A Plio‐Pleistocene eustatic and storm‐controlled mixed carbonate–siliciclastic marine ramp deposit in south‐west Florida: An example of sediment homogenisation with maintenance of carbonate‐producing organisms

Mixed siliciclastic and carbonate sediments are common in the stratigraphic record, but fully homogenised mixes are not. Many occurrences of mixed sediment sequences are dominated by end‐members with stacking of ‘nearly pure’ lithfacies (e.g. cyclothems containing alternating sandstone, limestone and coal units). The Plio‐Pleistocene sediments within south‐west Florida provide insights into the occurrence of fully homogenised siliciclastic/carbonate deposits. In all defined environments from lagoon to supratidal to inner tidal to beach to offshore to coral reef, quartz sand coexists with carbonates. Perhaps the key feature that allowed full homogenisation of the sediments within all facies and subfacies was the relatively shallow water (<10 m), which facilitated mixing during low‐order eustatic sea‐level events and storms. However, four factors contributed to the full homogenisation of the sediment types without termination or inhibition of carbonate organism growth. These factors are (1) the shallow water allowing wave‐driven sediment transport (all environments within the wave orbital depth), (2) close proximity and perhaps irregular nature of the depositional environment boundaries, (3) low influx rate of quartz sand via longshore transport, and (4) the lack of significant terrigenous mud transport into the system. Mixing processes at the large‐scale included movement of sediments from one depositional environment to another during storms, mixing along facies boundaries, and in situ mixing within autochthonous and parautochthonous mollusc death assemblages. At the smaller scale, mixing occurred by bioturbation and diagenetic dissolution of carbonate skeletal grains during minor high sea‐level stands.


| INTRODUCTION
The entry of high volumes of siliciclastic sediments into carbonate depositional environments can inhibit, terminate or reduce the growth rate of various organisms (e.g.coralgal communities) that produce skeletal grains (Schwartz et al., 2018), or cause a complete succession of facies from carbonate to siliciclastic (Heldt et al., 2010;Sanchez et al., 2012).Survival of the carbonate sedimentproducing organisms is dependent in part on the rate of siliciclastic sediment input, water clarity and the tolerance of various carbonate organisms to mud input (particularly terrigenous mud) (Lohrer et al., 2006;Peterson, 1985;Schwartz et al., 2018;Snelgrove & Butman, 1994).Despite the effect of siliciclastic sediment influx on carbonate production, mixed carbonate-siliciclastic environments are relatively common in the geological record (Budd & Harris, 1990;Doyle & Roberts, 1988;Mount, 1984).In fact, some coral reefs occur within siliciclastic environments, such as in Belize and along the Great Barrier Reef (Choi & Ginsburg, 1982;Dunbar et al., 2000;Flood & Orma, 1988;Purdy & Gischler, 2003).Mount (1984) proposed four fundamental types of mixed siliciclastic-carbonate sediment deposits.He asserted that these deposits are caused by: (1) transfer of sediments from one depositional environment into another during storm events or other extreme events, such as tsunamis; (2) mixing along facies boundaries where sediments of nearly pure composition occur adjacent to each other; (3) in situ mixing wherein carbonate grains are derived from autochthonous or parautochthonous living and death assemblages and accumulate within the siliciclastic sediment (some storm deposits and winnowed shell); and (4) source mixing in locations characterised by local uplift and erosion of nearby carbonate terranes with deposition into a siliciclastic environment.Most of the examples used by Mount (1984) are associated with rimmed platform geometries.However, there are a number of other scenarios that produce siliciclastic-carbonate mixes, including short-distance fluvial transport into a carbonate environment (Coffey & Read, 2007;Coffey & Sunde, 2014;D'Agostini et al., 2015), long-distance fluvial transport of siliciclastics onto a predominantly carbonate platform (Missimer & Maliva, 2017), longshore or nearshore transport of quartz sand along shorelines of a carbonate platform or ramp (Coffey & Sunde, 2014;Cunningham et al., 2003;Evans et al., 1989;Gischler & Lomando, 2005;Missimer, 2001a;Sanders & Höfling, 2000;Schwartz et al., 2018;Zeller et al., 2015), and inner ramp or shelf mixing of nearshore cross-shelf transported siliciclastics during storms with carbonate infauna (Conklin, 1968;DuBar, 1958DuBar, , 1962;;Huang & Goodell, 1967;Jones et al., 1991;Missimer, 2001b).The resulting characteristics of various carbonate-siliciclastic mixes frequently show vertical cyclic repetition of relatively pure compositions, spatial separation of pure compositions with mixing at the boundaries, or full homogenisation of the siliciclastics and carbonates (Missimer, 2002;Missimer & Ginsburg, 1998;Roberts & Murray, 1988).As documented by Schwartz et al. (2018), carbonate/siliciclastic mixed sediment systems can be dominated by either end-member component that can produce variations in composition.
Investigations of Holocene siliciclastic-carbonate homogenised mixtures have been conducted in the Arabian Gulf (Kukal & Saadallah, 1973), the Red Sea (Friedman, 1982;Roberts & Murray, 1988), in Pacific island lagoons (Gussmann & Smith, 2002), the south-west Florida continental shelf (Holmes, 1988), and the southwest Florida lagoon system (Huang & Goodell, 1967).In ancient rocks and some younger deposits, the metastable carbonate components of mixed systems are commonly removed by diagenetic processes, leaving a larger part of the system as siliciclastic with minimal carbonate sediment preserved along with altered hydraulic properties (Franco et al., 2016).Mixed siliciclastic/carbonate sequences are typically investigated at relatively large scales, which fails to allow analysis at the parasequence scale, where detailed information can be obtained on the impacts of minor sealevel changes or storms on composition (Ferro et al., 1999;Roberts & Murray, 1988;Tcherepanov et al., 2008).
In southern Florida, a number of mixed siliciclastic/ carbonate units have been evaluated to assess the degree of mixing and the geometry of various sequences (Evans et al., 1989;Evans & Hine, 1991;Guertin et al., 2000;Missimer, 1999Missimer, , 2001aMissimer, , 2001bMissimer, , 2002;;Missimer & Ginsburg, 1998).These units range in age from Oligocene to Holocene.In the Plio-Pleistocene units, wherein the carbonate component of sediment is well preserved, past investigations have emphasised the stratigraphy and palaeontology with little or no description of the sedimentology or the siliciclastic component of the sediment other than collection of insoluble residue (Brooks, 1968;Conklin, 1968;Dall, 1890Dall, -1903;;Dall & Harris, 1892;DuBar, 1958DuBar, , 1974;;Jones et al., 1991;Parker & Cooke, 1944;Perkins, 1977).Today, the transition between siliciclastic and pure carbonate sediment lies near Cape Sable, north of Florida Bay on the Florida West Coast and near South Biscayne Bay on the Florida East Coast.Modern sediments north of the transition are mixed to varying degrees but are mostly siliciclastic.
The objective of this paper is to investigate the processes, facies distribution and depositional controls on a homogenised carbonate-siliciclastic ramp deposit in south-west Florida at the outcrop scale to ascertain why the carbonate sediment production was not terminated during influx of siliciclastic sediment and what processes produced homogenisation.The site investigated provides documentation of homogenised siliciclastic-carbonate depositional environments controlled by minor sea-level changes, storms, biological mixing and diagenetic processes at a very small scale compared to other investigations of mixed composition units.Minor sea-level events may influence the nearshore transport of siliciclastic sediments into carbonate environments, because marine siliciclastic sediment homogenisation favours shallow water with storm influence where the water depth is consistently <10-15 m.A depositional model for this type of system could be compared to other siliciclastic-dominated mixed systems that have minimal fluvial input of sediment.

| Site description
The Nelson Road Pit, located in Lee County, Florida, was utilised to mine shell and sand between 1989 and 1991 (Figure 1).It was maintained in a dewatered state during most of its operation, which allowed detailed stratigraphic, sedimentological and palaeontological analysis of the exposed Plio-Pleistocene sediments.The base of the dewatered section lies at the top of the Early Pliocene Peace River Formation, the uppermost formation in the Hawthorn Group of southern Florida (Missimer, 2002;Scott, 1988).This site is particularly significant in that no natural full exposure of this part of the Neogene-Quaternary section occurs at any other location in southern Florida.It contains one member of the Tamiami Formation of Late Pliocene age (Jones et al., 1995;Missimer, 2002), a condensed section of the Caloosahatchee Formation of Early Pleistocene age (DuBar, 1958;Missimer, 2002), and the full section of the Fort Thompson Formation of Late Pleistocene age (Brooks, 1968;Conklin, 1968;Dall, 1890Dall, -1903;;Dall & Harris, 1892;DuBar, 1958DuBar, , 1974;;Missimer, 2002;Parker & Cooke, 1944;Perkins, 1977).The dewatered section ranges in thickness from 12 m to 16 m.Detailed research was conducted on the sediments exposed in this pit with the specific purpose of describing a homogenised siliciclastic-carbonate unit that exhibits minor lithic units that illustrate the impacts of minor sealevel changes on the sediment composition.
Intra-facies variability was found in the units observed within the stratigraphic section that illustrate several types of mixing of siliciclastic and carbonate sediments both horizontally (at quite small scales, e.g. 10 m) and vertically.Also, the stratigraphic units were clearly divided by either exposure horizons or by freshwater limestone units that separate marine sediments.A shallow shelf-like feature was also exposed that contained a wave-cut notch.The site contains a variety of detailed sedimentological associations that have not been documented at other locations in southern Florida or in other studies of mixed carbonate-siliciclastic sediments (e.g. a coralline boundstone unit containing intra-granular carbonate mud and quartz sand).For reference purposes, a summary of the Plio-Pleistocene stratigraphy of southern Florida with the historical background is contained within the Supplemental information.

