sea-level banks, Gorsline simply suggested as processes that might be factor rise. Mud banks are remarkably resistant to erosion even during (1) the growth and spread of mangroves, marshes, and grass flats, violent storms (Ball and others, 1967; Perkins and Enos, 1968). (2) diminution of tidal currents, and (3) freshwater inflow. He Carpets of turtle grass contribute significantly to their stability demonstrated a counterclockwise gyral circulation pattern in sev- (Ginsburg and Lowenstam, 1958). eral "lake" basins which he suggested would tend to move the Nucleation of mud banks cannot be attributed to pre-existing re- abundant sediment in suspension toward the periphery of each lief. The Pleistocene rock surface upon which they were deposited is basin — that is, onto the mud banks. The circulation pattern is without significant relief, as is the basal Holocene shelly veneer. presumably a result, not a cause, of the establishment of the banks. Red algae and finger corals, which may nucleate banks in the Once banks are formed, waves generated in the basins, where depth Florida shelf margin (Enos, 1977), have not been found in cores of and fetch are at a maximum, appear to be adequate to winnow Florida Bay interior banks, nor would they be expected in this muddy sediment from the lake floors and to deposit it on banks semirestricted environment. The possibility of significant mud- where the waves are damped by the shoal and currents are baffled bank migration is a matter of conjecture. The "buttress" appear- by grass and algae. ance of some islands, the preferential preservation of swamp de- Price (1967) presented an extended discussion of the origin of the posits beneath stable islands, and the asymmetry of banks all tend submarine morphology of Florida Bay proceeding from the as- to favor migration. An attempt to document migration by detailed sumption that the stratigraphic sequence in the bay "is doubtless age dating was unsuccessful (Fig. 16). the same in general terms" (1967, p. 375) as in idealized sections One of the first to attempt an explanation of the unusual pattern from the south Florida mainland north of Cape Sable presented by of sediment distribution in Florida Bay was Gorsline (1963), who Scholl and Stuiver (1967, Fig. 3). Those sections show a transgres- suggested that different forces may be dominant in different parts sive sequence with marine deposits overlying mangrove peat and

Figure 16. Radiocarbon ages (years before present) of coarse shell fractions (>1 NW SE mm) from mud bank west adjacent to Bot- F2.4 F2.I F2.2 F2.3 SL tle Key (see Fig. 5 for location). Dates from :V~ n" • various size fractions of a surface sample Co

southeast-inclined "foreset" effect. Data Co < 140 could be interpreted to reflect (1) extensive f6IO± 150 fCo<260 f830± 140 homogenization of sediment by mechanical \8I0± 150* \400± 150* reworking (possibly migration) or by or- -690±I50 - - \j080 ±140i * ganic reworking or (2) very recent growth of bank. Contact of island on bank sedi- 1 1 ment at 1 m (3.3 ft) subsea on north end of I ' I ' I ' I I I ' I ' T^^ * = FINE SHELLS, 62-IOOO^i.m Bottle Key (Fig. 11) suggests that accretion 100 25 Co = CONTEMPORARY of this bank reached sea level at a position 1.2 m (4 ft) above rock floor by about 3,000 FEET METERS yr ago (Fig. 2). 0.5

