SINKHOLE ACTIVITY IN WEST-CENTRAL FLORIDA:
A GEOGRAPHIC INFORMATION SYSTEMS ANALYSIS
By
KATHLEEN COMMINS
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2004
ACKNOWLEDGMENTS
Although there are many to whom I owe thanks, it is Dr. Randazzo to whom I give
special thanks. It was his outstanding patience and guidance that kept me on the right
track. I thank Dr. Smith for being a committee member. Also, during my tenure in
Geological Sciences at the University of Florida, both Dr. Randazzo and Dr. Smith have
always shown enthusiasm for geology and far exceeded their call of duty. I would also
like to thank Dr. Meert for being such a conscientious committee member, and due to his
input, the final product is a better one! Special thanks go to GeoHazards, Inc, for
allowing me to access proprietary materials for this database, and for gainful employment
while working on it.
I would like to acknowledge all those at the Southwest Florida Water Management
District for their patience and for supplying some of the data within this thesis; and to
Sam Palmer for his assistance, and for sharing his ArcView expertise with me. Without
this selfless cooperation I would still be trudging on!
TABLE OF CONTENTS
page
A CK N OW LED G M EN TS ................................................................................................ ii
LIST O F TA BLES ...............................................................................................................v
LIST O F FIGU RES ..................................................................................................... vi
A B STRA CT..................................................................................................................... viii
CHAPTER
1 IN TRO D U CTIO N ................................................................................................... 1
B ackground........................................................................................................... ..... 1
Sinkhole A architecture ................................................ ........................................... .2
Purpose .................................................................................................................. 5
2 GEOMORPHOLOGY OF THE STUDY AREA............................................. ............ 7
Structural and D epositional H history ..................................................... ...............7
G eom orphology.................................................................................................... 9
H ydrogeology ............................................................................................................. 11
Clim ate.............................................................................................................14
3 M ETH OD O LO G Y .....................................................................................................17
M ethodology of Sinkhole Investigation ..................................... ............................ 17
Previous Studies..........................................................................................................18
Criteria for Sinkhole Recognition...................................................... ... ...........18
Soil Testing..........................................................................................................19
G eophysics ....................................................................................................21
D ata.............................................................................................................................22
A acquisition of D ata..............................................................................................22
Processing of D ata ...............................................................................................25
4 RESU LTS ..................................................................................................................28
H ydrogeology .............................................................................................................28
Overburden Thickness to the top of the Floridan Aquifer System...........................28
Physiographic Provinces .......................................................................................34
Recharge/Discharge in the Floridan Aquifer System...................................... ...40
Potentiometric Surface..........................................................................................48
G eology................................................................................................................. 51
Major Wells in the Study Area ..................................................................... 54
Population Density................................................................................................58
5 A N A LY SES ......................................................................................................... 64
6 CONCLUSIONS ..................................................................................................77
APPENDIX
A GLOSSARY OF ARCMAP TECHNICAL TERMS ...................................... ...83
B HISTORICAL RAINFALL FOR STUDY AREA.......................................... ...85
REFERENCES CITED................................................................................................86
BIOGRAPHICAL SKETCH ....................................................................................... 92
LIST OF TABLES
Table page
2.1 Relationship of regional hydrogeologic units to measure stratigraphic units..........16
3.1 Parameters of the counties of the study area ....................................................27
4.1 Thickness of the overburden to the Floridan aquifer system, compared to various
sinkhole param eters..................................................................................... ..........29
4.2 Sinkhole parameters of the Physiographic Provinces ............................................36
4.3 Area of the discharge and recharge zones of the Floridan aquifer system in relation
to various sinkhole param eters................................... ..........................................40
4.4 Subsurface geology of the study area in relation to various sinkhole parameters ...52
4.5 Highest annual water pumpage (gallons) for top wells in each county of study area
for selected years................................................................................................55
4.6 Population density in relation to various sinkhole parameters...............................61
5.1 Results from GIS analyses of sinkhole distribution and significant themes ............67
LIST OF FIGURES
Figure page
1.1 Human-induced sinkhole development in the study area, Pasco and Hernando
counties (after Tihansky 1999)........................................ ..................................3
1.2 Types of sinkholes (from Tihansky 1999) ............................................ .............4
1.3 Map of the five-county study area (from Southwest Florida Water Management
D district 1990)........................................................................................................ 6
2.1 Structural features of the Florida peninsula (from Scott 1997)..............................8
2.2 Location of sinkholes in physiographic provinces of the study area .....................12
3.1 Counties of the study area showing locations of sinkhole activity of both the old
database and new ................................................................................................27
4.1 Distribution of sinkholes and overburden thickness to the Floridan aquifer
system ................................................................................................................. 30
4.2 Sinkhole density in relation to thickness of overburden ........................................31
4.3 Relative frequency of sinkholes in relation to thickness of overburden to the
Floridan .............................................................................................................. 32
4.4 Confining units of the Floridan aquifer system (after U.S. Geological Survey
"Ground Water Atlas of the United States, modified from Miller 1990) ................33
4.5 Sinkhole density in each of the Physiographic Provinces of the study area ............36
4.6 Relative frequency of sinkholes in each physiographic provinces of the study
area ..................................................................................................................... 37
4.7 Physiographic provinces versus sinkhole relative frequency and percentage of
total area ............................................................................................................. 38
4.8 Distribution of sinkholes and the Subsurface geology of the study area, with
overlay of rivers ................................................................................................. 39
4.9 Counties map with sinkhole locations, major roads and major rivers....................43
4.10 Recharge/Discharge zones measured in inches/year and sinkhole distribution.......44
4.11 Recharge/Discharge zones of the study area, (a) Sinkhole density (b) Relative
frequency of sinkholes .......................................................................................45
4.12 Overlay of the Thickness of the Overburden theme and the Recharge/Discharge
theme with the sinkhole distribution theme .....................................................47
4.13 The potentiometric surface values with sinkhole locations ...................................49
4.14 Sinkhole density and lithostratigraphic units of study area....................................53
4.15 Relative frequency of sinkholes in lithostratigraphic units of study area ...............54
4.16 Sinkhole distribution on the geology theme, with the 30 most highly pumped
wells in each county for the study area ................................................................57
4.17 Distribution of sinkholes and the US Census Bureau's Block Groups showing
population density for 2000..............................................................................60
4.18 Population density (km2) versus sinkhole density using US Census Bureau block
groups 2000 ........................................................................................................ 62
4.19 Sinkhole relative frequency for population density 2000 (km2) ............................63
5.1 GIS overlay of all the areas of highest sinkhole relative frequency for all digital
parameters used in the analyses.......................................................................... 68
5.2 All digital parameters with the greatest sinkhole density for the study area...........70
5.3 Recharge and Dishcarge zones with sinkhole distribution............................ ...72
5.4 Recharge/Discharge theme overlaid with the Depth to the Floridan Aquifer
System them e ..................................................................................................... 73
5.5 GIS intersection of all the parameters that displayed the highest sinkhole relative
frequency ............................................................................................................... 76
6.1 Proximity to the nearest sinkhole, indicated by one, two, and three kilometer
buffer zones ........................................................................................................ 81
6.2 Proximity to the nearest sinkhole, indicated by 0.5, 1.0, and 1.5 kilometer buffer
zones........................................................... ....................................................... 82
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
SINKHOLE ACTIVITY IN WEST-CENTRAL FLORIDA:
A GEOGRAPHIC INFORMATION SYSTEMS ANALYSIS
By
Kathleen H. Commins
May 2004
Chair: Anthony F. Randazzo
Major Department: Geological Sciences
Geographic Information Systems (GIS), involving spatial analysis, was used to
examine correlations between locations of sinkhole activity and the factors influencing
sinkhole distribution in West Central Florida. For the analyses, new data were collected
from private sources and combined with the existing database of the Florida Geological
Survey. The new data include 648 sites that were investigated from 1990-2001. Factors
that were suspected to be influencing sinkhole distribution were overlain and examined
for correlations. These factors were compared using sinkhole density and sinkhole
relative frequency and included the following themes: thickness of the overburden
overlying the Floridan aquifer system, physiographic provinces, discharge/recharge of the
Floridan aquifer system, potentiometric surface data, Florida's subsurface geology, the
major wells in each county, and the U.S. Census Bureau's population density block
groups. Analyses showed some strong and some weak correlations.
Certain factors appear to be of significance. Within the study area, the greatest
sinkhole activity is where the thickness of the overburden to the Floridan aquifer system
is 75 feet or less, where the near surface geology consists of the Hawthorn Group, the
Quaternary undifferentiated sediments, or the Suwannee Limestone, and within the Gulf
Coastal Lowlands and Coastal Swamps province.
The new database was used to construct maps showing locations of sinkhole
activity, and the various themes that could affect sinkhole frequency and distribution. In
an effort to address prediction of areas of future sinkhole activity, a GIS application of
nearest neighbor occurrences of sinkholes was conducted. It produced a map showing
areas where future sinkhole activity can be expected. These include most of Northern
Pinellas, southwest Hernando, the western coastline of Pasco, and northwestern
Hillsborough counties.
CHAPTER 1
INTRODUCTION
Background
The most commonly recognized feature of karst topography is the sinkhole.
Sinkholes are a significant geologic hazard and lead to a loss of many millions of dollars
each year. Sinkhole-related land subsidence problems are global and occur within 45
states in the United States, affecting over 17,000 square miles (Galloway et al. 1999).
A better understanding of sinkhole development in Florida is of importance to
planners at all levels. These include homeowners; architects; and engineers who design
both private and commercial structures; land developers and realtors; insurance agencies;
bankers; and local governments that issue permits for construction, waste disposal, and
water use. Awareness of sinkhole development and areas most sensitive to its occurrence
can be turned into a tool to be used to help avoid damage and financial loss caused by this
geologic feature.
Florida's rain water has a pH of 4.77 on average (Upchurch and Randazzo 1997),
and is an integral participant in the dissolution of limestone. As rain percolates through
the soil, it obtains CO2 from the metabolization of humus (from organic soils), lowering
the pH of the water entering the aquifer and dissolving Florida's Eocene to Miocene
carbonate bedrock, particularly during sea-level lowstands. Dissolution has given the
carbonate bedrock its high permeability or "swiss-cheese characteristic." One direct
benefit of this is the Floridan aquifer system (FAS), but a negative effect is the sinkhole
activity and the traveling of overburden into the voids left in the bedrock by the ongoing
dissolution process.
Many sinkholes are anthropogenically induced and are mainly caused by either
water withdrawal from wells or construction activities (Lane 1986). It is these activities
that can trigger instability, leading to sinkhole formation. For example, Sinclair (1982)
documented that in Tampa, Florida over 30 small sinkholes occurred north of a municipal
well-field within 1 year of the massive pumping from this well. One well field increased
from 5 mgal/d to 14 mgal/d, and 64 new sinkholes were reported within 1 month. Figure
1.1 shows some anthropogenically induced sinkholes.
Sinkhole Architecture
Sinkholes form by two processes (Sinclair et al. 1985): traveling, which is
downward erosion of unconsolidated sediments into preexisting cavities; and collapse, a
process where weak acidic solutions dissolve carbonate bedrock, creating cavities that
can collapse, or in which unconsolidated overburden sediments can migrate. Lithology
and thickness of the overburden determine sinkhole morphology (Sinclair 1985; Schmidt
and Scott 1984). According to Schmidt and Scott (1984) the frequency of sinkhole
occurrences in Florida, and their sizes are associated with the thickness of the overburden
(i.e., depth below land surface) to the FAS. In other words, both frequency and karst
feature size are directly related to stratigraphy. The three classic sinkhole types in
Florida are Dissolution, Cover-collapse, and Cover-subsidence (Figure 1.2).
Dissolution sinkholes are formed by weak acidic solutions infiltrating areas where
carbonate is exposed or only a thin layer of overburden is present (Tihansky 1999). The
acidic solution dissolves the underlying carbonates, creating cavities. In many cases, the
water flow is focused along fractures, bedding planes, or joints. This type of sinkhole
usually forms gradually as a bowl-shaped depression (Beck and Sinclair 1986).
Figure 1.1. Human-induced sinkhole development in the study area, Pasco and Hernando
counties (after Tihansky 1999).
Cover-subsidence sinkholes usually develop where the sediments are permeable
and contain sand (Tihansky 1999). Gradual traveling (downward migration of granular
sediments into openings in underlying carbonates) is known as piping, and the structures
formed are called solution pipes. As dissolution and infilling continue, a small surface
depression forms (Tihansky 1999). The overburden thickness may range from 15.2
meters (50 feet) or greater (Sinclair and Stewart 1985).
Cover-collapse sinkholes are usually in areas where he e ere is a thick covering of
sediments that contain a large amount of clay (Tihansky 1999). The clays usually
separate the sandy overburden from the underlying carbonates. Clays can provide some
cohesion to the overburden, and when a solution cavity develops below, the overburden
can act as a "bridge." The collapse occurs when the solution cavity can no longer support
the overburden and the "bridge" gives way. This kind of sinkhole can develop rapidly
and cause catastrophic damage (Beck and Sinclair 1986).