| Field measurements and sample collection
Field sampling at the Nelson Road Pit was conducted between mid-1989 and the end of 1991.During this entire period the pit was dewatered allowing exposure of the entire Plio-Pleistocene section during most of that time.Periodic wall collapses and slumps occurred that obscured some locations.Some of the sections exposed in the pit were described by Missimer (2001b).Collections of sediment and fossils were archived during the dewatering period.All fossils discussed herein are now in the collections of the Paleontological Research Institution in Ithaca, NY.
The stratigraphic section was measured along the pit walls at locations with significant exposure.Detailed measurements were made at three locations (A, B, C) with the locations shown in Figure 2. A detailed description of the north wall was made.A secondary section was measured at the bottom of the pit to the top of a secondary shelf capped by a laminated crust indicating a subaerial exposure horizon (Figure 2).
Bed collections of sediment and fossils were made at four of the five locations where the sections were described in detail.All beds were sampled at least once at the middle of the bed.Some of the thicker units were sampled at two or three vertical locations, especially where changes in composition were observed.Fossils were mostly handpicked in situ from the beds, especially where larger molluscs were observed protruding from the wall.Bulk samples of at least 0.03 m 3 of sediment with fossils were collected from each unit in the measured sections (where possible).
The site was visited at least 20 times during the period of dewatering.During each visit, fossils were collected from spoil piles and on the ground surrounding the pit.Frequent rainfall occurred during the summer months and aided in washing of the shell material to allow better collection.In addition, fossils were also collected from the walls as new exposure was opened during excavation.

| Grain-size distribution and moment calculations
The grain-size distribution was measured using standard sieves based on the methods described by Folk and Ward (1957) and Tanner and Balsillie (1995).Detailed analyses were conducted using 14 sieve increments at 0.25 phi intervals.Where the sample was a mix of carbonate and siliciclastic material, the mixed sample was sieved first, the carbonate material was removed using hydrochloric acid and the residue was re-sieved when sufficient sediment was available.The residue was carefully washed with freshwater and dried in an oven before the second round of sieving was performed (no fines lost).This was done to protect the sieves from being damaged by any acid residue.
Upon completion of the sieving, the data were analysed using a spreadsheet program developed by Rosas et al. (2014) to calculate the mean grain diameter, dispersion (sorting), skewness and kurtosis of the size distribution based on the Folk and Ward (1957) equations.Standard statistical methods were used to determine the mean values of the measured parameters and the standard deviation.

| Total carbonate percentage determination
The total carbonate percentage was measured by treating the mixed samples after grain-size analyses using 10% hydrochloric acid.The samples were dried and weighed.Then the acid was used to dissolve the carbonate.The residue was washed, dried and weighed.The residue weight was subtracted from the total weight to yield the carbonate percentage.

| Composition of the siliciclastic sediments
During the process of grain-size analyses, the siliciclastic sediments were separated into four fractions when possible (only two or three fractions occurred at some locations).These fractions were granular (above 2 mm in diameter), coarse sand (2-0.5 mm), medium sand (0.5-0.25 mm) and fine sand (0.25-0.0625 mm) following the method of Grantham and Velbel (1988).Standard grain mount thin sections were made for each fraction.The sections were stained with sodium-cobaltinitrate to allow identification of potassium feldspar and potassiumrhodizonate to facilitate the identification of plagioclase grains.Whenever possible, 500 grains were point-counted to determine composition and quartz properties on each section.Most samples contained only sand-sized fractions.

| Measurement of the quartz properties
The quartz properties were point-counted based on the Folk (1968) method and recorded.The quartz properties of extinction, inclusions, single or multi-crystalline grains, the distribution number of crystals in a grain and the roundness of each counted grain were recorded.The roundness assessment was made using the Powers (1953) photographic scale and the six classifications of Folk (1968).The roundness of each sample was classified using the angularity index described by Missimer and Maliva (2017).This index is calculated by giving the six roundness classifications of the Folk (1968) numbers from 1 to 6 (1 = very angular to 6 = well rounded).The class number is multiplied times the number of grains counted in that classification.Then, the values for all of the classes are summed and divided by the total number of grains counted.

| Petrography and composition of the carbonate sediments
Samples of the lithified units in each formation were collected and archived.Thin sections were made of select unlithified and lithified sediments within the stratigraphic section.The unlithified grains were incorporated into a resin before the thin sections were made.The skeletal and non-skeletal grains within each of the thin sections were point-counted to yield an approximate composition of the carbonate component of the sediment or rock.The counts were made using preserved grains and the geometry of pores left from dissolution such as molluscs, which left only large pores.Full point counts of all thin section components were done to assess the macroporosity, the cements, preserved skeletal grains and siliciclastic grains.The point count grid size was 0.5 × 0.5 mm and a minimum of 1000 points were counted.To obtain a representative composition, the point count number was increased to as high as 1600.Percentages of all properties were tabulated into tables and all data are given in a Supplemental Materials file.The rock types were described using the Dunham (1962) classification of carbonate sediments with modifiers for the siliciclastic components.

| Measurement of the strontium isotope and age estimation
Eight strontium isotope analyses were measured to determine 87 Sr/ 86 Sr ratios to estimate the age of the stratigraphic units.Primarily calcitic mollusc shell was collected for analysis (mostly pectinid bivalves).The shell was scraped and cleaned with HCl and broken to expose the inner part of the shell that was not exposed to diagenetic alteration.In most cases, pieces of the inner shell were removed using dental tools.The surface of some shell fragments was drilled to the middle, with the surface material discarded.Only the drilled inner part of the shell was collected and sent for analysis.
The strontium ratios were determined to the sixth decimal place using an MC-ICP-MS unit (Thermo Fisher Netune Plus) using the methods described by Pourmand et al. (2014) and Pourmand and Dauphas (2010).The samples were run by Isobar Science of Dublin, Ireland.These ratios were then compared to the look-up table developed by McArthur (complied by McArthur from the referenced works), supported by data in Farrell et al. (1995), Howarth andMcArthur (1997), andMcArthur et al. (2001) to estimate ages.In addition, the estimated ages were compared to the δ 18 O Plio-Pleistocene record of MIS events compiled by Lisiecki and Raymo (2005) to ascertain if various units could be correlated to specific MIS events, which could further constrain the ages of the sub-units.