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 79 freshwater lime mud. According to the hypothesis of Price (1967), on actual core profiles. These cores show that the entire sequence is Florida Bay developed from flooding of a coastal mangrove swamp, peat except for marine mud in the immediate vicinity of Cape enlarging the existing freshwater lakes by erosion and preserving Sable. the sediment banks or "ridges," which would necessarily be com- Longitudinal growth of "spits" appears to be effective in composed of mangrove peat. North-trending banks would represent partmentalizing the bay into smaller lake basins, if not in originat- "topographic inversion" of mangrove-bordered streams draining ing mud banks, although the mechanisms of spit growth are un- the Everglades. clear (R. N. Ginsburg and E. A. Shinn, 1964, personal commun.). Cores of numerous mud banks (Figs. 4, B, and 5), showing that There are no less than 26 spits or dead-end projections from banks the banks are composed of marine sediment, make this hypothesis into "lake" basins (Figs. 3, 6, A, and 17). On the basis of morphol- untenable. Moreover, the transgressive sequence on the south shore ogy, the growth of spits appears to have been responsible for the of the mainland, used as a point of departure by Price, is found only subdivision of some basins (Whipray Basin and unnamed basin north of Cape Sable in the Ten Thousand Islands area, where northwest of Russell Key, Fig. 1). quartz sand and molluscan banks have produced a thin marine se- The question of variations in bank width may hold clues to the quence overlying peat along a highly embayed coastline (Scholl, genesis of mud banks. Do the broad banks of western Florida Bay 1964, Figs. 3, 7; Shier, 1969). A much more accurate picture of the and the narrow, linear banks of eastern Florida Bay reflect funda- Holocene stratigraphy of south Florida adjacent to Florida Bay is mentally different processes of origin, or have the same growth presented in the cross sections of Spackman and others (1964 , processes simply been more effective in the western part of the bay? Figs. 13, 25, 34, and 58) and Scholl (1964, Fig. 5), which are based The apparent gradational increase in bank width from the eastern part to the western part suggests that the growth processes may be similar but the rates drastically different. The fundamental differ- ence is probably in the rate of skeletal production. Ginsburg (1965, personal commun.) termed the eastern part a "starved area" to indicate low biotic productivity resulting from shallow depths and isolation from open-marine water. Consequently the mud banks are narrow, and the large "lake" basins, which could contribute winnowed sediment to the banks, generally have not reached the critical sediment thickness required to support a grass carpet which could result in a dramatic increase in biogenic sediment production. In the western part of the bay, both basins and banks are carpeted by grass, have a relatively large and varied biota, and are apparently very productive. Carbon-14 dates from cores of a broad mud bank in western Florida Bay suggest appreciable rates of both vertical and lateral accretion (Fig. 18). Lateral accretion has occurred on the open, gulf side of the bank and at a slightly lower rate on the bay side. In situ production on the bank doubtless accounts for part of the accreted material. Material accreted on the side of the bank toward the Gulf of Mexico, if it exceeds in-place production, is probably derived from the broad open shelf. Sediment accreted to the bay side may be derived from the adjacent "lake" basin. It is apparent that the origin of mud banks cannot yet be pre- cisely formulated, but the facts and most reasonable inferences at least limit the range of possible origins.

CONTRASTS IN ISLAND DENSITY

Islands are much more numerous in the eastern and, particularly, the central parts of Florida Bay than in the western part. This appears anomalous because the greater areal extent of mud banks in the western part should offer higher potential for island development. Although this question could not be resolved in this study, several possibilities are suggested. Bank edges, rather than bank tops, appear to be the favored po- Figure 17. Aerial views of Florida Bay. A. High oblique looking sitions for island nucleation. Heaping up of sediment is favored at northwest across Russell Key (Fig. 1). Prominent "spits" are visible the bank edge, particularly the windward edge, because energy is in right foreground and left background. B. High oblique looking focused in response to abrupt shoaling. Drifting mangrove seed- south along emergent bank separating Blackwater Sound (left) lings, the pioneers of most islands, will also tend to take root on the from Florida Bay. Sedimentary processes in lighter colored bank windward edges of banks, particularly if a shoal of sediment is areas are comparable to those in open interior tidal flats of Florida present. Because of the greater width of the banks in the western Bay islands. Mangrove forests are more continuous on other emer- part of the bay, the length of bank edge per unit area of bank is less gent banks. than in the eastern part. Some islands in the western part, notably

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 80 ENOS AND PERKINS

Dildo Key, are located in the interior of broad banks. Whether ment (Davis, 1940). Mangroves flourish on existing islands, how- these islands nucleated in the interior of a bank or whether lateral ever, and throughout south Florida some of the most luxuriant accretion simply moved the bank edge away is unknown. Bank stands are found along waterways in which sea water flows, at least edges in the western part are in general less pronounced than in the periodically. The answer may simply be that mangrove seedlings eastern part, and the slopes of the flanks are flatter. This means that are more abundant in the eastern area. Mangroves line the main- sediment accumulations and mangrove seedlings would tend to be land coast, the continuous sediment ridges that close off the eastern dispersed over a broader area. Sediment accumulations would be area (Fig. 17, B), and the bay side of the Florida Keys, particularly less likely to become emergent and mangrove clumps less likely to where sediment forms the shoreline, as on the tidal deltas. The attain the critical density necessary for survival and island nuclea- numerous islands also contribute seedlings. By contrast, the mar- tion. gins of western Florida Bay are open to the Gulf of Mexico, the It is possible that the banks of western Florida Bay are younger more widely spaced middle Florida Keys, and the supratidal car- than those of the eastern part, that is, that they approached sea bonate flat west of Flamingo, which lacks large stands of man- level more recently because of the greater depth to Pleistocene bed- groves. Add to this the great distance between the north and south rock. Mangrove colonization and other island-forming processes margins of the western part of the bay and the low island density, would simply have had less time to develop. However, this and it would appear that the availability of mangrove seedlings hypothesis is difficult to reconcile with the much greater mud-bank would be considerably reduced. widths and the relative maturity of the islands in northwestern Florida Bay. It is also contrary to the thesis advanced above that ISLAND STRATIGRAPHIC CYCLE banks have existed throughout Florida Bay history. It could also be argued that the slightly more normal marine conditions of the The typical sedimentary record of a Florida Bay island is a cycle western part of its bay are less to the liking of mangroves, which consisting of a transgressive sequence overlain by a regressive se- reportedly favor salinities of 10%o to 20°/oo for optimum develop- quence (Fig. 19). Both parts of the cycle were deposited during the