TYPES OF SINKHOLES -.-- .---.-.- I... ..
Dissolution of the limestone or dolomite
is most intensive where the water first con-
tacts the rock surface. Aggressive dissolution
also occurs where flow is focused in pre-
existing openings in the rock, such as along
joints, fractures, and bedding planes, and
in the zone ofwater-table fluctuation
where ground water is in contact with the
atmosphere.
Thin overburden Rain
--_z ]. Rainfall and surface water percolate through
L- joints in the limestone. Dissolved carbonate
I rock is carried way from the surface and a
i .L small depression gradually foa s.
' ,>o !_- _
Pt-I
L `-~--
On exposed carbonate surfaces, a depression
may focus surface drainage accelerating the
dissolution process. Debris carried into the
developing sinkhole may plug the outflow
ponding water and creating wetlands.
Gently rolling hills and shallow depressions caused by solution sinkholes
are common topographicfeatures throughout much of Florida.
Cover-subsidence sinkholes tend to
develop gradually where the covering sedi-
ments are permeable and contain sand.
Granular sediments spall A column of overlying sedi- Dissolution and infilling con- The slow downward erosion
into secondary openings ments settles into the tinue forming a noticeable eventually forms small sur-
in the underlying carbonate vacated spaces (a process depression in the land face depressions I inch to
rocks. termed "piping. surface- several feet in depth and
diameter.
m''n--i '-- -- '-- --'
In areas where cover material is thicker or sediments contain more clay, cover-subsidence
sinkholes are relatively uncommon, are smaller, and may go undetected for long periods.
Cover-collapse sinkholes may develop
abruptly (over a period of hours) and cause
catastrophic damages. They occur where the
covering sediments contain significant
amount of day.
Sediments spall into a cavity. As spelling continues the The cavity migrates up- The cavity eventually
cohesive covering sedi- ward by progressive roof breaches the ground sur-
ments form a structural collapse. face, creating sudden and
arch dramatic sinkholes.
-J -I
Overtime, surface drainage, erosion, and deposition ofsediment transform the steep-walled
sinkhole into a shallower bowl-shaped depression.
Figure 1.2. Types of sinkholes (from Tihansky 1999)
Purpose
The purpose of this study is to broaden existing sinkhole database inventories and
to assemble information into a Geographic Information System (GIS), making it
accessible to the many sectors of the community affected by the environmental dangers
and economic impact that sinkhole activity causes. The database covers 5 Florida
counties where sinkhole activity is high (Figure 1.3). Results in this study will be
compared to results of a similar study done by Dodek (2003) for any correlations.
A further objective is to determine whether or not increased sinkhole occurrence
is correlated to geologic features and anthropogenesis (i.e., aquifer withdrawal rates). To
achieve these objectives, the data are compared as spatial thematic coverage in separate
GIS layers, allowing data to be easily retrieved and compared for correlations. The
software used is ArcMap, GIS software that was developed by the ESRI company.
Specialists can use this information as a tool for identifying areas with high
potential for sinkhole activity, or to more accurately access the environmental impact that
can be endured safely. A Florida Department of Insurance (FDI) study in 1992 (Butler et
al. 1992) pointed out the importance of identifying and predicting sinkhole occurrence.
This database and study should be very beneficial for better understanding the rates and
distribution of karst evolution.
STUDY AREA
Figure 1.3. Map of the five-county study area (from Southwest Florida Water
Management District 1990)
CHAPTER 2
GEOMORPHOLOGY OF THE STUDY AREA
Structural and Depositional History
Early Mesozoic rifting and sea-floor spreading between the African and North
American plates (Hine 1997) formed the basement for the Florida peninsula, a pre-
Jurassic crystalline rock body overlain unconformably by Jurassic and Cretaceous
sediments (Smith and Lord 1997) with the thickest sediments occurring in the southern
part of the peninsula (Heatherington and Mueller 1997). Situated on the North American
Plate, Florida's coastline is currently a passive margin, but there are features within the
basement that indicate Florida has undergone some structural changes throughout its
geologic history.
The structural features (highs and lows) in Florida that influenced deposition of
sediments include the Ocala Platform (Uplift); the South Florida Basin; the Jacksonville
Basin, and the Chattahoochee Anticline in the northwest; the Sanford High in the east;
the Peninsular Arch; and the Georgia Channel System (which collectively, is the
Apalachicola Embayment, the Gulf Trough, and the Southeast Georgia Embayment)
(Schmidt and Scott 1984). Figure 2.1 shows some of the structural features of Florida
that suggest that the Florida platform responded to tectonic forces during the Mesozoic
Era (Scott 1997). During the Cenozoic, the Peninsular Arch appears to have been the
major high that affected sedimentation in the Paleogene; and the Ocala Platform had the
most influence on the deposition of late Paleogene and Neogene sediments.
LEGEND \ \
AXIS OF POSITIVE
FEATURE -- SCE
AXIS OF NEGATIVE LOW
FEATURE
-APPROXIMATE UPDIP LIMIT
AND AREA UNDERLAIN BY THE
FLORIDAN AQUIFER SYSTEM
0 50 100 150 200 MILES
0 100 200 300 KILOMETERS
SCALE
Figure 2.1. Structural features of the Florida peninsula (from Scott 1997)
The deposition of the thick carbonate sequence covering the platform began in the
Jurassic, after the African plate rifted from the North American plate, with intermittent
periods of siliciclastic deposition until the Cenozoic. The first 35 million years of the
Cenozoic were dominated by deposition of carbonates, and evaporites (Scott 2001), until
the beginning of the Miocene, when transport of siliciclastic sediments over the peninsula
predominated (Walker et al. 1983). Up until then, the Gulf Trough's currents diverted
the Appalachian sediments away from the Platform. The reason for the renewed
siliciclastic influx is debatable, but its deposition, occurring on the platform in the Late
Oligocene, was due to the Gulf Trough's currents no longer occupying the Georgia
Channel System after the Late Oligocene sea level drop. The Appalachian siliciclastics
prevailed, covering the Florida Platform by the middle of the Pliocene. The supply
diminished by Late Pliocene (Scott 1997).
Hawthorn sediments are represented in almost the entire Miocene and part of the
lower Pliocene sections of the study area. These include the Arcadia Formation
(Hawthorn Group, with the Tampa and Nocatee members) and the Peace River
Formation (with the Bone Valley Member and Wabasso beds) (Scott 1997). Subaerial
exposure occurred during low-level stands in the Neogene and Quaternary (Scott 1997).
Deposition of Eocene or lower Oligocene, through Miocene sediments from the
Appalachian Mountains was blocked due to the Gulf Trough/Apalachicola Embayment.
The sediments exposed in the study area are mainly the Pleistocene-Holocene
undifferentiated sands, and the Hawthorn Group.
Geomorphology
Marine forces were the dominating factor in creating the landforms of Florida,
(Schmidt 1997). When the Florida Platform was covered by the sea, erosion and
deposition from the shallow marine currents left behind flat plains and scarps. Whenever
the sea covered the Florida Platform, the shallow marine currents and their associated
erosion and deposition shaped the shallow seabed, leaving subsequent erosional forces to
modify the geometry.
The exposed part of the Florida Platform is the State of Florida. It is approximately
half of the Platform and is about 650 km east to west and 800 km north to south. The
highest point in Florida is 104 meters above sealevel, in Walton County, northwest of the
study area. Land forms help to interpret the geologic history. Dissolution of bedrock has
formed basins for lakes in many parts of central Florida (Miller 1987). Both sea-level
and land-level changes influence the rate of deposition and erosion (Schmidt 1997),
continually changing the geomorphology. Varying sea-levels have alternately flooded
and exposed the Florida platform, and eroding and depositing sediments. When the water
table is low and the limestone bedrock is close to the surface, drainage is usually
"internal" (Schmidt 1997) with streams "disappearing" into the subsurface. This is karst
topography and is a common drainage characteristic in Florida which has developed from
dissolution of the carbonates that exist throughout the platform.
Water can drain through sinkholes rather than running off through stream systems.
Florida's extensive aquifer systems were "carved" out by the movement and chemical
action of the acidic waters. The reaction of the water with the limestone and dolostone, is
summed up by the following, which show the chemical reactions when rainwater comes
in contact with the land surface. The carbon dioxide needed for this reaction to occur is
obtained from the atmosphere and the soil. The disassociation of the carbonic acid
allows it to react with calcite to form karst features.
* CO2 gas + H20 -+ H2C03 carbonic acid
* H2CO3 + H+ +HCO3
* CaCO3 calcite + H2C03 < Ca2+ + 2HCO3
Figure 2.2 shows that the study area lies within several physiographic provinces
which are described by White (1970). The largest are the Gulf Coastal Lowlands,
occupying, 3350 km2, and the Polk Upland, 3119 km2. The smaller provinces are the
Coastal Swamps, Brooksville Ridge, and Zephryhills (Florida Department of
Environmental Protection (FDEP) 2000).
The Gulf Coastal Lowlands in the study area are mostly underlain by the Suwannee
Limestone and the more deeply buried Ocala Limestone (Green et al. 1995). The
lithology of the Suwannee Limestone, early Oligocene in age, is described as a packstone
or grainstone carbonate with trace amounts of sand and clay in the upper portions (Green
et al. 1995; Arthur et al. 2001).
The Polk Upland occupies the western portion of Hillsborough County, and most of
Polk County (Figure 2.2). Exposed Bone Valley Member, Peace River Formation
(Middle Miocene to Lower Pliocene ) sediments make up the western part of the Polk
Upland in eastern Hillsborough County, and the southeastern section of the Polk Upland
in Polk County. These sediments of quartz sand, silt, and clay contain phosphorite
grains, generally with poor consolidation. The Bone Valley Member has been
economically important in the phosphate industry of the United States (Scott 2001). On
the eastern edge of the Polk Upland lies the Lake Wales Ridge and the Osceola Plain,
both undifferentiated sands.
Hydrogeology
Hydrogeologic framework plays a large role in when and where sinkholes are
formed. In west-central Florida the framework is made up of three aquifer systems that
consist of carbonates and siliciclastics rock. The ground water levels (hydraulic head)
and chemistry determine where dissolution of the limestone will occur. Sinkholes also
influence the hydrogeologic framework (Tihansky 1999).
The Floridan (FAS), the Intermediate (IAS), and the Surficial (SAS) aquifers are
present in the study area. The FAS underlies most of the State (Hyde 1975; Wetterhall
12
1964) and is the main source of potable water in the study area. Table 2.1 shows the
regional hydrogeologic units' relationship to the major stratigraphic units in Florida.
N F Counties Lake Hcmny Ridge
SSinkholes Lake Upland
SPhysiopraphic Regions hFL- Wales dge
F-]Brooksvill- Ridge __LakelandRidge
Hernando Brook ee PolkUpland
S Coastal Swa ps opk. Pl
S "= GulfBarner Chain T. IerH i
C. ,C 9? Gulf Coastal LowlCds
;I hr r* r _Zephrl Gp
J 7''- --. --
V o
'-.'.. a
s :: II
imk j" .I *r... !.a *
I .I ... .Kilom e r
Figure 2.2. Location of sinkholes in physiographic provinces of the study area (after
Green et al. 1995), and is made up of undifferentiated sediments in the study area. These
consist mainly of quartz sand and varying amounts of clay and shell material. And in
some of the area, it may also include the Bone Valley Member of the Peace River
Formation/Hawthorn Group. The Bone Valley Member contains significant amounts of
phosphorite (Scott 1992) When present, the SAS either overlies the IAS, or when this
Si a it vris t S. I t S ria r g
K, ..* .,.1 I' -
.It'" "Hillsborough Polk Upland e
0 20 40 80
Figure 2.2. Location of sinkholes in physiographic provinces of the study area (after
White 1970).
Most of the study area contains the SAS in varying thicknesses (Arthur et al. 2001;
Green et al. 1995), and is made up of undifferentiated sediments in the study area. These
consist mainly of quartz sand and varying amounts of clay and shell material. And in
some of the area, it may also include the Bone Valley Member of the Peace River
Formation/Hawthorn Group. The Bone Valley Member contains significant amounts of
phosphorite (Scott 1992). When present, the SAS either overlies the IAS, or when this
system is absent, it overlies the FAS. In the Southwest Florida Water Management
District (SWFWMD) the SAS is a limited water resource in the northern section, but of
increasing importance to the south of the study area (Scott 1992).
The Southeastern Geological Survey (SEGS 1986) defined the IAS as the rocks
that lie between the Surficial and Floridan aquifers, retarding water exchange between the
two. The IAS consists of siliciclastics of the Hawthorn Group interlayered with Miocene
and younger carbonates (Scott 1992). Cross-sections developed by Arthur (2001) and
Green et al. (1995) indicate that the IAS is present in part of Pinellas, Hillsborough, and
Polk counties and is mostly thin to absent in Hernando and Pasco counties.
Sinclair and Stewart (1985) note that the southern portion of the Southwest Water
Management District (SWFWMD has limited karst development and few karst conduits
penetrate the intermediate aquifer system. The Florida Sinkhole Research Institute
(FSRI) and this database confirm this pattern in karst development. The FAS lies deeper
in these regions.
