Mineralogical model of the Floridan aquifer in the southwest Florida water management district

MISSING IMAGE

Material Information

Title:
Mineralogical model of the Floridan aquifer in the southwest Florida water management district
Series Title:
Florida Water Resources Research Center Publication Number 68
Physical Description:
Book
Creator:
Randazzo, A. F.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Notes

Abstract:
The distribution of geological and hydrochemical parameters in cores and wells penetrating the Floridan aquifer along portions of the west coast of peninsular Florida is related to the occurrence of dolomite. Trace element analyses of the cores indicate correlations between strontium and sodium concentrations and the particular carbonate phase. Chloride, sulfate, and conductivity values indicate the position and extent of the freshwater-saltwater interface. Thick sequences of carbonate rocks in western peninsular Florida have been dolomitized in the freshwater-saltwater mixing zone of the coastal aquifer. A multivariate computer analysis was made of the petrographic data from three of the cores studied. The resulting multidimensional scaling diagram revealed several trends of environment of deposition parameters. Energy levels and faunal diversities helped to reconstruct the paleoenvironments of these Tertiary carbonate rocks. These data were then coordinated with the stratigraphy of the study area. Correlations between the occurrence of dolomite and specific stratigraphic formations suggest that dolomitization was rock-selective. Evidence of continuous formation of dolomite indicates that the process is actively occurring. The nature and distribution of dolomite in the Floridan aquifer is significant in the development of carbonate rock porosity. The dissolution and replacement of minerals affect groundwater movement as the hydrologically dynamic system involves a continual interaction between water and rocks. Understanding these interactions will aid hydrologists to inventory more precisely present-day water supplies and to predict changes to be expected. More efficient water management can be achieved by integrating the mineralogical model with hydrologic data.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution and holding location.
System ID:
AA00001544:00001


This item is only available as the following downloads:


Full Text














1'.


~P~a~' S ~i~ i Zbt~P~.
''
r- .:
'* 'X' 5-
~r
-e rr
~;
--. _:lL
1:
~ba~ ~p~ .rIPiWla
I~ ~ .~
1

:. "
~ ,i


Publication No. 68



Mineralogical Model of the Floridan Aquifer in the

Southwest Florida Water Management District



By



A. F. Randazzo


Geology Department

University of Florida

Gainesville


"'
'' "
r

i,.-
r
;U( (+
~- ~:1-5-. IC i
~~-
'-"
.. ~. s-
-~tr~ r ..
:~
L i. ~~~ ?."L
) iL` Ib.
-~ ~
''; 1" 1'1
r
: .. .~..
;1
'
3~
'`
"
.r 'L :X-
r- ., ~" :li
Y ~~L~
;;. I
-~ i ?-
"
?*-` Lr. r..
`""
.i .. r,
I, ~I
.' :h -"s
~- ri
"-- 'i: I "-~ -~ .~. -
I F L.; rf r
r '~ '
'
:I ..
"- '" ''
:-.-I -~:r'''7-; ; .-:l:--t
.r ~-- -C--
-; ?-.
C- ~~; ~
~
i '
.;'


... .
IL;-, ~

:r;-
'
L-
-
i'
.: :
c
-~' ,...:
R-z~

?::
.. ---~i
..s~c-
-"
; .:.::.~ ...~..-.,_': -'


~$
' i
'
r z "


r~.
: ri'
t
i .~..
'- ~ -.-.

c~i~-~B~a
Q
'
-,
-S ::


i
r.
r

I i
;




C
r


~i4~p


r


I;. v~au

: ...


"""" *" ,.-. ~"r. :;~"", 4
r 4
~ :
1 li
I ~ v:i
i.





ii


TABLE OF CONTENTS

Abstract...... ............................... ...................... 1
Introduction ................................ ....................... 2
Geologic History ................................................... 3
Diagenetic Solutions ............................................... 7
Study Area ......................................................... 9
Methods of Analysis ................................................ 9
Distribution of Parameters........................................11
Environments of Deposition...................... ..... .......11
Multidimensional Scaling Model ..............................11
Cluster 1 .............................................16
Cluster 2................ ............................. 16
Cluster 3 ............................................ 16
Cluster 4 ............................................ 16
Cluster 5 ............................................ 16
Cluster 6 ............................................ 16
Cluster 7 ............................................18
Interpretation ..........................................18
Stratigraphy...................................................20
Mineralogy..................................................... 20
Trace Elements and Geochemistry of Rocks........................24
Strontium..................................... .............24
Sodium.......................................................24
Mole-percent MgCO3 ........................................ 28
14C........................................................... 30
Distribution of Hydrogeochemical Parameters.....................30
Chloride...................................................30
Conductivity...............................................33
Sulfate.....................................................33
Discussion....... .............................. ....................33
The Model-A Summary of Data and Interpretations.....................41
Acknowledgements...................................................44
References Cited.......................................... ........45









ABSTRACT

The distribution of geological and hydrochemical parameters in cores
and wells penetrating the Floridan aquifer along portions of the west coast
of peninsular Florida is related to the occurrence of dolomite. Trace
element analyses of the cores indicate correlations between strontium and
sodium concentrations and the particular carbonate phase. Chloride, sulfate,
and conductivity values indicate the position and extent of the freshwater-
saltwater interface. Thick sequences of carbonate rocks in western penin-
sular Florida have been dolomitized in the freshwater-saltwater mixing zone
of the coastal aquifer.
A multivariate computer analysis was made of the petrographic data from
three of the cores studied. The resulting multidimensional scaling diagram
revealed several trends of environment of deposition parameters. Energy
levels and faunal diversities helped to reconstruct the paleoenvironments
of these Tertiary carbonate rocks. These data were then coordinated with
the stratigraphy of the study area.
Correlations between the occurrence of dolomite and specific stratigraphic
formations suggest that dolomitization was rock-selective. Evidence of
continuous formation of dolomite indicates that the process is actively
occurring.

The nature and distribution of dolomite in the Floridan aquifer is
significant in the development of carbonate rock porosity. The dissolution
and replacement of minerals affect groundwater movement as the hydrologic-
ally dynamic system involves a continual interaction between water and
rocks. Understanding these interactions will aid hydrologists to inven-
tory more precisely present-day water supplies and to predict changes
to be expected. More efficient water management can be achieved by
integrating the mineralogical model with hydrologic data.










INTRODUCTION

The Southwest Florida Water Management District relies upon the Floridan
aquifer for its principal supply of potable water. The geologic nature
of this aquifer and its relationship to groundwater flow systems is of
fundamental importance in proper water management techniques. A model
was established in order to express the mineralogical distributions within
the aquifer and to demonstrate mineral interactions with groundwater.
The mineralogical model of the Floridan aquifer in the Southwest Florida
Water Management District relates specific geochemical, hydrologic, and geologic
data to the occurrence of carbonate minerals in the aquifer. This model
provides insight to the way carbonate aquifers develop. The aquifer repre-
sents a dynamic system in which waters of varying chemistry have or are
reacting with subsurface rock sequences to produce porosity and permeability
changes and the formation or destruction of certain minerals.
One major process which affects porosity and permeability in carbonate
systems is dolomitization. Weyl (1960) showed through conservation of mass
requirements that mole-for-mole replacement of calcium by magnesium will
result in a 13 percent volume shrinkage, increasing porosity by joining
smaller pores and thereby increasing permeability. Schmidt (1965) stated
that in the normal sequence of neomorphism (Folk, 1965) of calcium carbonate
grains, a mole-for-mole replacement of aragonite by calcite will increase
the mineral volume (decrease the porosity) by 8.7 percent. Subsequent
dolomitization of the calcite will result in a 13 percent volume decrease
(increase the porosity). Dolomitization of an original aragonite matrix,
however, will result in a volume decrease of only 5.4 percent.
Of course, in whichever sequence dolomitization occurs, such factors
as original packing, compaction, and introduction of materials by percolating
waters will affect the overall porosity and permeability. A better under-
standing of the processes that control dolomitization in a carbonate aquifer
system can provide valuable knowledge to the way aquifers develop. Previous
studies have related the occurrence of a mixing zone, where seawater and
groundwater meet in a coastal aquifer, to dolomitization (Kohout, 1965, 1967;
Runnells, 1969; Hanshaw et al., 1971; Badiozamani, 1973; Land, 1973; Folk
and Land, 1975; Hanshaw and Back, 1979).









The parameters considered in this study traditionally have been shown
to be either of direct significance to the dolomitization process, or to
be useful in defining a freshwater-saltwater interface. Trace element
analyses of cores were conducted on strontium and sodium. Correlating
their concentrations and distribution of the various mineralogies present
led to a hypothesis of dolomitization by solutions of diluted seawater.
Chloride, sulfate, and conductivity measurements from interstitital pore
waters were used to define the present position and extent of the freshwater-
saltwater interface.
Groundwater composition and flow characteristics are directly affected
by these water-rock interactions. Zones of high transmissivity, as well
as areas of water with high or low percentages of total dissolved solids,
can be better explained by utilization of a mineralogical model. Likewise,
this model can serve as a predictive tool for aquifer evolution.
This study involved a number of graduate students who made significant
contributions to the total research effort. The investigation expanded upon
two earlier endeavors (Randazzo, 1976, 1980) and represents a greater delinea-
tion and understanding of the Floridan aquifer. The reader is directed to
the works of Saroop (1974), Stone (1975), Hickey (1976), Liu (1978), Zachos
(1978), Fenk (1979), and Sharpe (1980) for specific details on the lithologic
and paleontologic characteristics of the various lithofacies recognized.
Hickey (1976), Sarver (1978), Zachos (1978), and Metrin (1979) discussed the
diagenetic and geochemical aspects of the important carbonate rock-forming
minerals, calcite and dolomite. Bloom (1982) summarized the data and conclu-
sions from these earlier reports and integrated them with new data presented
here.


GEOLOGIC HISTORY

At the beginning of the Tertiary Period a broad, stable carbonate bank
existed over the Florida Platform (Chen, 1965). This bank was bounded by
submarine escarpments on both the Atlantic and Gulf of Mexico sides, and
was separated from the continental shelf to the north by the Suwannee Channel
(Applin and Applin, 1944; Jordan, 1954; Chen, 1965) (Fig. 1). A warm shallow
sea environment resulted in the deposition of the Cedar Keys Formation in the
Paleocene, followed in superposed order by deposition of the Oldsmar, Lake
City, Avon Park, and Ocala carbonate sequences during the Eocene. There









..*.. GEORGIA
1- .( \


South
Florida
Shelf


-----.. Major Axis of Peninsular Arch
- Major Axis of Ocalo Arch


South
Florida
Basin


0 50 0 miles
! [__ J-


Major structural features of peninsular Florida.


ATLANTIC


OCEAN


GULF
OF
MEXICO


Figure 1.










were short episodes during the Eocene when great parts of the northern and
central portions of the platform were emergent and subjected to nondeposition
and subaerial erosion (Chen, 1965). These emergences are evidenced by
numerous unconformities (Randazzo et al., 1977).
The carbonate faces deposited during the early Tertiary dip and thicken
towards the south and the coasts, away from a subsurface structural high known
as the Peninsular Arch (Vernon, 1951; Chen, 1965). The updip sections of
the Floridan aquifer crop out at the contact with the Cretaceous Coastal
Plain sediments in southern Georgia and Alabama. Vast quantities of water
flow off of the southern Appalachians and infiltrate the porous carbonates
here, creating an area of principal recharge to the aquifer.
Deposition of the Oligocene Suwannee Limestone was followed by a marine
regression and an extensive period of post-Oligocene erosion (Vernon, 1951).
The post-Oligocene seas exposed a structural high in western peninsular
Florida. Originally described as a breached dome, this "Ocala Uplift"
(Cooke, 1945; Vernon, 1951) has more recently been termed a "blister dome"
(Winston, 1976). Winston believed that it formed when an increase in the
rate of deposition of the Lake City Formation produced a curvature, reflected
on the surface of deposition. Subsequently deposited carbonate sequences
also reflected this curvature, and Oligocene or post-Oligocene tilting of
the Florida Platform served to accentuate the dome.
The Ocala "blister dome" was exposed in the early and middle Miocene
during which time shallow brackish water and marine faces of the Tampa
Formation (early Miocene) and the Hawthorn Formation (middle Miocene) were
deposited around it. Cooke (1945) found the Tampa-Hawthorn contact to be
unconformable, and believed that the Hawthorn was deposited in an "expanded"
Tampa sea. The upper part of the Hawthorn Formation, which consists of a
heterogeneous mixture of lime mud, clays, marls, and phosphoric sands, forms
a confining unit over the Floridan aquifer, and thus marks its upper boundary.
At the close of the Miocene, the Hawthorn covered all of peninsular Florida
except the area over the Ocala "blister dome." Artesian circulation probably
began in the Floridan aquifer at that time, with recharge being confined pre-
dominantly to the areas of outcrop of the limestone (Stringfield and LeGrand,
1966; Vernon, 1951).











Today, local recharge occurs directly into the aquifer in areas of high
elevation along the crest of the Ocala "blister dome" where the Hawthorn was
not deposited. Where the Hawthorn is present, recharge can occur where it
is breached, either by erosion or sinkholes. Discharge occurs mainly out into
the Gulf of Mexico (Stringfield and LeGrand, 1966), and to a lesser extent
through coastal springs (Wetterhall, 1965; Fetter, 1980).
Stringfield and LeGrand (1966) agreed with Cooke (1945) that the distri-
bution of cavities within the Floridan aquifer, with their general decrease
in size and numbers with distance from recharge and discharge areas, suggests
that the present pattern of circulation developed in the Pleistocene during
low sea level stands. Ranges in sea level during the Pleistocene are estimated
to have been 270 feet above the present level (Cooke, 1945), and 500 feet
below the present level (Donn et al., 1962). Thus, when sea level was at its
lowest stand, it is conceivable that the groundwater level in western
peninsular Florida was a few hundred feet lower than it is at the present
time (Stringfield and LeGrand, 1966). During these stages, the Hawthorn was
extensively eroded allowing more widespread recharge into the aquifer, ac-
celerating dissolution of the limestone beneath.
The present-day nature of the carbonate aquifer is dependent upon the
interrelations of its geologic history and the hydrologic conditions to which
it has been subjected. Four geologic periods of aquifer development were
cited by Stringfield and LeGrand (1966, p. 7-8). The "depositional period"
establishes the volumetric extent of the carbonate unit and its primary
permeability. During the "immediate post-depositional period" the carbonate
deposit is elevated above sea level where it is exposed to meteoric weathering
and possibly subaerial erosion. In addition to diagenetic changes which
affect porosity, dissolution by infiltrating groundwater can greatly increase
the permeability of the rock during this stage. The "subsequent pre-recent
period" is marked by increased dissolution and permeability (assuming the
deposit is still exposed to groundwater infiltration), compaction, consolida-
tion, and recrystallization of the original limey deposits, and possibly the
occurrence of some folding or faulting which might restrict groundwater
circulation in some areas and promote it in others. Deposition of a sequence
of sediments of low permeability will hinder groundwater dissolution of the
underlying carbonates. The last period is marked by the "existing geologic
setting and structure." Karst :-. ]ii. develops where the limestone crops
out and is in the present groundwater circulation system.









The overall geologic structure is important in relation to topography
and recharge-discharge areas. In western peninsular Florida recharge pre-
dominantly occurs north of Tampa Bay where porous carbonate units crop out,
while south of Tampa Bay a layer of relatively impermeable clays inhibits
recharge to the aquifer except where sinkholes breach this aquiclude.


DIAGENETIC SOLUTIONS

Carbonate diagenesis appears to occur more rapidly in the phreatic
meteoric water environment than in the vadose meteoric or marine phreatic
environments. Ginsburg (1957) noted that the Miami Oolite is better cemented
where it presently exists below the modern water table than where it occurs
in the vadose or marine phreatic environments. Friedman (1964), in observing
that freshwater more commonly promotes mineralogical and textural changes
which lead to lithification than marine waters, ran a series of experiments
and concluded that low pH and low salinity of meteoric waters favor the
diagenetic changes that occur.
In studying a fossil water table in Bermuda, Land (1970) judged that the
phreatic meteoric water zone was where diagenesis of skeletal carbonates to
coarse-grained sparry limestone was most rapid. These rocks were more
highly altered and more stable mineralogically than rocks which spent more
time in the vadose or marine zones. Also, he found larger crystals of
cement in this zone where water could remain in pore spaces for long periods
of time, as opposed to the vadose zone where flushing groundwater resulted
in very little cement.
Land and Epstein (1970) discussed the mineralogical and isotopic
changes that occur during meteoric water diagenesis. They stated that,
generally, magnesium calcites incongruently dissolve rapidly to calcite,
followed by dissolution of aragonite and reprecipitation as sparry calcite
cement. As diagenesis proceeds, the unstable minerals dissolve, changing
the cation and anion chemistry of the solution. Thus, the newly formed
carbonates have lower concentrations of Sr+2, Na+, and Mg+2, representing
the interaction between meteoric water and the original marine sediments.
Steinen and Matthews (1973), reporting on a cored borehole on Barbados,
observed that the upper part of their core had been in the vadose zone since
its initial emergence from the marine environment. Although the carbonate
sediments in this zone were composed of the unstable phases aragonite and
high-magnesium calcite, they had not been affected to any great degree by
dissolution. That part of the core that had occupied the freshwater phreatic











lens at least once during sea level fluctuations appeared to be mineralogi-
cally stable (composed of low-magnesium calcite), extensively cemented, and
showed evidence of earlier dissolution. The lowest part of the core that
was subjected to vadose meteoric diagenesis during glacial epochs and marine
phreatic diagenesis during interglacial epochs, showed only minor cementation
and dissolution, and retained most of its depositional mineralogy.
The diagenetic changes that occurred within the lithofacies described
in this study that involve neomorphism of one carbonate mineral phase to
another probably occurred upon introduction to a phreatic meteoric water
environment.
The dolomite found within the cores in the study area has been categorized
according to origin as penecontemporaneous dolomite formed in a supratidal
environment; and secondary dolomite formed as a replacement of calcite by
diagenetic solutions. Table 1 summarizes the characteristics of dolomite types
as related to origin.


Table 1. Characteristics of dolomite types as related to origin.




Characteristics of Penecontemporaneous
Dolomite (Primary)

1. The dolomite is fine-grained (1-5 microns).
2. The dolomite occurs in association with supratidal sediments and
structures, such as algal stromatolites, birdseye vugs, root casts,
dessication features, and evaporite minerals
3. The dolomite is nonstoichiometric.
4. The dolomite forms by replacing lime muds, usually accompanied by a
preservation of original sedimentary textures.
5. The dolomitized areas are thin and/or localized.
6. One would expect to find alternating layers of gypsum-dolomite-
limestone in alternating supratidal-subtidal environments.
7. Trace element concentrations (Sr+2 and Na+) are high, reflecting the
hypersaline nature of the dolomitizing brines.

Characteristics of Dolomite Which Has Formed
in a Freshwater-Saltwater Mixing Zone
(Replacement)

1. The dolomite consists of larger crystals generally between 40-200
microns (Zenger, 1972). The more dilute the saline solution is, the
larger the crystals and the more euhedral the crystal fabric.
2. The dolomite can occur in sediments associated with any depositional
environment.









3. The dolomite is more-stoichiometric if the solution is quite dilute
and the rate of crystallization is slow; and less-stoichiometric if
it is only moderately dilute and accompanied by a faster rate of
crystallization.
4. Depositional texture is preserved more frequently in more-stoichiometric
dolomite, and suggests slow recrystallization in less saline solutions.
5. The dolomitized areas may vary in thickness, depending on the extent
of migration of the mixing zone.
6. More-stoichiometric dolomites are associated with lower Sr+2 and Na+
concentrations than less-stoichiometric dolomites, indicating formation
in less saline solutions. Both of these dolomite types, however,
contain lower trace element concentrations than penecontemporaneous
dolomites because of freshwater dilution in the mixing zone.
7. Coalescive neomorphism of original penecontemporaneous supratidal
dolomites leads to rhombic crystals. Previously undolomitized limestones
are recrystallized by a porphyroid neomorphic process that results in
euhedral, rhombic dolomite crystals.
8. More-stoichiometric dolomite shows a good clustering of points when
Sr+2 and Na+ concentrations are plotted against mole-percent MgCO3.
This indicates a longer residence time for the diagenetic solutions,
and a greater approach to equilibrium between crystal and solution than
occurs in less-stoichiometric dolomite (which characteristically produces
a scattering of points when Sr+2 and Na+ are plotted against mole-
percent MgC03).




STUDY AREA

The area under investigation is the Southwest Florida Water Management
District which occurs along the Gulf Coast of western peninsular Florida
(Fig. 2). All of the cores and wells studied lie within this district
except the Bell and Manatee Springs cores, which are within the Suwannee
River Water Management District. A complete north-south traverse covers
approximately 200 miles, and includes 29 wells, some as close together as
2 miles and others as distant as 47 miles.