| Stratigraphy
The Nelson Road Pit provides a dewatered stratigraphic section ranging from the uppermost part of the Early Pliocene Peace River Formation at the base to a thin section of the Late Pliocene Tamiami Formation to a condensed section of the Early Pleistocene Caloosahatchee Formation to a full section of the Late Pleistocene Fort Thompson Formation.Stratigraphic sections were measured on the east wall (A) and the north wall (B) (Figure 3).The Caloosahatchee Formation was not present in the measured section at points A and B, but was measured along the centre on a 'ledge' feature at location C.
At both locations A and B, the excavation penetrated into the Peace River Formation of the Hawthorn Group.This material is silty and sandy, dark yellowish brown mud (6/3) (Munsell Colour Chart).It contains some 'floating' dolomite rhombs, quartz sand and silt, and some lime mud containing foraminifera.It is disconformably overlain by the Tamiami Formation.The boundary is an erosional unconformity based on the extreme change in lithology, undulations and some minor deposits of phosphate grains along the boundary.
The Tamiami Formation at Section A is <1 m in thickness and contains two quartz sand beds at the base with the lowest containing calcitic shell fragments (Figure 3).There is a soft limestone near the top, which contains rounded clasts that appear to be intraclasts of partially lithified mud.At the top of the limestone there are some laminations, which are 1-3 mm in thickness.The unit may be a calcrete.The exposed Tamiami Formation section at location B is close to 1.5 m in thickness and contains four beds (Figure 3).The basal unit is a marl consisting of lime mud and quartz sand with a few unidentifiable mollusc shell fragments.The marl is overlain by a soft, chalky limestone, which is a wackestone to mudstone containing a few preserved shell fragments.A quartz sand and sandstone bed overlies the limestone.This unit contains soft concretions and some nodular phosphate.It is devoid of shells.The top of the unit is a laminated limestone.The Tamiami Formation units are colour-coded to show stratigraphic equivalence.
The Fort Thompson Formation sits disconformably upon the Tamiami Formation at locations A and B with the Caloosahatchee Formation missing.The Fort Thompson Formation shows significant differences at the two measured sections, which are located approximately 450 m apart.At Section A, there are seven lithofacies within 5.3 m of exposed section, while at Section B there are six units exposed within the 5.7 m Section.Bed H in Section A corresponds to Beds F and G in Section B and colour coding in Figure 3 was used to show equivalent stratigraphic units.Bed H in Section A is a shell and sand unit that contains an increased amount of mud with depth.Bed G in Section B is a sand and shell bed that corresponds to the basal part of unit H in Section A. However, it can be carefully differentiated from Bed F, which is predominantly shell with quartz sand.Units E, F and G in Section A correlate with unit E in Section B. The break in lithology between units H and G in Section A and between F and E in Section B appear to be related to a sea-level event based on a stratigraphic discontinuity, which is probably the same event.These units all contain predominantly well-preserved, aragonitic molluscs (numerous species along with Chione elevata) in a quartz sand matrix.The percentage of carbonate in these units varies from 0% to 100%.The shell and quartz sand units are capped by a thin, sandy limestone unit.This unit is predominantly marine at its base, but sometimes contains freshwater molluscs (see Palaeoecology section) near the top and some laminations at the very top.The top of the unit is a distinctive disconformity.The base of the unit may also be an unconformity based on the abrupt change in lithology.Two sand units (B and C) in Sections A and B, devoid of shells or any fossils overlay the sandy limestone.This sand is considered to be part of the Fort Thompson Formation.An artificial fill unit lies at the top of both measured sections.
A 2.5 m thick 'shelf' occurred in the north-east part of the pit and contained a ledge (Figure 2).This feature occurred in the pit because of the hard, laminated crust at its top, which would have required blasting and crushing to remove it as useful fill.The 'shelf' only occurred in the north-east part of the pit because this was the only occurrence of the Caloosahatchee Formation.The boundary of the 'shelf ledge' contained a wave-cut or bioerosion notch at its base.It is continuous along the entire length of the 'shelf ledge' and has a concave inward geometry.Section C was measured along the edge of this feature (Figure 4).The Caloosahatchee Formation section is missing in Sections A and B. Despite the thinness of the section, it contains numerous distinctive stratigraphic units and appears to exhibit evidence of as many as four sea-level events within a condensed section (see sequence stratigraphy section for more details).The basal unit is made up of quartz sand and shells (C j ).Some of the underlying mud is intercalated into unit C j .A shell and quartz sand unit (C I ) lies immediately above it with the contact appearing to be transitional from C j upwards into C i .Unit C I does not contain any mud and contains a higher percentage of quartz sand.The top of the unit is a disconformity, with the overlying unit being a hard, indurated wackestone containing predominantly mouldic pores (C H ). Unit C H is capped by a laminated crust of mudstone/wackestone (C G ) that contains some moulds and casts of freshwater molluscs (probably Planorbella sp.).The crust marks a disconformity between C G and C F .Another hard, indurated wackestone (C F ) unit overlies the disconformity.This unit has a very similar lithology to the wackestone underlying the disconformity.The top of C F is also a disconformity and has a very uneven surface showing signs of uneven erosion.A coralline boundstone unit lies above unit C F .It contains some lime mud, and quartz sand between the coral heads (coral types are defined in the palaeontology section and shown in Table S1) and ranges from 0.25 m to 0.35 m in thickness.The overlying unit (C D ) is another indurated wackestone containing mostly mouldic pores.It is overlain by a mixed shell and quartz sand unit (C C ), which separates the wackestone into two different units C D and C B .Unit C B is capped with a laminated mudstone (C A ) containing freshwater gastropods (Planorbella sp.).
Within the north-west corner of the pit, corresponding to the approximate location of the underlying Caloosahatchee Formation, a facies change occurs with the lowermost part of the Fort Thompson Formation.Unit H in Section A and units F and G in Section B are lithified and contain a variety of carbonate cement types, instead of being unlithified, sandy marine shell units.This part of the pit was very difficult to map because it was not fully excavated and was prone to collapse of the overlying unlithified quartz sand units.Some samples from this lithified unit were collected for thin section analysis (see Petrography section).

| Sedimentology
4.2.1 | Carbonate and siliciclastic percentages in the unlithified sediments The carbonate mud contribution in the mixed sediments ranges between 0% and 100% (Table S-1).This range was observed in the Caloosahatchee and Fort Thompson formations.Within the Tamiami Formation, the carbonate fraction was observed to be no greater than about 60%, which occurred only in some indurated limestone or sandstone beds.
Measured carbonate percentages within the unlithified part of the Fort Thompson Formation in Section A had a range of 4.5%-46.1% (δ = 24.5%,σ = 14.1).In Section B, the measured carbonate values in the Fort Thompson Formation ranged from 8.2% to 65.6% (δ = 34.4%,σ = 12.5).The Caloosahatchee Formation showed 12.5% carbonate in one sample.Data from the measured samples are presented in Table S-2. 4.2.2 | Siliciclastic sedimentology of unlithified sediments

Grain-size analyses
Samples were collected and analysed from Sections A, B and C to determine the grain-size characteristics of the mixed carbonate and siliciclastic sediments (raw data in Table S-2).A comparison of all grain-size characteristics based on the Folk parameters is given in Table S-3 and with depth in Table S4.
The mean grain diameter of the mixed carbonate and siliciclastic lithofacies (quartz sand plus skeletal grains) is larger than the pure siliciclastic lithofacies.The quartz sand from the Fort Thompson Formation is consistently finegrained in both Sections A and B and shows little variation and no vertical stratigraphic pattern (see Tables S-2, S-3 and S-4).The samples from the Caloosahatchee Formation (Section C) are also fine-grained and are similar, but slightly coarser than the Fort Thompson Formation samples.
Dispersion or sorting of the sediments was also determined.As expected, the pure quartz sand is better sorted compared to the mixed carbonate/siliciclastic mixes.Based on the scale of Folk and Ward (1957) and Folk (1968) the quartz sand component of the Fort Thompson and Caloosahatchee formations are well to moderately sorted with values very close to one (1).There is no significant difference between the sorting of the quartz sands in the two formations.The sorting of the mixed samples is classified as poor.
The skewness of all samples measured is positive.Based on the classification of Folk and Ward (1957), the samples from the Fort Thompson Formation with means ranging from 0.414 to 0.432 are strongly fine-skewed.The samples from the Caloosahatchee Formation are fine-skewed with a mean of 0.116.There is a significant difference between the skewness of the quartz sands in the Fort Thompson and Caloosahatchee formations.The distribution of the grains is unimodal in nearly all samples measured.
Kurtosis values in the mixed carbonate/siliciclastic sediment was significantly lower compared to the pure siliciclastic component.There is a considerable degree of consistency between the samples collected within the Fort Thompson Formation with the means of the two sections being 1.659 and 1.696.The Caloosahatchee Formation samples showed a mean of 1.571.According to the scale of Folk and Ward (1957), the sediments in both formations are leptokurtic.
The percentage of sediment that passes through the 0.0635 mm screen is termed the mud component.The average value in Sections A and B was 1.10% and 0.54% in the mixed sediments and 1.30% and 2.04% in the siliciclastic component (Table S-1).The samples had a large standard deviation.The mud percentage was lower in the Caloosahatchee Formation sediments at means of 0.37% and 0.06% in the mixed and siliciclastic components respectively (Table S-1).

Characteristics of quartz grains
Quartz grains contain some properties that are related to their origin, which includes extinction, inclusions and whether they are monocrystalline or polycrystalline.In certain cases, the occurrence of polycrystalline grains can also relate to their transport distance and mode in that some polycrystalline grains will tend to break down into monocrystalline grains during long-distance transport in either the fluvial or littoral mode.The measured extinction and inclusions are shown in Table S-5 and the polycrystalline/monocrystalline properties based on the aggregated grains counted is shown in Table S-6.The detailed measurements are given in Tables S-7a to S-7d.Note that in the Supplemental Material tables the measurements are shown for all three size-fractions counted.
The extinction data show that the quartz grains are over 95% single crystals with straight extinction with no difference in extinction between the sands in the Fort Thompson and Caloosahatchee formations (Table S

-5).
There is no distinction pattern in the extinction types based on the grain-size fractions measured (Tables S-7a to S-7d).The extinction data suggest that the origin of the quartz is a combination of volcanic and plutonic.
The inclusion data show that in most cases the grains have few vacuoles and no microlites.Again, there is no significant difference in the inclusions within the Fort Thompson and Caloosahatchee formations.There is a pattern in variation of the percentage of grains with grains having few vacuoles and no microlite with grain size.As grain size decreases the percentage of 'clean' grains increases.From the top of the Fort Thompson Formation (Section A, Bed B) to the base of Bed G, the variation in inclusion types varies in the following manner: few vacuoles and no microlite > microlites > abundant vacuoles > rutile needles (Tables S-7a to S-7b).This abundance ratio changes at the top of Section A, Bed H, which is: few vacuoles and no microlites > abundant vacuoles > microlites > rutile needles (Tables S-7a to S-7d).The same ratio occurs throughout the Caloosahatchee Formation.The inclusion data support the conclusion that the primary source of the quartz is volcanic and plutonic.
The vast majority (99.5% or greater) of quartz grains are monocrystalline (Table S-6).The number of grains in the polycrystalline grains is predominantly two with only a few being three or four grains.Detailed data on the monocrystalline/polycrystalline grain characteristics are contained in Tables S-7a to S-7d.
Based on the grain types, the data suggest that the sources of the quartz are volcanic and plutonic and that the transport distance was long with probable reworking.This can be further verified by analysis of the grain composition (mineralogy).