The FAS underlies the entire study area and is dominated by Paleocene to Miocene
carbonate sediments (Miller 1986). The thickness of the Floridan within the state varies
from 100 feet to over 3,500 feet and within the study area, is between 1,400 and 3,400
feet thick ( USGS 1990). Its major units are the Avon Park and Oldsmar formations, and
also include the Ocala and Suwannee limestones (Scott 1992). The upper Floridan within
peninsula Florida, is made up mostly of the Ocala Limestone (Scott 1992). The FAS
supplied more than 60% of the total groundwater used in the state in 1985 (USGS 1990),
which was 2,503 million gallons of water per day (Mgal/d).
The Ground Water Atlas of the United States' (USGS 1990) publication shows that
within the study area, in 1980, daily withdrawal from the FAS in Polk County was 200 to
350 million gallons, Hillsborough withdrew 100 to 300 million Mgal/day, Pasco, 50 to
100 Mgal/day, and Pinellas and Hernando both withdrew 20 to 50 Mgal/day. Of the total
FAS withdrawal, about 30% is from 5 counties, and 4 of these are within the study area -
Polk, Hillsborough, Pasco, and Pinellas (USGS 1990).). Sinkholes are induced by
aggressive pumping (draw-down of wells). This is due to abrupt changes in ground water
levels which disturbs the equilibrium of karst features and overburden (Tihansky 1999).
Climate
Florida is located between latitudes 24.50 and 31.00 degrees north, (Henry et al.
1994) and the study area is within latitudes 27.65 and 28.70 degrees north. Florida is
classified as having both a subtropical and tropical climate, but all of the study area is
within the subtropical region ["humid subtropical" according to the K6ppen classification
system of 1918 (Henry et al 1994)]. It is also characterized as low relief and high rainfall
(Miller 1997). Precipitation recharges the aquifers in Florida. The average annual
rainfall for the twenty year period from 1981 thru 2001 for the study area ranges from 52
to 55 inches (132 to 140 cm) by county. Historically, between 1915 and 2001, the
average annual rainfall for the same areas ranges from 49 to 53 inches (124 to 135 cm),
(SWFWMD 2003; NOAA 2003). This illustrates that the trend for rainfall has varied
little over the past 85 years. The average temperature in the study area over the past 80
years has ranged from 72 to 740, with the average low during this period being 61 to 62
degrees and the average high between 82 to 840 (SWFWMD 2003).
Continentality and oceanicity are the terms used to classify an area as to whether
land mass or oceans (respectively) has the most influence over its climate. According to
Currey (1974), Florida is classified as Oceanic, and has the least continentality of the
eastern United States (Henry et al. 1994). This has an influence on the precipitation
15
occurring in the state, which is approximately 150 billion gallons per day (Henry et al
1994). Movement of these great quantities of acidic water through the
limestone/dolostone aquifer system has promoted the dissolution process.
Table 2.1. Relationship of regional hydrogeologic units to measure stratigraphic units
Ir J iI -I -i rL i
PANHANDLE FLORIDA
SYSTEM SERIES LITHOSTRATIGRAPHIC UNIT HYDROSmATI-
GORAPHIC UNIT
i -... ...
QUARTERNARY
TERTIARY
CRETACEOUS
AND OLDER
HOLOCENE
PLEISTOCENE
PULIOCENE
MIOCENE
UNDIFFERENTIATED
PLEISTOCENE-HOLOCENE
SEDIMENTS
CITRONELLE FORMATION
MICCOSUKEE FORMATION
COARSE CLASTICS
ALUM BLUFF GROUP
PENSACOLA CLAY
INTRACOASTAL FORMATION
HAWTHORN GROUP
BRUCE CREEK LIMESTONE
ST.MARKS FORMATION
CHATTAHOOCHEE FORMATION
CHICKASAWHAY LIMESTONE
SUWANNEE LIMESTONE
OLIGOCENE MARIANNA LIMESTONE
BUCATUNNA CLAY
EOCENE
OCALA LIMESTONE
CLAIBORNE GROUP
UNDIFFERENTIATED SEDIMENTS
----
PALEOCENE
I II I I I I I
UNDIFFERENTIATED PALEOCENE ROCKS
UNDIFFERENTIATED
8URFICIAL
AQUIFER
SYSTEM
INTERMEDIATE
CONFINING
UNIT
FLORIOAN
AQUIFER
SYSTEM
BUB-FLORIOAN
CONFINING
UNIT
-, 5
NORTH FLORIDA
SOUTH FLORIDA
UTHOSTRATIGRAPHIC HYDROSTRATI- LITHOSTRATIGRAPHIC HYDROSTRAn-
UNIT GRAPnIC UNIITUNIT GRAPHIC UNIT
iI i. ..
UNDIFFERENTIATED
PLEISTOCENE-HOLOCENE
SEDIMENTS
MICCOSUKEE FORMATION
CYPRESHSEAD FORMATION
NASHUA FORMATION
HAWTHORN GROUP
STATENVILLE FORMATION
COOSAWHATCHIE FM.
MARK8HEAD FORMATION
PENNY FARMS FORMATION
ST MARKS FORMATION
SUWANNEE LIMESTONE
OCALA LIMESTONE
MON PARK FORMATION
OLDSMAR FORMATION
CEDAR KEYS FORMATION
UNDIFFERENTIATED
SURFICIAL
AQUIFER
SYSTEM
INTERMEDIATE
AQUIFER
SYSTEM OR
CONFINING
UNIT
'--I. -
FLORIDAN
AQUIFER
SYSTEM
CONFINING, ,
TUNIpT
UNDIFFERENTIATED
PLEISTOCENE-HOLOCENE
SEDIMENTS
MIAMI LIMESTONE
KEY LAROO LIMESTONE
ANASTASIA FORMATION
FORT THOMPSON FORMATION
CALOOSAHATCHEE FORMATION
TAMIAMI FORMATION
HAWTHORN GROUP
PEACE RIVER FORMATION
BONE VALLEY MEMBER
ARCADIA FORMATION
TAMPA- NOCATEE
MEMBERS
SUWANNEE LIMESTONE
OCALA LIMESTONE
AVON PARK FORMATION
LOSMAR FORMATION
CEDAR KEYS FORMATION
UNDIFFERENTIATED
.- .. p 3
SURFICIAL
AQUIFER
SYSTEM
INTERMEDIATE
AQUIFER
SYSTEM OR
CONFINING
UNIT
FLORIOWAN
AQUIFER
SYSTEM
SUB-FLORIDAN
CONFININGs ,,
UNIT, P
(from Fernald and Purdum 1998)
I I I =i i
L c
, ,
CHAPTER 3
METHODOLOGY
Methodology of Sinkhole Investigation
When there is an unexplained distress to a structure, subsurface evaluations are
usually requested by a property owner (or his representative) or an insurance company.
In areas that have high incidence of karst features, subsurface evaluations should also be
performed prior to land development. The evaluation is to determine the potential or
probability for sinkhole formation and damage to a future structure. Since a provision in
a homeowner's insurance policy covers the peril of sinkhole induced damage or loss,
determination of the cause of damage is pertinent. The rapid growth in Florida's
population may be responsible for the large increase in claims to insurance companies for
sinkhole related damage to property. The FDI Sinkhole Standards Summit could not
develop a uniform set of criteria for the investigation of subsurface disturbance and
concluded that no one uniform set of criteria could be universally applied to the
investigation of sinkhole claims, but is "site specific" (Butler et al. 1992). A list of
minimum standards that must be employed during a sinkhole investigation was compiled
for the geotechnical professionals involved in subsurface evaluations. Since
"shrink/swell clay activity may be the largest single cause of building damage in the
U.S." (Frank and Beck 1990 p.1), this Summit set forth criteria to help differentiate
between damage that occurred from sinkhole activity or shrink/swell clays. Other
deleterious soil conditions (e.g., organic) were also addressed.
Previous Studies
Many sinkhole studies in the past have been proprietary. Some have not included
studies of homes and other structures that have resulted in damage due to sinkhole
activity. One of the studies currently available to the public was conducted by the Florida
Sinkhole Research Institute (FSRI). The FSRI was founded in 1982 (and housed at the
University of Central Florida). The FSRI compiled an inventory of sinkholes that were
formed between 1960 and1996. These data are included in a 2001 report by Burwell and
Wilson. In 1992, when many issues concerning sinkhole damage were arising, in
compliance with a Florida Legislature directive, a study of sinkhole occurrence was
conducted by the Florida Department of Insurance (FDI) (Butler et al., 1992). One major
concern at that time was the role that the insurance companies would have to play in
restitution to the homeowner of a house damaged by sinkhole activity. There was no
centralized repository established for a database of sinkhole and traveling occurrence in
Florida. It was agreed upon by geologists, engineers, and academia that there was a great
and growing need for ongoing research for sinkhole occurrence in Florida, and a
centralized repository (Butler et al.1992). Although the Florida Geological Survey (FGS)
was appointed as the repository for the database, this depository has not been well
maintained, due mostly to lack of funding. The FGS provides a limited database. The
new data from this study significantly broadens the FGS database. Also, the database
from a 2003 study by Dodek is available to expand this new database.
Criteria for Sinkhole Recognition
* Criteria recommended by the FDI included the following:
* An on-site investigation that includes background history of the house, property,
surrounding properties, environment, and quality and workmanship of the structure.
* Familiarity with all the geological characteristics pertaining to the area. The
stratigraphy of an area is of major importance, since soil and rock type, and
thickness of the units is pertinent in a geological site analysis.
* Onsite soil sampling for shrink/swell clay and organic.
* Although geophysical studies are not a requirement, it has been the practice of the
geologic profession to consider geophysical surveys pertinent in an investigation to
meet the Florida Statutes in denying or verifying the presence of sinkhole activity.
The most common geophysical techniques employed in an investigation are ground
penetrating radar (GPR) and electrical resistivity (ER). The protocol for ER is
prescribed by the American Society for Testing and Materials (ASTM) and is
protocol G57-95a. The ASTM protocol for GPR methods is 6432-99.
Soil Testing
Soil samples are collected by two methods, either by hand augering or by a
standard penetration test. Standard penetration tests (SPT) involve a two-fold procedure.
Firstly soil samples are obtained and secondly, N values are recorded. An SPT is
conducted in accordance with ASTM standard D1586-99, entitled "Standard Test Method
for Penetration Test and Split-Barrel Sampling of Soils". The samples are usually
retrieved at 1.5 meter intervals. Samples are described by an on-site geologist and are
retained for further lab analysis when a need is indicated. In compliance with the ASTM
standards, an SPT boring is performed by advancing a drill bit into the ground using a
rotary drill rig. A sample is obtained by attaching a 2 inch long split-barrel to a
succession of drilling rods. The rods are advanced by the force of a 140 pound hammer
that is dropped freely from a 30 inch height. The number of blows required to advance
the bit six inches is recorded. To compute an "N" value, the sum of the blows for
advancing the second and third 6 inch increments provides the "N value". The
significance of the N value is for assigning a relative density to a soil, thereby
determining the relative stability of the subsurface material. The N value for each type of
soil has a different range of behavioral densities. For example, sand with an N value
between 0 and 4 is very loose but a limestone with an N value of a much wider range of
0 and 19 is very soft.
When appropriate, the samples might be subjected to laboratory analysis. Tests
include, but are not restricted to, clay swelling potential, grain size analysis, natural
moisture content, and percent organic. Randazzo and Smith (2003) explain the
importance of soil analyses during site investigations where property damage is suspected
to be sinkhole related.
* Expansive clay. The water content of clayey soil can affect the deformational
behavior of that soil. Atterberg limits (ASTM D4318) are a group of tests used in
determining the water content boundaries between the semi liquid and plastic states
(known as the liquid limit) and between the plastic and semi solid states (known as
the plastic limit) of a soil. This is also known as the plasticity of the soil mass.
Where the soil just begins to crumble when it has been rolled into a 1/8 inch thread,
it has reached its plastic limit, that is, it contains the percentage of moisture where
it can just retain this characteristic. When enough moisture has been added to a soil
to make it just begin to flow, it has reached its liquid limit. This range in moisture
content represents the plasticity index (PI). It is calculated by subtracting the
moisture content of a soil in the plastic state (PL) from the moisture content of that
soil in a liquid state (LL) (i.e., PI = LL PL). The United States Department of
Army (1983) recognizes a low PI to be less than 25, marginal PI to be between 25-
35, and greater than 35 to be highly plastic. Also, possible volume change of a
wetted soil should be determined. The volume change is the shrinkage factor limit
and the smaller the shrinkage limit, the greater the chance of volume change of the
soil.
* Grain Size. Clays must also be analyzed for grain size. This analysis is critical in
determining whether or not the clay's PI is relevant in causing damage to a
structure. The percentage of colloidal sized particles that make up a soil is a major
factor in the soils volume change and its plastic characteristics (Nelson and Miller
1992). Sridharan and Prakash (2000) have determined that a strong attenuation
factor in the shrinkage limit of a soil is the presence of coarse grain sizes (sand).