METHODS OF ANALYSIS

The data compiled in this study came from various sources. X-ray
analyses of cores 4-2, 6-3, 11-2, 17-1, 17-3, 19-3, and 21-2 were used to
identify the presence of calcite and dolomite. Atomic absorption spectro-
photometry was utilized to determine sodium and strontium concentrations in
the rocks. Other stratigraphic, mineralogical and trace element data were
obtained from the studies by Randazzo et al. (1977), Randazzo and Hickey
(1978), Sarver (1978), Liu (1978), Zachos (1978), Fenk (1979), Metrin (1979),











-10-


Figure 2.


0o 0 lO 20
SCALE IN MILES

- HORIZONTAL PROJECTION OF CORES AND WELLS
TO PRODUCE LEGS OF PANEL DIAGRAM
O WELL LOCATION


LOCATION OF CORES AND WELLS IN STUDY AREA






-11-


Sharpe (1980), and Bloom (1982). Sulfate, chloride, and conductivity
measurements were obtained from the Southwest Florida Water Management
District. The Dicarb Radioisotope Company analyzed several dolomite samples
for 14C.
Panel diagrams are used to illustrate the variations of individual
parameters across the study area. The panel diagrams were produced in
conjunction with a United States Geological Survey potentiometric map
(Fig. 3). The points where panels intersect were made to coincide with local
recharge areas of 50 feet above mean sea level in Levy County, and 80 feet
above mean sea level in Pasco County. In this way, the distribution of
each parameter can be compared to the general flow of groundwater (Fig. 4).


DISTRIBUTION OF PARAMETERS

Environments of Deposition

Multidimensional Scaling Model

One of the most difficult problems in the utilization of a large quantity
of data is the development of a combination of statistical methods for
facilitating the environmental interpretation of carbonate rocks. In
dealing with thin sections and modal analyses examination is made of tables
of data which depict abundances and paucities of variable parameters. Inter-
pretation of large amounts of data by "eye" is difficult. The final results
are always partly intuitive. The capacity of computers to make comparisons
provides a modern capability to establish objectively significant relation-
ships among variables. Computers can compare variables within samples and
recognize similarities, making interpretations more accurate.
Cluster analysis is a simple form of correlation analysis, a method of
searching for relationships in a large symmetrical data matrix (Hayes, 1980).
The advantage of cluster analysis is simplicity (Valentine and Peddicord,
1967, p. 502) because of the uncomplicated nature of the input and the re-
sultant visibly distinct clusters. However, the numerical methods should not
obviate the need to analyze the data objectively; they only aid in reducing
the dimensionality of the data and when fully utilized can provide various
insights not seen without computations.
A combination of statistical methods (multivariate analysis) were
applied to the thin section data of the lithofacies to permit an environ-
mental interpretation of the Tertiary carbonate rocks. Cluster analysis was
applied in order to group-together variables such as allochems, lime mud,











-12-


Figure 3.


-LNORTH


FMEXICO ISO S 20
OF scatLE ii -ES



6 WELL LOCATIONS
NOTE CONTOURS SHOW APPROXIMATE ALTITUDE
ABOVE MEAN SEA LEVEL IN FEET
REVISE FROM U S GEOLOGICAL SURVEY
OPEN FILE 81-40A


ESTIMATED POTENTIOMETRIC SURFACE FOR
THE TERTIARY LIMESTONE AQUIFER SYSTEM (1980)








-13-


Figure 4.


U 19-3
19') (600')

RS LEVY CO 107
(320 ) RECHARGE AREA
1 1 390')


VERTICAL
SSCALE
SIN FEET

G UF MEXICO IF

NORTH 200
1o o 10 20
PROJECTED FLOW OF GROUNDWATER IN TERTIARY SCALE IN MILES
LIMESTONE AQUIFER SYSTEM
NOTE NUMBERS iN PARENTHESES DENOTE
DEPTH IN FEET BELOW SURFACE







-14-


calcite, dolomite, etc., (Table 2) that are most similar genetically.
The variables considered are the product of environmental processes;
therefore, variables most similar probably formed in the same depositional
environment under similar processes. Multidimensional scaling (MDS), based
on petrographic similarity, was used to order the samples on a two dimen-
sional graph. Non-metric MDS is a technique illustrated by Kruskal (1964a)
applied to geological problems (Whittington and Hughes, 1972; Smosna and
Warshauser, 1979).
Multivariate analysis in carbonate petrology was employed by Imbrie and
Purdy (1962) using factor analysis on data from Bahamian sediments. In
factor analysis, however, the original data is ultimately lost in vectors,
whereas cluster analysis does not abandon the original data and is immediately
interpretable. In factor analysis mathematical techniques are used to
promote correlation coefficients or secondary and tertiary data banks for
and patterns. Therefore, the input data is processed and the final visible
data is intelligible and correlatible but complicated because the worker is
not viewing the original data. However, in cluster analysis similarities
between the variables are processed and presented as groups of most similar
variables showing the original data. Parks (1966) combined an R-mode cluster
analysis and Q-mode cluster analysis to describe Bahamian sediments. Ekdale
et al. (1976) used multivariate analysis for paleoecological interpretation
of Cretaceous rudists. Comparisons sample by sample is a Q-mode clustering
and variable by variable (within the sample) is an R-mode clustering (Sokal
and Sneath, 1963). This computer application combines a variety of methods
that are nonparametric (not based on normal distribution). Any method that
reduces the dimensionality of the data causes distortion. Clustering imposes
a hierarchical structure to the data. Almost inevitably the method introduces
distortion in representing the multidimensional relations between the
localities on a two dimensional dendogram. This distribution of distortion
is well known; the relations at the tips of the dendogram are well-represented
and distortion is greater in the later formed clusters (Rowell et al., 1973).
The distortion of stress has to be taken into account when analyzing the
clustering output. Because of the existence of this distortion RoWell et al.
(1973, p. 3430) suggest that some form of ordination should be used as an
alternative or complementary means of displaying structure in the data.
Ordination is a comparative process. All samples or variables are compared
to each other in a symmetrical matrix. Dissimilarity is measured and this
produces a graphic geometric display. Cluster analysis often obscures the





-15-


Table 2. List of variables used in the multivariate analysis of three
cores of mid-Tertiary carbonate rocks.


Allochems Matrix Detrital

Foraminifera Peloid Calcite Quartz
Echinoderm Intraclast Micrite
Mollusca Pellet Dolomite


overall relationship, whereas ordination shows gradational relationships.
The amount of distortion may be estimated by calculating a correlation
between the interlocality distances in the reduced dimensional space and the
corresponding distances in the original similarity matrix. The distortion
that is typically present is distributed differently than that in a
dendogram. The distortion is greater in smaller distances and not larger
distances between the data points. Ekdale et al. (1976) used an ordination
algorithm versus the nonmetric algorithm used by Kruskal (1964a, 1964b).
Kruskal's algorithm includes a built-in accommodation of stress (distortion).
Basically the method seeks to find a geometric configuration of n points
(variables) in a reduced space of k dimensions (containing the variables)
such that interpoint distances correspond to the similarities between points.
The solution is the best fitting configuration that minimizes stress.
Nonmetric MDS distributes distortion better than clustering.
The carbonate rocks of this study were lithologically distinct and were
deposited under different environmental influences and processes. Ordering
the thin section samples on a MDS plot reflects gradients of these processes.
Multivariate analysis was performed on the thin section data from the litho-
facies described in three of the cores studied (Sharpe, 1980) to illustrate
the gradational trends of carbonate rock variables across a marine carbonate
shelf.
An R-mode cluster analysis and a Q-mode cluster analysis were run on 176
thin-section samples for 52,800 observations. Seven clusters from the Q-mode
were produced and scanned. Each cluster revealed characteristic variables
that distinguish that individual cluster. They were: 1) dolomite, 2) dolomite/
mud, 3) mud, 4) skeletal/mud, 5) intraclast, 6) quartzose mud and 7) skeletal/
sparite. Because of the large number of samples (176) the three dimensional







-16-


MDS plot was extremely cluttered and gave no clear isometric projection.
However, the two dimensional MDS plot shows well the gradations and is
illustrated here (Fig. 5). The cluster boundaries displayed on the plot
were delineated based on the Q-mode data. Some thin section points fell
between the cluster boundaries. They were transitional samples with charac-
teristics of the two nearest clusters. This illustrates the gradational
nature of the MDS plot.

Cluster 1

Dolomite microfacies. A mosaic of dolomite which completely obliterates
all allochems, while preserving some megascopic sedimentary textures (e.g.
faint laminations). These sediments are part of a Lithofacies of dolomitic
mudstones and wackestones from the Avon Park Formation.

Cluster 2

Dolomite/mud microfacies. The samples are dolomitized mudstones and
and wackestones with textures and allochems preserved. Within this cluster
also falls a dolomitic sandstone.

Cluster 3

Mud microfacies. These samples represent laminated mudstones and
wackestones with sparse fauna. They include peat layers and algal laminated
sediments.

Cluster 4

Skeletal/mud microfacies. The samples are characterized by high
faunal diversities and a high mud matrix content. This cluster contains
the samples with large foraminifera (e.g. Nummulites sp.) from the Ocala
Limestone.

Cluster 5

Intraclast microfacies. Abundant mudstones and wackestone occur here.
These samples have large intraclast contents, mud and sparse fauna.

Cluster 6

Quartzose mud microfacies. The samples are predominantly wackestones
and mudstones with sparse fauna but have very abundant detrital quartz
(20 80 percent).
























Supratidal


Figure 5.


Subtidal


Shallow


Two-dimensional MDS configuration displaying the relationships
among 176 thin-sectioned samples based on ten variables (Table 2).







-18-


Cluster 7

Skeletal/sparite microfacies. The samples are all skeletal grainstones
with abundant void-filling spar and high faunal diversities.

Interpretation

Clusters 1, 2 and 3 correspond to those lithofacies interpreted as
being deposited in the supratidal zone. The samples were independently
described as supratidal deposits. Thus the multivariate analysis and
clusters verified the similarity interpreted by conventional petrographic
methods. The inclusion of a dolomite sandstone into these clusters is
interesting. This unit was initially interpreted as a plastic bar deposit
and yet the clustering indicates a supratidal (possibly beach) environment.
The sandstone is very coarse and has associated organic material (dark
color); therefore, this could very well represent a beach or supratidal
channel deposit. Conversely, the presence of dolomite may have caused the
cluster analysis to misclassify the two thin section samples containing
this particular lithology (2 samples out of 176).
Cluster 4 is interpreted as representing deep subtidal deposits, with
low wave energy allowing the mud to accumulate and a high faunal diversity,
characteristic of the deep subtidal.
Both clusters 5 and 6 are interpreted as representing shallow water
intertidal depositional regimes because of the abundant quartz and mud
(restricted quiet water lagoon).
Samples found in cluster 7 represent shallow subtidal material seen
predominantly in several lithofacies. These samples have a high faunal
diversity, high faunal content and abundant sparry calcite.
Characteristics of the offshore depositional environments are gradational
perpendicular to shore. Therefore, a plot of the various variables charac-
teristic to one or more subenvironments should show this gradation. The multi-
dimensional scaling configuration (Fig. 5) demonstrates the gradational nature
of the 176 samples, allowing for an environmental interpretation. The clusters
(containing the samples) seem closer to one another, showing geometrically
the gradation.
In the interpretation of the MDS clusters, analysis must be made of
chemical, physical and biological aspects of the clusters. The dolomite and
mud-dominated clusters (1, 2 and 3) are interpreted as containing samples
representing the supratidal zone. The intertidal zone is represented by
samples in clusters 5 and 6, and samples of clusters 7 and 4 represent shal-
low subtidal and deep subtidal zones, respectively.









The cluster showing the hi.'-.t energy (wave action) is cluster 7.
Therefore, after placing the shoreline to the left of the diagram (supra-
tidal), curve A can be constructed showing -._.-- normal to the shore
(Fig. 5), culminating in cluster 7, the shallow *i--dal zone. The shallow
subtidal zone shows the least mud and the most .*..., and so ie-?;ents
the highest wave energy zone (cluster 7). The transect of hydrodynamic
energy starts in the supratidal and gradually increases through the deep
subtidal, through the intertidal and reaches the hi;i -.t level in the shallow
subtidal zone. This is illustrated by mud abundance varying from 81 percent,
in the supratidal samples, to 39 percent in the deep subtidal samples, to
27 percent in the intertidal samples, and 0-8 percent in the shallow subtidal.
Conversely the amount of sparite increases from 0-3 percent in clusters 1, 2
and 3 to 35 percent in cluster 7 (shallow subtidal). The pattern of fabric
support also parallels this trend passing from mudstones in clusters 1, 2


and 3 to grainstones in cluster 7.
of increasing energy and substrate
Faunal diversity is highest in
subtidal (7). Therefore, parallel
a faunal diversity trend (curve B)
clusters 1 and 2 to clusters 4 and
The two intertidal clusters (5
cluster (3), indicate that the MDS
change (advent of detrital quartz)


These ,:...-,ties parallel the direction
mobility.
the deep subtidal (4) and shallow
ing the hydrodynamic trend (curve A) is
showing increasing faunal diversity from
7 (Fig. 5).
and 6), together with the mud/quartz
analysis also realized the regional
in sediment characteristics. Consequently


clusters 3 and 6 are interpreted as indicating more plastic intertidal and
supratidal zones as shown by another hydrodynamic (energy) curve (C) (Fig. 5).
Diagenetic trends can also be shown on the MDS plot (Smosna and Warshauer,
1979). The susceptibility of supratidal sediments to dolomitization is
illustrated by the dolomite being concentrated in the -.:-.. ndal deposits
and this is supported by previous work (Randazzo and Hickey, 1978). The
trend does not show phases of dolomitization because there are different
dolomitic textures within cluster 2.
The presence of two distinct clusters of dolomite is thought to illustrate
different salinity conditions fluctuating from below normal to above normal
(Smosna and Warshauer, 1979). The two dolomite clusters may indicate
phreatic waters of different concentrations of Na, :., and Ca indicating the
possibility that different groundwater systems may have acted on the samples







-20-


of these clusters. Another hypothesis is that the clusters may well
indicate a difference in the time of contact with groundwater fluids.
Therefore, from this data presentation can be derived a hydrodynamic
and paleoecological interpretation of the Tertiary carbonates of west-
central Florida (Fig. 6). Figure 6A shows the carbonate platform with
lowest water energy at the supratidal and highest energy in the shallow
subtidal. Low faunal diversity is characteristic of the supratidal zone
increasing in diversity offshore, and the most pronounced dolomitization
has occurred in supratidal sediments. Figure 6B shows a slightly steeper
gradient in a clastic influenced tidal regime, shown by samples in clusters
2, 3 and 6.


Stratiqraphy

Figure 7 shows the stratigraphic distribution of the geologic units
to a depth of 600 feet. From Hernando County to the south, the formations
can be seen dipping to the south where they progressively become overlain
by younger units. This feature reflects the deeper lying structures of
the South Florida Shelf and the South Florida Basin (Fig. 1). The regional
effects of the Ocala "blister dome" can be seen in the northern half of
the diagram where the Ocala Limestone crops out.


Mineralogy

The distribution of limestone and dolomite are depicted in Figure 8.
Analyses were conducted by x-ray diffraction, and only those samples in
which the carbonate fraction was found to be 100 percent dolomite are
illustrated here. One of the most distinguishing observations is that
the Avon Park Formation has been significantly dolomitized in this area,
while the Ocala and Suwannee Limestones lying above .have not been. The
limestone-dolomite boundary follows the Ocala-Avon Park boundary, and dips
to the south. South of the Pasco County recharge area most of the Miocene
age deposits contain dolomite. The dolomite zone here thickens southward and
seaward, as do those stratigraphic units.







-21-


S7 5 1,2


Clusters
Increasing h-,-, ;.-amic ;,-.-


Increasing faunal diversity


Dolomite










.:+. ~~ "-
'.4..


V


V


6 2,3


Clusters

Increasing energy

Increasing diversity
WM_ -------i


Quartz


I


Schematic interpretation of Tertiary carbonate rocks illus-
trating the paleoenvironments. Plotted are the positions of
the seven Q-mode clusters and environmental gradients. Part
A shows the carbonate platform and the environmental inter-
pretation. Part B is the clastic-influenced interpretation.


low tide
wave base


Figure 6.


r__~_____(~s____P_____~___~e_~________j_


---- ---" ----"----"----.


h--------------------


P-sLmPPPLI-------~-----I--p--------~ ~II-


1.


I







-22-


Figure 7.






-23-


Figure 8.







-24-


Trace Elements and Geochemistry of Rocks

Strontium

The distribution of strontium is shown in Figure 9. Sr+2
concentrations greater than the '!-rce.-cent reduction associated with
the formation of dolomite from calcite (Behrens and Land, 1972) have been
explained by dolomitization occurring in waters more saline than those in
which calcite formed (Randazzo and '-i.-'. 1978). The Sr+2 content in
the northern cores (B, MS, CO, RS, 124, 21-2, HS, and 107) does not drop
to lower values at the limestone-dolomite boundary, but rather changes farther
down into the dolomite zone. Core 19-3 shows an increase in Sr+2 concen-
tration in its;deeper regions to over 800 ppm. Hce-, than average Sr+2
concentrations also occur at certain horizons in cores 21-2 and 17-1. These
areas of high Sr+2 concentration indicate where dolomite formed from solutions
more saline than did the surrounding rock. Across the entire transect of
the northern cores, it appears that the dolomites of the uppermost part of
the Avon Park formed in solutions more saline than the lower part.
The southern cores (19-3, 101, 17-1, 17-3, 11-2, 6-3, and 4-2) all
contain higher Sr+2 concentrations, on a whole, than do the northern cores.
The calcite samples in core 101 average twice the concentration of Sr+2
as do the calcites in the northern cores. The dolomite samples in core 101
contain approximately half the concentration of Sr+2 as do the calcite
samples, implying that both carbonate ., H.- ..'' l1y formed in similar
solutions.
Dolomite samples analyzed for cores 4-2, 6-3, and 11-2 occur in
Miocene deposits which are considerably younger than the Eocene carbonates
of the Ocala Limestone and Avon Park Formation. The Sr +2~ *.. -ations
of dolomites in the Miocene rocks are ',ei.: than 'i,- .: else in the study
area, ranging from less than 200 ppm to greater than 800 ppm. Figure 9
indicates that there is a steady increase in Sr+2 concentration in dolomite
of these cores with depth. It should be noted that small amounts of Sr+2
may have been contributed by associated clay minerals which were sometimes
present in the southern cores.

Sodium

According to Veizer et al. (1977) and Land (1980), sodium interpretations
should be used cautiously because methods of analysis cannot disti-7;'ish
lattice-bound sodi sodium sodium or absorbed in the crystal struc-








-25-


-w -' -

















STRONTIUM CONCENTRATIONS IN ROCK SAMPLES
STRONTIUM CONCENTRATIONS IN ROCK SAMPLES


VERTICAL
SCALE
CONCENTRATIONS OF Sr2 IN PPM IN FEET


Hi8 1
M S200(
e I meomo
g >.t __o o 10 20
SCALE IN MILES
NOTE: NUMBERS IN PARENTHESES DENOTE
DEPTH IN FEET BELOW SURFACE.


Figure 9.







-26-


ture as inclusions of NaCl. Because Na+ is a small ion, it can substitute
with equal facility into Ca or Mg lattice positions (Land and Hoops, 1973).
Thus calcite and dolomite that have formed from similar solutions should
contain similar aQantities of sodium.
Figure 10 depicts sodium distribution in the rocks studied; and those
wells that show incomplete data contain analyses on dolomite samples only.
Again some Na+ may have been contributed by associated clay minerals which
sometimes were present in the southern cores.
All of the calcite samples within the diagram contain sodium in
concentrations less than 500 ppm, which indicates diagenesis in relatively
fresh waters. The dolomite samples show a range of sodium values from
less than 500 ppm to over 2,000 ppm. The increase in sodium values exactly
at the limestone-dolomite boundary in the northern cores and the decrease in
sodium concentrations in the lower regions of the Avon Park Formation in those
cores agrees well with the conclusion drawn from the strontium data that
the uppermost dolomites in the Avon Park formed in solutions more saline than
the lower parts.
The cores adjacent to and south of the Pasco County recharge area
(cores 4-2, 6-3, 11-2, 17-1, and 17-3) contain significantly higher levels
of sodium than the other cores. This is in agreement with the Sr+2
distribution in Figure 9. However, the Sr+2 concentrations showed an increase
with depth in the Miocene deposits, whereas the sodium concentrations
appear to show no predictable alternation in pattern. Zones of high Na+
values (> 2,000 ppm) occur among zones of low Na+ values (< 500 ppm). Also,
the base of core 19-3, which contains abnormally high strontium values, showed
relatively low sodium concentrations in comparison.
It should be noted again that the dolomites in the cores south of Pasco
County recharge occur in rocks younger than those to the north, and are
associated with clay minerals. The sodium content of ancient dolomites
have been found by many authors to be depleted relative to recent dolomites
(Weber, 1964; Behrens and Land, 1972; Fritz and Katz, 1972; Land and Hoops,
1973). Land and Hoops (1973) believed that this was the result of re-equili-
bration with a meteoric reservior. If this is true, then dolomite, originally
equilibrated in saline solutions, could lose weakly held sodium from its
lattice and bring about new equilibrium conditions when flushed by fresh-
water. :. :-, all of the cores within this st'."; have been exposed in
some -'!? to meteoric water, the sodium concentrations analyzed are probably
all less than what ,':', were originally.