Roundness of quartz grains
Roundness of the quartz grains falls mostly into the subrounded classification of Folk (1968).This classification has a value of 4.0 in the Angularity Index of Missimer and Maliva (2017).The index values for the Fort Thompson Formation ranged from 3.80 to 3.97.The two Caloosahatchee Formation samples had values of 3.65 and 3.66, which indicates they were significantly more angular or less rounded.In all cases the roundness of the quartz grains declined with grain size, which is expected.The roundness data for the full number of counted gains is shown in Table S-8 and the data, broken down into the three grain-size classifications, are shown in Tables S-7a to S-7 d.

Composition of unlithified siliciclastic sediments excluding carbonate components
Point counts of the siliciclastic component demonstrated that it was predominantly quartz in the Fort Thompson Formation; quartz ranged from 97.7% to 100%.A minor percentage of k-feldspar was present (0.1%), which had a few twined grains.Plagioclase occurred as only one grain in one sample.A small percentage of the quartz grains contain large inclusions of other minerals.These minerals included muscovite, biotite, hornblende, staurolite, opaque, k-feldspar and zircon.A very minor presence of accessory minerals was found, including hornblende, rutile, sillimanite, tourmaline, titanite and zircon.Based on the composition of the siliciclastic sediments, it is considered to be very mature.A data summary is presented in Table S-9 with detailed data of individual samples in Tables S-10a to S-10d. 4.2.3 | Petrography and diagenesis of the lithified units (mixed siliciclastics and carbonates)

Rock types
A variety of rock types occur in the stratigraphic section investigated.All of the lithified sediments share a common component in that some percentage of quartz sand occurs in every sample.Using the terminology of Dunham (1962), the rock types are given in Table S-11.

Tamiami Formation
Four stratigraphic units were sampled to assess the petrology of the units.All of the rock types are sandy wackestones, but with some significant differences.Quartz grains are fully homogenised into the rock as grains within the matrix material or within infilled burrows.
In Section A, Bed I, two samples were collected and point-counted (Figure 5; Table S-12).The unusual feature of this rock type is that it contains a matrix containing three cement or matrix types, including a mixed lithology containing carbonate mud and clay, a very-fine-grained pure calcitic mud, and calcitic micro-spar, which has a grain size of 3-5 μm (Figure 5A).The sediment was burrowed and the burrows were filled with one or both of the mud types and quartz sand.An unusual burrow was found that fully penetrated a mollusc shell (Figure 5B).Another example of a burrow running through a mollusc shell is shown in Figure 5A at the location marked 'B'.Most of the skeletal grains are fully dissolved, micritised, or partially altered.The macro-porosity is low in both samples at 0.6% and 1.3%.Mouldic and cavity porosity is predominant.Many of the pores are partially or fully infilled or lined with blocky calcite cement.There are a number of quartz grains that contain an isopachous rim of microcrystalline calcite cement.Some of the rims have been recrystallised into blocky or granular calcite.Some of the cement may have occurred in the marine environment as isopachous aragonite and then was recrystallised in a meteoric environment.Three different beds were sampled from Section B, including beds A T , B T and C T (Figure 5; Table S-12).Bed A T has the least quartz in the formation at 0.9% and contains a granular calcite matrix with the particles being about 10 μm in diameter (Figure 5C).The pores are cavities (solution) and mouldic with many of the mouldic pores filled with blocky calcite cement.The Bed A T sample has the highest macroporosity of any of the Tamiami Formation units.The samples from beds B T and C T have a similar composition with each having a matrix of fine-grained micro-spar (Table S -12).Most of the skeletal grains have been fully dissolved, micritised, or replaced by blocky calcite (Figure 5D).Most of the skeletal and quartz grains contain full or partial rims of isopachous microcrystalline calcite.Some of this cement has been recrystallised into blocky or granular calcite.A small percentage of the skeletal grains maintain the original structure of the organisms with molluscs being predominant.

Composition of the Caloosahatchee Formation lithified samples
A discontinuity crust (or duracrust) lies very close to the top of the Caloosahatchee Formation (Figure 4; Beds A-1 to A-3).Three samples of this unit were point-counted and show similar compositions (Table S-13).Below the crust, the lithology is predominantly a sandy wackestone (Figure 6A).The quartz grain contribution ranged from 24.3% to 26.3%.Four different types of skeletal grains were found in each sample with molluscs being the most common component.All of the skeletal grains have been altered to a degree, either partially or fully micritised, partially or fully dissolved, or fully replaced.A few peloids were found in the unit.
Micritic calcite formed a matrix from the top of the discontinuity downward.This is some combination of cement and mud (Figure 6A).Blocky and granular calcite cement fills most mouldic pores as a replacement for the skeletal grains (Figure 6B).The overall macroporosity varied greatly from 4.9% to 23.7%.There are several different types of pores occurring mostly below the crust (Table S -13).
Two samples (B-1 and B-2) were point-counted from Section B, unit B of the Caloosahatchee Formation (Table S7).Both samples are predominantly sandy wackestones.There is a matrix of microcrystalline calcite, which is a combination of mud and cement (Figure 6C).The unit contains between 26% and 29.6% quartz sand, which is unevenly distributed within the rock.Many of the mouldic and cavity pores contain some microspar cements lining them or are fully filled with blocky calcite (Figure 6D).A few mollusc grains are not altered.
Unit C is an unlithified unit consisting of quartz sand and skeletal grains.A large variety of molluscs occurs in this unit.It does not contain any significant percentage of mud and has a very high percentage of quartz sand.
Unit D is a sandy wackestone that contains skeletal and quartz grains in a matrix of microcrystalline mud.The quartz sand in the two samples described was 5.1% and 18.1%.The distribution of quartz sand is quite variable.Some of the mouldic pores are filled with either blocky or granular calcite (Figure 6E).Macroporosity ranges from 11.9% to 12.4% and occurs as mouldic and cavity (dissolution) pores.
Unit E is a coralline boundstone with a muddy matrix.It contains four to 10 species of corals in the genera Dichocoenia, Diplora and Siderastrea.It is essentially  unlithified and contains some intra-granular lime mud and quartz sand occurring between coral heads.
Bed F is a sandy skeletal packstone with the predominant skeletal grains being molluscs (18.4%).Many of the mouldic and interparticle pores are partially-filled with blocky calcite (Figure 6F).Parts of the rock contain microcrystalline cement and some microcrystalline mud captured in concave upwards mollusc fragments.The quartz content of the rocks is 18%.
Unit G is a crust of sandy wackestone/packstone.Some mud lumps occur in parts of the rock, which contain quartz grains and interparticle and mouldic pores filled with blocky calcite.The mud lumps could be intraclasts composed of calcitic mud (Figure 6G).The mixed occurrence of sandy skeletal packstone with wackestone is caused by bioturbation as shown in Figure 6H.The macroporsity of the rock is 12.6%.
Unit H is a grainstone/packstone containing quartz grains and skeletal grains cemented by microcrystalline calcite.Most grains contain isopachous microcrystalline cement (Figure 6H).Some of the mollusc grains are infilled with quartz grains and other skeletal grains cemented with either microcrystalline or blocky calcite cement (Figure 6I).Most of the mollusc grains are unaltered.Non-skeletal grains, peloids and intraclasts are partially dissolved.The unit has a high macroporosity at 27.2%.
Unit I is a sandy shell bed that contains some friable sandy packstone/wackestone layers (Figure 6J).The larger mollusc grains are not altered, but the small molluscs and some of the other skeletal grains are completely dissolved in the partially lithified units.There is 35.1% mud and micritic cement in the thin layers.The thin layers have a high macroporosity at 25.1%.