The acceptable sieve size for this analysis is the -200 mesh (ASTM D1140). A -
200 mesh filters out the clay and silt size particles and hence determines what
percentage of fines is contained in the material. There is more attenuation of the
shrink/swell behavior of the clay, the larger the grain size and the higher the
percentage of coarse material.
* Moisture. The natural moisture content of a soil is relevant when investigating a
site for sinkhole activity. The soil moisture percentage will indicate the relative
water saturation state. Soils with relatively low soil moisture percentages,
occurring during the rainy season or below the local water table, are indicative of
non-expansive soils (ASTM D2216).
* Hand auger borings. They are usually advanced to an approximate depth of 1.5
meters, in accordance with ASTM standards D1452-80, entitled "Standard Practice
for Soil Investigation and Sampling by Auger Borings. This results in a continuous
vertical sampling profile of the near surface materials. The technician looks for
clays and organic, and notes the water table, if it is penetrated. When clays are
present, it is very important to note at what depth they occur. Shrink/swell clay
within the upper 1.5 meters can be the cause of, or a contributing factor to,
structural damage (Chen 1988). Field identification is limited and can be
subjective, so soil samples are retained for lab analysis, which is more definitive.
* Organics. High organic content can be a contributing factor for structural damage
to both commercial buildings and homes. Organic material, especially when in a
dry environment, decomposes, decreasing the original volume of the soil. The
decrease in volume allows subsidence to occur. It is the subsidence that allows
movement of the structure and hence damage. The threshold for organic content
(ASTM D2974) in a soil before damage can take place is 5% (Frank and Beck
1990, p.20).
Geophysics
a) Ground penetrating radar (GPR) is a geophysical technique used by geologists
as an aid in interpreting the subsurface. The GPR device transmits microwave radiation
into the subsurface as it is dragged along a selected configuration on the surface of the
ground. When the radiation encounters a reflective surface, the waves are reflected back
and are collected by the GPR receiver. Materials are identified by the intensity of the
reflections they produce when the electromagnetic wave passes through them. Contacts
between rock types, foreign objects (debris and garbage for example), and most
importantly, voids and soil disturbance, can be detected. An anomaly in the subsurface is
indicated by the contrasting dielectric intensities of materials. Anomalous areas might be
indicators of raveled zones, or in filled paleosinks. These areas can be further
investigated by drilling. Where shallow clay layers are present, GPR signals can be
attenuated or absorbed, therefore limiting the effectiveness of the technique, and thus
necessitating other geophysical techniques to be used.
The GPR field data is downloaded to a computer system where it can be displayed
as a profile. A qualified geologist then interprets the profile to determine whether or not
anomalies exist and hence, whether or not sinkhole activity is present.
b) Electrical resistivity (ER) is another geophysical technique that can be
employed during a subsurface investigation. This technique is performed by placing
electrodes into the ground, using the configuration in accordance with the ASTM
standards. The Wenner electrode configuration, which is a 4 electrode procedure (ASTM
G57-95a), utilizes the four electrodes by connecting them to a direct current battery that
passes an electrical current into the ground. The resistance to flow is measured. Since
different earth materials, cavities, and manmade materials are characterized by their
resistance to flow, an evaluation of subsurface materials can be made. The depth to
which the current reaches is determined by the distance between the electrodes, and the
greater the distance, the greater the depth of flow.
One or both of these geophysical techniques was employed for all of the sinkhole
investigations included in this study.
Data
Acquisition of Data
Geotechnical related data from sites that have been investigated for sinkhole
activity from 1986-2001, provide the database for this study. This database includes 665
new sites within the study area, including 15 in Hernando, 75 in Hillsborough, 212 in
Pasco, 320 in Pinellas, and 43 in Polk (Table 3.1). The new data were obtained by
reviewing over 500 reports of geophysical investigations conducted by GeoHazards, Inc.,
a geotechnical consulting firm in Gainesville, Florida, and from the records of legal firms
within the study area. Most of the investigations were conducted at home sites, and the
remainders were located at private businesses. In order to develop a more comprehensive
study, these data have been combined with the Florida Geological Survey's database,
which documents cases from 1960 thru 1996, then displayed in a geographic information
system (GIS). This gives a total of 1,664 sinkholes and sites where sinkhole activity has
occurred.
A large percentage of these new (unpublished) cases are insurance-related sites
where sinkhole activity has occurred and sinkhole investigations have been performed.
As discussed in Randazzo and Smith (2003), and Zisman (2003), a geotechnical survey is
used to determine if sinkhole activity is occurring at a site. Geotechnical results are then
reviewed by the responsible insurance company, and if a settlement can not be negotiated
between the insurance company and the insured, litigation may ensue. When sinkhole
activity is recognized, an engineering company will be retained to evaluate mitigation of
the structure's damage.
The following is a description of the data used in this study, how it was obtained
and processed for review and analysis in GIS.
* More than 368 subsurface investigations performed by GeoHazards, Inc., from
1986 through 2001. The firm uses geophysics and other techniques to evaluate
subsurface conditions for the occurrence of sinkhole activity. The reports reviewed
contained detailed information of the investigation including on site descriptions
and diagrams of the site, hand auger results, lab analysis when applicable,
geophysics, and SPT's when applicable. Each site was identified by an
alphanumeric address and in order to display the sites on a map of Florida within
ArcView, the Excel file (Appendix 1) had to be put into a database format (dbf) and
then geocoded (converted into a latitudinal and longitudinal coordinate system).
The program used for geocoding was EZlocate and was downloaded from Teleatlas
at Teleatlac.com. The newest version of ArcView 8.3 (ArcMap) has included a
geocoding extension, which was utilized for geocoding additional addresses. An
additional 287sites reported by a legal firm are included, though they did not all
include all the parameters that were in the GeoHazards, Inc. reports.
* The Florida Sinkhole Research Institute's database of sinkhole occurrences,
between 1960 and 1996, is included in this study as an ArcView layer and occurred
in multiple Florida counties. In 1995 Spencer and Lane published the data in the
FGS Open File Report 58, "Florida Sinkhole Index". At that time it was available
in dBase format and is now current through 2001. It is expected to be updated with
additional cases that have been reported by private citizens, the Water Management
Districts, Department of Transportation, Sheriff's departments and the Departments
of Community Affairs. The FDEP verified all the reports prior to adding them to
the database. In recent years a private company, "Subsurface Evaluations, Inc."
converted the data into a Microsoft Excel spreadsheet and it is now available on the
FGS website. The file has been projected into Albers Spheroid Global Reference
System (GRS) 1983 HPGN (Central Meridian, -84 degrees; Central Parallel, 24
degrees; Standard Parallel 1, 24 degrees; and Standard Parallel 2) and was
downloaded from the Florida Geological Survey web site,
http://www.dep.state.fl/us/geology/gisdatamaps.
* The Geologic map of Florida published by the FDEP was downloaded from
http://www.dep.state.fl/us/geology. This map was in Albers HPGN projection. It
is included as an ArcView layer to enable the overlay of other data to identify
correlations that might exist between sinkhole occurrence and stratigraphy or
geology.
* The Florida base map and county maps were obtained from the Florida Geographic
Data Library (FGDL). These are available on the GeoPlan website,
http://www.fgdl.org/ or at the GeoPlan office at the University of Florida on CD.
The digital maps are in Albers projection and are based on HPGN.
* The County Potentiometric surface maps for 1975, 1985, 1994, and 2000 were
obtained from SWFWMD website at,
http://www.swfwmd.state.fl.us/data/dataonline/. This data layer was in UTM Zone
17, North American Datum 83/90 (NAD83) projection so had to be reprojected into
Albers HPGN, which was completed by a SWFWMD GIS technician. They were
then imported into this study as an ArcView layer.
* Thickness of the overburden to the top of the FAS maps were obtained from
ftp.dep.state.fl.us/pub/gis/data. These data had not yet been published and were
only available in Raster format. Therefore, to make them compatible with the other
themes, it was necessary to convert them into Polygon format and project them into
Albers.
* A Physiographic Provinces map was obtained from http://www.fgdl.org/. Like
other FGDL data, its projection was Albers, therefore compatible with the other
study themes.
* A map indicating areas of the recharge and discharge zones of the FAS was
obtained from SWFWMD. It was projected in UTM Zone 17, North American
Datum 83/90 (NAD 83), so in order to import into ArcView as a theme compatible
with the other data, it was converted to Albers by using the ArcView Projection
Utility. The values were then reclassified into workable categories.
* A map that shows the 30 top pumped wells in each county of the study area for
1980, 1985, 1990, 1996, and 2001was constructed from data supplied by
SWFWMD. The data were in Excel spreadsheet format and degree/minute/second
coordinates. The coordinate system was converted to decimal degrees and then
imported into ArcView as a theme.
* Other data that are included in this study but not available in GIS format, are
individual maps addressing ground water and the Floridan aquifer. The maps were
taken from the online version of "The Ground Water Atlas of The United States",
at http://www.capp.water.usgu.gov/, and include (i) a map of the thickness of the
Floridan aquifer system, (ii) a map showing the 3 areas in the state with the largest
withdrawal from the Floridan aquifer in the state. One of these lies within the
study area. (iii), a map showing the amount of freshwater withdrawal from the
Floridan aquifer, and (iv) a map of transmissivity of the Floridan aquifer.
The reports reviewed for this study contained physical addresses for each site. The
addresses were geocoded (assigned longitude and latitude), to enable projection of the
points onto a map within ArcView. By converting the data to a GIS format, multiple
layers of different variables were constructed and overlain. This allowed the layers to be
"turned-off and turned-on" so that any pattern of sinkhole occurrence could be more
easily recognized. Features that have correlations such as geomorphologic features and
sinkhole distribution, soil types, geology, and hydrogeologic parameters can be
recognized. The geology, stratigraphic data, rainfall data, potentiometric surfaces, and
water withdrawal from the Floridan aquifer can then be integrated with these features.
Processing of Data
After each set of data was collected it was then reviewed and converted to an
ArcView theme, when possible, or compared side-by-side with physical maps when
digital data was not available. By putting the data in a GIS format, layers could be
integrated, overlain, or removed. This made it possible to make correlations and identify
patterns in the distribution of occurrence of sinkholes/raveling, the geology, stratigraphy,
potentiometric surface, high magnitude springs (not in digital format), rainfall (not in
digital format), and the recharge and discharge of the Floridan Aquifer System. Each of
these layers is shown in the chapters that follow. For the data that were not in a digital
format (Springs, Rainfall) a side-by-side comparison was made to the digital data. Table
3.1 provides a summary of the main parameters of each county's data.
With the data in ArcView, analyses were made and exported to Excel, where tables
and graphs were generated for further analyses. Sinkhole distribution is the parameter to
which all other data was compared. Sinkhole density was calculated by dividing the
number of sinkholes by the total area for each parameter and sinkhole relative frequency
was determined by dividing the number of sinkholes per parameter by 1,664, the total
number of sinkholes in the study area (Figure 3.1). When analyzing the data, both
sinkhole density and sinkhole relative frequency have been considered. This was an
attempt to address some obvious biases. One is called a "collection bias". Collection (or
reporting) of data can be affected by certain factors, such as where population is greatest.
Where population is greatest reporting would be expected to be greatest. Use of relative
frequency can reduce reporting biases. Instead of simply using the number of sinkholes
per square kilometer (sinkhole density) utilization is made of the percentage of the
number of times the sinkholes occur within a given parameter (sinkhole relative
frequency). Despite the inherent biases of databases such as that developed in this study,
actual sinkhole locations have been identified on a regional scale and allow for causal
correlations.
Table 3.1. Parameters of the counties of the study area
Number of Relative Sinkholes
County Area km2 Sinkholes Frequency / km2
Pinellas 748.97 369 22.9% 0.49
Pasco 1980.95 431 25.9% 0.29
Hillsborough 2773.54 404 24.8% 0.15
Hernando 1276.41 203 12.2% 0.16
Polk 5208.30 257 15.4% 0.05
Totals 11,988.17 1664
I1lometers
0 20 40 80
Figure 3.1. Counties of the study area showing locations of sinkhole activity of both the
old database and new (after Florida Geographic Data Library 1990)
CHAPTER 4
RESULTS
The set of maps, charts, and tables presented in this chapter provide an effective
visualization of each parameter used for analyses with sinkhole distribution. Each
parameter was overlain on the sinkhole distribution map, calculations made and exported
to Excel where tables were constructed. The parameters in digital format include
hydrogeology, major well fields, physiographic provinces, recharge and discharge,
overburden to the FAS, subsurface geology, potentiometric surface (for 4 years),
thickness of the FAS, and population density for 2002. The non-digital data used for
determining correlations included rainfall and springs. The tables and charts show both
sinkhole density and its relative frequency.
Hydrogeology
Overburden Thickness to the top of the Floridan Aquifer System
Data for this theme were provided by the FDEP. The FDEP constructed the
model by subtracting the top of the FAS coverage from the Floridan digital elevation
model (DEM). They obtained the Top of the Floridan maps from multiple sources and
spliced the maps together in order to make a state-wide theme. Categories of
overburden thickness reflect generally acceptable divisions that provide a more useful
resolution of sinkhole distribution for this theme. The sinkhole theme was joined to the
overburden coverage and the results are in Table 4.1 and Figure 4.1.