Figure 18. Radiocarbon ages (years before present) of bulk sediment samples from cores of Ninemile Bank, western Florida Bay (unpub. data of S. D. Kerr). Patterns at 20 cm (8 in.) below sediment surface (above) and 80 to 90 cm (3 ft) both suggest lateral accretion. Central bank marked "oldest" is bounded by lineations of unknown origin visible on vertical air photos. Superposition of the two levels also indicates vertical accretion at a rate well in excess of rate of sea-level rise, but scatter is too great for definite conclusions.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 81

continuous Holocene rise in sea level. The cycle is asymmetric in overlie bank sediments, but there is no record of subtidal sedimen- the sense that the two parts are not mirror images and that the re- tation being re-established over island sediments.4 Calusa Key (Fig. gressive sequence is generally much thicker. An idealized sequence, 12) and Crane Key (Fig. 13) each spread over the adjacent mud in ascending order, is (1) light-colored calcitic mud, (2) mangrove banks from a small nucleus. Bottle Key (Fig. 11) and perhaps Rab- peat, (3) gray, very shelly lime mud (wackestone and packstone), bit Key (Fig. 14) shows evidence of lateral migration. Island area (4) gray, sparsely shelly lime mud (wackestone), and (5) tan or could conceivably be reduced during lateral migration, but the gen- cream, laminated or crumbly lime mud (wackestone and mudstone) eral trend is clearly toward more and larger islands. with peat lenses. Transgressive and regressive sequences within an island cycle The basal calcitic mud, deposited in freshwater ponds, discon- may develop quite independently of the position of the mainland formably overlies Pleistocene marine limestone. Mangrove peat re- shoreline. The shallowing or regressive sequence reflects bank and cords the encroachment of salt water and the development of man- island sedimentation rates in excess of the rate of sea-level rise. The grove swamps. Maximum transgression produced shelly lime mud mainland shoreline may continue to migrate inland while islands that typically contains a slightly more open-marine biota than in are being established. The history of the south shore of the Florida the present "lake" basins. Muddier bank sediment, with a more mainland and the projected future development of Florida Bay in- restricted biota, indicates regressive sedimentation, although sea dicate that continuation of the present patterns of sedimentation level continued to rise. Island sedimentation began when vertical ultimately will result in progradation, pushing the shoreline south accretion reached the intertidal zone. Mature islands accreted into and west from its present position. the supratidal zone. The sequence of depositional environments is thus (1) freshwater pond, (2) coastal mangrove swamp, (3) marine "lakes," (4) mud bank, (5) island. The cycle may be incomplete at 4 some sites, with any or several of members 1 through 4 absent, but This conclusion is based on about 50 bank cores in addition to the is- their order of appearance is always the same. land cores reported here. An apparent minor exception is in Core F1.6 on Calusa Key (Figs. 10, 11). At the north end of Bottle Key, island sediment is Islands may develop at any stage of bay history, including the being eroded below sea level (Fig. 11), but bank sediments are not being initial inundation of the shoreline. Island sediments commonly deposited.