Composition of the Fort Thompson Formation lithified samples
Most of the Fort Thompson Formation is unlithified.However, a sandy limestone 'caprock' or duricrust occurs in Section A (Unit D) and in Section B (Unit D) and the lower part of the formation located along the north wall from 200 m east of Section B to the north-east corner of the excavation is lithified.Detailed compositional data are contained in Table S-14.
The 'caprock' or duricrust (Unit D) is present throughout south-west Florida, commonly occurring near the land surface where it is a laminated sandstone at the top and transitions into a marine limestone with depth (Figure 3).Three samples were collected from Section A, Bed D (D-1, D-2, D-3).The uppermost unit is a laminated sandstone cemented tightly by microcrystalline and microspar calcite (Fe-stained) (Figure 7A).With depth, the quartz sand decreases from 41.4% to 22.5% (Table S8).The quantity of blocky calcite that infills pores increases with depth from 9.5% to 40.7%, and some granular calcite cement is present (Figure 7B,C).The diversity of skeletal fauna increases with depth and corals occur in the basal section (Bed D-3).The unit is very hard throughout the region.
The lithified section occurring in unit H L (200 m east of Section B towards the north-east corner of excavation) is separated from the overlying unlithified part of the unit by a laminated limestone crust and from the underlying Caloosahatchee Formation by another laminated crust, which is believed to occur at the top of the underlying formation.The upper part of the unit is a sandy skeletal molluscan packstone (Figure 7D).The unit transitions with depth to a more heterogeneous limestone with greater faunal diversity (Figure 7E).The quartz sand declines from the upper part of the unit at 31.9% to the middle at 9.8% and 3.2% and increases at the base to 40.7% based on the point counts (Table S7 and Figure 7F).The matrix of the unit is predominantly micritic calcite mud and cement (27%) near the top with some blocky calcite infilling pores (17%).In the middle section, the blocky calcite cement is predominant at 53% and 64.6%.Near the base of the unit blocky calcite declines to 25.6% and micritic mud and cement increase to 19.7%.The total macroporosity is 14.9% in the upper part of the unit and declines to between 1.4% and 4.5% in the middle to 75% in the lower part.The lowermost section is quite heterogeneous and the lithology is a mixed wackestone/packstone.
At the base of the Fort Thompson Formation 80 m east of Section B, there is another lithologic unit that occurs immediately above the Tamiami Formation.Only this unit was exposed at this location, so no full section could be described.This unit is a skeletal wackestone that contains very little quartz sand (3.2%).The matrix is micritic calcite mud (71.6%).Mouldic pores in the matrix are infilled with blocky and granular calcite cement with many open mouldic pores.It contains freshwater molluscs (probably Planorbella sp.).The unit is a freshwater limestone that likely occurs in the lowermost part of the Fort Thompson Formation, but cannot be correlated to any specific stratigraphic unit.

Skeletal and non-skeletal grains in the lithified sediments
Most of the lithified samples have undergone extreme diagenesis and most of the skeletal carbonate grains have been removed by dissolution or are unidentifiable based on micritisation or full recrystallisation.Therefore, many of the skeletal grains are point-counted as some type of cement based on recrystallisation or as mouldic pores.Visual scans of thin sections were made to estimate the semi-quantitative populations of the skeletal and nonskeletal grains based on the geology of the pores and the remnant structures of the grains.The detailed data on the scans are contained in Table S-15.Within the Fort Thompson Formation molluscs are the predominant skeletal grain type and these contribute between 68.2% and 99%.The highest percentage occurs in a freshwater limestone unit that occurs at the base of Fort Thompson Formation 80 m east of Section B. Up to five additional skeletal grain types occur in the formation with foraminifera and echinoids having the highest percentages.Three types of non-skeletal grains occur in the formation and include intraclasts, peloids and phosphate nodules.
The predominant skeletal grain type in the Ca lo osahatchee Formation is also mollusc derived, contributing between 70.4% and 95.3%.Up to eight additional skeletal grain types are found in this formation, with foraminifera and echinoids having the highest percentages.There is a bed of coralline boundstone, which is not included in the analysis of the lithified rock.There is slightly more skeletal grain type diversity in the Caloosahatchee Formation compared to the Fort Thompson Formation.Seven different types of non-skeletal grains were found.The predominant types were intraclasts and phosphate (phosphorite) nodules (note that phosphate (phosphorite) nodules are not carbonate).A minor occurrence of ooids and coated grains was also found.
Skeletal grains in the Tamiami Formation are less diverse than the overlying units and are dominated by molluscs ranging from 82.1% to 90.6%.Echinoids and foraminifera are the most abundant secondary grains.Phosphate nodules are the predominant non-skeletal grain type.The upper two calculated ages for the Fort Thompson Formation suggest an age of 1.15 ± 0.3 and 1.2 ± 0.3 Ma.The third suggested age is 0.36 ± 0.8 Ma, which is within the normally accepted range of age for the unit.The range in age in this analysis based on strontium isotopes is 0-1.5 Ma including the error range.The one older suggested age is very much higher than expected.It is likely that the calcite shell and limestone samples were contaminated with older material.

| Strontium isotope ages of units
Previous age estimates for the Fort Thompson Formation by Puri and Vanstrum (1969) estimated it to range from 120 000 to 140 000 yr based on 234 Th/ 238 U measurements.An assessed age range of Fort Thompson Formation deposits at Oldsmar in eastern Pinellas County was 120 000-130 000 yr at Oldsmar 1 and 200 000 yr at Oldsmar (Karrow et al., 1996).
The three strontium isotope ages calculated for the Caloosahatchee Formation are 1.2 ± 0.2, 1.3 ± 0.2 and 1.65 ± 1.3 Ma (Table 1), yielding a range in age from 1.0 to 1.8 Ma, exclusive of the error range.The ratios were in close proximity to stratigraphic order.
A single age for the Tamiami Formation was determined from the lowest sample in Section A. This yields an age of 4.9 ± 2.7 Ma.

| Caloosahatchee Formation
The mollusc fauna of the Caloosahatchee Formation includes at least 495 species of marine bivalves and gastropods (Allmon et al., 1993(Allmon et al., , 1996)), of which only 35 species were identified from the Nelson Road Pit (Table S-16).The Caloosahatchee fauna is not dominated by a single species, although the small bivalve Transenella conradina is abundant in Section C, unit C J .This species is common today in south-west Florida in assemblages influenced by freshwater discharge (Perlmutter, 1982).
At least six species of corals were found in the Caloosahatchee Formation at the Nelson Road Pit.Siderastrea siderea and Solenastrea bournoni inhabit a wide range of reef environments in Florida and/ or throughout the Caribbean today (Foster, 1979;Gilliam, 2012), and so are not particularly indicative of a particular environment of deposition.Solenastrea hyades is a non-hermatypic species occurring today off the south-east coast of Florida in patch reefs and outer reef platforms, in depths of 9-20 m (Goldberg, 1973).The three remaining corals found within the Caloosahatchee Formation (Unit C J ) are Dichocoenia eminens, Dichocoenia caloosahatcheensis and Pseudodiploria/Diploria sp.The first two species of this group are extinct.Dichocoenia tends to occur in 2-72 m of water (Jaap et al., 1989).

| Fort Thompson Formation
The mollusc fauna of the Fort Thompson Formation includes at least 81 species of marine bivalves and one freshwater example, one species of scaphopod, and 146 species of marine, 17 freshwater, and 20 terrestrial species of gastropods (Kittle & Portell, 2010).The Fort Thompson fauna at the Nelson Road Pit includes at least 107 species (Table A-8) within a number of beds (especially Sect.A, Bed H and Section B, Beds E, F) dominated numerically by the bivalve Chione elevata.The small bivalve Macrocallista maculata is also abundant, especially in Section B, Beds E and F. Corals are rare and small in the Fort Thompson Formation at Nelson Road, and consist mostly of S. hyades.Barnacles and bryozoans are most common in Section B, Bed F.
Chione elevata is a shallow-burrowing, suspensionfeeder frequently found today in shallow depths associated with seagrasses (Daley et al., 2007;Moore & Lopez, 1969).Only a very small proportion of the shells of this species show significant bioerosion or encrustation, suggesting rapid burial without extended exposure on the sea floor (cf., Anderson & McBride, 1996).The bivalve M. maculata is today characteristic of the shallow shoreface (Anderson & McBride, 1996).The lowermost unit in the Fort Thompson Formation located about 80 m east of Section B contains freshwater gastropods (Planorbella, Viviparus).
Previous studies of the Fort Thompson Formation have interpreted its depositional environment as a sandy, shallow-water setting with both grass cover and open sandy areas (Daley, 2001;Daley et al., 2007).This fauna clearly indicates very shallow water, probably less than 10-15 m depth.
Fossil molluscs were used in some depositional environment designations, particularly for verification of freshwater limestone.However, many of the mollusc species occur in shallow water across depositional environment boundaries.In addition, a massive number of molluscs are commonly contained in death assemblages and are no longer located in their original water depths and depositional environments, so they have limited use in defining water depths.

| Tamiami Formation
The sequence stratigraphy was assessed for the exposed section of the Tamiami Formation and the other stratigraphic units by carefully defining each lithofacies unit and assigning a depositional environment based on the lithologic characteristics and fossils based on the literature (Table 2).It is important to note that some quartz sand is present in all of the 27 defined lithologic units.
The thickness of the Tamiami Formation at the Nelson Road Pit is small at <1 m in Section A and about 1.5 m in Section B (Figure 3).Two thin parasequences can be defined in Section B, where a chalky soft limestone sits atop a sandy marine marl.The marl appears to be a shallow nearshore deposit or outer lagoonal deposit where some quartz sand entered a primarily carbonate depositional environment.The limestone also appears to be a finegrained muddy unit, also derived from a restricted circulation environment.A modern analogue is the area south of Cape Sable to the north part of Florida Bay (Gebelin, 1977;Sussko & Davis Jr., 1992).The unit is a shoaling-upward sequence, likely deposited during a minor rise in sea level; at its top is a minor discontinuity surface.The overlying unit contains an unlithified quartz sand unit topped with a laminated limestone, which is likely an exposure horizon.It may be a restricted circulation nearshore deposit with an exposed sandy carbonate unit above it, similar to the uppermost unit in the Fort Thompson Formation at this site.This is another shoaling-upward parasequence.There is a single parasequence exposed in Section A (Figure 3), which appears to be similar to the upper parasequence in Section B. An unlithified quartz sand and shell unit is topped by a limestone containing intraclasts.It has a disconformity at the top.The parasequence seems to be shoaling-upward with a low energy beach or a shallow offshore deposit topped by an intertidal/supratidal marine unit.The two parasequences in Section B and the one in Section A are transgressive.
This section occurs near the base of the Tamiami Formation and likely occurs within the Tan Clay and Sand facies (see Background Material on Plio-Pleistocene Stratigraphy of Southern Florida in the Supplemental Materials) as defined by Missimer (1992) and the sequence stratigraphy of the full section of the formation as described by Cunningham et al. (2003).The section is contained within the lower Tamiami Formation as described by Zullo and Harris (1992).They defined it as a transgressive highstand condensed interval.