-27-


Figure 10.










Cores 21-2 and HS show higher sodium concentrations in comparison to
neighboring cores. of these cores occur where the potentiometric
surface (Fig. 3) is lower than in "'y of the other cores in the study area.
The Ghyben-Herzberg principle (Reichenbaugh, 1972). states that saltwater
intrusion occurs to a ,':.arr degree, and affects rocks closer to the surface,
in areas of low potentiometric head. The strontium data from the dolomites
in core HS correlate with the northern cores, illustrating a less saline
dolomitizing solution for the lower part of the Avon Park. However, the
sodium data for this core does not agree with its northern counterparts.
The zone (< S' ppm) illustrate':. this relationship in the lower parts of
the northern cores is replaced by zones indicating dolomitizing solutions
of much higher salinities in the HS core.
The distribution of sodium in Figure 8 indicates that while it sub-
stantiates the conclusions drawn from strontium analyses in some instances,
in other cases it can lead to ambiguous interpretations. Because trapped
and absorbed sodium in the carbonates can contribute to the total sodium
detected in analysis, used alone it is not a reliable indicator of the
salinity of diagenetic solutions. However, where there is agreement
between sodium distributions and other trace element distributions, more
confident conclusions may be reached regarding the salinity of diagenetic
solutions.


Mole-percent MgCO3

Dolomites are chemically characterized by their mole-percent MgCO3
content. Randazzo and ('-i::.y (1978) -*.::-' ,zed the less-stoichiometric
dolomite in their study as being composed of 44-48 mole-percent MgCO3,
and the more-stoichiometric dolomite 49-51 mole-percent MgCO3. Figure 11
shows the distribution of mole-percent O''3 of some of the dolomite in this
study. The numerical divisions used in the index do not exactly correspond
to the boundaries used by Randazzo and Hickey (1978), but offer a greater
distinction of the degree of dolomite stoichiometry. More-stoichiometric
dolomites generally indicate slow formation in waters less saline, with fewer
competing ions to disrupt the resulting structure (Folk and Land, 1975;
Randazzo and Hickey, 1978). It should be noted that other factors such as
changing Mg/Ca ratios of pore waters, rechargeable sources of Mg+2, crystal
sizes, and .: -::."ility can -- the relationship between dolomite stoichio-
metry and the salini>-. -' the dolomitizing solution (Lumsden and Chimahusky,
1980).






















































































MOLE.PERCENT MGCO, OF DOLOMITE SAMPLES


VERTICAL
SCALE
IN FEET




MOLE-PERCENT MgrCO,
FigC1 i 0 i0 20

B 44 SCALE IN MILES
- NOTE~ NUMBER IN PARENTHESES DENOTE
00PTH IN FEET BELOW SURFACE


Figure 11.


/~Li~


__mlllll_.lllll___ ~1_-_











^ -..: son of the di -.-..- of mo I-- .. -~. MgCO3 with the distribution
of Sr+2 for the cores in the northern portion of the study area (Figs. 9 and
11) shows a marked decrease in Sr+2 occurring at :rr :-.,,")>*-,.tely the same
e': '- as a marked increase in mol-:-r->-* :, :: This inverse distribution
of more-stoichiometric dolomite and low concentrations of Sr+2 suggests the
formation of dolomite in so-'",r..-, less saline than those acting upon the
:'1;,S .. section of the carbonate :' .: The di .--..;1: of Na+ in
these same cores (Fig. 10), .'".:; the limitations cited earlier, further
substantiates the conclusion drawn about the salinity characteristics of
the dolomitizi:-'.-, solutions.

14C

Ten dolomite samples *,.. cores 4-2, 6-3, 19-3, 21-2 and MS were analyzed
for 14C. The 14C content of the :-.r1es was appreciable and would translate
into "ages" of 26,470 38,760 years B.P. Although these "ages" cannot be
interpreted rigorously as geologic ages, they clearly demonstrate that
carbon has been exchanged between dolomite and the atmospheric reservior in
the geologically recent past. Though not conclusive, these data suggest that
the formation, or at least partial recrystallization of dolomite occurred
within the last 30,000 years.



Data were obtained on chloride and sulfate concentrations, and specific
conductance from selected wells. At intervals during drilling of the cores,
water samples were collected from flow'"- water at the base of the wells.
Readings were taken ',-,.: these --..:;.,les. Tight casing prevented waters from
higher areas mixing with waters .~: the base -. the well.
These data can be helpful in defi.;- .; the position of a freshwater-
saltwater ini^..',--: that presently exists. If the present pattern of
circulation within the Floridan :. 'e. deve~~l.j-: during low stands of
Pleistocene seas (Stringfield and t' *-.- 1 -i.), then the present position
of the .'i-. may be related to the distribution pattern of dolomite in
the vt, ..

Chloride

The most reliable indicator saltwater intrusion is the chloride
content of the water 1 1'- ). The chloride ion concentration
in pure seawater is '..* ".-tely 19, mg/1 (V- ....- 1972; ',.-.!. .. 11'" )





-31-


Fresh Florida groundwater, on the other :. .. contains chloride ion concen-
trations less than 50 mg/1 (Wetterhall, 1964). L :It..'biih (1972) used
a chloride concentration of greater than 250 mg/1 to indicate that a well
had contacted the freshwater-saltwater iTsr: Thus Figure 12 provides
a good indication of the salinity of encroaching solutions.
The values from wells 17-1, 17-3, and 21-2 denote the .re; of
highly saline solutions within the aquifer at shallow depths. A dramatic
increase in chlorinity occurs from freshwater concentrations just below
the surface, to concentrations near that of seawater at only 200 feet of
depth. The chlorinity in wells 19-3 and 18-1 are under similar potentiometric
heads as those wells which occur at close distances, .", they constitute
freshwater all the way to their base. Wetterhall (1964) found that intruded
saltwater may not be vertically continuous in the aquifer. He identified
wells within the same area of Hernando County as 19-3, where zonation of the
aquifer allowed freshwater to occur between layers of relatively salty
water.
Wells 11-2, 6-3, and 4-2 all give data which indicate that saltwater
has intruded into the southern region, but not to the degree that it has in
some of the other wells. One reason for the low chloride concentrations in
these wells is that the wells occur in the confining layers of low permeabilib
overlying the aquifer. Water is not able to flow as freely into these
formations either from the sea or from recharge as it can in the carbonate
aquifer units which lie below. Another explanation, concerning the height
of the potentiometric surface, can explain the distribution of chloride
concentrations in all of the wells (except 19-3, 124, 97, and 18-1 -
for which the previous interpretation will hold). The potentiometric levels
in wells 17-1, 17-3, 21-2, and 124 are less than 10 feet (above mean sea
level), while in wells 11-2, 6-3, and 4-2 the heights range from approximately
25-35 feet (Fig. 3). Assuming uniform permeability, the '.I;--.-I:rzberg
principle denotes that the freshwater-saltwater ii ~, .:. should lie at Is
than 400 feet in those northern wells, while it should occur between ,..)-
1,400 feet in the southern wells. Allowing for the low permeability of the
Hawthorn Formation, saltwater encroachment would still not be expected to
be as great at the depths penetrated by the southern ,=1s, as at the ::; .
penetrated by those wells north of the Pasco County recharge area.



















































101
CHLORIDE CONTENT OF GROUNDWATER (w360


17-3
(52 1 )




rI





'I


"' MS
(49')


,j
I
.. :.

1-~ ~. ~-. i ~. ~ ~
'':'''

I i

-~-
.:-~iC
'5. .





I


L: ''
''
~


~
; ;~.
i?
L1_' I'
~6~ >lo.ooo


S 4-2


- --M_ n
'' ---- -














VERTICAL
SCALE
IN FEET






10 20
_. ....LES

i.urF. h..i,:..: il PARENTHESES DENOTE
DEPTH IN FEET BELOW SURFACE,


-- --


S .e 12.









Conductivity

In his geohydrologic reconnaissance of Pasco and southern Hernando
Counties, Wetterhall (1964) found chloride content and specific conductance
(conductivity) of the waters to be generally related. Fi.i:Le 13 shows
conductivity distribution in the waters from the same wells as F:.. 12.
These values correlate extremely well with the chloride values, .:,r. -!.ivity
measurements for the waters (in microohms per centimeter) are approximately
an order of magnitude greater than their .c~ f chloride concentrations
(in mg/1). Conductivity measurements rei.\. the conclusions conce~:',i.!
salinity of the water inferred from the chloride data. Al ,'.!Y the conductiviv
of seawater is temperature sensitive (Duxbury, 1971), the temperatures in the
waters of these wells all range within a few degrees of 25C (Plummer and
Back, 1980).


Sulfate

Sulfate ion concentrations in pure seawater have been recorded at
2,511 mg/l (Gross, 1972) and 2,649 mg/l (Duxbury, 1971). Wells 6-3 and
4-2 contain low sulfate values (Fig. 14) in the Hawthorn Formation (less
than 500 mg/1), as expected. Below the Hawthorn, in the Tampa and Suwannee
where aquifer waters flow, an increase in sulfate concentrations occur,
reflecting an advancing freshwater-saltwater interface. Wells 17-1, 17-3,
97, and 124 give high sulfate concentrations. Overall, the sulfate values
are in agreement with the model of saltwater intrusion originally deduced
from the chloride data.



DISCUSSION

In attempting to formulate a mineralogical model of the Floridan ,i. ,.'i
several factors must be considered: (1) the age of the carbonate units
and their original lithologies and environments; (2) the mineralogical
and physical evolution of the aquifer; (3) the nature and variations of the
hydrologic regimes to which the aquifer rocks were ejectedted ; and (4) the
distribution of the various geochemical and hydrologic h.:i.-:: ..











































CONDUCTIVITY OF GROUNDWATER


II



Iii






.-..,"
,." '




; -4

.. .ir '^ ,,!:l< *'
,,, ; '^ / -*

! : i'i i .


101
(3G0I)




17-3
11-2




,,:, ----.- I




ILI
J ,


JI' I

-- -- -
'.UT5


VERTICAL
SCALE
IN FEET



rORTH 200
0 S10 20
XCALE IN MILES
-S: .IMBERS IN PARENTHESES DENOTE
..CIT,, ,. FEET BELOW SURFACE.


7 -,e 13.


I


---------


i"~
,, hi



~~:~i ..., ..- .....
'~~? ~ -
;.~ :;.



































































































Figure 14.


~ ~_I







-36-


In order to explain the water-rock interactions that led to the present
minera ..-ical distribution within the study area, the evolution of the
aquifer must first be considered. Fol'>.,ing deposition and early dia-
genesis of the various carbonate units, a lowering of sea level exposed the
sediments to meteoric -.-.:.'-,-.ter solutions. v.. carbonate sediments
first :l-r.e- from the marine environment and .,,:... flushing by freshwater,
the dissolved solids content of the phreatic waters decrease, with the major
ions in solution : *-ing .--: Na+, C1, '.+2, and S04-2 to Ca+2 and HCO~3.
Meteoric waters are enriched in a'-::'...-,'-ived CO2 and are undersaturated
with .*-: to CaCO3. Therefore, the 7... :1,.A-J- a will tend to dissolve
CaC03. In this ",-. .:. fractures and :.,:..;. spaces in the rocks were enlarged
as CaCO3 was dissolved, init~~'-." the fo ...-t.' k of the aquifer.
Sea level fluctuated many times since the deposition of these carbonate
rocks. Each time the sea level was lowered, meteoric groundwaters enriched
in CO2 gas enlarged and extended existing flow-networks by a process that
led to further physical development of the aquifer. Calcite dissolution is
a relatively fast process in C02-enriched groundwaters. The flow-network
within the aquifer probably developed by enlarging and extending pre-existing
channels, rather than by developing new ones.
In western peninsular Florida, Hanshaw et al. (1971) found that by the
time groundwater reaches the zone of mixing with seawater, it is saturated
with respect to calcium carbonate. Seawater in the coastal mixing zone
may also be saturated with "-'-.-*:t to calcium carbonate. However, the seawater
would be saturated under a d'i-:e"r, partial pressure of CO2 gas than the
groundwater. The mixing of the two .'... both saturated with respect to
calcium carbonate, but havi -.. -~.= =i-t. partial .-... O.! s of CO2, will result
in a range of mixed waters undersaturated with .-:. : to calcium carbonate.
It is this control on calcite saturation that al' calcite to be dissolved
in the freshwater-saltwater i r.,".;-:o. 's: et al. (1979) described this
process occurring today in the brackish zone of Xel Ha lagoon in Mexico.
Another condition ., .-:.r. .ng dolomite formation in the freshwater-
saltwater '"i.'': .= is a ,.--."'ciently h.;. Mg+2/Ca+2 ratio. Hanshaw et al.
(1971) have shown that the ". :--',.-. equilibrium (calcite-dolomite-solution)
occurs at a Mg+2/Ca+2 ratio of I-., :-~.'ely 1 in the Floridan aquifer.
Dolomitization can occur provided -i'.:icient time and Mg+2 ions are available.
The Mg+2,-..+2 ratio is .-.:- -. than I in the mixing zone of the Floridan
aquifer in peninsular oride ( and 1977). Thus the chemical
and kinetic *'.:. -rements are met for do --itization to occur within the
freshwater-saltwater mixi .- zone of the study area.





-37-


One of the most significant observations made in i.-' ...~'f;'., the data
presented in the panel diagrams was that the Avon -'' Formation is a.,,.:;.t
completely dolomitized in the study area, while the Ocala and Suwannee are
mostly limestones. There must have been some factor (or factors) related to
the lithologic character of the Avon Park that resulted in its -.,[,..- :ill
dolomitization.
Murray and Lucia (1967) described water-controlled dolomitization and
rock-controlled dolomitization. In the case of water-controlled selectivity,
the distribution of dolomite is directly related to the availability and
access of the dolomitizing solution. They stated that dolomite may form in
carbonate rocks that underlie supratidal sediments, as a result of the pro-
duction of hypersaline brines on the supratidal ,d..lat. ;',,,c', dolomite
may be absent in carbonates of similar lithology that do not underlie supra-
tidal sediments.
Rock-controlled dolomitization is directly related to the physical and
chemical characteristics of the rock at the time of dolomitization. In their
study, Murray and Lucia (1967) found a preference for dolomite to replace lime
muds. A chemical factor they believed might be of importance involved the
relative solubilities of the particles. They stated that because recent lime
muds contain a higher percentage of aragonite than most other carbonate
deposits, the difference in solubility between calcite and aragonite may be
sufficient to cause some selectivity. They concluded that another likely
physical factor involved the small particle size. The greater surface area
of the micrites would enable more dissolution to occur and provide a greater
opportunity for replacement by dolomite.
Similar rock-selective dolomitization of mudstones has been is::itL;-d
by Choquette and Steinen (1980), of micrites by Land (1973), of wackestones
by Inden and Koehn (1979), and of siliceous clays by J. l idt (I:5). Land
(1973) demonstrated the near contemporaneity of aragonite dissolution and
dolomitization of the micrites of the '-op Gate r',..tion of Jamaica. Sibley
(1980) concluded that high-magnesium calcite and .-a'y ite are susceptable
to dolomitization, while low-magnesium calcite is not in his study of
rock-controlled dolomitization on Bonaire. Sibley stated that if the
metastable aragonite and high-magnesium calcite can remain preserved until
exposure to the freshwater-saltwater mixing zone, then they will be selectively
dolomitized.










The dolomitized "- .. *: in the .-:'* area consists predominately of
mudstones and wackestones which were initially .'-:r-.--.ted as aragonite lime
muds in supratidal and intertidal .: while the .',,Y~aie and Ocala con-
sist to a greater extent of highly 7.il'- ._... packstones and grainstones.
If the dolomite selectivi- were water-controlled, it would probably have
been restricted to the extent .-" the mix-':. zone. -i'-.,.'r, Pleistocene
terraces in ;:'o.*.*'' 'p Florida t. : that sea level once stood 270 feet
higher than its present level (Cooke, 1945). As sea level fell, the mixing
zone would have --.-' "''.. .1' :-la and ".r.o.- .' rocks. Because those
units are not do'I.:-'itized in the '. area, dolomite selectivity was more
likely related to the ''--ained, high surface area nature of the lime
muds in the Avon ". The formation remained unlithified until it was
exposed to a meteoric water ".':- ("*.*.*" .- et al., 1977). This supports
the theory that the ..:- micrites remained as the metastable, more
soluble aragonite phase until their ;. .-..* :. to freshwater solutions. Direct
dolomitization of arp:c., te precursors would then have been an additional
rock-selective factor, as ;-:,,: .d byi:'.' and Lucia (1967) and Sibley
(1980).
The parameters presented in the -r :1'.irams support a mechanism of
dolomitization by a "-.:.:..- ter-saltwater mixing zone for the Avon Park
lithofacies in the .... area. The parameters also show that the composition
and salinity of the waters in the mixi zone varied and resulted in the
formation of dolomite with a ': .:. ; mo' *-,- cent ,:C''-.
The mole-percent MgCO3 d*. ". (Fig. 11) indicates that dolomite of a
more-stoichiometric form developed in the lower reaches of the Avon Park
within the study area. This would have been related to a freshwater-saltwater
mixing zone of lower salinity than that which formed the less-stoichiometric
dolomites lying above. Sarver (19T ) ..,o .:"., that the less-stoichiometric
dolomites of the ~,"; '? formed in a saline, coastal mixing zone that moved
laterally in ,'.-:F.,e to sea level uctuations. The more-stoichiometric
dolomites, he believed, formed in a less-saline, inland mixing zone which
moved vertically, mainly in response to atmospheric conditions and ground-
water recharge. It is --.---- here that sea level changes could also cause
considerable vertical movement the coastal mixing zone. During sea level
regressions the coastline would have moved farther out to sea, exposing new
areas to freshwater *. s added ';..- :- plus the movement of
inland :;.'.'.' -'ters fl1' beneath the : to a discharge point farther
westward, could be -ient to caus7 the seaward migration of the mixing









zone. Thus, the distribution of less-stoichiometric dolomite over
more-stoichiometric dolomite could be a resu:, a mixing zone controlled
by fluctuations in sea level.
Strontium and sodium data (Figs. 9 and 10) .*1;.-..':, additional evidence
for the model described. The strontium content of the more-stoichiometric
dolomites in the Avon Park averages less than _."i' .. while the concen-
trations in the less-stoichiometric do -....,ites range between 200-400 ppm,
indicating a more saline dolomitizing solution. It was also ,-.i..:. that the
Sr+2 concentration in the Avon -';..; dolomite, overall, is more than 50
percent higher than that of the calcite. As stated earlier, ::.-ver (1 )
believed that this was the result ,. do.,....itization .' ..:.; taken place in
a more saline environment than calcite ..-,.. ism. Another explanation,
however, may be related to the minerall:] of the precursor sediments. Veizer
and Demovic (1974) obtained hyih er .:+2 concentrations in sediments that
were originally aragonite muds than in other sediment T :... The various
carbonate mineral phases occur in two -':c&.. crystal classes: (1) rhombo-
hedral(calcite and dolomite), and (2) orthorhambic ( ;-..:eite). The ortho-
rhombic structure is larger than the rhombohedral :structure, and ..::..... :.
large cations, such as Sr+2, fit more readily into the orthorhombic aragonite
structure than in the rhombohedral calcite structure (Hanshaw et al., 1:;).
If the majority of the Avon Park carbonates remained as .. Ujo,'ite until
dolomitization, as proposed, then a larger concentration of Sr+2 would have
been available for incorporation into the dolomite lattice than dolomitization
of calcite would have allowed.
Sodium concentrations are less than 500 ppm for the calcite samples of
the Ocala analyzed. Thus, the sodium data, in ...:. -. with the other data,
denote that the calcites of the Ocala Limestone within this .. .,. area were
stabilized to low-magnesium calcite by a relatively .~ ..ter solution.
The more-stoichiometric dolomites also reveal low Na'' concentrations (< 500 .
while the less-stoichiometric dolomites analyze between '..1,000 .... with
a layer of higher concentrations (1,.9"i-1,500 ;.. occurr-i,: within it. The
sodium data, therefore, agrees with the strontium and mol-- ..... :.,...
data concerning the paleoenvironmental i.t;.,. ,.A~.tion ., the do itizi.,
solutions.
Rock-selective dolomitization is also indicated '.. the .. ..... its
of the study area. The Hawthorn Formation in western peninsular Florida










contains beds :. clay, sand, and carbonates i. u.' ",-:,i;. with one another
(Weaver and 1977), In peninsular Florida the uppermost part of the
Tampa Formation contains green-c'.-.. interbedded with carbonates, while
the lower part contains only limestone *; --;.: and >i,, 1977). X-ray
analyses done in this -'r- (cores 4-2, 6-3, and 11-2) indicate that dolomite
occurs only in the .;-.: and upper Tampa. T -.: '- ':, dolomite selectivity
here is :., :-...ly related to the occurrence : clay minerals in the formations.
Dolomite selectivity has been related to clay minerals by Schmidt (1965)
in Germany; and 'le (1. T:) be- i.. that clay minerals may contribute
Mg+2 ions needed -"; dolomitization.
If dolomitization of the Hawthorn and .--:. T---,: occurred in the
freshwater-saltwateri : then the older units lying below should
have been dolomitized. .; in, some type of lithologic control must have
been exerted by the :'- ".' sediments.
Other factors related to do:-.:ite selectivity may be the fine grain
size of the carbonates, and the low -;:i.:,ility of the sediments. Most
of the dolomites ::l* .:; in the .- sediments are clay-sized and,
together with the siliceous cl6. minerals, account for the low permeability
of various parts within these units. Zones of lower permeability in the
Hawthorn and 'r"e;- Tampa result in longer residence time for the groundwater
solutions. This -.- o ..r would the ,:..-.*tion of dolomite, because the
I-:'.?'" of time that the dolomitizing so. :'. remains in contact with the
precursor sediment is an .i....: control over whether dolomitization will
occur. These factors ...".ly enhanced the secondary formation and growth
of dolomite.
The trace element data (F'-.- 9 and 10) for the Miocene dolomites are
markedly di".-'-' .: than .' the Eocene dolomites. Strontium and sodium
concentrations in the c-- :. rn and .- T ,-:= dolomites are both higher
and vary more with depth than in the ...1 Park dolomites. This is likely
a result of the .-. of the rocks and the flow pattern within the aquifer.
Given enough time, trace elements can be exorcized from the dolomite by
ub..',,;renr. recrystallizations (Kinsman, I"1-'),. ;"-:e- ive recrystallizations
would lead to lower va :. .:- Sr+2 in the rock. The fact that the Na+
content of Holocene dolomites is much higher than the Na' content in ancient
dolomites supports the bel' that .+ is also selectivity removed during
:-.:.i.-:... t :.lizations (Land, "I..^). The low concentrations of Sr+2
and Na+ in the '. dolomi'teh 1. the -.' graphic evidence reported
by 1 -- ~ and ',-- ~19--) that extensive .- l '-.llization occurred.