HS Figure 19. Island stratigraphie cycle. Environmental interpretations are indicated. Core D127.4, Crane Key (Fig. 13; for log, see Figure 10).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 82 ENOS AND PERKINS

FUTURE EVOLUTION OF FLORIDA BAY have flooded the Everglades. Because a high stand is now considered unlikely (Scholl and others, 1969), such widespread destruc- Island stratigraphy and analogy with modern environments bor- tion of the evidence need not be postulated. However, progradation dering Florida Bay provide an insight into present evolutionary of sediment has shifted the shoreline seaward more than 8 km (5 trends and the future development of the bay. The problem of the mi) in the Flamingo—Cape Sable area during the continuous origin of mud banks is far from resolved, but given the presence of Holocene rise in sea level. mud banks the trend is clearly toward development of islands, The stratigraphic record of the filling of Florida Bay might be as primarily through colonization by mangroves. Colonization of the little as 4 or 5 m (12 to 15 ft) in maximum thickness (the present narrow, sublinear banks of the eastern part of the bay and coales- maximum thickness is about 4 m; Fig. 3) if filling occurred rapidly cence of adjacent islands would lead to continuous mangrove- and the rate of sea-level rise remained at less than 30 cm/1,000 yr. covered ridges like those that now enclose the "sounds" of north- Maximum thickness would be found in southwestern Florida Bay, eastern Florida Bay (Figs. 3, and 17, B). These intervening basins where the fill would be almost entirely limestone, with insignificant would become more isolated and "starved" as skeletal productivity coal lenses near the base and, possibly, substantial ones near the declines. Organic sediment would form an increasingly large pro- top. The limestone would have a thin (few decimetres or less), portion of the basin fill. coarse molluscan packstone ("lake") at the base grading up into a The broad banks of the western bay may, with continued vertical molluscan wackestone (bank). Supratidal mudstone or wackestone accretion, develop into a continuous marsh very similar to the pres- might occur as discontinuous lenses restricted to the top of the se- ent southwestern Florida mainland between Cape Sable and quence or locally extending through the entire thickness (islands). If Whitewater Bay (Fig. 3). Supratidal carbonate flats could develop broad supratidal flats were established, a continuous upper on the areas that receive sediment from the adjacent shelf during mudstone unit, with local thickening to include the entire bed, storm tides. Shelly beach ridges like those of Cape Sable would would result. likely develop along the exposed western border, as can be ob- The entire bed of "Holocene limestone" would thin abruptly or served at Sandy Key (Fig. 15, A). Mangrove swamps would develop disappear west of Florida Bay and thin gradually eastward and in the more isolated areas. The "lake" basins would become ponds northeastward. The upper, supratidal unit would be replaced east- alternating between fresh and salt water, depending on the relative ward by isolated lenses. Coal and carbonaceous limestone lenses,5 incidence of rain and of flooding by storm tides. perhaps associated with freshwater limestone, would become more At this last stage, Florida Bay would look very much like the abundant overlying and laterally replacing the limestone to the present-day adjacent Florida mainland. The broad area of thick northeast. The thin wedge of "Holocene limestone," recording a sediment between Cape Sable and Whitewater Bay (Fig. 3) would transgressive and a regressive sequence, would be time equivalent have a counterpart in the banks of western Florida Bay. The isola- to the upper part of the seaward-thickening, transgressive wedge of tion of eastern Florida Bay would be almost complete at this stage, limestone, as much as 15 m (50 ft) thick being deposited on the roughly comparable to Whitewater Bay at present. The water shelf margin seaward of the keys (Enos, 1977). Time-equivalent might seasonally vary from saline to fresh, as the balance between limestone with thickness and facies more comparable to the Florida mainland runoff, marine flooding, and evaporation shifted. Marine Bay fill would be the tidal-flat carbonates of northwest Andros Is- recharge would occur through any remaining tidal passes, such as land, also a transgressive sequence with local regression (Shinn and Travernier Creek, and perhaps by flow through the extremely por- others, 1969). The detailed record of transgressive and regressive ous Key Largo limestone where it is not blanketed by muddy sedi- sedimentation, even in a tectonically stable area with steadily rising ment. Runoff into this area would be much less than into Whitewa- sea level, can be very complex. ter Bay because the drainage area is smaller. Sediments would include much plant material from the mangroves with marine and ACKNOWLEDGMENTS freshwater shell accumulations in the former basins. Thus, the entire area of Florida Bay, as defined here, could be accreted to the The impetus for systematic study of the stratigraphy of Florida Florida mainland as an extension of the salt-water fringe of the Bay islands appears to have come from detailed coring of Bottle Everglades. This could occur durung a stationary or continually ris- Key by R. M. Lloyd, R. L. Walpole, and N. D. Jeffries in 1964. ing sea level. These cores produced a much different story than had been de- The southernmost part of the Florida mainland from Cape Sable duced earlier from a few cores on Crane Key and Cluett Key by at least to Flamingo (Fig. 1) has a history like that proposed for the R. N. Ginsburg, E. A. Shinn, K. W. Stockman, and M. M. Ball. future evolution of Florida Bay, as shown by our cores and by those These discrepancies led to the more detailed coring and analysis re- of Scholl (1964), Spackman and others (1964, Fig. 34), P. T. Lucas ported here, between 1964 and 1968. (unpub. data), and F. C. Craighead (1964; 1965, personal com- The considerable background "lore" essential for reasonable in- mun.). Marine carbonate mud deposited in a restricted environ- terpretations of the historical geologic significance of the cores in- ment similar to that of Florida Bay is overlain by supratidal lime cludes contributions from most of the geologists involved in Shell mud or mangrove peat. Craighead (1964) clearly recognized that Development Company's Holocene carbonate project between Florida Bay could be accreted to the mainland and suggested that 1956 and 1968. R. N. Ginsburg and, subsequently, R. J. Dunham most of the present area of the Everglades had evolved in this fash- were directors of the project; E. A. Shinn, R. M. Lloyd, M. M. Ball, ion. He hypothesized that solution of carbonate sediment beneath John MacCallum, and K. W. Stockman were long-time partici- mangrove swamps destroyed the record of marine sedimentation beneath the Everglades. This suggestion was based on the postu- 5 The peat layer, which has about 90% initial porosity (Table 1) and lated higher-than-present stand of sea level in south Florida during would be subject to loss of volatiles, would be reduced to one-tenth its orig- Holocene time (Parker, and others, 1955, p. 124), which would inal thickness in the stratigraphic record.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021 EVOLUTION OF FLORIDA BAY 83