Data are also presented in histogram format (Figures 4.2 and 4.3). Figure 4.2
indicates that there is a correlation between the number of sinkholes per km2 (sinkhole
density) and the thickness of the overburden (i.e. depth to the aquifer). There is a greater
distribution of sinkholes where the depth to the aquifer is less 75 feet or less. Though
sinkholes do occur in areas where the overburden is thicker than this, the occurrence is
much less. For example, where the thickness of the overburden is greater than 300 feet,
sinkhole occurrence is less than 1%. Sinkhole activity in these more thickly covered
areas may be due to extraordinarily large cavity systems residing below the overburden.
The sinkhole relative frequency results (Figure 4.3) indicate an even stronger
correlation between sinkhole distribution and overburden thickness. It indicates that
approximately 73% of the sinkholes occur in overburden less than 75 feet thick and 27%
in areas where the overburden is greater than 75 feet.
Table 4.1. Thickness of the overburden to the Floridan aquifer system, compared to
various sinkhole parameters
Overburden # of Sinkholes/ Relative
Depth in feet Sinkholes Area (km2) km2 Frequency
0-15 358 1,655 0.22 21.5%
16-30 314 960 0.33 18.9%
31-50 318 1,284 0.25 19.1%
51-75 226 589 0.38 13.6%
76-100 62 449 0.14 3.7%
101-150 95 1,532 0.06 5.7%
151-200 115 1,208 0.10 6.9%
201-300 162 2,439 0.07 9.7%
301-400+ 14 1,676 0.01 0.8%
S*Sinkoles Ravelling 51-75
S Rivers 76-100
Floridan Overburden
101-150
0-15 ft
151-200
M16-30
31530 201-300
31-50
301400+
Kilometers
0 20 40 80
Figure 4.1. Distribution of sinkholes and overburden thickness to the Floridan aquifer system (after Florida Department Environmental Protection raster data 2003)
N
_ _~~__~~~ _~~_~
I
0.40 0.38
0.35- 0.33
0.30
025
E 0.25-
0.27
M 0.20-
C)
,. 0.14
C 0.15-
.10
0.10-,
.07
0.06
0 01
Figure 4.2. Sinkhole density in relation to thickness of overburden
32
25%
a) 1501 13 6%
t-
2.
U.
). 9.7"%
The 6.9Fh
5.70
591- 3.7
0 8'a
__- I-- II I -"
0-15 16-30 31-50 51-75 76-100 101-150 151-200 201-300 301-400+
Overburden Thickness (feet)
Figure 4.3. Relative frequency of sinkholes in relation to thickness of overburden to the
Floridan
The FAS is confined within some geographic areas and unconfined in others. It is
noted in the USGS Ground Water Atlas of the United States (1990) that large solution
cavities are present in areas where the confining unit is thin or absent. The map (Figure
4.4) representing the confining units, was not available in digital format, so an overlay
with the Sinkhole and Overburden themes was not possible. A side by side comparison
of Figure 4.1 and Figure 4.4, shows that the study area has three categories of
confinement (1) the upper confining unit thin or absent (2) upper confining unit generally
less than 100 feet thick, breached, or both (3) and areas where the confining unit is
generally greater than 100 feet and not breached. Where the sinkhole occurrence is high,
Figure 4.1 shows that the thickness of the overburden is 75 feet or less. Figure 4.4 also
indicates that in these areas, the upper confining unit of the Floridan is either absent or
thin, less than 100 feet or breached, or it is both.
In these situations where the overburden is thin, and the confining unit is thin or
absent, acidic waters can more easily leach units of the FAS and cause dissolution of the
limestone beds. As could be expected, the comparison also shows that where the
Floridan is confined and generally greater than 100 feet thick, that sinkhole occurrence is
less.
_- W
...-. ,' I' n ..' -- ,
EXPLANATION
SArea here Froridon aquifer systern is uronfined- icpr-
c ,rl.nmin L-n, ,% ahbs.rl r- hn
R Area here Floridan aquUer system is thInl) ronned- -pp*.r
,..'i'lirgn 1 Jr i s V y rc lallv Ic" i= 1 00 klct c'h br cch'-.d.
or both
Area where Fioridan aquifer system Is confined-- pper confin-
ing unit is qc rrally greater then 100 feet thick end urnrfe4ched
Lower Floridan aquifer confined by more than 200 feet of low-
permeabllty rocks
------ Approximate rnit of upper confining unit
The ciaycy rocks of the upper cnhinb~q unit of
the Hortdan aquifer system have been eroded aw~ay cornpletWy in
ptere~s and arr tes.t than 1() feet Ihkk in nilwt 4caceI ltarV
solution openings. some of which cause shnkhoks, we dtciloped
in the Floridan chiefly where this confining unit is thin or absent.
SCALE 1 5.000.00O
S LO 100 MILES
0 10 100 %jLWOlE1U4S
Figure 4.4. Confining units of the Floridan aquifer system. (after U.S. Geological
Survey "Ground Water Atlas of the United States, modified from Miller 1990)
,,
I~i;" ~''iMn(K1a(rPICU~t t98a
I)-
Physiographic Provinces
Fourteen Physiographic Provinces are joined with the sinkhole density theme in
ArcView (Figure 2.2). Table 4.2 indicates that the physiographic province with the
highest sinkhole density is the Coastal Swamps province (approximately 0.5/km2), an
area underlain by the Tertiary Suwannee Limestone. It is a bedded pure to slightly sandy
limestone (Randazzo 1997). The sinkhole density and relative frequency data from Table
4.2 are shown in Figures 4.5 and 4.6 respectively. The table indicates that the highest
sinkhole density is in the Coastal Swamps Province, 0.49 sinkholes per km2, and the next
is the Gulf Coastal Lowlands at 0.33 sinkholes per km2. Following the Coastal Swamps
Province and the Gulf Coastal Lowlands provinces, the next highest occurrence of
sinkholes (sinkholes per km2) is within the Lakeland Ridge and Winter Haven Ridge,
each with a density of approximately 0.20/km2. Although the greatest density of
sinkholes is in the Coastal Swamps province, the greatest relative frequency for sinkholes
(Table 4.2) is within the Gulf Coastal Lowlands region (66% vs only 7.6% for the
Coastal Swamps).
Within the database there are 1,106 documented occurrences of sinkholes in the
Gulf Coastal Lowlands and 234 in the Polk Uplands. Since there is a great difference in
total area for the provinces, the total relative frequency of sinkholes (% sinkholes per
kilometer) versus the total square kilometers in each physiographic province was
compared (Figure 4.7). Although the Gulf Coastal Lowlands and the Polk Uplands are
almost identical in size, the relative frequency of sinkholes for the Gulf Coastal Lowlands
is almost 5 times greater than it is for the Polk Uplands. The Polk Upland covers 26% of
the study area and has 14% of the total number of sinkholes. The Gulf Coastal Lowlands
covers 28% of the total study area but contains 66% of the total number of sinkholes.
Also, the Gulf Coastal Lowlands has a much greater density of sinkholes (Figure 4.5)
than the Polk Uplands (0.33 /km2 and 0.08/ km2 respectively).
For further analyses other parameters of the physiographic provinces were
compared. An overlay of maps depicting the physiographic provinces (Fig. 2.2) and the
subsurface geology (Figure 4.8) indicates that the Gulf Coastal Lowlands are mostly
covered by Quaternary age beach ridge dunes and undifferentiated sediments. And the
Bone Valley Member of the Peace River Formation (Hawthorn Group) underlies most of
the Polk Uplands (Campbell 1984; White 1970). The Bone Valley sediments are sand
and clayey fine sand, with montmorillonite and some concentration of phosphorite grains.
Within the Polk Uplands lie three north to south trending ridges, the Lakeland
Ridge, Lake Wales Ridge, and the Winter Haven Ridge. It was described by White
(1970) and later by Lane (1986) that the Lakeland Ridge and the Brooksville Ridge
(which is also within the study area) have star-shaped sinkholes on their edges. This
current GIS analysis also indicates that the Lakeland Ridge and the Winter Haven Ridge
have high sinkhole concentrations along their edges (Figure 2.2). But when sinkhole
relative frequency was computed, it indicated that they have only a relative frequency of
less than 3% each (Table 4.2). Figure 4.8 indicates that all of these ridges are
characterized by the Cypress Head Formation, surrounded by Bone Valley Member of
the Peace River Fm (Hawthorn Group) and reworked Cypress Head Undifferentiated
Sediments.
Table 4.2. Sinkhole parameters of the Physiographic Provinces
Area # of Sinkholes/ Relative Percent of
Province km2 Sinkholes km2 Frequency Total Area
Bombing Range Ridge 111 0 0.00 0.0% 0.9%
Brooksville Ridge 1,111 49 0.04 2.9% 9.3%
Coastal Swamps 256 126 0.49 7.6% 2.1%
Gulf Coastal Lowlands 3,351 1106 0.33 66.5% 28.0%
Lake Upland 541 2 0.00 0.1% 4.5%
Lake Wales Ridge 791 15 0.02 0.9% 6.6%
Lakeland Ridge 232 41 0.18 2.5% 1.9%
Osceola Plain 798 0 0.00 0.0% 6.7%
Polk Upland 3,119 234 0.08 14.1% 26.1%
Tsala Apopka Plain 156 1 0.01 0.1% 1.3%
Western Valley 362 5 0.01 0.3% 3.0%
Winter Haven Ridge 260 43 0.17 2.6% 2.8%
Zephyrhills Gap 680 42 0.06 2.5% 5.7%
Other 187 0 0.00 0.0% 9.2%
I
Figure 4.5. Sinkhole density in each of the Physiographic Provinces of the study area
S 0.30-
S 0.25-
0.20- 0.18
0.15-
0.10-
0.04
0.05 0.02
0.00 .0I00.. 0.000
Physc 0.00iog hic vin
Physiographic Province
0.00
~t~"~k~* g94
F
70% 66.5%
60%-
50%-
30% -
20% -
14.1%
10% 7.6%
00%/ 0 25% 2.6%
0.0 0.1 3% 0.0%
0%
Physiographic Province
Figure 4.6. Relative frequency of sinkholes in each physiographic provinces of the study
area
38
70% 66.5%
%- E0 Relative Frequency
60%-
E % of Total Area
50%-
O 40%-
28.0%
30%- 2 26.1%
20%
14.1%
9.3% 92%
10% 7.6% 6.6% 6.7% 2.6% 5.7%
2.9%
.24.5% 2.5/o 3% 3.0% 2.5%
0 0%9 0.0
Physiographic Province
Figure 4.7. Physiographic provinces versus sinkhole relative frequency and percentage
of total area
To
Ts
.**f .-
I.
.
* '
.1* *.".
Sibsltiiface Geoltlg\
Foi zin alltn
.. i I*.ene :e i
-. .mjEcr~rr~d,tf ~XirJ
T T -* !. '-: i ,.i -',e. seds
STQu-Undiff. seds
STQuc-Reworked Cypress Hd
C Tc-Cypress Head
S Th-Hawthorn Gp
C- That-Hawthor Gp/Arcadia Fm/Tampa Mem
C3 Thp-Hawthorn Gp/Peace River
7 Thpb-Hawthom GplPeace River/Bone Valley Mem
To-OcalaLS
li Ts-Suwanee LS
".0~ ~
4 .4 :.**E.~
4.
S.j
I-
* V
TQNu I~. *- i
LJ
&-
3 20
To -r
ZU.
'Ici n
Tm~
pbo.
:i~~ 4J ...;
I- .)
IT
S iN
9~~~j "'i 5N .f
= 2
S- Kilometers
40 80
Figure 4.8. Distribution of sinkholes and the Subsurface geology of the study area, with overlay of rivers (after Florida Department of Environmental Protection 2000)
AN
Recharge/Discharge in the Floridan Aquifer System
Annual recharge and discharge data are collected by the water management
districts, and represent the amount of water that enters and leaves the FAS annually. The
digital data for this theme were downloaded from the SWFWMD's website. The annual
data is reported in inches and has been categorized into generally accepted discharge and
recharge zones (Table 4.3). The statewide theme was clipped to represent the counties of
the study area, and then a spatial join with the sinkhole theme made it usable for
calculating how many sinkholes fall within each zone (Figure 4.10). Overlays with other
themes allowed for an overall examination of parameters that show correlations to the
amount of water entering or leaving the aquifer.