| Caloosahatchee Formation
The sequence stratigraphy of the Caloosahatchee Formation was again defined by the interpretation of the depositional environments of the measured lithofacies listed in Table S10 and shown in Figure 4. Four parasequences occur within the section, which are: units C j , C I , C H and C G , units C F and C E , units C D and C C , and units C B and C A .Based on the petrographic analysis that includes the rock classifications and the amount of quartz sand, the C J , C I , C H and C G parasequence appears to transition upwards from offshore to nearshore to beach to intertidal and may be regressive (light green colour in Figure 4).The C F and C E parasequence transitions from offshore (middle shelf) to reef (coralline boundstone) (tan colour in Figure 4).The C D to C C parasequence transitions from offshore to beach or shallow subtidal based on the molluscs and the relative amounts of sand and mud (yellow colour in Figure 4).The C B to C A parasequence transitions from offshore wackestone to packstone to shallow marine and to freshwater (light blue colour in Figure 4).The basal parasequence (green in Figure 4) and the upper two parasequences (yellow and blue in Figure 4) are considered to be transgressive, while the orange parasequence topped with a coralline boundstone is believed to be regressive (Figure 4).Mud, mixed quartz silt, calcitic, and dolomitic Deltaic, mixed marine and brackish water (Missimer, 1999, Scott, 1988) A/L Peace River Formation Unconsolidated mud, dark yellowish brown (6/3).Mud consists of clay with lime mud, quartz silt and fine sand, dolomite rhombs (floating), and some nodular phosphate Marl, sandy Lagoonal mud (Huang & Goodell, 1967;Missimer, 2002) B/D T Tamiami Formation Slightly lithified lime mud with medium to fine grained quartz lenses, light grey to white colour, no skeletal grains Limestone, mudstone, slightly sandy Intertidal (Gebelin, 1977;O'Connell et al., 2021;Sussko and Davis Jr., 1992) B/C T Tamiami Formation Mudstone, slightly sandy (fine quartz), chalky texture, friable, light tan, some mollusc casts Quartz sand with concretions Intertidal/Supratidal (Hoffman et al., 2004;Khalifa, 2012;Missimer, 1970;Shinn, 1983) B/B T Tamiami Formation Unconsolidated fine quartz sand, and sandstone, partially cemented with calcite, concentrations (intraclasts?),fine nodular phosphate, grey Limestone, wackestone Supratidal (Gomez & Astini, 2015;Shinn, 1983) B/A T Tamiami Formation Limestone, wackestone, moderately hard, laminations, sandy Calcitic shell with fine quartz sand Intertidal (Gebelin, 1977;Sussko & Davis, 1992) A/K Tamiami Formation Unconsolidated calcitic shell with up to 40% very fine grained quartz sand.May contain small, oyster fragments, grey Quartz sand with trace of sand-sized skeletal fragments Intertidal (Gomez & Astini, 2015;Shinn, 1983) A/J Tamiami Formation Unconsolidated quartz sand, fine to very fine, dark brownish grey, organics Limestone, wackestone, intraclasts, sandy Supratidal (Hardie, 1977;(Hardie & Shinn, 1988;Missimer, 1970) A/I Tamiami Formation Limestone, wackestone, moderately indurated, intraclasts Unlithified, quartz sand and shell Shallow offshore, shelf (Meeder, 1987;Zeller et al., 2015) C/C  (Meeder, 1987;Perkins, 1977) C/C D Caloosahatchee Formation Predominantly sandy skeletal wackestone, some packstone, distribution of quartz sand highly variable Unlithified quartz sand and shell Beach or very proximal subtidal (offshore) (Allmon, 1992) C/C C Caloosahatchee Formation Numerous mollusc types, no mud, may be death assemblage Sandy molluscan wackestone Offshore, shallow shelf (Holmes, 1988;Meeder, 1987) C/C B Caloosahatchee Formation Sandy wackestone, variable composition with sand shell, and mud, not proximal to beach, but shallow water Sandy wackestone/ packstone Discontinuity crust, onshore (Gerdes & Krumbein, 1987;Goudie, 1973;Multer & Hoffmeister, 1968) C/C A Caloosahatchee Formation Sandy wackestone/packstone, sandstone in place, very hard, some laminations of quartz sand and mud Sandy, wackestone, freshwater Interior wetland marl (Pederson et al., 2019) B/80 east base Fort Thompson, Formation Sandy wackestone, lime mud with freshwater molluscs, slightly sandy

Muddy sand and shell
Offshore open shelf (Conklin, 1968;Holmes, 1988;Stanley, 1986) B/G Fort Thompson Formation Unlithified muddy sand and shell, quartz sand fine with some clay

Shell and quartz sand
Beach/proximal offshore (Conklin, 1968;Holmes, 1988;Missimer, 1973;Stanley, 1986 Offshore (Conklin, 1968) A/F Fort Thompson Formation Fine quartz sand with shell, may be localised subfacies The depositional model implied from analyses of the various lithologies with associated water depths from the Caloosahatchee Formation is shown in Figure 8A.Note that the subfacies lie in shallow water above the waveorbital depth and are essentially parallel, using a modern day analogue for the southern part of the western Florida Shelf from perhaps Cape Sable towards the south-east.