pants. Studies that contributed background data, individually ac- Parker, G. G., Ferguson, G. E., and Love, S. K., 1955, Water resources of knowledged in the text where possible, include those of Ernest southeastern Florida: U.S. Geological Survey Water Supply Paper Brownson, T. D. Cook, Philip Fickman, L. D. Hamburg, George 1255, 964 p. Perkins, R. D., 1977, Pleistocene depositional framework of south Florida, Herman, S. D. Kerr, P. T. Lucas, Roger Middendorf, James Rod- in Enos, Paul, and Perkins, R. D., Quaternary sedimentation in south gers, P. R. Rose, and W. J. Turney. Florida: Geological Society of America Memoir 147, p. 131-198. This report was drafted in 1971 after both participants had Perkins, R. D., and Enos, Paul, 1968, Hurricane Betsy in the Florida- moved to other positions. We thank Shell Development Company, Bahama area — Geologic effects and comparison with Hurricane and in particular P. T. Lucas, manager, Geology Department, for Donna: Journal of Geology, v. 76, p. 710-717. Pray, L. C., 1966, Hurricane Betsy (1965) and nearshore carbonate sedi- releasing this report for publication. ments of the Florida Keys: Geological Society of America Special Paper 101, Abstracts for 1966, p. 168-169. REFERENCES CITED Price, W. A., 1967, Development of the basin-in-basin honeycomb of Florida Bay and the northeastern Cuba lagoon: Gulf Coast Associa- Ball, M. M., 1967, Carbonate sand bodies of Florida and the Bahamas: tion of Geological Societies Transactions, v. 17, p. 368-399. Journal of Sedimentary Petrology, v. 37, p. 556—591. Scholl, D. W., 1963, Sedimentation in modern coastal swamps, southwest- Ball, M. M., Shinn, E. A., and Stockman, K. W., 1967, The geologic effects ern Florida: American Association of Petroleum Geologists Bulletin, of Hurricane Donna in South Florida: Journal of Geology, v. 75, v. 47, p. 1581-1603. p. 583-597. 1964, Recent sedimentary record in mangrove swamps and rise in sea Craighead, F. C., 1964, Land, mangroves, and hurricanes: Fairchild Tropi- level over the southwestern Florida coast: Part I: Marine Geology, cal Garden Bulletin, v. 19, p. 5-32. v. 1, p. 344-366. Davis, J. H., Jr., 1940, The ecology and geologic role of mangroves in 1966, Florida Bay: A modern site of limestone formation, in Fair- Florida: Carnegie Institution of Washington Publication 517, Papers bridge, R. W., ed., Encyclopedia of earth sciences: New York, Tortugas Laboratory, v. 32, no. 16, p. 305-412. McGraw-Hill Book Co., p. 282-288. Dunham, R. J., 1962, Classification of carbonate rocks according to depo- Scholl, D. W., and Stuiver, Minze, 1967, Recent submergence of southern sitional texture, in Ham, W. E., ed., Classification of carbonate rocks Florida: A comparison with adjacent coasts and other eustatic data: — A symposium: American Association of Petroleum Geologists Geological Society of America Bulletin, v. 78, p. 437-454. Memoir 1, p. 108-121. Scholl, D. W., Craighead, F. C., Sr., and Stuiver, Minze, 1969, Florida sub- Enos, Paul, 1977, Holocene sediment accumulations of the south Florida mergence curve revised: Its relation to coastal sedimentation rates: shelf margin, in Enos, Paul, and Perkins, R. D., Quaternary sedimen- Science, v. 163, p. 562—564. tation in south Florida: Geological Society of America Memoir 147, Shier, D. E., 1969, Vermetid reefs and coastal development in the Ten p. 1-130. Thousand Islands, southwest Florida: Geological Society of America Fleece, J. B., 1962, The carbonate geochemistry and sedimentology of the Bulletin, v. 80, p. 485-508. keys of Florida Bay, Florida [M.S. thesis]: Tallahassee, Florida State Shinn, E. A., 1968a, Practical significance of birdseye structures in carbon- University, 112 p. (also available from Armed Services Technical In- ate rocks: Journal of Sedimentary Petrology, v. 38, p. 215-223. formation Agency, Arlington, Va., AD 286 266). 