The higher trace element concentrations in the r...-. ."- -its -: '.':,.ect
the relative youthfulness of those carbona-.-, with a lesser .;e
recrystallization having occurred.
The distribution of the hydrogeochemical (....t-. (Figs. 1-. 13, and
14) reveal that the freshwater-saltwater i .... does not encroach in a
uniform manner upon the coastal lirq?.; Wells .'2, 17-1, 17-3, and
124 indicate that saltwater is close to the .-,.r. in their particular areas,
but the remaining-area is currently being flushed relatively w.. .water
solutions. The discrepancies in the he"e'.l that the saltwater has attained
in those wells that penetrate the Ocala and .. are a result !- the
physical structure and flow network of the i..,fer. Dissolution conduits
enable seawater to penetrate the .:i;l. in some places, while elsewhere the
channels contain freshwater under .?r;';,
The greater abundance of impermeable ,..',= sediments results in a
higher potentiometric surface than in the other units (Fig. 3). As a result,
the Ghyben-Herzberg principle predicts that the freshwater-saltwater inter-
face should be at a lower depth under the ;.;. ., ;, :... The '.,l' -logic
parameters signify this, because nowhere in the cores -,' the southern part
of the study area does the saltwater lens rise to the level that it does in
places in the northern part.
Slow percolation through the Hawthorn Formation is i.flected in the gradual
increases in salinity with depth. On the other :,!..'. the development .., the
aquifer, itself, produced a network of large dissolution channels in which
the groundwater lens and the saltwater lens can enter into each other's
realm producing an irregular boundary along the coast. T..;-le e, the majority
of the dolomitization of the Avon Park in the '.r.... area took place :. ::
times when sea level was higher than it is ?.. and seawater was able to
infiltrate the aquifer through channels that are currently in the ':- =itc
meteoric zone. Also, as a result, the .' -". -. Formation was ,.ected to
greater interaction with the coastal and inland mixing environments than
the Hawthorn and upper Tampa.



THE MODEL-A S ,-.. ".' OF
DATA AND i. ; ..' ..:

Deposition of carbonate sediments occurred in the .;.,..'-idal, inter-
tidal, and subtidal environments of a carbonate bank along the west coast
of peninsular Florida during the Terti :i..',: T:. .-...* tion and extent











of these environments changed with fluctuations in sea level, and uplift
and down% -.ing of the Florida P1.. ; ;. '5.;idal sediments were likely
to have been originally ..'--, -.J.:.-.?um calcite predominantly, while inter-
tidal and -..::. ''-1 ::- .;: were mostly aragonite. Lime muds were more
common in the supratidal and deep subtidal zones. The higher energy environ-
ment in which mud was winnowed deve. *.:,: high primary interparticle porosity.
A multivariate i-.: anal- is, utilizing multidimensional scaling,
revealed several trends that correspond well with interpretations based on
-".'. :-' :. 'c and ...-- :.':.:; ic examinations of three cores representing the
.; : ;-; and .a:-,la carbonate sequences. MDS was used to determine environ-
mental characteristics normal to the shoreline for these sediments. The
analysis revealed the r .--c:. in the carbonate bank depositional environments.
This was depicted by two hydrodynamic curves produced by MDS which showed
variations from ':'...., dolomitic, :'.,it.-';, supratidal sediment clusters
to a shallow subtidal, 'V''':h ;, sediment cluster.
During the :''!i:.- carbonate muds and siliceous clays were deposited in
brackish water lagoons, along with sands and phosphates. This resulted in
the formation of a 1l"' of low permeability which today defines the
upper boundary of the Floridan ,*.u. The Avon Park Formation probably
remained unli:-.; '-: until exposure to a meteoric water regime. Fine-grained
dolomite ;.':- .1s .-' -:-. in the supratidal lithofacies from aragonite mud
precursors.
During periods of subaerial ....:.i-- of the marine sediments, flushing
by meteoric waters occurred. The unstable high-magnesium calcite sediments
inverted to low-magnesium calcite, accompanied by the inversion of aragonite
sediments. Lime mud inverted ':. ,...i;,;-:.:.d neomorphism to microspar, and in
doing so, aided in lithification. Shell F-.,!-,mts and other sediments pro-
bably underwent dissolution-reprecipitation and recrystallization, which
destroyed original textures and resulted in the formation of pseudospar.
As the unstable minerals were dissolved, their cation and anion chemistry
were contributed to the solution. The newly formed carbonates had lower
concentrations of Sr+2, :: +, and .+2
The dissolution-reprecipitation process was also responsible for the
formation of sparry calcite cement. In the vadose meteoric zone, finely
crystalline cement was ..-. pitted from meniscus solutions in interparticle
pore ',;e. and at grain contacts. .'.. rhombohedral sparry crystals
slowly precipitated in in the phreatic meteoric zone.





-43-


While groundwater was responsible .: ,.- .-ipitating cements, it was
also a major factor in the production of secondary 7,..: .ties. >, ial
dissolution of fossil tests resulted in moldic ..:; ., and vugs. ...
and wackestones developed less secondary porosil.- than ;--. :tones and
grainstones because of their fewer allochems and lower : -.ility.
During periods when sea level was lowered, 1,; t.r:ter dissolution the
limestone enlarged existing pore spaces as the ?,..;I.:. <-. .. :. .p: i,';?
towards discharge out into the Gulf of Mexico. Succeed ; marine trans-
gressions resulted in inland migration of the fh:ii.,.ter-saltwater i:::.
Selective dolomitization took place at the ',.:.; -..,ter-saltwater 2.- :.
in the Avon Park, Hawthorn, and upper Tampa Formation. The dolomite selectivity
was related to (1) the fine grain size of the sediments with their high
surface area/volume ratio; (2) the occurrence of earlier formed dolomite;
(3) the precursor carbonate being aragonite; and (4) the occurrence of clay
minerals.
Dolomitization proceeded in the freshwater saltwater mixing zone
because (1) mixing freshwater and saltwater can result in a zone where the
waters become undersaturated with respect to calcite, yet remain super-
saturated with respect to dolomite; (2) the salinity of the mixing zone
is lower than it is in seawater, therefore there is less foreign ion
competition inhibiting dolomite formation; and (3) the Mg+2/Ca+2 ratio
remains well above the level of unity required for dolomitization.
Where the dolomitized precursor sediments were calcite, the dolomite
crystals grew by coalescive neomorphism. Where penecontemporaneous dolomite
crystals were already present in supratidal lithofacies, the fine crystals
underwent porphyroid neomorphis, resulting in larger, more t-.... *-7:.,;U.!
crystals. In both cases dolomitization proceeded by a process of dis-
solution-reprecipitation, accompanied by some loss of Sr+2 and Na the amount
depending upon the mineralogy of the precursor ...Adi:i,,t the role of clay
minerals as a contributor of trace elements, the salinity of the dolomitizing
solution, and the duration of time in which sediments were sije...:.-: to
dolomitizing solutions.
The aggrading neomorphic growth of dolomite crystals resulted in the
redistribution of many small pores into fewer, but larger, ;,.;. ..,:.
although the porosity may not have appreciably -!,. ; .::..ility could
have greatly increased. In the less saline reaches of the mixing zone,
more-stoichiometric dolomite formed slowly without the in-e .;_ : of many
foreign ions. This resulted in large, well-developed, rhombohedral crystals











with low Sr+2 and Na+ concentrations. Less-stoichiometric dolomite, with
associated higher Sr+2 and Na+ values *:-. -d more rapidly in the more saline
areas of the mixing zone.
Diagenesis of carbonate rocks is a continuous process. Both groundwater
and saltwater solutions in contact with the rocks are constantly exchanging
ions with the rock in an i:..:.: ': to :;'.. the most stable mineralogical
conditions : --'ble. The "'-o'--tic ...r'--.-..ic :.n-:"tr. of dolomite crystals
to larger dolomite rhombs is added evidence of this continuing process
occurring in dolomites. The '.'. ter-saltwater i'~tr;face is probably the
most .. Lc of the environments in which '*,;- occur. 14C data suggest
-:n t.-on or recrystallization of dolomite occurred in the last 30,000 years.
Each time that sea : -.1 --. .- moved the mixing zone back into dolomitized
units, recrystallization of the rocks could have occurred, with an accompanying
trace element '..rletion. T'- ,-.:-, it is reasonable to assume that dolomites
which :-tiin. 'lly formed in the .-.**'-.:" during earlier sea level fluctuations
are still striving towards more-stoichiometric conditions in the present
freshwater-saltwater mixing zone. Carbonate systems tend towards equilibrium
with age and time (Lumsden and ..-, .i -, 1980), and the expected trend for
dolomite is towards a more nearly stoichiometric compound.





I am deeply grateful for the many suggestions, interpretations, points
of guidance, and cooperation :..;r, -1.. by -: -I Spangler, Paul Mueller,
James Eades, and i'..-1 Wahl of the ;:!versity of Florida, to William Back
of the United States Geological S- ---., to Tom Scott of the Florida Bureau
of Geology, and to ,. New and Kim :---.'-, of the Southwest Florida Water
Management District. trial l acknowledgement must be made to Jon Bloom who
compiled and synthesized much of the data in this report while a graduate
student under my direction. The i -,-1ion provided in this paper must be
credited to a large p :- to his dedication and professional effort to the
tasks confronting him. The work *:.'rE which this report is based was sup-
ported in ;:, by funds r.- ..ded ., the United States Department of the Interior
as authorized under the Water -.. i.. and r.--elopment Act of 1978.





-45-


REFERENCES CIT)
Applin, P.L., and E. R. Applin, 1944, : .-' :4 -..1 ..;. : : stratigr.:
and structure of Florida and southern Georgia: .- Bull., v.
p. 1673-1753.
Back, W., B. B. Hanshaw, T. E. Pyle, L. N. Plummer, and A. E. Weidie,
1979, Geochemical significance of : :...':.. ter di~. -..: carbonate
solution to the formation of Cael:. Xel Ha, .,..: ;:.., .. co:
Water Resources Research, v. 15, p. 1521-1':-
Badiozamani, K., 1973, The dorag dolomitization model, .,:rlication to the
middle Ordovician of Wisconsin: Jour. Sed. .*'.. : ', v. '.. p. -. 984,
Behrens, E. W., and L. S. Land, 1972, Subtidal Holocene dolomite, ..-,-.n
Bay, Texas: Jour. Sed. Petroi,;,j', v. 42, p. 155-161,
Bloom, J. I., 1982, A mineralogical model of the idan-aquifer in the
Southwest Florida Water Management ii::.., ct: : -.. 's thesis, Univ.
of Florida, 140 p.
Chen, C. S., 1965, The regional lithostrati.; .- anal, -is of -:leocene
and Eocene rocks of Florida: Florida Geol. ,..:, 11 45, 105 p.
Choquette, P. W., and R. P. Steinen, 19'J:, Mississi,:-.:.- l.:..-.;.. ratidal dolo-
mite, Ste. Genevieve Limestone, Illinois Basin, evidence '- mixed-
water dolomitization, in Concepts and models of dolomitization:
SEPM Spec. Pub. 28, p. 163-196.
Cooke, C. W., 1945, Geology of Florida: Florida Geol. Survey Bull. 29, p
Donn, W., W. Ferrand, and M. Ewing, 1962, Pleistocene ice volumes and sea
level lowering: Jour. Geology, v. 70, p. :i-214.
Duxbury, A. C., 1971, The' earth and its oceans: Readi'-,, I -.::. .,}.,.:tts,
Addison-Wes1ey Pub. Co., 381 p.
Ekdale, A. A., S. F. Ekdale, and J. L, Wilson, Iv-~ i.; :,ical analysis
of carbonate microfacies in the Cupido Limestone (Neocomian-Aptian),
Coahuila, Mexico: Jour. Sed. Petrology, v. 46, p. 362--'
Fenk, E. M., 1979, Sedimentology and strati:;.'h.. of middle and upper '.
carbonate rocks, Lake, Hernando and Levy Counties, Florida: Master's
thesis, Univ. of Florida, 133 p.
Fetter, C. W., 1980, Applied hydrogeology: Columbus, z-'., Charles ',.:.'1
Pub. Co., 488 p.
Folk, R. L., 1965, Some aspects of recrystallization in ancient limestones,
in Dolomitization and limestone diage:-s.is: :. *... o Pub. 13,
p. 14-48.
Folk, R. L., and L. S. Land, 1975, Mg/Ca ratio and salinit.,: two controls
over crystallization of dolomite: A Bu', Bull., v. 59, p. :i-.:.,
Friedman, G. M., 1964, Early diagenesis and lithification in carbonate
sediments: Jour. Sed. Petrology, v. 34, p. 777-'-3.
Fritz, P., and A. Katz, 1972, The sodium distribution of dolomite .. '. s:
Chem. Geology, v. 10, p. 237-244.
Ginsburg, R. N., 1957, Early diagenesis and lithification in shallow-water
carbonate sediments in south Florida, in : .- :. -1 .:. ::; of carbonate
deposition: SEPM Spec. Pub. 5, p. 80-100.
Gross, M. G., 1972, Oceanography, a view of the Earth: *^. .lewood Cl--.'
New Jersey, Prentice-Hall Pub. Co., .:.;: p.






-46-

-i-.- .', B. B., and W. :.- 197 :.-.or .--.. :. ical processes in the evolution
of carbonate-aqui .-.-: Jour ;-.'. -logy, v. 43, p. 287-312.
Hanshaw, B. B., W. =. and R. G. Dieke, 1971, A geochemical hypothesis for
dololitization '* ground water: Econ. -.C-'.-, v. 66, p. 710-724.
., B. J. R. '.: A cluster anal i. -'-.-:.. : on of middle
.,..:.' ician b.: i.es, southern : .'ie Mountains: Canadian Jour.
Earth Sci., v. 17, p. 1377-1.-
Hickey, E. W., 1976, Sedimento`-:. and dolomitization in Eocene carbonate
rocks, Gilchrist and lionn Counties, Florida: Master's thesis, Univ.
of Florida, 74 p.
T.'- e, J. and E. G. -- ,i -'ation of modern -.-.;i?., .- carbonate
sediments, in Cl- action Carbonate .'..: AAPG Mem. 1, p. 253-272.
..'-P, .R. F., and H. '-'- '. 19-^, Dolomitization of offshore carbonate
'::'"-ts in Harmmett Shale, lower Cretaceous, Texas: (abs.) AAPG Bull.,
v. 63, p.
L., :. L Oil .- -3 : cities in Florida: Oil and Gas Jour., v. 53,
p. 370-375.
Kahle, C. F., 1.', '::- ,le roles :^ clay minerals in the formation of
dolomite: ;-.. Petrology, v. '", p. *1..*.. :.
Kinsman, D. J. J., 1969, Interpretation of Sr" concentrations in carbonate
minerals and rocks: Jour. Sed. '..:. ;-,'..', v. 39, p. 486-508.
Kohout, F. A., 1965, A ,.:..'_!- is concerning cyclic flow of salt water related
to .se.,,ieil heat in the Floridan aquifer: N.Y. Acad. Sci., Trans.,
v. 28, p. .:-.:71.
Kohout, F. A., 1'37, ?.. .-. -'-water flow and the geothermal regime of the
Floridan Plateau, in :.-ium on *..:-...ic history of the Gulf of
.K ,.P.:. Caribbean "-. .- a Gulf ...:- Assoc. Geol. Soc. Trans.,
v. 17, p. -.'-354.
Kruskal, J. B., 1'^-., Multidimensional scaling by optimizing goodness of
fit to a nonmetric .:. -~: ;.:: *. *;.. .:- 'ica, v. 29, p. 1-27.
Kruskal, J. B., 1?''' :-, '.:.. ;:.'-. multidimensional scaling: a numerical
method: -..- ca, v. ', p. 11. 129,
Land, L. S., 19 ." :-.tic vs. vadose meteoric *:': s.:.'sis of limestone,
evidence *.-:-. a ..I water tab .: Sedimentology, v. 14, p. 175-185.
Land, L. S., 1973, :".. do:. ,itization of middle Pleistocene
,-fi. by meteoric water, north Jamaica: '. 11. Marine Sci., v. 23,
p. 64-' .
Land, L. S., 10'', The i- .c and trace element geochemistry of dolomite,
the state of the art, in .::. ots and models of dolomitization: SEPM
Spec -.'. ".. p. 87-1-.-
Land, L. S., and S. :-;.'tein, 1 Late 'istocene diagenesis and dolomitiza-
tion, north ---ica: :--.'-imentoi '1 v. 14, p. 187-200.
1.;:.., L. S., and G. K. K. -- 19.7.. 'ium in carbonate sediments and rocks,
a possible index to the salini:- of di.--.., -:ic solutions: Jour. Sed.
Petrology, v. .", p. 61--~17.
Liu, J. S., 1973 Petrol l. 1ast int, %-;, and Duette drill
cores, Hill -" and tee Counies, Florida: Master's thesis,
:'or 1 p.





-47-


Lumsden, D. N., and J. S. -h.!..,- .y, 1- ',.. .-:lationshi: between dolomite
non-stoichiometry and carbonate faces parameter, in r...;-:.'.. and models
of dolomitization: SEPM Spec. Pub. 28, p. :1-.1 :
Metrin, D. B., 1979, Geochemical significance of selected ions in .... .
carbonate rocks of peninsular Florida: Master's thesis, Univ. of
Florida, 64 p.
Murray, R. C., and F. J. Lucia, 1967, Cause and control of dolomite distri-
bution by rock selectivity: Geol. Soc. Amer. Bu11., v. 78, p. 21-36.
Parks, J. M., 1966, Cluster analysis applied to multivariate geo",',ic
problems: Jour. Geology, v. 74, p. 703-715.
Plummer, L. N., and W. Back, 1980, The mass balance -. .. : application
to interpreting the chemical evolution of : :; .. .- teamss : ..
Jour. Sci., v. 280, p. 130-142.
Puri, H. S., 1957, Stratigraphy and zonation of the Ocala Group: Florida
Geol. Survey Bull. 38, 248 p.
Randazzo, A. F., 1976, Petrographic and c.'.c1 .,&i. logic model of -e.ifer
limestones in Florida: Florida Water Resources Research Center, -...
No. 35, 51 p.
Randazzo, A. F., 1980, Geohydrologic model of the -'.ridan ,..i.--- in the
Southwest Florida Water Management District: Florida Water '-:-..:.
Research Center, Pub. No. 46, 79 p.
Randazzo, A. F., and E. W. Hickey, 1978, Dolomitization in the Floridan
aquifer: Amer. Jour. Sci., v. 278, p. 1177-1184.
Randazzo, A. F., G. C. Stone, and H. C. Saroop, 1977, Diagenesis of middle
and upper Eocene carbonate shoreline :equv...c:, central Florida:
AAPG Bull., v. 61, p. 492-503.
Reichenbaugh, R. C., 1972, Sea-Water intrusion in the upper o-.- of the
Floridan aquifer in coastal Pasco County, Florida, 1.'-.- Florida Bur.
Geology Map Series 47.
Rowell, A. J., D. J. McBride, and A. R. .;:mer, 1973, n.;.,;- tative : .'.i, of
Trempealearin (latest Cambrian) trilobite distribution in v-"..'-,i America:
Geol. Soc. America Bull., v. 84, p. '4. 1-~ ': 1..
Runnells, D. D., 1969, Diagenesis, chemical sediments, and the mixi... ..
natural waters: Jour. Sed. Petrology, v. 39, p. 1188-1201.
Saroop, H. C., 1974, Sedimentology, paleoecology and ....i. ..:is -r middle
and upper Eocene carbonate shoreline :e;.;::.": .., C -:'1 Florida:
Master's thesis, Univ. of Florida, 165 p.
Sarver, T. J., 1978, Geochemical analysis of selected .... .- carbonate
rocks of peninsular Florida: Master's thesis, i.' of Florida, 77 p.
Schmidt, V., 1965, Facies, diagenesis, and related reservoir properties in
the Gigas beds (upper Jurassic), northwestern Germany, in Dolomitization
and limestone diagenesis: SEPM Spec. Pub. 13, p. 124-168.
Sharpe, C. L., 1980, Sedimentological interpretation ,., Tertiary carbonate
rocks from west central Florida: Master's thesis, '..:v. of ..ida,
170 p.