Table 4.3. Area of the discharge and recharge zones of the Floridan aquifer system in
relation to various sinkhole parameters
Recharge/ Relative
Discharge No. of AREA Sinkholes Relative Frequency
inches/yr Sinkholes km2 /km2 Frequency total
Discharge < 1 24 506 0.05 1.4%
Discharge 1 5 785 2,932 0.27 47.2% 49.5%
Discharge > 5 15 155 0.10 0.9%
Recharge < 1 84 1,817 0.05 5.1%
Recharge 1 10 501 4,804 0.10 30.1% 50.5%
Recharge > 10 255 1,773 0.14 15.3%
The study area has 1,664 sinkhole sites with 824 (0.23 sinkholes/km2) in the
discharge zones and 840 (0.10 sinkholes/km2) in the recharge zones. In Table 4.3 the
zone with the highest sinkhole density (0.27) and the highest relative frequency of
sinkholes is the discharge zone 1 to 5 inches/year, 47.2%. This is shown in histogram
format in Figures 4.10 (a) and (b). The second highest density of sinkholes falls where
recharge is >10, and the second greatest relative frequency of sinkholes is 30.1%, where
recharge is 1 to 10 inches/year (Table 4.3).
The FSRI noted in its report of sinkhole occurrence that there is a high correlation
between the density of sinkhole occurrence and the recharge rate in some Florida areas
(Frank and Beck 1990), but the new sinkhole data do not reflect this relationship. Since
dissolution of limestone from acidic water is a precursor to sinkhole formation, and
considering Frank and Beck (1990), it might be expected that the areas with the highest
amount of recharge, i.e., greater than 10 inches annually, would have the greatest number
of sinkholes. However the new data show that zones where the recharge and discharge is
less than 1, the sinkhole occurrence is least.
Findings by Dodek (2003) concur with Frank and Beck (1990) that there is a
relationship between high sinkhole activity and the recharge zones. Dodek (2003) found
that within his study area, the greatest relative frequency of sinkholes occurred where the
recharge is greater than 10 inches per year, and the highest sinkhole density is within the
zone of recharge 1 to 10 in/yr. Dodek (2003) concluded that the recharge zones are more
acidic than discharge areas, and hence, more dissolution should occur. When evaluating
these findings, location of the discharge areas must be considered. Within the study area,
the discharge zones with high sinkhole activity lie on the west coast, which may be a
contributing factor to their formation. According to Upchurch and Randazzo (1997), in
Florida, dissolution of limestone develops not only in recharge areas, but in the
saltwater/freshwater mixing zones, i.e. coastal areas.
When the Depth to the FAS theme (Figure 4.1) and the theme for
Recharge/Discharge (Figure 4.9) were overlain onto one another (Figure 4.11), it
indicated that the greatest sinkhole distribution occurred where (1) the FAS is 0 to 30 feet
and the discharge is 1 to 5 inches annually. (2) the depth to the FAS is 0 to 30 feet and the
recharge is > 10 inches per year, and (3) the thickness to the FAS ranged from 0 to 50
feet and the recharge to the FAS is 1 to 10 inches per year. Since there are no data
available regarding changes in discharge/recharge with time, the averages of
discharge/recharge were relied upon. The total relative frequency for the 3 discharge
zones is 49.5% and the total relative frequency for the 3 recharge zones is 50.5 %. Based
on this limitation, these time-averaged values may indicate that the natural recharge or
discharge to the aquifer systems is not a dominating factor in sinkhole development and
that the thickness of the overburden may have more of an influence.
When the major rivers and roads themes were overlain with the sinkhole theme
(Figure 4.9) and proximity to sinkhole locations was examined, no specific trend was
evident between the sinkhole locations and either the roads or the rivers. The rivers are
not included in the discharge theme created by FGS, but six of the seven major rivers of
the study area are located within discharge zones. Dodek (2003) also found that
discharge zones coincided with the rivers. His data showed a directional trend with one
of the major highways (US Highway 27), but a reporting bias was considered to have
affected the data. Much of his data were reported by the Florida Department of
Transportation (FDOT), hence, the sites were near roads. Within this study area, no
correlation was recognized between the sinkhole locations and the rivers and major roads.
A
n 20 40 80
- ajor Roads
Sinkholes
/ \/ Rivers
Counties
E Hernando
B Hillsborough
[ Pasco
Pinellas
D Polk
oiueters
Figure 4.9. Counties map with sinkhole locations, major roads and major rivers (after Florida Geographic Data Library 1990)
Figure 4.9. Counties map with sinkhole locations, major roads and major rivers (after Florida Geographic Data Library 1990)
_ ~ _~
olneters
/N' Major Rivers Discharge and Recharge
I !i -. ( *I.
inr^
0 20 IK
0 20 40 80
Ar
Figure 4.10. Recharge/Discharge zones measured in inches/year and sinkhole
distribution (after Southwest Florida Water Management District 2002)
.27
0.30
0.25-
E 0.20 .14
0.15- .0
) .04
0.10 .05 .05
0.00-
Recharge / Discharge inches / yr
(a)
Figure 4.11. Recharge/Discharge zones of the study area, (a) Sinkhole density (b)
Relative frequency of sinkholes
46
47.2%
50%
45%-i
40%-
%30.1%
35%
S 30% -
25%-
S2% 15.3%
20%-
S 15%
5.1%
10% 1.4% 0.9%
5 5%
0%'
Discharge/Recharge inches / year
(b)
Figure 4.11. Continued
I "-
; .
T .J' *. -.-
S*"' \ .
.S. .
J .. I ---. I -
*1, SI 5 o
.....-* ," .
*
5- 5A ** '
-a ~
-
+-. -.i ,
Ii
,. ,. 1 .. .
.- :, "1.I~ I'' I;
W Counties
Sinkholes
Overburden to Floridan
S0-15 feet
16-30 feet
Recharge/Discharge
S
QJ DISCHARGE/1 TO 5
SRECHARGE/1 TO 10
.aJFT Kilometers
f.l 0 10 20 40
.I I \
,I \
**
,I,; -. -.
1 p .
. *..
.* ..r : ." '* I *I *
*.,~,- .I. *.
%I. J I I.
: .. i m* ". .
S .. I. .
w ., .*' \ -
/ ''^ '* N" : '."y /
Figure 4.12. Overlay of the Thickness of the Overburden theme and the Recharge/Discharge theme with the sinkhole distribution theme
Potentiometric Surface
Potentiometric surface represents the total head of groundwater and is the level to
which water will rise in a tightly cased well. Potentiometric surface changes result from
recharge and discharge of the aquifers. This theme was constructed using data supplied
by the SWFWMD. Figures 4.13(a) and (b) display potentiometric surface data for 1975,
1985, 1994, and 2000. Categories of potentiometric surface span the years that the
sinkhole data was recorded and reflect years when rainfall was considered to be average
rainfall, low rainfall, and high annual rainfall (SWFWMD 2003). The rainfall values are
in Appendix (a).
Potentiometric surface maps were overlain with the sinkhole theme and
correlations identified. Figures 4.13(a) and (b) show that the highest density of sinkholes,
for all four years, is located where the elevation of the potentiometric surface is low (5 to
10 feet and 20 to 50 feet). Sinkhole formation is accelerated during periods of drought or
during artificial lowering of the potentiometric surface by over pumping of the aquifer
(Tihansky 1999). When potentiometric surface is low, there is a loss of buoyancy and a
reduced strength of the overburden; hence it can make the area conducive to sinkhole
formation. This suggests that sinkhole formation occurs during a drop in the
potentiometric surface. A lowered potentiometric surface may be a triggering factor in
sinkhole development (Upchurch and Randazzo 1997).
1974
(a)
Figure 4.13. The potentiometric surface values with sinkhole locations (a) for 1974 and
1985 (b) for 1994 and 2000 ( after Southwest Florida Watermanagement
District 2003)
POTENTIOMETRIC SURFACE
Kiloineteas
Figure 4.13. Continued
A
2000
80
',
x
\.
Y
Geology
Themes for both the surficial and the subsurface geology were constructed. The
Polygon coverage was obtained from the FGDL (http://www.fgdl.org/). Geology
coverage for the counties in the study area were merged and then joined with the sinkhole
database, and analyses were made. The digital coverage for the subsurface and the
surficial geology showed very little variation from one another in terms of the geology.
Dodek (2003) also found this in his study area. Therefore, only the subsurface geology
was analyzed for this study.
Table 4.4 and Figure 4.14 indicate that the greatest sinkhole density occurs in the
Beach Ridge and Dune lithostratigraphic unit, which has 0.46 sinkholes per km2. The
really largest lithostratigraphic unit (Table 4.4) is the Hawthorn Group (3,654 km2),
which is also where the greatest relative frequency of sinkholes occurs (Figure 4.15).
The Hawthorn Group includes its Arcadia Formation (Miocene), Peace River Formation
(Miocene-Pliocene), and it's Bone Valley Member (Miocene-Early Pliocene). The
lithostratigraphic unit with the second highest relative frequency is the Quaternary
Undifferentiated Sediments lithostratigraphic unit (32.2%).
The Hawthorn Group's Arcadia Formation is predominately siliciclastic-bearing
carbonates, and the Peace River Formation, which overlies it, is a plastic unit (Compton
1997; Scott 1997). It contains variable amounts of carbonate with clay and quartz sand
(Compton 1997). This area contains 572 sinkholes and has a relative frequency of 34.4%
(Table 4.4), the greatest for all the geologic units.
By utilizing ArcView, an overlay of the sinkhole locations, the overburden
thickness theme (Figure 4.1), and the geology theme (Figure 4.8) were viewed. The GIS
representation reveals that the areas with the greatest density of sinkholes occur in the
Beach Ridge and Dune geologic unit and the Suwannee Limestone unit. The Beach
Ridge and Dune geologic units are mainly situated where the thickness of the overburden
to the FAS is 0 to 30 feet, and the Suwannee Limestone is occurs where the FAS is less
than 75 feet thick. Dissolution of limestone and dolomite is the precursor to sinkholes
and traveling. The Suwannee Limestone is principally a carbonate unit. The coverage
suggests a correlation between the high incidence of sinkholes, the geologic units and the
thickness of the overburden.
Table 4.4. Subsurface geology of the study area in relation to various sinkhole
parameters
Lithostratigraphic Unit
Holocene Sediments
T-Q Undiff. Sediments
Peace River Fm/Hawthorn Gp
Dunes
Plio-Pleist Shelly Sediments
Beach Ridge & Dunes
Arcadia Fm/Hawthorn Gp
Ocala Limestone
Suwannee Limestone
Cypress Head Fm
Hawthorn Group
Reworked Cypress Head
Bone Valley Mem/Peace River
Quaternary Undiff. Sediments
All Hawthorn (combined)
Area
km2
57
180
208
230
297
367
401
420
668
778
1,135
1,750
1,910
3,542
3,653
# of
Sinkholes
2
5
16
1
17
170
61
3
199
72
318
87
177
536
572
Sinkholes /
km2
0.04
0.03
0.08
0.00
0.06
0.46
0.15
0.01
0.30
0.09
0.28
0.05
0.09
0.15
0.16
Relative
Frequency
0.1%
0.3%
1.0%
0.16
1.0%
10.2%
3.8%
0.2%
12.0%
4.3%
19.1%
5.2%
10.6%
32.2%
34.4%
% Total
Area
0.5%
1.5%
1.7%
1.9%
2.5%
3.1%
3.4%
3.5%
5.6%
6.5%
9.5%
14.7%
16.0%
29.7%
30.6%
0.50- 0.46
0.45
0.40
0.35 -
0.30
2 0.30-
0.25-
i
0.20-
0.15 0.16
0.10.15
0.09
0.050. 004 003 06 0.05
S0.01
00.00 i-i R i ti 0.0
Lithostratigraphic Unit
Figure 4.14. Sinkhole density and lithostratigraphic units of study area
35%- 32.2%
30%-
25%-
10%-
S 5%12.0%
4%5.2%
5%-
Lithostratigraphic Unit
Figure 4.15. Relative frequency of sinkholes in lithostratigraphic units of study area
Major Wells in the Study Area
A theme was constructed within ArcView, showing the locations of the 30 most
highly pumped wells in each county of the study area in relation to the sinkhole
distribution (Figure 4.16). The data were supplied in Excel spreadsheet format by
SWFWMD and each well contained a longitude and latitude. These were then projected
onto the sinkhole distribution map in the same projection, Albers GCS North American
1983. In chapters 2 and 3 it was noted that increased sinkhole activity may result from
over pumping of the aquifer. An overlay of the sinkhole theme and geology theme were
used to identify correlations between the known sinkhole activity and the location of the
well. The data that are included were taken from collections in 1980, 1985, 1990, 1996,
and 2002. Some of the wells may have been active (and still active) in all of these years,
while other wells may have been capped (abandoned) and new wells put into existence.
Since 1980 there has been a significant increase in annual pumpage. The highest annual
pumpage in 1980 was in Polk County and was 1,784,834,700 gallons, and in 2002 the
highest annual pumpage occurred in Polk County, 5,928,020,000 gallons (SWFWMD
2002) (Table 4.5).