| Fort Thompson Formation
The sequence stratigraphy of the Fort Thompson Formation shows three parasequences as shown in Figure 3 based on the lithofacies interpretation in Table S10.Three sealevel events appear to occur within the Fort Thompson Formation section.There is a discontinuity between units H and G in Section A and between units F and E in Section B. This is the same discontinuity.The occurrence of some muddy sediment near the base of the section may indicate a shallow offshore depositional environment (perhaps similar to that occurring today offshore from Sanibel Island along the Florida West Coast).Along the north wall of the pit, the lowest part of the Fort Thompson Formation within Bed H (same as in Section A) shows a progressive change in lithology from sandy wackestone/packstone to sandy skeletal wackestone to sandy packstone, which indicates a slight increase in energy and supports the shoaling upward nature of the unit (samples HL-4, HL-3, HL-2 and HL-1).The occurrence of quartz sand and shells in Section A, upper Bed H and in Section B, Bed F appears to be a beach deposit, similar to the Holocene barrier islands currently occurring in south-west Florida (Missimer, 1973;Stapor et al., 1991).This is a shoaling-upward parasequence and may be regressive.Beds G, F and E appear to be a nearshore deposit (just off the beach) to a beach deposit.It may be essentially a repeat of the lower parasequence.Bed D may represent a parasequence by itself or could be associated with an offshore-beach-shallow lagoonfreshwater parasequence.The petrography of the unit shows that the diversity of grain types is highest at the base in a sandy wackestone containing corals.The percentage of quartz sand increases from the base to the top of the unit, where there is a laminated sandstone, which is a likely exposure horizon.This unit may represent a single sea-level event.
A fine-quartz sand unit lies atop unit D in Sections A and B. This unit is unconsolidated and contains no skeletal or non-skeletal carbonates.It is interpreted to be an erosional part of the Fort Thompson Formation that may be associated with a late low-stand sea-level event.There is no evidence for long distance littoral transport of the quartz sand.The unit commonly occurs above Bed D in locations with altitudes below 3.5 m and is absent in many locations at altitudes up to 7.6 m.In the literature, it is commonly termed the Pamlico Sand, but this unit in southern Florida has never been properly defined and it is likely part of the Fort Thompson Formation (Missimer, 1984).
The two lower parasequences shown in Figure 3 (yellow and tan) are believed to be transgressive.The duricrust (bed D) in both sections is an exposure horizon marking the top of a sea-level event.Another sea-level event produced a medium to fine quartz sand deposit (brown colour in both Sections A and B).This unit is believed to be regressive and currently contains carbonate fossil material, which was likely removed by dissolution (Franco et al., 2016).
A model for parallel facies as an analogue for deposition with the Fort Thompson Formation is shown in Figure 8A.Note that the depositional environments found within the formation are quite similar to those observed offshore to onshore for Sanibel Island to Charlotte County along the Florida West Coast.The water depths range from 5 m to 10 m onshore, above sea level on the barrier island beaches, at depths of 0 m to 2 m in the supratidal and intertidal environment with no facies representing a lagoonal environment.Between water depths of 5 m and 10 m offshore, the sediment is still predominantly quartz sand with shells, but contains higher mud content with distance offshore.The area between the barrier island beaches contains shell and sand with little mud and typically can contain a death assemblage of molluscs at specific locations.The barrier island sands contain more quartz sand than shell and are devoid of mud.Landward of the barrier islands is a shallow supratidal/intertidal area that contains mud, quartz sand and shell in varying proportions.There is a shallow water sandy beach occurring along the land mass.The siliciclastic component moves along the barrier island beaches from north to south and mixes based on storms and minor sea-level events.This parallel facies belt can produce all of the depositional environments and lithologies observed in the Plio-Pleistocene sediments at the Nelson Road Pit.
The facies belt conceptual model for the Caloosahatchee Formation is more complex than that of the Fort Thompson Formation (Figure 8B).The analogue area would be further to the south offshore from Cape Sable.The water depth range is believed to be 0-20 m which creates a larger number of lithologies.The offshore area contains a combination of carbonate mud, shell and some quartz sand.The offshore content of quartz sand decreases from north to south and increases from west to east, particularly parallel to the barrier islands.The siliciclastic component is a series of barrier islands which terminates in a predominantly carbonate environment.On the southern end of the facies belt, some small coral reefs (or patch reefs) occur in 3-10 m of water depth.Some supratidal and subtidal areas occur between the offshore reef areas and the shoreline.A supratidal or very low slope land area occurs parallel to a very shallow lagoonal belt.The shoreline quartz sand supply is limited, thereby sending irregular pulses (e.g.storm-generated) of quartz into the varied depositional environments from the beaches both offshore and inshore.This general model serves as an analogue that can produce all of the lithic units found in the Caloosahatchee Formation at the Nelson Road Pit.
A key observation in all of the sediment facies observed and described is the shallow water depth associated with all of them (Figure 8).Water depth is likely 10-20 m or less in all depositional environments from this Plio-Pleistocene stratigraphic section.With water depths less than 10 m, the siliciclastic sediment fraction is mobile in all marine environments from lagoon to beaches to offshore.The storm water base that allows sediment movement in the current Gulf of Mexico is at least 20 m (Bernard et al., 1959(Bernard et al., , 1962;;Missimer, 2002).Maximum near bottom wave orbital velocities of 200 cm/s have been measured at 30 m in the Gulf of Mexico during Hurricane Andrew (Stone et al., 1995).
Since the water depth is 10 m or less in all environments, quartz sand is easily transported into all environments and incorporated or homogenised within carbonate sediments.
5.3 | Estimated ages of the stratigraphic units and correlation to sea-level events 5.3.1 | Tamiami Formation   As suggested in the introduction, a re-analysis of Missimer (2002) data suggests that the top of the Pinecrest Sand member of the Tamiami Formation is constrained to occur between 2.3 and 2.59 Ma and the base of the formation occurs between 2.8 and 2.9 Ma.The base of the formation may be older based on the research done by Tao and Grossman (2010) (estimated age 3.5 Ma).The one age determination from the Tamiami Formation in the present study was from a limestone sample collected at about 5.5 m below surface in Section B (Table S9).This yields an age of 4.9 ± 2.7 Ma.This age appears to support an older age for the base of the formation.
The exposed stratigraphic section of the Tamiami Formation is quite thin and the two parasequences defined likely occur near the base of the formation.If this is the case and the age of the formation base based on re-analysis of Missimer (2002) data, MIS G15 and/or G14 could correlate to these parasequences (Lisiecki & Raymo, 2005).Considerably more data are needed to better define the age of the Tamiami Formation. 5.3.2| Caloosahatchee Formation   Missimer (2002) constrained the age range of the Caloosahatchee Formation to be between 2.14 and 0.7 Ma with an estimated error range of ±0.1 Ma.The ages found in this investigation range were 1.2 ± 0.2, 1.3 ± 0.2 and 1.65 ± 1.3 Ma, which all fit within the estimated age range of Missimer (2002).Four parasequences were defined in the Caloosahatchee Formation section shown in Figure 4.It is not clear how much time elapsed during the disconformities in this section, including that between the Caloosahatchee and Fort Thompson formations, which might be as much as 1 Myr.
Within the range of 1.6 and 1.2 Ma, there are 38 MIS events defined on the basis of oxygen isotopes by Lisiecki and Raymo (2005).Therefore, it is not possible to clearly relate the parasequences at Nelson Road Pit to specific sea-level events associated with the isotope events.The entire formation is now defined to be early Pleistocene based on the redefinition as defined in Finney (2010) and Gibbard et al. (2010). 5.3.3| Fort Thompson Formation   Past age estimates for the Fort Thompson Formation made by Puri and Vanstrum (1969) suggested that it ranged from 120 ka to 140 ka based on 234 Th/ 238 U measurements.An assessed age range of Fort Thompson Formation deposits at Oldsmar in eastern Pinellas County was 120-130 ka at Oldsmar 1 and 200 ka at Oldsmar (Karrow et al., 1996).Jones et al. (1995) estimated the age range to occur between 600 and 900 ka based on 87 Sr/ 86 Sr ratios.The range age in this analysis based on strontium isotopes is 0-1.5 Ma including the error range.One sample produced a very old age and is considered to be a reworked shell.These data are not particularly useful in constraining the age of the unit.
Based on the known stratigraphic relationship of the Fort Thompson Formation to the Miami Limestone and Key Largo Limestone (units in south-east Florida directly correlated to the Fort Thompson Formation), the younger age seems more likely (Broecker & Thurber, 1965;Missimer, 1984;Perkins, 1977).The lowest parasequence in Sections A and B may correlate to the MIS event 7b of Lisiecki and Raymo (2005).The second parasequence may correlate to sea-level event 5.The limestone crust is either part of the second parasequence or another parasequence, which would correlation to event 5a or 5b, which are the only MIS events that could correlate to it.The uppermost clean quartz sand unit correlates to either event 5b or 3.This unit is regressive, which suggests that MIS event 3 may be the best correlation.

| Origin and maturity of the siliciclastic fraction of the sediment
A very high percentage of quartz was found in the siliciclastic sediment with very few other minerals or rock fragments.A small percentage of siliceous grains contained sizable inclusions of muscovite, biotite and hornblende.The siliciclastics sediment is therefore classified as 'supermature' or a quartz arenite (Folk, 1968;McBride, 1963).In addition, it contains virtually no trace grains of the typical modern beach sands of southern Florida, such as titanite (sphene), monzanite, zircon and others (Hsu, 1960;Martens, 1935).The primary source of the siliciclastic sediment is the southern Appalachian Mountains, which is located nearly 1800 km from the Nelson Road Pit.In addition, the unlithified quartz sand appears to have been reworked numerous times, thereby eliminating feldspars and other trace minerals.The quartz sands within the Caloosahatchee and Fort Thompson formations may also have been eroded and reworked from Pliocene-age deposits occurring on the Trail Ridge area of central Florida depending on the ability to transport them to the west and south (Pirkle et al., 1964).Late Miocene fluvial sands occurring in stream channels to the east could also have been reworked up-section to produce some of the quartz sand (Missimer & Maliva, 2017).However, these Late Miocene sands are quite immature and larger in grain size.Therefore, extensive reworking would be required to produce the composition and grain size found at the Nelson Road Pit.The siliciclastic component of the lithified rock does contain trace amounts of feldspars and a greater percent of grains containing large inclusions.
Quartz is the predominant grain type with straight extinction and only a few composite grains containing multiple grains.Few grains contain abundant vacuoles or microlites.Based on the roundness scale of Power (1963) and the angularity index of Missimer and Maliva (2017), the quartz has an index of between 3.5 and 4 (subrounded to rounded).These properties suggest that the quartz originates from igneous and volcanic sources either in the southern Appalachian Mountains or from the drainage area of the Mississippi River (Missimer & Maliva, 2017).In either case, the transport distance with episodic storage was quite long, which resulted in the super-mature nature of the quartz.The siliciclastic sediments at the Nelson Road Pit in the Pleistocene are more mature than the modern beach sands of southern Florida (Hsu, 1960;Martens, 1935).