Sibley, D. F., 1980, Climatic control of dolomitization, Seroe Domi Formation
(Pliocene), Bonaire, N.A., in Concepts and models of dolomitization:
....-' .. ,:-.' ':.. p. 247-258.
Smosna, R. and S. M. Warshauer, 1979, A scheme for multivariate analysis in
carbonate :':.;.:-, with an example from the Silurian Tonoloway Limestone:
Jour. Sed. ;:-,:logy, v. 49, p. 257-271.
Sokal, R. R. and P. H. A. Sneath, 1963, The Principles of Numerical Taxonomy:
San Francisco, W. H. Freeman and Co., 359 p.
Steinen, R. P., and R. K. Matthews, 1973, Phreatic vs. vadose diagenesis,
stratigraphy and mineralogy of a cored borehole in Barbados, W. I.:
Jour. Sed. '-.:.1l:", v. 43, p. 1012-1020.
Stone, G. C., 1975, Paleoecology and diagenesis of the Avon Park Formation,
Gulf : :", l,:' ''" ife '' :-. ': ... P Levy County, Florida: Master's
thesis, '"'.. of Florida, 128 p.
Stringfield, V. T., and H. E. LeGrand, 1966, Hydrology of limestone terranes
in the coastal plain of the southeastern United States: Geol. Soc.
America Spec. .;.e 93, 46 p.
Valentine, J. W. and R. G. Peddicord, 1967, Evaluation of fossil assemblages
by cluster analysis: J.,'r;. Paleontology, v. 41, p. 502-507.
Veizer, J., and R. Demovic, 1974, Strontium as a tool in facies analysis:
Jour. Sed. Petrology, v. 44, p. 93-115.
Veizer, J., J. Lemieux, B. Jones, M. R. Gibling, and J. Savelle, 1977,
Sodium: Paleosalinity indicator in ancient carbonate rocks: Geology,
v. 5, p. 177-179.
Vernon, R. 0., 1951, Geology of Citrus and Levy Counties, Florida: Florida
Geol. Survey Bull. 33, 256 p.
Weaver, C. E., and K. C. Beck, 1977, Miocene of the S. E. United States,
a model for chemical sedimentation in a peri-marine environment:
Amsterdam, Elsevier Pub. Co., 214.p.
Weber, J. N., 1964, Trace element composition of dolostone and dolomites
and its bearing on the dolomite problem: Geochim. et. Cosmochim. Acta,
v. 28, p. 1817-1868.
Wetterhall, W. S., 1964, Geohydrologic reconnaissance of Pasco and southern
Hernando Counties, Florida: Florida Geol. Survey Rept. of Investigation
34, 28 p.
Wetterhall, W. S., 1965, Reconnaissance of springs and sinks in west-
central Florida: Florida Geol. Survey Rept. of Investigation 39, 42 p.
Weyl, P. K., 1960, Porosity through dolomitization, conservation-of-mass
requirements: Jour. Sed. Petrology, v. 30, p. 85-90.
Whittington, H. B. and C. P. Hughes, 1972, Ordovician geography and faunal
provinces deduced from trilobite distribution: Philos. Trans. Royal
Soc. London, v. 263, p. 235-278.
Winston, G. 0., 1976, Florida's Ocala Uplift is not an uplift: AAPG Bull.,
v. 60, p. 992-994.
Zachos, L. G., 1978, Strat-':.~;r?, and petrology of two shallow wells, Citrus and
Levy Counties, Florida: Master's thesis, Univ. of Florida, 105 p.
Zenger, D. H., 1972, S- chancee supratidal dolomitization in the geologic
record: Geol. -: ." '11., v. ': p. 1-12.




Full Text

PAGE 1

WATER IiRESOURCES researc center Publication No. 68 Mineralogical Model of the Floridan Aquifer in the Southwest Florida Water Management District By A. F. Randazzo Geology Department University of Florida Gainesville UNIVERSITY OF FLORIDA

PAGE 2

i i TABLE OF CONTENTS Abstract ........................................... ................ 1 Introduction ....... ................................................. 2 Geologic History ..................................................... 3 Diagenetic Solutions ....................................... ......... 7 Study Area........................................................... 9 Methods of Ana lysi s .................................................. 9 Distribution ofPa rameters .......................................... 11 Environments of Deposition ....................................... ll Multidimensional Scaling Model ............................... 11 Cluster 1 ................................................. 16 Clus ter 2 ............... ,; ................................ 16 Cluster3 ................................................. 16 Cluster 4 ................................................ 16 Cluster 5 ................................................. 16 Cluster 6 ............................................ 16 C1 uster 7 ....................... 18 Interpretation ............................................ 18 Stra t i graphy ..................................................... 20 Mi nera logy ...................................................... 20 Trace Elements and Geochemistry of Rocks ......................... 24 Stront i um ................................................... 24 Sodium ...................................................... 24 Mo 1 e-percent MgC03 ........................................... 28 14C 30 Distribution of Hydrogeochemical Parameters ...................... 30 Chloride ................................................. 30 Conductivity ................................................. 33 Sulfate ..................................................... 33 Discussion .......................................................... 33 The Model-A Summary of Data and Interpretations ..................... 41 Ac know1 edgements .................................................... 44 References Cited ..................................................... 45

PAGE 3

-1-ABSTRACT The distribution of geological and hydrochemical parameters in cores and wells penetrating the Floridan aquifer along portions of the west coast of peninsular Florida is related to the occurrence of dolomite. Trace element analyses of the cores indicate correlations between strontium and sodium concentrations and the particular carbonate phase. Chloride, sulfate, and conductivity values indicate the position and extent of the freshwater saltwater interface. Thick sequences of carbonate rocks in western penin sular Florida have been dolomitized in the freshwater-saltwater mixing zone of the coastal aquifer. A multivariate computer analysis was made of the petrographic data from three of the cores studied. The resulting multidimensional scaling diagram revealed several trends of environment of deposition parameters. Energy levels and faunal diversities helped to reconstruct the paleoenvironments of these Tertiary carbonate rocks. These data were then coordinated with the stratigraphy of the study area. Correlations between the occurrence of dolomite and specific stratigraphic formations suggest that dolomitization was rock-selective. Evidence of continuous formation of dolomite indicates that the process is actively occurring. The nature and distribution of dolomite in the Florid'an aquifer is significant in the development of carbonate rock porosity. The dissolution and replacement of minerals affect groundwater movement as the hydrologically dynamic system involves a continual interaction between water and rocks. Understanding these interactions will aid hydrologists to inven tory more precisely present-day water supplies and to predict changes to be expected. More efficient water management can be achieved by integrating the mineralogical model with hydrologic data.

PAGE 4

-2INTRODUCTION The Southwest Florida Water Management District relies upon the Floridan aquifer for its principal supply of potable water. The geologic nature of this aquifer and its relationship to groundwater flow systems is of fundamental importance in proper water management techniques. A model was established in order to express the mineralogical distributions within the aquifer and to demonstrate mineral interactions with groundwater. The mineralogical model of the Floridan aquifer in the Southwest Florida Water Management District relates specific geochemical, hydrologic, and geologic data to the occurrence of carbonate minerals in the aquifer. This model provides insight to the way carbonate aquifers develop. The aquifer repre: sents a dynamic system in which waters of varying chemistry have or are reacting with subsurface rock sequences to produce porosity and permeability changes and the formation or destruction of certain minerals. One major process which affects porosity and permeability in carbonate systems is dolomitization. Weyl (1960) showed through conservation of mass requirements that mole-for-mole replacement of calcium by magnesium will result in a 13 percent volume shrinkage, increasing porosity by joining smaller pores and thereby increasing permeability. Schmidt (1965) stated that in the normal sequence of neomorphism (Folk, 1965) of calcium carbonate grains, a mole-for-mole replacement of aragonite by calcite will increase the mineral volume (decrease the porosity) by 8.7 percent. Subsequent dolomitization of the calcite will result in a 13 percent volume decrease (increase the porosity). Dolomitization of an original aragonite matrix, however, will result in a volume decrease of only 5.4 percent. Of course, in whichever sequence dolomitization occurs, such factors as original packing, compaction, and introduction of materials by percolating waters will affect the overall porosity and permeability. A better under standing of the processes that control dolomitization in a carbonate aquifer system can provide valuable knowledge to the way aquifers develop. Previous studies have related the occurrence of a mixing zone, where seawater and groundwater meet in a coastal aquifer, to dolomitization (Kohout, 1965, 1967; Runnells, 1969; Hanshaw et al., 1971; Badiozamani, 1973; Land, 1973; Folk and Land, 1975; Hanshaw and Back, 1979).

PAGE 5

-3-The parameters considered in this study traditionally have been shown to be either of direct significance to the dolomitization process, or to be useful in defining a freshwater-saltwater interface. Trace element analyses of cores were conducted on strontium and sodium. Correlating their concentrations and distribution of the various mineralogies present led to a hypothesis of dolomitization by solutions of diluted seawater. Chloride, sulfate, and conductivity measurements from interstitital pore waters were used to define the present position and extent of the freshwater saltwater interface. Groundwater composition and flow characteristics are directly affected by these water-rock interactions. Zones of high transmissivity, as well as areas of water with high or low percentages of total dissolved solids, can be better explained by utilization of a mineralogical model. Likewise. this model can serve as a predictive tool for aquifer evolution. This study involved a number of graduate students who made significant contributions to the total research effort. The investigation expanded upon two earlier endeavors (Randazzo. 1976, 1980) and represents a greater delinea tion and understanding of the Floridan aquifer. The reader is directed to the works of Saroop (1974). Stone (1975), Hickey (1976). Liu (1978), Zachos (1978), Fenk (1979), and Sharpe (1980) for specific details on the lithologic and paleontologic characteristics of the various lithofacies recognized. Hickey (1976), Sarver (1978), Zachos (1978), and Metrin (1979) discussed the diagenetic and geochemical aspects of the important carbonate rock-forming minerals, calcite and dolomite. Bloom (1982) summarized the data and conclu sions from these earlier reports and integrated them with new data presented here. GEOLOGIC HISTORY At the beginning of the Tertiary Period a broad, stable carbonate bank existed over the Florida Platform (Chen, 1965). This bank was bounded by submarine escarpments on both the Atlantic and Gulf of Mexico sides, and was separated from the continental shelf to the north by the Suwannee Channel (Applin and Applin, 1944; Jordan, 1954; Chen, 1965) (Fig. 1). A warm shallow sea environment resulted in the deposition of the Cedar Keys Formation in the Paleocene, followed in superposed order by deposition of the Oldsmar, Lake City. Avon Park, and Ocala carbonate sequences during the Eocene. There

PAGE 6

-40'-._ GEORGIA $ \ .". 0 ___ .. ,( _o_.-. .a,.. -\(j' .(::;,...... \ ) ... \ ..... eO \ \ ...., ATLANTIC OCEAN .' \ \ \ GULF OF MEXICO \. ... \ .. \ .. \ "tI" \ \ 'II .. ". '\ \. .. -. .... Souih Florida Shelf ..... ,. ....... Major Axis of Peninsular Arch' Major Axis of Ocala Arch Sheff Rim "-South Floridq 8asin o 50 L oa(J.4 ... .. 9.,Figure 1. Major structural features of peninsular Florida. '\ .,fI

PAGE 7

-5-were short episodes during the Eocene when great parts of the northern and central portions of the platform were emergent and subjected to nondeposition and subaerial erosion (Chen, 1965). These emergences are evidenced by numerous unconformities (Randazzo et al 1977). The carbonate facies deposited during the early Tertiary dip and thicken towards the south and the coasts, away from a subsurface structural high known as the Peninsular Arch (Vernon, 1951; Chen, 1965). The updip sections of the Floridan aquifer crop out at the contact with the Cretaceous Coastal Plain sediments in southern Georgia and Alabama. Vast quantities of water flow off of the southern Appalachians and infiltrate the porous carbonates here, creating an area of principal recharge to the aquifer. Deposition of the Oligocene Suwannee Limestone was followed by a marine regression and an extensive period of post-Oligocene erosion (Vernon, 1951). The post-Oligocene seas exposed a structura1 high in western peninsular Florida. Originally described as a breached dome, this "Ocala Upliftll (Cooke, 1945; Vernon, 1951) has more recently been termed a "blister dome" (Winston, 1976). Winston believed that it formed when an increase in the rate of deposition of the Lake City Formation produced a curvature. reflected on the surface of deposition. Subsequently deposited carbonate sequences also reflected this curvature, and Oligocene or post-Oligocene tilting of the Florida Platform served to accentuate the dome. The Ocala "blister dome" was exposed in the early and middle Miocene during which time shallow brackish water and marine facies of the Tampa Formation (early Miocene) and the Hawthorn Formation (middle Miocene) were deposited around it. Cooke (1945) found the Tampa-Hawthorn contact to be unconformable. and believed that the Hawthorn was deposited in an "expanded" Tampa sea. The upper part of the Hawthorn Formation, which consists of a heterogeneous mixture of lime mud, clays, marls, and phosphoric sands, forms a confining unit over the Floridan aquifer, and thus marks its upper boundary. At the close of the Miocene, the Hawthorn covered all of peninsular Florida except the area over the Ocala "blister dome." Artesian circulation probably began in the Floridan aquifer at that time, with recharge being confined pre dominantly to the areas of outcrop of the limestone (Stringfield and LeGrand, 1966 ; Vernon, 1951).

PAGE 8

-6-To.day, 1Dca1 recharge o.ccurs directly into. the aquifer in areas o.f high e1evatiDn alo.ng the crest Df the Ocala "blister dDme" where the Hawtho.rn was no.t depDsited. Where the Hawtho.rn is present, recharge can o.ccur where it is breached, either by erDsiDn Dr sinkhD1es. Discharge o.ccurs mainly Dut into. the Gulf Df Mexico. (Stringfield and LeGrand, 1966), and to' a lesser extent thrDugh co.astal springs (Wetterha11, 1965; Fetter, 1980). Stringfield and LeGrand (1966) agreed with CDDke (1945) that the distributiDn Df cavities within the Flo.ridan aquifer, with their general decrease in size and numbers with distance fro.m recharge and discharge areas, suggests that the present pattern Df circu1atio.n develDped in the Pleisto.cene during lo.W sealevel stands. Ranges in sea level during the PleistDcene are estimated to. have been 270 feet abo.ve the present level (Co.Dke, 1945), and 500 feet belDw the present level (Do.nn et al., 1962). Thus, when sea level was at its lo.west stand, it is cDnceivable that the gro.undwater level in we$tern peninsular F1o.rida was a few hundred feet 1o.wer than it is at the present time (Stringfield and LeGrand, 1966). During these stages, the Hawtho.rn was extensively ero.ded al1o.wing mo.re widespread recharge into. the aquifer, ac celerating dissDlutio.n Df the limestDne beneath. The present-day nature Df the carbo.nate aquifer is dependent upon the interrelatiDns Df its geD1Dgic histo.ry and the hydrD10gic CDnditions to' which it has been subjected. FDur geD10gic peri Dds of aquifer development were cited by Stringfield and LeGrand (1966, p. 7-8). The IIdepositiDnal periDd" establishes the vDlumetric extent Df the carbonate unit and its primary permeabil ity. Duri ng the "immedi ate post-depDsitiDna 1 peri Dd" the carbo.nate depDsit is elevated abDve sea level where it is eXPo.sed to meteo.ric weathering and possibly subaerial erosion. In addition to diagenetic changes which affect pDrosity, dissolution by infiltrating groundwater can greatly increase the permeabil ity of the rock during thi s stage. The "subsequent pre-recent periodll is marked by increased disso.lution and permeability (assuming the deposit is still exposed to groundwater infi ltration), compaction, consol idation, and recrystallization of the original limey deposits, and possibly the occurrence of some folding or faulting which might restrict groundwater circulation in some areas and promote it in others. Deposition of a sequence of sediments Df low permeability will hinder groundwater dissolution of the underlying carbonates. The last period is marked by the lIexisting geologic setting and structure." Karst topography develops where the limestone crops out and is in the present groundwater circulation system.

PAGE 9

-7-The overall geologic structure is important in relation to topography and recharge-discharge areas. In western peninsular Florida recharge pre dominantly occurs north of Tampa Bay where porous carbonate units crop out, while south of Tampa Bay a layer of relatively impermeable clays inhibits recharge to the aquifer except where sinkholes breach this aquiclude. DIAGENETIC SOLUTIONS Carbonate diagenesis appears to occur more rapidly in the phreatic meteoric water environment than in the vadose meteoric or marine phreatic environments. Ginsburg (1957) noted that the Miami Oolite is better cemented where it presently exists below the modern water table than where it occurs in the vadose or marine phreatic environments. Friedman (1964), in observing that freshwater more commonly promotes mineralogical and textural changes which lead to lithification than marine waters, ran a series of experiments and concluded that low pH and low salinity of meteoric waters favor the diagenetic changes that occur. In studying a fossil water table in Bermuda. Land (1970) judged that the phreatic meteoric water zone was where diagenesis of skeletal carbonates to coarse-grained sparry limestone was most rapid. These rocks were more highly altered and more stable mineralogically than rocks which spent more time in the vadose or marine zones. Also, he found larger crystals of cement in this zone where water could remain in pore spaces for long periods of time, as opposed to the vadose zone where flushing groundwater resulted in very little cement. Land and Epstein (1970) discussed the mineralogical and isotopic changes that occur during meteoric water diagenesis. They stated that, generally, magnesium calcites incongruently dissolve rapidly to calcite, followed by dissolution of aragonite and reprecipitation as sparry calcite cement. As diagenesis proceeds, the unstable minerals dissolve, changing the cation and anion chemistry of the solution. Thus, the newly formed carbonates have lower concentrations of Sr+2 Na+, and Mg+2, representing the interaction between meteoric water and the original marine sediments. Steinen and Matthews (1973), reporting on a cored borehole on Barbados. observed that the upper part of their core had been in the vadose zone since its initial emergence from the marine environment. Although the carbonate sediments in this zone were composed of the unstable phases aragonite and high-magnesium calcite, they had not been affected to any great degree by dissolution. That part of the core that had occupied the freshwater phreatic

PAGE 10

-8-lens at least once during sea level fluctuations appeared to be mineralogically stable (composed of low-magnesium calcite), extensively cemented, and showed evidence of earlier dissolution. The lowest part of the core that was subjected to vadose meteoric diagenesis during glacial epochs and marine phreatic diagenesis during interglacial epochs, showed only minor cementation and dissolution, and retained most of its depositional mineralogy. The diagenetic changes that occurred within the lithofacies described in this study that involve neomorphism of one carbonate mineral phase to another probably occurred upon introduction to a phreatic meteoric water environment. The dolomite found within the cores in the study area has been categorized according to origin as penecontemporaneous dolomite formed in a supratidal environment; and secondary dolomite formed as a replacement of calcite by diagenetic solutions. Table 1 summarizes the characteristics of dolomite types as related to origin. Table 1. Characteristics of dolomite types as related to origin. Characteristics of Penecontemporaneous Dolomite (Primary) 1. The dolomite is fine-grained (1-5 microns). 2. The dolomite occurs in association with supratidal sediments and structures, such as algal stromatolites, birdseye vugs, root casts, dessication features, and evaporite minerals 3. The dolomite is nonstoichiometric. 4. The dolomite forms by replacing 1 ime muds, usually .accompanied by a preservation of original sedimentary textures. 5. The do1omitized areas are thin and/or localized. 6. One would expect to find alternating layers of gypsum-do1omite limestone in alternating supratidal-subtidal environments. 7. Trace element concentrations (Sr+2 and Na+) are high, reflecting the hypersaline nature of the do1omitizing brines. Characteristics of Dolomite Which Has Formed in a Freshwater-Saltwater Mixing Zone (Replacement) 1. The dolomite consists of larger crystals generally between 40-200 microns (Zenger, 1972). The more dilute the saline solution is, the larger the crystals and the more euhedral the crystal fabric. 2. The dolomite can occur in sediments associated with any depositional environment.