Table 4.5. Highest annual water pumpage (gallons) for top wells in each county of study
area for selected years
County 1980 1985 1990 1996 2002
Hernando 900,847,108 724,062,000 820,175,046 1,305,510,000 1,015.348.000
Hillsborough 818,505,000 2,001,658,016 1,813,487,502 861,005,600 1.241,900.000
Pasco 804,145,137 653,945,909 518,648,000 482,182,000 995,260,000
Pinellas 193,497,000 141,684,000 306,273,000 325,033,000 491,370.000
Polk 1.784,834,700 2,263,694,854 2,892,071,000 2,105,493,200 5,928,020.000
Within ArcView, it was possible to identify where sinkhole activity was high and
multiple wells existed. The four circles on Figure 4.16 show these areas. A correlation
between the large withdrawal from these wells and sinkhole distribution is generally
indicated. It appears to demonstrate that the wells may be a contributing factor, but since
some areas show a concentration of wells but not a concentration of sinkholes, other
parameters must also be considered as possible contributing factors.
When overlays of different parameters were viewed within ArcView, the
Subsurface Geology theme indicated (Figure 4.16) that the stratigraphic units in these
four areas of high sinkhole distribution are the Quaternary undifferentiated sediments,
Quaternary Beach Ridge and Dune, and the Suwannee limestone. Another important
parameter that may be a contributing factor to sinkhole formation is the thickness of the
overburden. By viewing the ArcView overlays, four areas were identified where there is
a high occurrence of sinkhole formation, a large number of highly pumped wells (Figure
4.16) and the overburden to the top of the Floridan aquifer is mostly 75 feet thick or less
(Figure 4.1).
Some biases to consider that may be having a large influence here are the
collection bias, and the limited information about the actual drawdown of each well. The
collection bias would indicate that the larger the population, the greater demand for water
and hence, more wells would be expected in the vicinity of larger populations. More
sinkholes are expected to be reported where population is greatest.
SW0 |__ HoWkeem se dime ts Cyp.IeaId Fm
S U d" ff- SdiM ri.s Hnarth Group
.T -a D.a.s Ha tH-GApl tr dbaFm
S rR Phao-RPeitslllyseds HarthomGpe River
.' .. -- \_ C.] Bm ay
SGgyd ffsedirw t f Hwatn eGpPee River
.41
Beach *dgld um *- RedxkidCygpashead : OcaIaLimestei
0 .
.* re .
oTo
i' .
I*
e *. \
f_ .s
***
J 4..
"TQu
J '' '*
-. .:Q
AI
( 0 20 40 80
Figure 4.16. Sinkhole distribution on the geology theme, with the 30 most highly pumped wells in each county for the study area. The circles indicate areas of high sinkhole
density and a large number of highly pumped wells (after Southwest Florida Water Management District 2003)
:,, ; --'# fll "" "
., -41,
2' 4"- .i' ",.
Figure~ 4.1.Snhl d :. o nth elg hm, ihte3 os ihypme wlsi ahcunyfrt 'ud ra h iclsidct raso ihs] l
dn it n ag ub o !ihl pupe well" '"te So:: .:s Flo:id, Wae Mangen ,.tc .
~
Population Density
Population density data, when applied to sinkhole distribution pattern, suffers from
biases that limit meaningful correlations. The 2000 US Census data is used in the
Population Density theme. The data were obtained from the FGDL. The US Census
Bureau delineates geographic areas into Census tracks which are made up of Census
block groups. The Census block groups are defined by the US Census Bureau as having
approximately 600 to 3,000 people, with an optimum size of 1,500 people in each Census
block group. The area of each track is in square meters. This area was then converted to
square kilometers and then broken into its associated Blocks. The population per square
kilometer was then calculated (Table 4.6). It is hypothesized that the areas with the
highest population density (people per square kilometer, Figure 4.17) would have the
greatest sinkhole occurrence. This is to be expected for many reasons. For example,
anthropogenesis is likely to be a strong contributing factor, and as mentioned before, can
be causing a reporting bias. For example many of the cases included in this thesis are
reported by homeowners that have structural damage to their home, and if the damage is
verified as being due to sinkhole activity, it might lead to other neighbors having similar
damage investigated. Hence, confirmation of sinkhole activity for several homes in one
neighborhood can increase the reported sinkhole density.
The US Census Bureau's block data (Figure 4.17) of population density was used
to compute sinkhole density and sinkhole relative frequency for specified ranges of
population (Table 4.6). Figure 4.18 shows a general trend of increased sinkhole activity
with an increase in population. Sinkhole relative frequency for the area does not indicate
as strong a correlation, but this may be explained by the U.S. Census Bureau's block
sizes. The blocks decrease in size as the population increases in size. This would
decrease the relative frequency for those blocks. This analysis agrees with Dodek (2003)
who also recognized a general trend between higher sinkhole densities and an increase in
population density. Where population is high, many anthropogenic factors may be
influencing sinkhole activity. For example, an increase of population means an increase
in water usage, which increases withdrawal from the aquifer. Figure 1.1 demonstrates the
affect of large water withdrawals. Population may also affect water drainage patterns
which may play a role in sinkhole distribution when certain other factors are present.
Table 4.6 indicates that the population range with the highest sinkhole density, 0.95
sinkholes per sq km, is 1,501 to 2,000 people/km2 range. The next three population
groups in descending order of sinkhole density are the 5,001 to 6,000 range, 3,001 to
4,000 range, and 2,001 to 3,000 range. Both the sinkhole density and the relative
frequency of sinkholes are within the 1,501 to 2,000 range (15.0%).
Analyses also show that the most populous areas, 3,000 to 6,000 people/sq km have
a much lower relative frequency of sinkholes than expected. This is the result of the
biases previously mentioned. The area where population is 3,000 to 6,000 people/sq km
represents only 33 km2 of the 11,987 km2. This is only 0.3% of the study area and might
be considered anomalous. The 0 to 50 and 101 to 200 ranges represent 67.5% of the total
study area with less than 0.02 to 0.14 sinkholes density, but have an average relative
frequency of 10.5% (Figures 4.18 and 4.19).
Overall, population density data are inadequate in addressing sinkhole distribution
issues, but they do provide interesting patterns for future analyses.
N _Population Density
II 0-50 pop km2
j50-150
S j.3. 50-500
S; .' 500-1000
'----'' oo1000-2000
S '- I 2000-3000
3000-6000
S.. .. 6000-27000
__- -,.^ -*I ','** :_ --. '\ --
-f \ -
-4,
S, -.-- 4 .4"
I j -. i 4. /
I"-L' -- ...... I --- + ..- t
II-- ., Kilometers
0 20 40 80
Figure 4.17. Distribution of sinkholes and the US Census Bureau's Block Groups showing population density for 2000 ( from US Census Bureau 2000)
.. ',p-:1" I i ,-:-, .- _: -
.. ... -4 'j \
Table 4.6. Population density in relation to various sinkhole parameters
Range Number of Area Sinkholes / Relative
(pop/km2) Sinkholes km2 km2 Frequency
0-50 164 6,884 0.02 9.9%
51-100 87 1,090 0.08 5.2%
101-200 172 1,207 0.14 10.3%
201-300 96 588 0.16 5.8%
301-450 134 564 0.24 8.5%
451-600 98 362 0.27 5.9%
600-750 77 189 0.41 4.6%
751-1000 187 262 0.71 11.2%
1001-1500 233 380 0.61 14.0%
1501-2000 249 261 0.95 15.0%
2001-3000 140 167 0.84 8.4%
3001-4000 25 30 0.84 1.5%
4001-5000 1 3 0.33 0.1%
5001-6000 1 1 0.88 0.1%
Population Density /km2
Figure 4.18. Population density (km2) versus sinkhole density using US Census Bureau
block groups 2000
63
16%
15.0%
14.0%
14%- -
12%- 11.2%
>, 10.3%
,) 9.9%
C 10%-
S8.4%
8.1%
LL 8%-
.5.8% ,5.9%
m 6% 55.2% /
4%
... .. 1.5%.
2% 1.5%
i 0.1% 0.1%
Population Density I km2
Figure 4.19. Sinkhole relative frequency for population density 2000 (km2)
CHAPTER 5
ANALYSES
Utilization of a Geographic Information Systems analysis has made it possible to
test several hypotheses at once, recognizing contributing factors to sinkhole distribution.
In some cases, parameters could be evaluated without GIS themes, but in order to make a
stronger evaluation, a combination of factors hypothesized to be triggering factors to
sinkhole formation were overlaid by using themes a method for validating the results.
It is also a means to detect existing spatial biases. The complexities of sinkhole
distribution are demonstrated in Figures 5.1(a) and (b) and 5.2(a) and (b), which indicates
that sinkhole formation, is influenced by a combination of factors. The figures are a
complex representation of the parameters that exist in areas where the sinkhole relative
frequency and sinkhole density are greatest. Throughout this analysis, all parameters
were overlain on each other and each layer was "turned off and on" to examine for
correlations. All of the parameters addressed in Figures 5.1(a) and (b) and 5.2(a) and (b)
have been separately clipped from figures throughout this thesis to show the areas with
the greatest sinkhole activity and sinkhole relative frequency occur. The data are
summarized in Table 5.1.
Figure 5.1(a) represents the greatest sinkhole relative frequency for all the
parameters. It indicates that the relative frequency of sinkholes is greatest in the areas
where the overburden to the Floridan aquifer system is either missing or less than 30 feet
thick. The overburden thickness appears to play a large role in sinkhole formation, as
indicated by these results. This figure also shows that the physiographic province
containing the most sinkholes (greatest relative frequency) is the Gulf Coastal Lowlands.
The Gulf Coastal Lowlands has an overburden thickness that varies between 0 to 50 feet.
However, Table 5.1 displays that the highest density of sinkholes is in the Coastal
Swamps, which also has a very thin overburden to the FAS of only 0 to 15 feet (Figure
4.1 and Figure 2.2). Also, the differing results in the relative frequency and the density of
sinkholes is a reflection of the sizes of the provinces. The Gulf Coastal Lowland
represents 28% of the total study area, and the Coastal Swamps only 2%. This size bias
must be considered when making analyses. Also, the Polk Uplands represents 26% of the
total area and has the next highest relative frequency of sinkholes of only 14%, and a
sinkhole density of only 0 .08 per km2.
Figures 5.3(a) and (b) illustrate that both the relative frequency and the density of
sinkholes are greatest in the Discharge zone of 1 to 5 inches annually within the study
area. This does not correlate with the general idea that Recharge zones are the areas that
are expected to have the greatest number of sinkholes (Dodek 2003). ArcView allowed
the Recharge/Discharge theme to be overlaid by the Overburden theme (Figure 5.4) and it
shows that most of the discharge areas occur where the overburden is only 0 to 30 feet
thick, and that the discharge zones with the greatest sinkhole density are located along the
coast line (Figure 5.3) where saltwater/freshwater coastal mixing occurs (Upchurch and
Randazzo 1997). It is in these coastal areas that dissolution of limestone, and hence karst
formation, is preferentially developed (Upchurch and Randazzo 1997). The overlap of
figures also demonstrates that in many areas where the recharge zones occur, the
overburden is 100 to 300 feet thick. This suggests that the lower number of sinkholes
occurring in the recharge areas is due to the increased thickness of the overburden, and
that the high occurrence of sinkholes in the discharge areas could be related to the thin
overburden to the FAS.
Rock type has always been considered to be an important factor in sinkhole
formation. An example is the chemical weathering of carbonates (Lane 1986, p12; Beck
and Sinclair 1986; Galloway et al. 1999). As noted in Table 5.1, high sinkhole activity is
occurring where the Suwannee Limestone is at or near the surface and also a high relative
frequency of sinkholes occurs in the Hawthorn Group sediments, which is also carbonate
bearing (Compton 1997; Scott 1997). Although the Beach Ridge and Dunes and the
Suwannee Limestone units actually have the greatest sinkhole densities (Figure 4.14), the
Hawthorn Group and the Undifferentiated Quaternary Sediments units have the greater
sinkhole relative frequencies (Figure 4.15). This might be reflective of a size bias. The
Beach Ridge and Dunes and Suwannee Limestone are significantly smaller really than
both the Hawthorn Group and the Quaternary Undifferentiated Sediments (Table 4.4).
Relative frequency is a statistical operation that addresses the size bias of a sampling
population and it represents a valuable statistic accompanying sinkhole density. The
Hawthorn Group has a sinkhole relative frequency of 34.4%, but the Quaternary
Undifferentiated Sediments has a sinkhole relative frequency almost as great at 32.2%.
The Hawthorn Group contains a relatively higher percentage of carbonates (Scott 1997)
than the Quaternary Undifferentiated Sediments unit.
Other factors may be contributing to sinkhole distribution. The age (Oligocene,
Brooks 1981) and carbonate purity of the Suwannee Limestone, promote its dissolution.
Most of the Suwannee Limestone in the study area lies under an overburden of 0 to 15
feet along the coastline (Figure 4.1). The Beach ridge and Dune unit is composed of
sand, which is highly permeable, and also has an overburden of less than 30 feet thick
(Figures 4.1, 4.13).
As previously noted, population density data are extremely biased by many factors
and is temporal, but the results indicate (Table 4.4) that the greatest distribution of
sinkholes is in the zone where population density is 1,000 to 2,000. It was expected to
occur in the most highly populated areas, 3,000 to 6,000. It is noted that the US Census
Blocks are designed by population rather than area. This means that some blocks have a
much higher population density than others. This skews sinkholes per population
numbers, and when relative frequency is calculated the bias is exacerbated.