| Processes causing homogenisation of siliciclastic and carbonate sediments
The entire Plio-Pleistocene section at the Nelson Road Pit contains quartz sand and a few facies contain a limited amount of clay mixed with carbonate mud.All of the depositional environments occurring in the stratigraphic section had water depths of 20 m or less and in most cases <10 m.A key feature of the siliciclastic component within all of the environments is that it is virtually all quartz sand with no significant mud content to adversely impact the growth of many carbonate-secreting organisms.
Although the relationship between terrigenous sediment input and benthic fauna is complex (Lohrer et al., 2006;Peterson, 1985;Snelgrove & Butman, 1994), it is generally true that most suspension-feeding benthic organisms have limited tolerance for high sediment input, especially mud.The sedimentological analysis suggests the environments allowed the carbonate factory to continue to produce without being overwhelmed by the fine siliciclastic sediment influx, as exemplified by the occurrence of a coralline boundstone unit containing quartz sand in the Caloosahatchee Formation and abundant molluscs in the upper part of the Fort Thompson Formation within a sandy mud environment.Recent studies of corals show that there is a degree of species-dependent tolerance to turbidity and low light tolerance (Van Oppen et al., 2005).In addition, coral growth has occurred on the Great Barrier Reef under conditions of long-term, fine-grained, terrigenous sediment accumulation (Perry et al., 2009).In the section studied, the continued success of carbonate-producing organisms occurred not only due to the low abundance of terrigenous mud, but also because of the generally low rate of influx of even coarser sediment, which did not overwhelm the benthic organisms.With the exception of the corals, periodic winnowing of mud during storms does not inhibit the growth of the molluscan infauna, because the deposition of mud is a common occurrence within the lagoonal and shallow offshore environments.The coral development indicates that the occurrence of turbid water during that time period was sufficiently short not to fully curtail growth.In the shallow offshore environment found in the Fort Thompson Formation, small spherical corals grew on hard (e.g.large molluscs) substrate, but were moved around on the bottom during storms and not fully buried.
The only means of transport of quartz sand to the southern end of the Florida Peninsula during the Plio-Pleistocene was via longshore and nearshore transport with minimal fluvial transport.This is unlike the Late Miocene/Early Pliocene time in southern Florida when fluvial influx of sediment did occur (Missimer, 2002;Missimer & Maliva, 2017).In the Early Pliocene, deltaic sedimentation occurred that contained massive quantities of mud precluding the growth of many carbonate skeletal organisms (Peace River Formation) (Missimer, 1999(Missimer, , 2002;;Scott, 1988).Fluvial sediment entering into carbonate environments is known to produce mixed sediments during many periods of the geological record (Coffey & Read, 2007;Coffey & Sunde, 2014;D'Agostini et al., 2015).
Of the four external mixing methods suggested by Mount (1984), three appear to be represented in the Plio-Pleistocene at the Nelson Road Pit and throughout southern Florida (Table 3).Episodic mixing occurs between differing depositional environments, but it is not rare as suggested by Mount (1984).Tropical storm-induced sediment transport occurs in water shallower than 20 m, which encompasses all of the depositional environments in the Plio-Pleistocene in southern Florida and the frequency of storm occurrence was likely under 10 years, if similar to modern conditions or during the Holocene as found in barrier island sediments in south-western Florida (Missimer, 1973).Mixing of bounding facies is also common because of the shallow water depth and the common storm occurrence, which has resulted in large death assemblages in the Fort Thompson Formation.Some of the shell beds in these units appear to be storm deposits.Autochthonous growth of many carbonate organisms occurred in the shallow subtidal offshore, the intertidal/supratidal and lagoonal depositional environments, similar to current conditions (Figure 8).
The thin parasequences show that mixing of mature quartz sand and a variety of carbonate organisms was episodic and repeated many times over the past 3 Myr based on relatively small changes in eustatic sea levels (class 6 sea-level events; Lisiecki & Raymo, 2005).The key issues in the homogenisation of the siliciclastic and carbonate sediments is the combination of shallow water, significant numbers of differing marine and freshwater environments occurring in close proximity, storm activity and the ability of the carbonate organisms to recover from storm events; thereby leaving virtually no end-member compositions.Direct evidence for storm activity occurs in the Fort Thompson Formation shell beds wherein large molluscs occur with minimal quartz sand at various increments in the section.
At a smaller scale, intense bioturbation is common in all of the intertidal and subtidal environments found in these sediments (Table 3; Figures 5, 6 and 7).This process, along with full dissolution of the metastable carbonate grains, exacerbated the homogenisation process by allowing movement of quartz sand into large pores (Franco et al., 2016).

| CONCLUSIONS
Detailed analysis of the Plio-Pleistocene stratigraphic section exposed in the Nelson Road pit demonstrates that the mixing of carbonate and siliciclastic sediments that today characterises the south-west coast of Florida north of Florida Bay was established no later than the beginning of the Pleistocene.Sedimentological, stratigraphic and palaeontological data indicate that during the Pleistocene the causes of this mixing included: (1) transfer of sediments from one depositional environment into another during storms or other extreme events; (2) mixing along facies boundaries; and (3) in situ mixing.The siliciclastic and carbonate depositional environments occurred in a series of subparallel belts with some differences between those found in the Caloosahatchee and Fort Thompson formations.
Homogenisation of carbonate and siliciclastic sediments occurred in virtually all depositional environments in the Plio-Pleistocene section, because the water depths were <10 m in nearly all cases.Therefore, storm mixing could occur unabated by wave orbital depth restrictions, causing extensive mixing of sediment types.
The rate of siliciclastic sediment influx into southwest Florida via nearshore or beach transport was low enough to allow continued growth of carbonate organisms, thereby producing mixed environments in every lithofacies.Since the primary component of the siliciclastic sediment was quartz sand, mud deposition did not produce poor water clarity, nor did it inhibit carbonate organism growth.
Based on the observations made of the Plio-Pleistocene stratigraphic section in south-west Florida, complete homogenisation of siliciclastic and carbonate sediments can occur without completing shutting down the production of carbonate sediments.In cases where bounding depositional environments contain water depths <10 m, homogenisation of carbonates and siliciclastic sediments occur uninhibited in these shelf or ramp environments.Therefore, occurrence of pure end-member compositions is uncommon in this type of homogenised system, whereas even in most siliciclasticdominated systems, nearly pure carbonate and siliciclastic subfacies are common.The Plio-Pleistocene

F
I G U R E 1 Location of the Nelson Road Pit in Lee County, Florida.

F
Aerial view of Nelson Road Pit showing the locations of the measured stratigraphic sections, A, B, C. Nelson Road Pit is now flooded and appears blue.The aerial photograph is a blowup of the north-east lake shown within the red box of Figure 1.The crosshatched area is a shelf at the top of the Caloosahatchee Formation where it was left intact.Locations A and B correspond to the stratigraphic sections in Figure 3 and the location C stratigraphic section is shown in Figure 4.

F
I G U R E 3 Stratigraphy of the Nelson Road Pit at Sections A and B (location shown on Figure 2).The transgressive (T) and regressive (R) parasequences are shown.Blue arrows are sampling locations of the unlithified sediments and the red arrow shows the location of sampling of rock for thin section construction.F I G U R E 4 Measured Section C at the Nelson Road Pit showing the Caloosahatchee Formation (in shelf above the floor of the pit).The darker lines highlight the discontinuities.

F
Thin sections showing lithologic characteristics of the Tamiami Formation.(A) Four different matrix types including mixed composition mud (A), lime mud (B), microspar (C), and isopachous microcrystalline calcite cement (D).(B) Burrow infilled with mixed composition mud and partially recrystallised into granular calcite.Note burrow cuts through a mollusc fragment.(C) Fully recrystallised mud matrix that did contain mouldic pores.The mouldic pores are infilled by blocky calcite (A) with the matrix graded from granular calcite (B) to microspar (C).(D) Microspar matrix with micritised skeletal grains (arrows) and quartz sand.

F
Photomicrographs showing lithologic characteristics of the Caloosahatchee Formation.(A) Sandy wackestone/sandstone at the top of Bed A. (B) Deeper in the uppermost unit (Bed A) mouldic pores from skeletal dissolution are filled with blocky calcite.(C) Bed B contains a matrix of calcitic mud and microcrystalline cement and many open mouldic pores.(D) Lower in Bed B the open mouldic pores are lined with microspar (may be marine).(E) Bed D contains blocky calcite infill of some mouldic pores and nearly all remaining skeletal grains are micritised (arrows).(F) Sandy skeletal packstone with many pores infilled with granular or blocky calcite in Bed F. (G) Compressed intraclasts or mud lumps form part of the matrix in Bed G. (H) Sandy skeletal packstone contain burrows infilled with sandy wackestone (A) and most skeletal grains are micritised (arrows) in another part of Bed G. (I) Sandy skeletal packstone with microcrystalline calcite cement (some isopachous cement) and skeletal grains infilled with cement (Bed H). (J) Sandy skeletal packstone/wackestone with some mud containing quartz sand trapped in larger pores (Bed I).

T A B L E 2
Facies constituting sequences and parasequences.Note that the location of the facies at the Nelson Road Pit are shown in the third column.

F
Facies belt models for the deposition of the Fort Thompson and Caloosahatchee formations.(A) Note that the parallel facies for the Fort Thompson Formation are predominantly siliciclastic dominated.(B) The Caloosahatchee Formation, however, has a siliciclastic component ending in a predominately carbonate depositional system.
Farrell et aMcArthur (1997), andMcArthur et al. (2001)made on samples collected from Sections A and C. The data are presented in Table1.Estimated ages based on the strontium isotopes were made using the Look-Up versus age data from McArthur (obtained from the author as a table) as supported by data inFarrell et al. (1995),Howarth  andMcArthur (1997), andMcArthur et al. (2001)(Table1).

Formation Depth below surface Measured 87 Sr/ 86 Sr Age and error (error includes combined measurement and age curve)
Sources of siliciclastic sediment and causes of homogenisation with carbonate sediments. of south-west Florida can be used as a model for unabated homogenisation of mixed siliciclastics and carbonates in ancient rocks based on the low-slope ramp environment, shallow water depth and tightly bounded depositional environments.
T A B L E 3 sediments