PAGE 11

-9-3. The dolomite is more-stoichiometric if the solution is quite dilute and the rate of crystallization is slow; and less-stoichiometric if it is only moderately dilute and accompanied by a faster rate of crystallization. 4. Depositional texture is preserved more frequently in more-stoichiometric dolomite, and suggests slow recrystallization in less saline solutions. 5. The dolomitized areas may vary in thickness, depending on the extent of migration of the mixing zone. 6. More-stoichiometric dolomites are associated with lower Sr+2 and Na+ concentrations than less-stoichiometric dolomites, indicating formation in less saline solutions. Both of these dolomite types, however, contain lower trace element concentrations than penecontemporaneous dolomites because of freshwater dilution in the mixing zone. 7. Coalescive neomorphism of original penecontemporaneous supratidal dolomites leads to rhombic crystals. Previously undolomitized limestones are recrystallized by a porphyroid neomorphic process that results in euhedral, rhombic dolomite crystals. 8. More-stoichiometric dolomite shows a good clustering of points when Sr+2 and Na+ concentrations are plotted against mole-percent MgC03' This indicates a longer residence time for the diagenetic solutions, and a greater approach to equilibrium between crystal and solution than occurs in less-stoichiometric dolomite (which characteristically produces a scattering of points when Sr+2 and Na+ are plotted against mole percent MgC03). STUDY AREA The area under investigation is the Southwest Florida Water Management District which occurs along the Gulf Coast of western peninsular Florida (Fig. 2). All of the cores and wells studied lie within this district except the Bell and Manatee Springs cores, which are within the Suwannee River Water Management District. A complete north-south traverse covers approximately 200 miles, and includes 29 wells, some as close together as 2 miles and others as distant as 47 miles. METHODS OF ANALYSIS The data compiled in this study came from various sources. X-ray analyses of cores 4-2, 6-3, 11-2, 17-1, 17-3, 19-3, and 21-2 were used to identify the presence of calcite and dolomite. Atomic absorption spectro photometry was utilized to determine sodium and strontium concentrations in the rocks. Other stratigraphic. mineralogical and trace element data were obtained from the studies by Randazzo et al. (1977), Randazzo and Hickey (1978), Sarver (1978). Liu (1978). Zachos (1978), Fenk (1979), Metrin (1979).

PAGE 12

-10-HILLSBOROUGH COUNTY OF LOCATION OF CORES AND WELLS IN STUDY AREA Figure 2. r.4E)(IC O 10 10 20 SCALE IN MILES -HORIZONTAL PROJECTION OF CORES AND WELLS TO PRODUCE LEGS OF PANEL DIAGAA", WEULOCATION

PAGE 13

-11-Sharpe (1980), and Bloom (1982). Sulfate, chloride, and conductivity measurements were obtained from the Southwest Florida Water Management District. The Dicarb Radioisotope Company analyzed several dolomite samples for 14C. Panel diagrams are used to illustrate the variations of individual parameters across the study area. The panel diagrams were produced in conjunction with a United States Geological Survey potentiometric map (Fig. 3). The points where panels intersect were made to coincide with local recharge areas of 50 feet above mean sea level in Levy County, and 80 feet above mean sea level in Pasco County, In this way, the distribution of each parameter can be compared to the general flow of groundwater (Fig. 4). DISTRIBUTION OF PARAMETERS Environments of Deposition Multidimensional Scaling Model One of the most difficult problems in the utilization of a large quantity of data is the development of a combination of statistical methods for facilitating the environmental interpretation of carbonate rocks. In dealing with thin sections and modal analyses examination is made of tables of data which depict abundances and paucities of variable parameters. Interpretation of large amounts of data by "eye" is difficult. The final results are always partly intuitive. The capacity of computers to make comparisons provides a modern capability to establish objectively significant relationships among variables. Computers can compare variables within samples and recognize similarities, making interpretations more accurate. Cluster analysis is a simple form of correlation analysis, a method of searching for relationships in a large symmetrical data matrix (Hayes, 1980). The advantage of cluster analysis is simplicity (Valentine and Peddicord, 1967, p. 502) because of the uncomplicated nature of the input and the resultant visibly distinct clusters. However, the numerical methods should not obviate the need to analyze the data objectively; they only aid in reducing the dimensionality of the data and when fully utilized can provide various insights not seen without computations. A combination of statistical methods (multivariate analysis) were applied to the thin section data of the lithofacies to permit an environ mental interpretation of the Tertiary carbonate rocks. Cluster analysis was applied in order to group-together variables such as allochems, lime mud,

PAGE 14

-12GUl. F OF MEXICO ESTIMATED POTENTIOMETRIC SURFACE FOR THE TERTIARY LIMESTONE AQUIFER SYSTEM (1980) 50 40 '" SCALE IN MilES WELL LOCATIONS NOTe-CONTOURS SHOW APPROXIMATE ALTITUDE ABove MEAN SEA lEVEL IN FEET REVISED FROM U S. GEOL.OGICAL SUAve ... OPEN FilE 80--406 -------------------------------------_._----_.--------Figure 3.

PAGE 15

-13B (360') j MS (496') CO (289') RS LEVY co (320') RECiARGE AREA 21-2 (29') HS I (484') 107 (390-) I I 101 (360') GULF PROJECTED FLOW OF GROUNDWATER IN TERTIARY LIMESTONE AQUIFER SYSTEM Figure 4. OF MEXICO 6-3 (600') I 4-2 (610') VERTICAL SCALE IN FEET J \0 10 20 SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET BELOW SURFACE.

PAGE 16

-14-calcite, dolom'ite, etc., (Table 2) that are most similar genetically. The variables considered are the product of environmental processes; therefore, variables most similar probably formed in the same depositional environment under similar processes. Multidimensional scaling (MDS), based on petrographic similarity, was used to order the samples on a two dimensional graph. Non-metric MDS is a technique illustrated by Kruskal (1964a) applied to geological problems (Whittington and Hughes, 1972; Smosna and Warshauser, 1979). Multivariate analysis in carbonate petrology was employed by Imbrie and Purdy (1962) using factor analysis on data from Bahamian sediments. In factor analysis, however, the original data is ultimately lost in vectors, whereas cluster analysis does not abandon the original data and is immediately interpretable. In factor analysis mathematical techniques are used to promote correlation coefficients or secondary and tertiary data banks for and patterns. Therefore, the input data is and the final visible data is intelligible and correlatible but complicated because the worker is not viewing the original data. However, in cluster analysis similarities between the variables are processed and presented as groups of most similar variables showing the original data. Parks (1966) combined an R-mode cluster analysis and Q-mode cluster analysis to describe Bahamian sediments. Ekdale et al. (1976) used multivariate analysis for paleoecological interpretation of Cretaceous rudists. Comparisons sample by sample is a Q-mode clustering and variable by variable (within the sample) is an R-mode clustering (Sokal and Sneath, 1963). This computer application combines a variety of methods that are nonparametric (not based on normal distribution). Any method that reduces the dimensionality of the data causes distortion. Clustering imposes a hierarchical structure to the data. Almost inevitably the method introduces distortion in representing the multidimensional relations between the localities on a two dimensional dendogram. This distribution of distortion is well known; the relations at the tips of the dendogram are well-represented and distortion is greater in the later formed clusters (Rowell et al., 1973). The distortion of stress has to be taken into account when analyzing the clustering output. Because of the existence of this distortion Rowell et al. (1973, p. 3430) suggest that some form of ordination should be used as an alternative or complementary means of displaying structure in the data. Ordination is a comparative process. All samples or variables are compared to each other in a symmetrical matrix. Dissimilarity is measured and this produces a graphic geometric display. Cluster analysis often obscures the

PAGE 17

-15Table 2. List of variables used in the multivariate analysis of three cores of mid-Tertiary carbonate rocks. Allochems Matrix Detrital Foraminifera Peloid Ca 1 cite Quartz I Echi noderm Intraclast Micrite I Mollusca Pell et Dol ami te overall relationship. whereas ordination shows gradational relationships. The amount of distortion may be estimated by calculating a correlation between the interlocality distances in the reduced dimensional space and the corresponding distances in the original similarity matrix. The distortion that is typically present is distributed differently than that in a dendogram. The distortion is greater in smaller distances and not larger distances between the data points. Ekdale et al. (1976) used an ordination algorithm versus the nonmetric algorithm used by Kruskal (1964a, 1964b). Kruskal's algorithm includes a built-in accommodation of stress (distortion). Basically the method seeks to find a geometric configuration of n points (variables) in a reduced space of k dimensions (containing the variables) such that interpoint distances correspond to the similarities between points. The solution is the best fitting configuration that minimizes stress. Nonmetric MDS distributes distortion better than clustering. The carbonate rocks of this study were lithologically distinct and were deposited under different environmental influences and processes. Ordering the thin section samples on a MDS plot reflects gradients of these processes. Multivariate analysis was performed on the thin section data from the lithofacies described in three of the cores studied (Sharpe. 1980) to illustrate the gradational trends of carbonate rock variables across a marine carbonate shelf An R-mode cluster analysis and a Q-mode cluster analysis were run on 176 thin-section samples for 52,800 observations. Seven clusters from the Q-mode were produced and scanned. Each cluster revealed characteristic variables that distinguish that individual cluster. They were: 1) dolomite, 2) dolomite/ mud, 3) mud, 4) skeletal/mud, 5) intraclast, 6}quartzose mud and 7) skeletal/ sparite. Because of the large number of samples (176) the three dimensional

PAGE 18

-16MDS plot was extremely cluttered and gave no clear isometric projection. However, the two dimensional MDS plot shows well the gradations and is illustrated here (Fig. 5). The cluster boundaries displayed on the plot were delineated based on the Q-mode data. Some thin section points fell between the cluster boundaries. They were transitional samples with characteristics of the two nearest clusters. This illustrates the gradational nature of the MDS plot. Cluster 1 Dolomite microfacies. -A mosaic of dolomite which completely obliterates all allochems, while preserving some megascopic sedimentary textures (e.g. faint laminations). These sediments are part of a Lithofacies of dolomitic mudstones and wackestones from the Avon Park Formation. Cluster 2 Dolomite/mud microfacies. -The samples are Qolomitized mudstones and and wackestones with textures and allochems preserved. Within this cluster also falls a dolomitic sandstone. Cluster 3 Mud microfacies. -These samples represent laminated mudstones and wackestones with sparse fauna. They include peat layers and algal laminate1 sediments. I Cluster 4 Skeletal/mud microfacies. -The samples are characterized by high faunal diversities and a high mud matrix content. This cluster contains the samples with large foraminifera (e.g. Nummul.ites sp.) from the Ocala limestone. Cluster 5 Intraclast microfacies. Abundant mudstone.s and wackestone occur here. These samples have large intraclast contents, mud and sparse fauna. Cluster 6 Quartzose mud microfacies. -The samples are predominantly wBckestones and mudstones with sparse fauna but have very abundant detrital (20 80 percent).

PAGE 19

Supratidal Figure 5. Two-dimensional MDS configuration displaying the relationships among 176 thin-sectioned samples based on ten variables (Table 2). I I-' --.J I

PAGE 20

-18-Cluster 7 Skeletal/sparite microfacies. -The samples are all skeletal grainstones with abundant void-filling spar and high faunal diversities. Interpretation Clusters 1, 2 and 3 correspond to those lithofacies interpreted as being deposited in the supratidal zone. The samples were independently described as supratidal deposits. Thus the multivariate analysis and clusters verified the similarity interpreted by conventional petrographic methods. The inclusion of a dolomite sandstone into these clusters is interesting. This unit was initially interpreted as a clastic bar deposit and yet the clustering indicates a supratidal (possibly beach) environment. The sandstone is very coarse and has associated organic material (dark color); therefore, this could very well represent a beach or supratidal channel deposit. Conversely, the presence of dolomite may have caused the cluster analysis to misclassify the two thin section samples containing this particular lithology (2 samples out of 176). Cluster 4 is interpreted as representing deep subtidal deposits, with low wave energy allowing the mud to accumulate and a high faunal diversity, characteristic of the deep subtidal. Both clusters 5 and 6 are interpreted as representing shallow water intertidal depositional regimes because of the abundant quartz and mud (restricted quiet water lagoon). Samples found in cluster 7 represent shallow subtidal material seen predominantly in several lithofacies. These samples have a high faunal diversity, high faunal content and abundant sparry calcite. Characteristics of the offshore depositional environments are gradational perpendicular to shore. Therefore, a plot of the various variables characteristic to one or more subenvironments should show this gradation. The multi dimensional scaling configuration (Fig. 5) demonstrates the gradational nature of the 176 samples, allowing for an environmental interpretation. The clusters (containing the samples) seem closer to one another. showing geometrically the gradation. In the interpretation of the MDS clusters. analysis must be made of chemical, physical and biological aspects of the clusters. The dolomite and mud-dominated clusters (1, 2 and 3) are interpreted as containing samples representing the supratidal zone. The intertidal zone is represented by samples in clusters 5 and 6, and samples of clusters 7 and 4 represent sha1low subtidal and deep subtidal zones, respectively.

PAGE 21

-19-The cluster showing the highest energy (wave action) is cluster 7. Therefore, after placing the shoreline to the left of the diagram (supratidal), curve A can be constructed showing energy normal to the shore (Fig. 5), culminating in cluster 7, the shallow subtidal zone. The shallow subtidal zone shows the least mud and the most spar, and so represents the highest wave energy zone (cluster 7). The transect of hydrodynamic energy starts in the supratidal and gradually increases through the deep subtidal, through the intertidal and reaches the highest level in the shallow subtidal zone. This is illustrated by mud abundance varying from 81 percent. in the supratidal samples, to 39 percent in the deep subtidal samples, to 27 percent in the intertidal samples, and 0-8 percent in the shallow subtidal. Conversely the amount of .sparite increases from 0-3 percent in clusters 1, 2 and 3 to 35 percent in cluster 7 (shallow subtidal). The pattern of fabric support also parallels this trend passing from mudstones in clusters 1, 2 and 3 to grainstones in cluster 7. These properties parallel the direction of increasing energy and substrate mobility. Faunal diversity is highest in the deep subtidal (4) and shallow subtidal (7). Therefore, paralleling the hydrodynamic trend (curve A) is a faunal diversity trend (curve B) showing increasing faunal diversity from clusters 1 and 2 to clusters 4 and 7 (Fig. 5). The two intertidal clusters (5 and 6), together with the mud/quartz cluster (3), indicate that the MDS analysis also realized the regional change (advent of detrital quartz) in sediment characteristics. Consequently clusters 3 and 6 are interpreted as indicating more clastic intertidal and supratidal zones as shown by another hydrodynamic (energy) curve (C) (Fig. 5). Diagenetic trends can also be shown on the MDS plot (Smosna and Warshauer, 1979). The susceptibility of supratidal sediments to dolomitization is illustrated by the dolomite being concentrated in the supratidal deposits and this is supported by previous work (Randazzo and Hickey, 1978). The trend does not show phases of dolomitization because there are different dolomitic textures within cluster 2. The presence of two distinct clusters of dolomite is thought to illustrate different salinity conditions fluctuating from below normal to above normal (Smosna and Warshauer, 1979). The two dolomite clusters may indicate phreatic waters of different concentrations of Na, Mg, and Ca indicating the possibility that different groundwater systems may have acted on the samples

PAGE 22

-20of these clusters. Another hypothesis is that the clusters may well indicate a difference in the time of contact with groundwater fluids. Therefore, from this data presentation can be derived a hydrodynamic and paleoecological interpretation of the Tertiary carbonates of west central Florida (Fig. 6). Figure 6A shows the carbonate platform with lowest water energy at the supratidal and highest energy in the shallow subtidal. Low faunal diversity is characteristic of the supratidal zone increasing in diversity offshore, and the most pronounced dolomitization has occurred in supratidal sediments. Figure 6B shows a slightly steeper gradient in a clastic influenced tidal regime, shown by samples in clusters 2, 3 and 6. Stratiqraph.y Figure 7 shows the distribution of the geologic units to a depth of 600 feet. From Hernando County to the south, the formations can be seen dipping to the south where they progressively become overlain by younger units. This feature reflects the deeper lying structures of the South Florida Shelf and the South Florida Basin (Fig. 1). The regional effects of the Ocala IIblister domell can be seen in the northern half of the diagram where the Ocala Limestone crops out. Mineralogy The distribution of limestone and dolomite are depicted in Figure 8. Analyses were conducted by x-ray diffraction, and only those samples in which the carbonate fraction was found to be 100 percent dolomite are illustrated here. One of the most distinguishing observations is that the Avon Park Formation has been significantly dolomitized in this area, while the Ocala and Suwannee Limestones lying above .have not been. The limestone-dolomite boundary follows the Ocala-Avon Park boundary, and dips to the south. South of the Pasco County recharge area most of the Miocene age deposits contain dolomite. The dolomite zone here thickens southward and seaward, as do those stratigraphic units.

PAGE 23

-21I OW tide : wave base ....... =--=--:::..-:. -:;.. I .... ..-1_ ....: 7 4 -. 5 1,2 I I :I A Clusters Increasing hydrodynamic energy ... Increasing faunal diversity Dolomite 1 6 2,3 B Clusters .. Increasing energy f Increasing diversity I .... Quartz Figure 6. Schematic interpretation of Tertiary carbonate rocks illustrating the paleoenvironments. Plotted are. the positions of the seven Q-mode clusters and environmental gradients. Part A shows the carbonate platform and the environmental interpretation. Part B is the clastic-influenced interpretation.

PAGE 24

STRATIGRAPHY OF STUDY AREA -22PLIO.PLEISTOCENE--f.[IJ] PLEISTOCENE SANDS {c:J TAMIAMI FORMATION VENICE CLAY MIOCENE HAWTHORN FORMATION TAMPA FORMATION OLIGOCENE SUWANNEE OCALA UMESTONE EOCENE AVON PARK FORMAnON LAKE CITY FORMATION """--...!jQl VERTICAL SCALE IN FEET J 10 10 20 I SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET BELOW SURFACE. ------------'-----------------------------_. Figure 7.

PAGE 25

-23DISTRIBUTION OF CARBONATE MINERALS Figure 8. lZ9 SAND &. C!.AY o UMElITONE OOI.OMITE "'" 10 .!!2B!I! VERTICAL SCALE IN FEET ,. .. SCALE IN MILES NOTE: NUMBERS IN PARENTHeses DENOTE DEPTH IN FEn 1SEL0'N SURFACE.

PAGE 26

-24-Trace Elements and Geochemistry of Rocks Strontium The distribution of strontium is shown in Figure 9. Sr+2 concentrations greater than the 50-percent reduction associated with the formation of dolomite from calcite (Behrens and Land, 1972) have been explained by dolomitization occurring in.waters more saline than those in which calcite formed (Randazzo and Hickey, 1978). The Sr+2 content in the northern cores (B, MS, CO, RS, 124, 21-2, HS, and 107) does not drop to lower values at the limestone-dolomite boundary, but rather changes farther down into the dolomite zone. Core 19-3 shows an increase in Sr+2 concentration in its:;deeper regions to over 800 ppm. Higher than average Sr+2 concentrations also occur at certain horizons in cores 21-2 and 17-1. These areas of high Sr+2 concentration indicate where dolomite formed from solutions more saline than did the surrounding rock. Across the entire transect of the northern cores, it appears that the dolomites of the uppermost part of the Avon Park formed in solutions more saline than. the lower part. The southern cores (19-3, 101, 17-1, 17-3, 11-2, 6-3, and 4-2) all contain higher Sr+2 concentrations,on a whole, than do the northern cores The calcite samples in core 101 average twice the concentration of Sr+2 as do the calcites in the northern cores. The dolomite samples in core 101 contain approximately half the concentration of Sr+2 as do the calcite samples, implying that both carbonate phases probably formed in similar solutions. Dolomite samples analyzed for cores 6-3, and 11-2 occur in Miocene deposits which are considerably younger than the Eocene carbonates of the Ocala Limestone and Avon Park Formation. The'Sr+ 2 toncentrations of dolomites in the Miocene rocks are higher than anywhere else in the study area, ranging from less than 200 ppm to greater than 800 ppm. Figure 9 indicates that there is a steady increase in Sr+2 concentration in dolomite of these cores with depth. It should be noted that small amounts of Sr+2 may have been contributed by associated clay minerals which were sometimes present in the southern cores. Sodium According to Veizer et al. (1977) and Land (1980), sodium interpretations should be used cautiously because present methods of analysis cannot distinguish lattice-bound sodium from sodium trapped or absorbed in the crystal struc-

PAGE 27

STRONTIUM CONCENTRATIONS IN ROCK SAMPLES Figure 9. -25CONCENTRATIONS OF Sr+2IN PPM <200 1iilii.ii400-600 I!IlIDIIIIJ 6....., .>eoo 10 VERTICAL SCALE IN FEET J 10 20 SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET BELOW SURFACE.