1968b, Burrowing in recent lime sediments of Florida and the Ginsburg, R. N., 1956, Environmental relationships of grain size and con- Bahamas: Journal of Paleontology, v. 42, p. 879-894. stituent particles in some south Florida carbonate sediments: Ameri- Shinn, E. A., Lloyd, R. M., and Ginsburg, R. N., 1969, Anatomy of a mod- can Association of Petroleum Geologists Bulletin, v. 40, p. 2384- ern carbonate tidal-flat, Andros Island, Bahamas: Journal of Sedimen- 2427. tary Petrology, v. 39, p. 1202-1228. 1957, Early diagenesis and lithification of shallow water carbonate sed- Spackman, William, Scholl, D. W., and Taft, W. H., 1964, Environments of iments in south Florida, in Le Blanc, R. J., and Breeding, J. G., eds., coal formation in south Florida: Geological Society of America Regional aspects of carbonate deposition: Society of Economic Guidebook, Field Trip 5, Miami Beach, Florida, 67 p. Paleontologists and Mineralogists Special Publication 5, p. 80-100. Stockman, K. W., Ginsburg, R. N., and Shinn, E. A., 1967, The production compiler, 1964, South Florida carbonate sediments: Geological Soci- of lime mud by algae in south Florida: Journal of Sedimentary Petrol- ety of America Guidebook, Field Trip 1, Miami Beach, Florida, 72 p. ogy, v. 37, p. 633-648. Ginsburg, R. N., and Lowenstam, H. A., 1958, The influence of marine bot- Taft, W. H., and Harbaugh, J. W., 1964, Modern carbonate sediments of tom communities on the depositional environment of sediments: southern Florida, Bahamas, and Espiritu Santo Island, Baja California: Journal of Geology, v. 66, p. 310-318. A comparison of their mineralogy and chemistry: Stanford University Gorsline, D. S., 1963, Environments of carbonate deposition, Florida Bay Publications in Geological Science, v. 8, no. 2, 133 p. and the Florida Straits, in Bass, R. O., ed., Shelf carbonates of the Turney, W. J., and Perkins, B. F., 1972, Molluscan distribution in Florida Paradox basin: Four Corners Geological Society Symposium, 4th Field Bay: Sedimenta III, Rosenthiel School of Marine and Atmospheric Sci- Conference, p. 130-143. ence, University of Miami, Florida, 37 p. Lloyd, R. M., 1964, Variations in the oxygen and carbon isotope ratios of Vaughan, T. W., 1935, Current measurements along the Florida coral reef Florida Bay mollusks and their environmental significance: Journal of tract: Carnegie Institution of Washington, Papers Tortugas Labora- Geology, v. 72, p. 84-111. tory, v. 24, p. 129-141. McCallum, John, and Stockman, K. W., 1961, Salinity of Florida Bay MANUSCRIPT RECEIVED BY THE SOCIETY JUNE 15, 1976 [abs.]: Society of Economic Paleontologists and Mineralogists Re- REVISED MANUSCRIPT RECEIVED MAY 18, 1977 search Committee Symposium, Denver 1961, Program, p. 78. MANUSCRIPT ACCEPTED JUNE 22, 1977

Printed in U.S.A.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/1/59/3434258/i0016-7606-90-1-59.pdf by guest on 28 September 2021