PAGE 28

-26-ture as inclusions of NaCl. Because Na+ is a small ion, it can substitute with equal facility into Ca or Mg lattice positions (land and Hoops, 1973). Thus calcite and dolomite that have formed from similar solutions should contain similar quantities of sodium. Figure 10 depicts sodium distribution in the rocks studied; and those wells that show incomplete data contain analyses on dolomite samples only. Again some Na+ may have been contributed by associated clay minerals which sometimes were present in the southern cores. An of the calcite samples within the diagram contain sodium in concentrations less than 500 ppm, which indicates diagenesis in relatively fresh waters. The dolomite samples show a range of sodium values from less than 500 ppm to over 2,000 ppm. The increase in sodium values exactly at the limestone-dolomite boundary in the northern cores and the decrease in sodium concentrations in the lower regions of the Avon Park Formation in those cores agrees well with the conclusion drawn from the strontium data -that the uppermost dolomites in the Avon Park formed in solutions more saline than the lower parts. The cores adjacent to and south of the Pasco County recharge area (cores 4-2, 6-3, 11-2, 17-1, and 17-3) contain significantly higher levels of sodium than the other cores. This is in agreement with the Sr+2 distribution in Figure 9. However, the Sr+2 concentrations showed an increase with depth in the Miocene deposits, whereas the sodium concentrations appear to show no predictable alternation in pattern. Zones of high Na+ values (> 2,000 ppm) occur among zones of low Na+ values 500 ppm). Also, the base of core 19-3. which contains abnormally high strontium values, showed relatively low sodium concentrations in comparison. It should be noted again that the dolomites in the cores south of Pasco County recharge occur in rocks younger than those to the north. and are associated with clay minerals. The sodium content of ancient dolomites have been found by many authors to be depleted relative to recent dolomites (Weber, 1964; Behrens and land, 1972; Fritz and Katz, 1972; Land and Hoops, 1973). Land and Hoops (1973) believed that this was the result of re-equilibration with a meteoric reservior. If this istrue, then dolomite, originally equilibrated in saline solutions, could lose weakly held sodium from its lattice and bring about new equilibrium conditions when flushed by fresh water. Because all of the cores within this study have been exposed in some degree to meteoric water, the sodium concentrations analyzed are probably all less than what they were originally.

PAGE 29

SODIUM CONCENTRATIONS IN ROCK SAMPLES 101 (360') Figure 10. -2711 (590') COMCENTRATtoNS OF Na+ IN PPM m _,oon 1000-1500 1500-2000 >2000 .... NQRTH VERnCAl SCALE IN FEET J 10 10 20 SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET BELOW SURFACE.

PAGE 30

-28Cores 21-2 and HS show higher sodium concentrations in comparison to neighboring cores. Both of these cores occur where the potentiometric surface (Fig. 3) is lower than in any of the other cores in the study area. The Ghyben-Herzberg principle 1972). states that saltwater intrusion occurs to a greater degree, and affects rocks closer to the surfa(:e, in areas of low potentiometric head. The strontium data from the dolomites in core HS correlate with the northern cores, illustrating a less saline do1omitizing solution for the lower part of the Avon Park. However, the sodium data for this core does not agree with its northern counterparts. The zone 500 ppm) illustrating this relationship in the lower parts of the northern cores is replaced by zones indicating dolomitizing solutions of much higher salinities in HS core. The distribution of sodium in Figure 8 indicates that while it substantiates the conclusions drawn from strontium analyses in some instances, in other cases it can lead to ambiguous interpretations. Because trapped 1 and absorbed sodium in the carbonates can contribute to the total sodium detected in analysis, used alone it is not a reliable indicator of the salinity of diagenetic solutions. However,where there is agreement between sodium distributions and other trace element distributions, more confident conclusions may be reached regarding the salinity of diagenetic solutions. Mole-percent MgC03 Dolomites are chemically characterized by their mole-percent MgC03 content. Randazzo and Hickey (1978) categorized the less-stoichiometric dolomite in their study as being composed of 44-48 mole-percent MgC03' and the more-stoichiometric dolomite 49-51 mole-percent MgC03' Figure 11 shows the distribution of mole-percent MgC03 of some of the dolomite in this study. The numerical divisions used in the index do not exactly correspond to the boundaries used by Randazzo and Hickey (1978),' but offer a greater distinction of the degree of dolomite stoichiometry. More-stoichiometric dolomites generally indicate slow formation in waters less saline, with fewer competing ions to disrupt the resulting structure (Folk and Land, 1975; Randazzo and Hickey, 1978). It should be noted that other factors such as changing Mg/Ca ratios of pore waters, rechargeb1e sources of Mg+2, crystal sizes, and permeability can affect the relationship between dolomite stoichiometry and the salinity of the dolomitizing solution (Lumsden and Chimahusky, 1980)

PAGE 31

MOLE-PERCENT MoCO, OF DOLOMITE SAMPLES 101 (3601 Figure 11. -29MOLE-PERCEN1 MgC0 3 0 ...... --> ... 8-3 10 ,. VERTICAL SCALE IN FEET 20 !!! IeALE IN MILES NOTE: NUMHRlIH PARENTHESES DENOTE IN fElT IELOW SURFACE.

PAGE 32

-30Comparison of the distribution of mole-percent MgC03 with the distribution of Sr+2 for the cores in the northern portion of the study area (Figs. 9 and 11) shows a marked decrease in Sr+2 occurring at approximately the same depth as a marked increase in mole-percent MgC03 This inverse distribution of more-stoichiometric dolomite and low concentrations of S.r+2 suggests the formation of dolomite in solutions less saline than those acting upon the uppermost section of the carbonate sequence. The distribution of Na+ in these same cores (Fig. 10), despite the limitations cited earlier, further substantiates the conclusion drawn about the salinity characteristics of the do1omitizing solutions. 14C Ten dolomite samples from cores 4-2, 6-3, 19-3, 21-2 and MS were analyzed for 14C. The 14C content of the samples was appreciable and would translate into lIages of 26,470 38,760 years B.P. Although these "ages" cannot be interpreted rigorously as geologic ages, they clearly demonstrate that carbon has been exchanged between dolomite and the atmospheric reservior in the geologically recent past. Though not conclusive, these data suggest that the formation, or at least partial recrystallization of dolomite occurred within the last 30,000 years. Distribution of Hydrogeochemical Parameters Data were obtained on chloride and sulfate concentrations, and specific conductance from selected wells. At intervals during drilling of the cores, water samples were collected from flowing water at the base of the wells. Readings were taken from these samples. Tight casing prevented waters from higher areas mixing with waters from the base of the well. These data can be helpful in defining the position of a freshwater saltwater interface that presently exists. If the present pattern of circulation within the Floridan aquifer developed during low stands of Pleistocene seas (Stringfield and LeGrand, 1966), then the present position of the interface may be related to the distribution pattern of dolomite in the region. Chloride The most reliable indicator of saltwater intrusion is the chloride content of the water (Wetterhal1, 1964). The chloride ion concentration in pure seawater is approximately 19.000 mg/1 (Gross, 1972; Duxbury, 1971)0

PAGE 33

-31-Fresh Florida groundwater, on the other contains chloride ion concentrations less than 50 mg/l (Wetterhall, 1964). Reichenbaugh (1972) used a chloride concentration of greater than 250 mg/l to indicate that a well had contacted the freshwater-saltwater interface. Thus Figure 12 provides a good indication of the salinity of encroaching solutions. The values from wells 17-1, 17-3, and 21-2 denote the presence of highly saline solutions wi/thin the aquifer at shallow depths. A dramatic increase in chlorinity occurs from freshwater concentrations just below the surface, to concentrations near that of seawater at only 200 feet of depth. The chlorinity in wells 19-3 and 18-1 are under similar potentiometric heads as those wells which occur at close distances, yet they constitute freshwater all the way to their base. Wetterhall (1964) found that intruded saltwater not be vertically continuous in the aquifer. He identified wells within the same area of Hernando County as 19-3, where zonation of the aquifer allowed freshwater to occur between layers of relatively salty water. Wells 11-2, 6-3, and 4-2 all give data which indicate that saltwater has intruded into the southern region, but not to the degree that it has in some of the other wells. One reasan for the low chloride concentrations in these wells is that the wells occur in the confining layers of low permeability overlying the aquifer. Water is not able to flow as freely into these formations either from the sea or from recharge as it can in the carbonate aquifer units which lie below. Another explanation, concerning the height of the potentiometric surface, can explain the distribution of chloride concentrations in all of the wells (except 124, 97, and 18-1 for which the previous interpretation will hold). The potentiometric levels in wells 17-1, 17-3, 21-2, and 124 are less than 10 feet (above mean sea level), while in wells 11-2, 6-3, and 4-2 the heights range from approximately 25-35 feet (Fi g. 3). Assumi ng uni form permeabil ity, the Ghyben-Herzberg principle denotes that the freshwater-saltwater interface should lie at less than 400 feet in those northern wells, while it should occur between 600-1,400 feet' in the southern wells. Allowing for the low permeabil ity of the Hawthorn Formation, saltwater encroachment would still not be expected to be as great at the depths penetrated by the southern wells, as at the depths penetrated by those wells north of the Pasco County recharge area.

PAGE 34

CHLORIDE CONTENT OF GROUNDWATER MS (496') -32-Figure 12. CONCENTRATIONS IN MGfL <50 m 50 llJO.500 500-1000 lOO1J.5OOO BI\!iI\II 5OIJO.l0,OOO ,,0,000 ,. ,. VERTICAL SCALE IN FEET J '" SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET PELOW SURFACE.

PAGE 35

-33-Conductivity In his geohydrologic reconnaissance of Pasco and southern Hernando Counties, Wetterha11 (1964) found chloride content and specific conductance (conductivity) of the waters to be generally related. Figure 13 shows conductivity distribution in the waters from the same wells as Figure 12. These values correlate extremely well with the chloride values. Conductivity measurements for the waters (in microohms per centimeter) 3'j"e appr'oximately an order of magnitude greater than their corresponding chloride concentrations (in mg/l). Conductivity measurements reinforce the conclusions concerning salinity of the water inferred from the chloride data. Although the conductivi of seawater is temperature sens i ti ve (Duxbury, 1971). the temperatures in the waters of these well s all range within a few degrees of 25.C (Plummer and Back, 1980). Sulfate Sulfate ion concentrations in pure seawater have been recorded at 2,511 mg/l (Gross, 1972) and 2,649 mg/l (Duxbury, 1971). Wells 6-3 and 4-2 contain low sulfate values (Fig. 14) in the Hawthorn Formation (less than 500 mg/l), as expected, Below the Hawthorn. in the Tampa and Suwannee where aquifer waters flow, an increase in sulfate concentrations occur, reflecting an advancing freshwater-saltwater interface. Wells 17-1, 17-3, 97. and 124 give high sulfate concentrations. Overall. the sulfate values are in agreement with the model of saltwater intrusion originally deduced from the chloride data. DISCUSSION In attempting to formulate a mineralogical model of the Floridan aquifer several factors be considered: (1) the age of the carbonate units and their original lithologies and environments; (2) the mineralogical and physical evolution of the aquifer; (3) the nature and variations of the hydrologic regimes to which the aquifer rocks were subjected; and (4) the distribution of the various geochemical and hydrologic parameters.

PAGE 36

CONDUCTIVITY OF GROUNDWATER B -34-Figure 13. IN JlOHMSICM <500 500-1000 1.QOO.2000 IIIIIlIIB 2000-5000 5000-10,000 1I!lIIII'0,QOO.2O,OOO >20,000 '"' 10 NORT!1 10 VERTICAL SCALE IN FEET J 20 SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOfE DEPTH IN FEET BELOW SURFACE.

PAGE 37

SULFATE CONTENT OF GROUNDWATER B (360') MS (496') -35CONCENTRATIONS IN MGll <100 g 100-500 II1II 500-1000 >1000 Figure 14. 10 4-2 (610') VERTICAL SCALE IN FEET J 10 20 SCALE IN MILES NOTE: NUMBERS IN PARENTHESES DENOTE DEPTH IN FEET BELOW SURFACE.

PAGE 38

-36-In order to explain the waterMr'ock interactions that led to the present mineralogical distribution within the study area, the evolution of the aquifer must first be considered. Following deposition and early dia genesis of the various carbonate units. a lowering of sea level exposed the sediments to meteoricgrmmdvlIater solutions. When carbonate sediments first emerge from the marine environment and undergo flushing by freshwater. the dissolved solids content of the phreatic waters decrease, with the major ions in solution changing from Na+. Cl-. Mg+2. and 504 2 to Ca+2 and HCO-3. Meteoric waters are enriched in atmosphere-derived C02 and are undersaturated with respect to CaC03. Therefore. the groundwaters will tend to dissolve CaC03 In thismanner, fractures and pore spaces in the rocks were enlarged as CaC03 was dissolved, initiating the flow-network of the aquifer. Sea level f:luctuated il1.any times since the deposi on of these carbonate rocks. Each time the sea level was lowered, meteoric groundwaters enriched in CO2 gas en"larged and extended existing flow-networks by a process that led to further physical development of the aquifer. Calcite dissolution is a relatively fast process in CO2-enriched groundwaters. The flow-network within the aquifer probably developed by enlarging and extending pre-existing channels, rather than by developing new ones. In western peninsular Florida, Hanshaw et al. (1971) found that by the time groundwater reaches the zone of mixing with seawater. it is saturated. with respect to co. lei urn carbonate. Seawater in the coastal mix; ng zone may also be saturated with respect to calcium carbonate. However. the seawater would be saturated under a different partial pressure of CO2 gas than the groundwater. The mixing of the two waters, both saturated with respect to calcium carbonate, but having different partial pressures of C02' will result in a range of mixed waters undersaturated tiilith respect to calcium carbonate. It is this control on calcite saturation that allows calcite to be dissolved in the freshwater-saltwater interface. Back et a1. (1979) described this process occurring today in the brackish zone of Xel Ha lagoon in Mexico. Another condition.fostering dolomite formation in the freshwater saltwater interface is a sufficiently high Mg+2/Ca+ 2 ratio. Hanshaw et al. (1971) have shown that the equilibrium occurs at a Mg+2/Ca+2 ratio of approximately 1 in the Floridan aquifer. Dolomitization can occur provided sufficient time and Mg+2 ions are available. The Mg+2/Ca+2 ratio is greater 1 in mixing zone of the Floridan aquifer in pen ar ( 1979). Thus the chemical and kinetic requi fres hwater-sa"1 twa are zation to occur within the n9 zone of the study area.

PAGE 39

-37One of the most significant observations made in inte'rpreting the data presented in the panel diagrams was that the Avon Park Forrnation is almost completely dolomitized in the study area, while the Ocala and Suwannee are mostly limestones. There must have been some factor (or factors) related to the lithologic character of the Avon Park that resulted in its preferential do 1 omit; za ti on. Murray and Lucia (1967) described water-controlled dolomitization and rock-controlled dolomitization. In the case of water-controlled selectivity, the distribution of dolomite is directly related to the availabil ity and access of the dolomitizing solution. They stated that dolomite may form in carbonate rocks that underlie supratidal sediments, as a result of the pro duction of hypersaline brines on the supratidal mudflat. However, dolomite may be absent in carbonates of similar lithology that do not underlie supratidal sediments. Rock-controlled dolomitization is directly related to the physical and chemical characteristics of the rock at the time of dolomitization. In their study. Murray and Lucia (1967) found a preference for dolomite to replace lime muds. A chemical factor they believed might be of importance involved the relative solubilities of the particles. They stated that because recent lime muds contain a higher percentage of aragonite than most other carboml.te deposits, the difference in solubility between calcite and aragonite may be sufficient to cause some selectivity. They concluded that another likely physical factor involved the small particle size. The greater surface area of the micrites would enable more dissolution to occur and provide a greater opportunity for replacement by dolomite. Similar dolomitization of mudstones has been described by Choquette and Steinen (1980), ofmicrHes by Land (1973), of wackestones by Inden and Koehn (1979), and of siliceous clays by Schmidt (1965). Land (1973) demonstrated the near contemporaneity of aragonite dissolution and dolomitization of the micrites of the Hope Gate Formation of Jamaica. Sibley (1980) concluded that high-magnesium calcite and aragonite are susceptable to dolomitization, while low-magnesium calcite is not in his study of rock-controlled dolomitization on Bonaire. Sibley stated that if the metastable aragonite and high-magnesium calcite can remain preserved until exposure to the freshwater-saltwater mixing then they will be se"'ectively dolomitized.

PAGE 40

The dolomitized AlVon Park in the study area consists predominately of mudstones and wackestones ch were i ally deposited as aragonite lime muds in supratidal and 1 e the Suwannee and Ocala consist to a greater extent Qf highly fossiliferous packstones and grainstones. If the dolomite were water-controlled, it would probably have been restricted to the extent of the mixing zone. However, Pleistocene terraces in peninsular Flori prove that sea level once stood 270 feet higher than its present level (Cooke. 1945). As sea level the mixing zone would have passed through Ocala and Suwannee rocks. Because those units are not dolomi zed in the study dolomite selectivity was more likely related to the ne-grained. high surface area nature of the lime muds in the Avon Parle The form.tl_tion remajned Lml ithified until it was exposed to a meteoric water regime (Randazzo et al" 1977). This supports the theory that the Avon Park micr-ites remained as the metastable, more soluble aragonite phase until their exposlIre to freshwater solutions. Direct dolomitization of aragonite precursor's would then have been an additional rock-selective factor, as described by Murray and lucia (1967) and Sibley (1980) The parameters presented in the panel diagrams support a mechanism of dolomitization by a freshwater-saltwater mixing zone for the Avon Park lithofacies in the study area. The parameters also show that the composition and salinity of the waters in the mtKing zone varied and resulted in the formation of dolomite with a range of MgC03' The mole-percent MgC03 diagram (Fig. 11) indicates that dolomite of a more-stoichiometric form developed in the lower reaches of the Avon Park within the study area. This would have been related to a freshwater-saltwater mixing zone of lower salinity than that en formed the less-stoichiometric dolomites lying above. Sarver (1978) suggested that the less-stoichiometric dolomites of the Avon Park formed in a coastal mixing zone that moved laterally in response to sea level uctuations. The more-stoichiometric dolomites, he believed, formed in a less-saline. inland mixing zone which moved vertically, mainly in t'esponse to atmospheric conditions and groundwater recharge. It is sea level changes could also cause considerable vertical movement the coastal mixing lone. During sea level regressions the coastline would areas to in1 B.nd groundwater'S westward. could be moved farther out to sea. exposing new recharge, plus the movement of CiUJs.e to a discharge point farther migration of the mixing

PAGE 41

-39zone. Thus, the distribution of less-stoichiometric dolomite overlying more-stoichiometric dolomite could be a result of a mixing zone controlled by fluctuations in sea level. Strontium and sodium data (Figs. 9 and 10) present additional evidence for the model described. The strontium content of the more-stoichiometric dolomites in the Avon Park averages less than 200 ppm, while the concentrations in the less-stoichiometric dolomites range between 200-400 ppm, indicating a more saline dolomitizing solution. It was also found that the Sr+2 concentration in the Avon Park dolomite, overall, is more than 50 percent higher than that of the calcite. As stated earlier, Sarver (1978) believed that this was the result of dolomitization having taken place in a more saline environment than calcite neomorphism. Another explanation, however, may be related to the mineralogy of the precursor sediments. Veizer and Demovic (1974) obtained higher Sr+2 concentrations in sediments that were originally aragonite muds than in other sediment types. The various carbonate mineral phases occur in two different crystal classes: (l) rhombohedral(calcite and dolomite), and (2) orthorhombic (aragonite). The orthorhombic structure is larger than the rhombohedral ;structure, and therefore large cations, such as Sr+2 fit more readily into the orthorhombic aragonite structure than in the rhombohedral calcite structure (Hanshaw et a1., 1971). If the majority of the Avon Park carbonates remained as aragonite until dolomitization, as proposed, then a larger concentration of Sr+2 would have been available for incorporation into the dolomite lattice than dolomitization of calcite would have allowed. Sodium concentrations are less than 500 ppm for the calcite samples of the Ocala analyzed. Thus, the sodium data, in agreement with the other data, denote that the calcites of the Ocala Limestone within this study area were stabilized to low-magnesium calcite by a relatively freshwater solution. The more-stoichiometric dolomites also reveal low Na+ concentrations 500 ppm), while the less-stoichiometric dolomites analyze between 500-1,000 ppm with a layer of higher concentrations (1,000-1,500 ppm) occurring within it. The sodium data, therefore, agrees with the strontium and mole-percent MgC03 data concerning the paleoenvironmental interpretation of the dolomitizing solutions. Rock-selective dolomitization is also indicated for the Miocene deposits of the study area. The Hawthorn Formation in western peninsular Florida

PAGE 42

contains beds of and carbonates interfingering vllith one another (Weaver and Beck, i ), ar Fl do. the uppermost part of the Tampa Formation contains green-eJays interbedded with carbonates, while the lower part cont,dns 1 imestone (t
PAGE 43

-41The higher trace element concentrations in tvliocene its may refi the relative youthfulness of those carbonates, with a lesser degree of recrystal1 ization. having occurred. The distribution of the hydrogeochemical parameters (Figs .. 129 13, and 14) reveal that the freshwater-saltwater interface does not encroach in a uniform manner upon the coastal aquifer. Wells 21-2. 17-1. 17-3, and 124 indicate that saltwater is close to the surface in their particular but the remai ning 'area is currently being flushed by reI atively fresher water' solutions. The discrepancies in the height that the salb/ater has attained in those wells that penetrate the Ocala and Avon Par'l< ay' a result of the physical structure and flow network of the aquifer. Dissolution conduits enable seawater to penetrate the aquifer in some places, while sewhere the channels contain freshwater under pressure. The greater abundance of impermeable Miocene sediments results in a higher potentiometric surface than in the other units (Fig. 3), As a result, the Ghyben-Herzberg principle predicts that the freshwater-saltwater inter face should be at a lower depth under the Miocene formations. The parameters signify this, because nowhere in the cores of the southern part of the study area does the saltwater lens Y'ise to the level that it does in places in the northern part. Slow percolation through the Hawthorn Formation is reflected in the gradual increases in salinity with depth. On the other the development of the aquifer, itself, produced a network of large dissolution channels inltJhich the groundwater 1 ens and the sa ltwa tel" lens can enter into each other IS realm producing an irregular boundary along the coast, Therefore. the majority of the dolomitization of the Avon Park in the study area took place during times when sea level was higher than it is today, and seawater was able to infiltrate the aquifer through channels that are currently in the phreatic meteoric zone. Also, as a result, the Avon Park Formation subjected to greater interaction with the coastal and inland mixing environments than the Hawthorn and upper Tampa. THE MODEL-A SUMMARY OF DATA AND INTERPRETATIONS Deposition of carbonate sediments occurred in the supratidal. intertidal. and subtidal environments of a carbonate bank along the west coast of peninsular Florida during the Tertiary Period. The position and extent

PAGE 44

-42-of these environments changed with fluctuations in sea level, and uplift and downwarping of the Florida Platform. Subtidal sediments were likely to have been originally low-magnesium calcite predominantly, while intertidal and supratidal sediments were mostly aragonite. Lime muds were more common in the supratidal and deep subtidal zones. The higher energy environment in which mud was winnowed developed high primary interparticle porosity. A multivariate computer analysis, utilizing multidimensional scaling, revealed several trends that correspond well with interpretations based on petrographic and megascopic examinations of three cores representing the Avon Park and Ocala carbonate sequences. MDS was used to determine environmental characteristics normal to the shoreline for these sediments. The analysis revealed the changes in ttJe carbonate bank depositional environments. This was depicted by two hydrodynamic curves produced by MDS which showed variations from muddy, dolomitic, quartzose, supratidal sediment clusters to a shallow subtidal, high energy sediment cluster. During the Miocene carbonate muds and siliceous clays were deposited in brackish water lagoons, along with sands and phosphates. This resulted in the formation of a layer of low permeability which today defines the upper boundary of the Floridan aquifer. The Avon Park Formation probably remained unlithified until exposure to a meteoric water regime. Fine-grained dolomite crystals formed in the supratidal lithofacies from aragonite mud precursors. During periods of subaerial exposure of the sediments, flushing by meteoric waters occurred. The unstable high-magnesium calcite sediments inverted to low-magnesium calcite, accompanied by the inversion of aragonite sediments. Lime mud inverted by porphyroid neomorphism to microspar, and in doing so, aided in lithification. Shell fragments and other sediments pro bably underwent dissolution-reprecipitation and recrystallization, which destroyed original textures and resulted in the of pseudospar. As the unstable minerals were dissolved, their cation and anion chemistry were contributed to the solution. The newly formed carbonates had lower concentrations of Sr+2 Na+, and Mg+2. The dissolution-reprecipitation process was also responsible for the formation of sparry calcite cement. In the vadose meteoric zone, finely crystalline cement was precipitated from meniscus solutions in interparticle pore spaces and at grain contacts. large rhombohedral sparry crystals slowly precipitated in pores in the phreatic meteoric zone.

PAGE 45

-43-While groundwater was responsible for precipitating it was also a major factor in the production of secondary porosities. Preferential dissolution of fossil tests resulted in moldic pores and vugs. Mudstones and wackestones developed less secondary porosity than paci<.stones and grainstones because of their fewer allochems and lower permeabili During periods when sea level was lowered. groundwater dissolution the limestone enlarged existing pore spaces as the groundwater opened paths towards discharge out into the Gulf of Mexico. Succeeding marine trans gressions resulted in inland migration of the freshwater-saltwater interface. Selective dolomitization took place at the freshwater-saltwater interface in the Avon Park, Hawthorn, and upper Tampa The dolomite selectivity was related to (1) the fine grain size of the sediments with their high surface area/volume ratio; (2) the occurrence of earlier formed dolomite; (3) the precursor carbonate being aragonite; and (4) the occurrence of clay minerals. Dolomitization proceeded in the freshwater saltwater' mixing zone because (1) mixing freshwater and saltwater can result in a zone where the waters become undersaturated with respect to calcite. yet remain super saturated with respect to dolomite; (2) the salinity of the mixing zone is lower than it is in seawater, therefore there is less foreign ion competition inhibiting dolomite formation; and (3) the Mg+2/Ca+2 ratio remains well above the level of unity required for dolomitization. Where the dolomitized precursor sediments were calcite, the dolomite crystals grew by coalescive neomorphism. Where penecontemporaneous dolomite crystals were already present in supratidal lithofacies, the fine crystals underwent porphyroid neomorphis, resulting in larger, more equant-shaped crystals. In both cases dolomitization proceeded by a process of dis-.solution-reprecipitation, accompanied by some loss of Sr+2 and Na+ the amount depending upon the mineralogy of the precursor the role of clay minerals as a contributor of trace elements, the salinity of the doiomitizing solution. and the duration of time in which sediments were subjected to dolomitizing solutions. The aggrading neomorphic growth of dolomite crystals resulted in the redistribution of many small pores into fewer, but larger, pores. Therefore. although the porosity may not have appreciably changed. permeability could have greatly increased. In the less saline reaches of the mixing zone, more-stoichiometric dolomite formed slowly without the interference of many foreign ions. This resulted in large, well-developed. rhombohedral crystals

PAGE 46

-44with low Sr+2 and Na+ concentrations. Less-stoichiometric dolomite, with associated higher Sr+2 and Na+ values formed more rapidly in the more saline areas of the mixing zone. Diagenesis of carbonate rocks is a continuous process. Both groundwater and saltwater solutions in contact with the rocks are constantly exchanging ions with the rock in an attempt to produce the most stable mineralogical conditions possible. The porphyri c neomorphic growth of dolomite crystals to larger dolomite rhombs is added evidence of this continuing process occurring in dolomites. The interface is probably the most dynamic of the in which changes occur'. 14C data suggest formation or recrysto.l1 ization of dolomite occurred in the last 30,000 years. Each time that sea level changes moved the mixing zone back into dolomitized units, recrystallization of the rocks could have occurred, with an accompanying trace element depl orL it is reasonable to assume that dolomites which originally formed in the aquifer during earlier sea level fluctuations are still striving towards more-stoichiometric conditions in the present freshwater-sa ltwater mi xi n9 lone. Carbonate systems tend towards equi 1 ibri um with age and time (Lumsden and Chimahusky. 1980). and the expected trend for dolomite is towards a more nearly stoichiometric compound. ACKNOWLEDGEMENTS I am deeply grateful for the many suggestions, interpretations, points of guidance, and cooperation provided by Daniel Spangler, Paul Mueller, James Eades, and Michael of the University of Florida, to William Back of the United States Geological Survey, to Tom Scott of the Florida Bureau of Geology, and to Greg New and Kim Preedom of the Southwest Florida Water ivlanagement District. Special acknowledgement must be made to Jon Bloom vlho compiled and synthesized much of the data in this report while a graduate student under my direction. The information provided in this paper must be credited to a large part to his dedication and professional effort to the tasks confronting him. The work upon which this report is based was supported in part by funds provided by the United States Department of the Interior as authorized under the Water Research and Development Act of 1978.

PAGE 47

-45REFERENCES CITED Applin, P.L., and E. R. Applin, 1944, Regional subsurface stratigraphy and structure of Florida and southern Georgia: AAPG Bull., v. 28, p. 1673-1753. Back, W., B. B. Hanshaw, T. E. Pyle, L. N. Plummer, and A. E. Weidie, 1979, Geochemical significance of groundwater discharge and carbonate solution to the fonnation of Caelta Xe1 Ha, Quintana Roo, Mexico: Water Resources Research, v. 15, p. 1521-1535. Badiozamani, K., 1973, The dorag dolomitization model, application to the middle Ordovician of Wisconsin: Jour. Sed. Petrology, v. 43, p. 865-984. Behrens, E. W., and L. S. Land, 1972, Subtidal Holocene dolomite, Baffin Bay, Texas: Jour. Sed. Petrology, v.42, p. 155-16l. Bloom, J. I., 1982, A mineralogical model of the Floridan'aquifer in the Southwest Florida Water Management Pi-strict: Master's thesis, Univ. of Florida, 140 p. Chen, C. S., 1965, The regional lithostratigraphic analysis of Paleocene and Eocene rocks of Florida: Florida Geo1. Survey Bull. 45, 105 p. Choquette, P. W., and R. P. Steinen, 1980, Mississippian non-supratidal dolo mite, Ste. Genevieve Limestene .. Illinois Basin, evidence for mixed-water dolomitization, in Concepts and models of dolomitization: SEPM Spec. Pub. 28, p.-r63-196. Cooke, C. W., 1945, Geology of Florida: Florida Geo1. Survey Bull. 29, 339 p. Donn, W., W. Ferrand, and M. Ewing, 1962, Pleistocene ice volumes and sea level lowering: Jour. Geology, v. 70, p. 206-214. Duxbury, A. C., 1971 .. The' earth and its oceans: Reading, Massachusetts, Addison-Weslleyc'pUb: Co., .381 Ekda1e, A. A., S. F. Ekda1e, and J. L. Wilson, 1976, Numerical analysis of carbonate microfacies in the Cupido Limestone (Neocomian-Aptian), Coahuila, Mexico: Jour. Sed. Petrology, v. 46, p. 362-368. Fenk, E. M., 1979, Sedimentology and stratigrapI\Y of middle and upper Eocene carbonate rocks, Lake, Hernando and Levy Counties, Florida: Master's thesis, Univ. of Florida, 133 p. Fetter, C. W., 1980, Applied hydrogeology: Columbus, Ohio, Charles Merrill Pub. Co., 488 p. Folk, R. L., 1965, Some aspects of recrystallization in ancient limestones, in Dolomitization and limestone diagenesis: SEPM Spec. Pub. 13, Po" 14-48. Folk, R. L., and L. S. Land, 1975, Mg/Ca ratio and salinity: two controls over crystallization of dolomite: AAPG Bull., v. 59, p. 60-68. Friedman, G. M., 1964, Early diagenesis and lithification in carbonate sediments: Jour. Sed. Petrology, v. 34, p. 777-813. Fritz, P., and A. Katz, 1972, The sodium distribution of dolomite crystals: Chern. Geology, v. 10, p. 237-244. Ginsburg, R. N., 1957, Early diagenesis and lithification in shallow-water carbonate sediments in south Florida, in Regional aspects of carbonate. deposition: 'SEPM Spec. Pub. 5, p. 80-100. Gross, M. G., 1972, Oceanography, a view of the Earth: Englewood Cliffs, New Jersey, Prentice-Hall Pub. Co., 498 p.

PAGE 48

-46Hanshaw. B. B., and Back, 1979, Major geochemical processes in the evolution of Hydrology, v. 43, p. 287-312. l-Iansha\'I. B. B W. Back, and R. G. Dieke,_1971, A geochemical hypothesis for dolopjtization ground water: Econ.Geology, v. 66, p. 710-724. Hayes, B. J. R 1 A cluster analysis interpretation of middle Ordovician bi ofacies, southern Mackenzie Canadian Jour. Earth Sci v. 17. p. 1377-1388. Hie'key. E. W 1976. Sedimentology and dolomitization in Eocene carbonate rocks, Gilchrist and Marion Counties. Florida: Master1s thesis, Univ. of Florida, 74 p. Imbrie, J. and E. G. Purdy. assifkation of modern Bahamian carbonate Carbonate Rocks: AAPG Mem. 1, p. 253-272. sediments, in Classifi on Inden. R. r., H. Koehn, deposits in Hammett v. 63, p. 472. Jordan, L., 1954. 1 passi p. 370-375. 1979, Dolomitization of offshore carbonate e, lower Cretaceous, Texas: (abs.) AAPG Bull., lities in Florida: Oil and Gas Jour v. 53, Kahle, C. F., 1965, Possi e roles clay minerals in the formation of dolomite: Jour. Sed. Petrology, v. 35. p. 448-453. Kinsman, D. J. J., 1969, Interpretation of Sr+2 concentrations in carbonate minerals and rocks: Jour. Sed. Petrology. v. 39, p. 486-508. Kohout, F. A., 1965, A hypothesis concerning cyclic flow of salt water related to geothermal heating in the Floridan aquifer: N.Y. Acad. Sci., Trans., v. 28. p. 249-271. Kohout, F. A 1967. Ground=water flow and the geothermal regime of the Floridan Plateau. in Symposium on geologic history of the Gulf of Mexico Caribbean Antillean Basin: Gulf Coast Assoc. Geol. Soc. Trans., v. 17. p. 339-354. Kruskal. J. B., 1964a, Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis: Psychometrica. v. 29. p. 1-27. Kruskal. J. B 19Mb. Nonmetdc mu1tidimensional scaling: a numerical method: v. p. 11 129. Land, L. S., 1970. Phreatic 'IS. vadose meteoric diagenesis of limestones, evidence from a i1 water Sedimentology. v. 14, p. 175-185. Land, L. S 1973$ Contemporaneous dolomitization of middle Pleistocene reefs by meteoric \'I/ater, north ea: Bull. Marine Sci., v. 23, p. 64-92. Land. L. S., 1980. The isotopic and trace element geochemistry of dolomite, the state of the art, in Concepts and models of dolomitization: SEPM Spec. Pub. 28. p. 87-105. Land. L. 5., and S. Epstein, 1970. late Pleistocene diagenesis and dolomitiza tion, north Jamaica: Sedimentology. v. 14. p. 187-200. Land. L. 5., and G. K. Hoops, 1 Sodium in carbonate sediments and rocks, a possible index to the sali of diagenetic solutions: Jour. Sed. Petrology. v. 43, p, 614.-617. U, J, S., 1978, cores. tee 1V. Point, Brandon. and Ouette drill es, Florida: Master's thesis.

PAGE 49

-47-Lumsden. D. N., and J. S. 1 Rel p between dol te non-stoichiometry and carbonate facies in Concepts and mode"ls of dolomitization: SEPM Spec. Pub. 28. p. 123-138--. Metrin, D. B., 1979, Geochemical significance of selected ions in Eocene carbonate rocks of peninsular Florida: Master's Univ. of Florida, 64 p. Murray. R. C.. and F. J. Luci a. 1967, Cause andcontro 1 of omite di s bution by rock Geol. Soc. Amero v. 78, p. 21-36. Parks, J. M., 1966, Cluster analysis applied to multivariate geologic problems: Jour. Geology, v. 74, p. 703-715. Plummer. L. N and W. Back, 1980, The mass balance approach: application to interpreting the chemical evolution of hydrologic systems: Amer. Jour. Sci., v. 280. p. 130-142. Puri, H. S., 1957, Stratigraphy and zonation of the Ocala Group: Florida Geol. Survey Bull. 38, 248 p. Randazzo, A. F 1976, Petrographic and geohydr'ologic model of aquifer limestones in Florida: Flotida Water Resources Research Center, Pub. No. 35, 51 p. Randazzo. A. F., 1980, Geohydrologic model of the Floridan aquifer in the Southwest Florida Water Management District: Florida Water Resources Research Center, Pub. No. 46, 79 p. Randazzo. A. F., and E. W. Hickey. 1978. Dolomitization in the F"loY'idan aquifer: Amer. Jour. Sci., v. 278, p. 1177-1184. Randazzo, A. F., G. C. Stone, and H. C. Saroop, 1977, Diagenesis of middle and upper Eocene carbonate shoreline sequences, central Florida: AAPG Bull., v. 61, p. 492-503. Reichenbaugh, R. C., 1972, Sea-water intrusion in the upper part of the Floridan aquifer in coastal Pasco florida, 1969: Florida Bur. Geology Map Series 47. Rowell, A. J., D. J. McBride. and A. R. Palmer, 1973, Quantitative study of Trempealearin (latest Cambrian) trilobite distribution in North America: Geol. Soc. America Bull., v. 84, p. 3429-3442. Runnells, D. D 1969. Diagenesis. chemical sediments. and the mixing of natural waters: Jour. Sed. Petrology, v. 39. p. 1188-1201. Saroop, H. C., 1974, Sedimentology, paleoecology and diagenesis of middle and upper Eocene carbonate shoreline sequences, River. Floy'ida: Master's thesis, Univ. of Florida, 165 p. Sarver. T. J., 1978, Geochemical analysis of selected Eocene carbonate rocks of peninsular Florida: Master!:; thesis, Univ. of Florida, 77 p. Schmidt, V., 1965, Facies, diagenesis, and related reservoir properties in the Gigas beds (upper Jurassic), Germany. in Dolomitization and limestone diagenesis: SEPM Spec. Pub. 13, p. 124-168. Sharpe, C. L 1980, Sedimentological interpretation of carbonate rocks from west central Florida: Masterls thesis, Univ. of Florida. 170 p.

PAGE 50

-4BSibley, D. F., 1980, Climatic control of dolomitization, Seroe Domi Formation (Pliocene), Bonaire, N.A., in Concepts and models of dolomitization: SEPM Spec. Pub. 28, p. 247-258. Smosna, R. and S. M. Warshauer, 1979, A scheme for multivariate analysis in carbonate petrology with an example from the Silurian Tonoloway Limestone: Jour. Sed. Petrology, v. 49, p. 257-271. Soka1, R. R. and P. H. A. Sneath, 1963, The Principles of Numerical TaxonomY: San Francisco, W. H. Freeman and Co., 359 p. Steinen, R. P., and R. K. Matthews, 1973, Phreatic vs. vadose diagenesis, stratigraphy and mineralogy of a cored Borehole in Barbados, W. I.: Jour. Sed. Petrology, v. 43, p. 1012-1020. Stone, G. C., 1975, Paleoecology and diagenesis of the Avon Park Formation, Gulf Hanmock Wildlife Mangement Area, Levy County, Florida: Master's Univ. of Florida, 128 p. Stringfield, V. T., and H. E. LeGrand, 1966, Hydrology of 1 imestone terranes in the coastal plain, of the southeastern United States: Geo1. Soc. America Spec. Paper' 93, 46 p. Valentine, J. W. and R. G. Peddicord, 1967, Evaluation of fossil assemblages by cluster analysis: Jour. Paleontology, v. 41, p. 502-507. Veizer, J., and R. Demovic, 1974. Strontium as a tool in facies analysis: Jour. Sed. Petrology, v. 44, p. 93-115. Veizer, J., J. Lemieux, B. Jones, M. R. Gib1ing, and J. Savelle, 1977, Sodium: Pa1eosa1infty indicator in ancient carbonate rocks: Geology, v. 5, p. 177-179. Vernon, R. 0., 1951, Geology of Citrus and levy Counties, Florida: Florida Geol. Survey BUll. 33, 256 p. Weaver, C. E., and K. C. Beck, 1977, Miocene of the S. E. United States, a model for chemical sedimentation in a peri-marine environment: Ams terdam, El sevi er Pub. Co., 214,. p. Weber, J. N., 1964, Trace element composition of dolostone and dolomites and its bearing on the dolomite problem: Geochim. et. Cosmochim.Acta, v. 28, p. 1817-1868. Wetterhall, W. S., 1964, Geohydrologic reconnaissance of Pasco and southern Hernando Counties, Florida: Florida Geol. Survey Rept. of Investigation 34, 28 p. Wetterha11, W. S., 1965, Reconnaissance of springs and sinks in west central Florida: Florida Geol. Survey Rept. of Investigation 39, 42 p. Weyl, P. K., 1960, Porosity through dolomitization, conservation-of-mass requirements: Jour. Sed. Petrology, v. 30, p. 85-90. Whittington, H. B. and C. P. Hughes, 1972, Ordovician geography and faunal provinces deduced from trilobite distribution: Phi10s. Trans. Royal Soc. London, v. 263, p. 235-278. Winston, G. 0., 1976, Florida's Ocala Uplift is not an uplift: AAPG Bullq v. 60, p. 992-994. Zachos, L. G., 1978, Stratigraphy and petrology of two shallow wells, Citrus and Levy Counties, Florida: Master's thesis. Univ. of Florida, 105 p. Zenger, D. H., 1972, Significance of supratidal dolomitization in the geologie record: Geol. Soc. Amer. Bul1., v. 83, p.1-12.