Table 5.1. Results from GIS analyses of sinkhole distribution and significant themes
Theme Highest Sinkhole Sinkholes/m2 Greatest Relative Relative
Density Frequency Frequency
Ov n 51-75 feet 0.38 0-15 feet 21.6%
Overburden
16-30 feet 0.33 31-50 19.1%
Physiographic Gulf Coastal
Physiographic Coastal Swamps 0.49 Lowlands 66.5%
Province
Gulf Coastal Lowlands 0.33 Polk Uplands 14.1%
Recharge/ Discharge 1-5 inches/yr 0.27 Discharge 1-5 47.9%
Discharge Recharge >10 0.14 Recharge 1-10 30.1%
Beach Ridge/Dune 0.46 All Hawthorn Seds 34.4%
Geology Suwannee LS 0.30 Quat. Undiff Seds 32.2%
Population 1501-2000 people/km2 0.95 1501-2000 15.0%
Density 5001-6000 0.88 1001-1500 14.0%
Sinkholes
E- Counties
Pop sqkm
1001 -100
50I1-00
Ph'. ooglz phIc Provinces Geology Formation
F7 Cww.LOvrr a d
0-IS feI
16-3feet
DischargeRechaige
Reckmgel. 10
&, ~ [I Other
- fl a'~t0'p M'. Vsky
* S
SI Kilometers
N
A
\
(a)
Figure 5.1. GIS overlay of all the areas of highest sinkhole relative frequency for all digital parameters used in the analyses, (a) with population density (b) without population density
I
S Smikholes
I Counties
Discharge/Rechaige
| Discharge/1 5 inches/yr
Recharge/ l- 10
SOther
Physiographic Provinces
W GulfCoastal Lowlands
Polk Uplands
OCverburden
0-15 feet
16-30 feet
Geology Fonnation
3 Undiff Seds
X Hawthorn Gp
SHawthorn/Arcadia Fm
SHawthom/Peace River
SHawthom/P River/Bone Valley
IU
SKilometers
0 20 40 80
Figure 5.1. Continued
'**
Geology Formation
] Beach Ridge/Dune
Hawthorn Gp
77 Hawthorn/Arcadia Fm
Hawthorn/Peace River
Hawthom/P River/Bone Valley
SSuwanee LS
( '
I]Kilometers
0 20 40 80
(a)
Figure 5.2. All digital parameters with the greatest sinkhole density for the study area (a) with the Census' blocks of population density, (b) without the Census' blocks of population density
z
71
S Sinkholes Physiographic Provinces Geology Formation
S; .. -- Counties Coastal Swamps Beach Ridge/Dune
Overburden Gulf Coastal Lowlands 3 Hawthorn Gp
S15-30 ft Dischare Recharge Hawthom/ArcadiaFm
i' --50-75 ft Discharge/1 to 5 Hawthom/Peace River
.L Recharge/Greater than 10 Hawthom/P River/Bone Valley
Other Suwanee LS
-" ..
F u 5" Co.t .-n.ed
13
"" *, I "
F r
J %*
Figure 5.2. Continued
0 20 40 80
Figure 5.2. Continued
Sinkhole Relative Frequency
Sinkholes
SCounties
Di shcarge/Recharge
mches/yr
SDischarge/l to 5
Recharge/l tolO
Kilometers
S-- 10 20 40
". ,
* i .! .rL.
i. Ti**
II =
..-r:.'
Figure 5.3. Recharge and Dishcarge zones with sinkhole distribution. Relative
frequency and density of sinkholes
Sinkholes
7 Counties
RechargeDischarge
mches/yr
Discharge/l to 5
Recharge/Greater Than 10
Kilometers
0 10 20 40
A
aI .
aaka
S** *t* .
aI a
a *f t a, a
SI" .r *
a. --
,a, ,.- ., .- -
-i- ,,. .i
N4 0
S"''
H <'..7 -/ -- __- ; ,-:
i I L M O
"- a "
.. .: ", ..... aa
\ 1-. i ....
r, \ l C. I "..
e .' i ?
Figure 5.4. Recharge/Discharge theme overlaid with the Depth to the Floridan Aquifer System theme
I IIII en I. Fi I
,1. lll ll i 1'.' h l i 0u
,, 1 I nr .
Rediall-u e Di-,. i-t "e cul \
m t'. :.:har..- 1 I .
F i .l hrr" I I I"'
i.
a
I __________
I a I
a I
i I a a I
a a
Sm 1 -. .
*I
I .
I. I,
a I lii
aKilomelers
0 10 20 4C
-
".'. ,. .
i If .
"I i 1 ..
";--,--
II
\1 *
\
,.
1.
Ma
-'
'i:
Figure 4.16 suggests that the location of the 30 highest pumped wells in each
county correlate with the occurrence of sinkholes in at least four areas. This may
illustrate that well fields are one more influencing factor in sinkhole formation and
distribution, while a conclusion can not be made on this analysis it is useful when
examining other local factors.
Figure 5.5 shows the result of overlaying all the GIS themes and clipping out the
areas that intersect all of them. These (small) areas are outlined in red. The result
demonstrates that the formation of sinkholes does not require all of the theme conditions
to be present, but instead a combination of factors controls sinkhole distribution. From
the overlays in this analysis, tabulated in Table 5.1, it is indicated that the strongest
influencing factors for sinkhole development are discharge areas of 1 to 5 inches annually
(occurring 47% of the time), areas where the aquifer is shallow (50 feet or less), and areas
where the geologic units consist of permeable and easily dissolved materials. These
parameters are found most often in the Gulf Coastal Lowland physiographic province,
where the aquifer is close to the surface or exposed, the discharge is 1 to 5 inches
annually, and the near-surface geologic units are the Suwannee Limestone and the
Quaternary Undifferentiated Sediments.
This database can be combined with other studies to increase the accuracy of
predicting future sinkhole development. For example, an enlargement of the database
would help to substantiate findings. This can be done by adding newly collected data and
non-reported data, which for example, could be obtained by maps developed through
remote sensing. Further studies could include additional factors for parameters in this
study, which showed weak correlations to sinkhole distribution, i.e. the theme showing
Well locations in each county. Additional information could include the extent of the
well field's draw-down.
These results suggest that the thickness of the overburden to the FAS plays the
strongest role in sinkhole formation, and when combined with other sinkhole triggering
factors, sinkhole density and sinkhole relative frequency may increase. For example,
since sinkhole relative frequency is greatest where the overburden is less than 75 feet
thick and the Hawthorn Group is present (Figure 4.13 and Table 5.1) sinkhole activity
should increase. The physiographic provinces that had the greatest number of sinkholes
(Table 5.1) also coincided with areas of thin overburden and geologic units that are prone
to dissolution, and hence sinkhole formation.
A similar study by Dodek (2003) utilized the current FSRI database to increase his
database of sinkholes for 11 north-central Florida counties. His projections of future
sinkhole occurrence for several counties were based on results he determined were
influencing sinkhole development. These parameters were areas of high recharge to the
aquifer and a minimal depth to the aquifer system. He found that sinkhole locations
coincided with the presence of near-surface Ocala Limestone, a carbonate unit.
The results in this study found that there was a high density of sinkholes and a high
relative frequency of sinkholes coinciding with areas where the geologic units consisting
of near surface carbonates occurred (Table 4.4). This study did not agree with Dodek's
findings that sinkhole distribution is greatest in areas of high recharge. Factors, such as
the depth the FAS in the discharge zones, have an influence on sinkhole distribution.
..* **
3 Hilisborougyh
.. .
6 7.
., Polk
,*, '
.. *
*. *. __ ** ,. .
**i Hillsborou*h *
*.. ..
.. ** ***
HPinellao ,- .
Figure 5.5. GIS intersection of all the parameters that displayed the highest sinkhole relative frequency. The red areas represent the polygons that contain all the parameters
CHAPTER 6
CONCLUSIONS
A new sinkhole inventory has been created and includes the previously existing
FGS database. From this database, factors that were hypothesized to influence sinkhole
formation have been analyzed using a GIS, Esri's ArcView 8.3 (ArcView). After
calculating the sinkhole density and sinkhole relative frequency for each factor, they were
then coalesced into maps that allowed trends and patterns to be recognized. Results
indicate a correlation between sinkhole distribution and other factors, and are controlled
by a combination of parameters. New data has been obtained that enlarges the existing
database, and as this database grows, it strengthens results that are produced by any
future analyses.
Table 5.1 represents GIS results and indicate that when all parameters are
considered, sinkhole relative frequency is greatest in (1) the Gulf Coastal Lowlands (2)
overburden is 0 to 15 feet thick (3) Hawthorn sediments are present in the near-surface,
and (4) the aquifer discharge is 1 to 5 inches per year. Population density of 1,501 to
2,000 people per km2 may also be a factor. All of these parameters were extracted from
their ArcView themes and presented in Figures 5.1 (a) and (b).
When sinkhole density was computed, the results, summarized in Table 5.1, show
that the parameters where sinkhole density is greatest are where (1) the Depth to the
Floridan is 50 to 75 feet, (2) the Coastal Swamps province, (3) the discharge is 1 to 5
inches per year, and (4) the Beach Ridge and Dune subsurface geology is present.
Population density of 1500 to 2000 people per square kilometer may also be significant.
Other factors that may have an influence on sinkhole density include areas of low
potentiometric surface and water withdrawal from the aquifer system, i.e., well fields.
Figure 4.16 shows four areas where the wells with the highest withdrawal rates are
located in the vicinity where sinkhole occurrence is greatest. Withdrawal of large
quantities of water from the aquifer, contributes to the lowering of the potentiometric
surface, and may also result in sinkhole activity (Figure 1.1). Correlations of sinkhole
location and low potentiometric surface, suggests that lowering of the potentiometric
surface is another factor contributing to the formation of sinkholes. In the four years
depicted in Figures 4.11 and 4.12, the areas of highest potentiometric surface are those
where fewer sinkholes have formed (or have been reported).
The Gulf Coastal Lowlands has a sinkhole relative frequency value almost 5 times
larger than any other physiographic province, and although the Coastal Swamps has the
highest density of sinkholes, 0.49 per square kilometer (Table 5.1), the Gulf Coastal
Lowlands is close behind at 0.33. The Gulf Coastal Lowlands is a region where
Quaternary undifferentiated sediments are the most prominent geologic formation, and it
has an overburden to the FAS of 30 to 50 feet, well within the major mode for the
greatest sinkhole density and the highest relative frequency of sinkhole occurrence.
Population density is generally 1,000 to 2,000 people per square kilometer.
Biases are an important aspect in recognizing the significance of the various
parameters and sinkhole frequency. When an overlay of the major rivers and roads in the
area was made (Figure 4.10), no significant pattern or correlation was noted, but this
might be one of the biases that were discussed in prior chapters (i.e., the "collection
bias"). Of the seven major rivers, five of these are in areas that have a very low
population density, 0 to 50 people (Figure 4.17), or where population density is of 50 to
150 people per square kilometer. Within these ranges (people per km2), sinkhole density
is very low. It is only towards the mouth of the Hillsborough River that the sinkhole
density increases, and the population density in these areas also increases to 1,000 to
6,000 people per square kilometer (Figure 4.17). One could assume that the low number
of sinkholes is due in part to the reporting bias. In other words, where population is low,
reporting will be minimal, but as noted previously, population density data contains many
biases.
Future studies could increase the database by including newly reported sinkholes,
and non-reported sinkholes. Non-reported sinkholes could be obtained through digital
topographic maps, obtained by the use of remote sensing. With the increase in the
database, trends and patterns of occurrence will be more easily identified and hence, as
the database "grows", making predictions of future sinkhole distribution should be more
accurate.
In order to utilize the sinkhole database for predicting the high risk areas for
sinkhole formation, a theme that representing the distance-to-the-nearest-sinkhole was
constructed, using buffer zones of one, two, and three kilometers. In light of the large
study area, these buffers represent a reasonable category-distance resolution. The theme
is presented in Figure 6.1. It is hypothesized that the higher potential for sinkholes is
where sinkhole density is greatest and where the distance between sinkholes is least. In
Figure 6.1, that would be the 0 to 1 km zone, indicated by the blue buffer zone. Figure
6.2 also represents "nearest neighbor" but with buffers of 0.5, 1.0, and 1.5 kms. The
figures demonstrate that the highest risk areas would likely be in the north-western
section of Hillsborough County, the western section of both Pinellas and Pasco counties,
and central Polk County. The areas within Hillsborough, Pinellas, and Pasco are areas
where the FAS is close to the surface, 0 to 75 feet (Figure 4.1), and in Polk County, the
area where sinkhole density is greatest, is where the overburden thickness is 75 feet to as
great as 300 feet (Figure 4.1).
As population density grows, there will be an increasing demand for ground
water, and hence greater withdrawal from the FAS and lowering of the potentiometric
surface. This trend will impact areas where the significant factors for sinkhole formation
are concentrated, and future sinkhole density and sinkhole relative frequency will be
greatest.
Nearest Sinkhole
1:;. 0-1 km
SI 1-2km
1 C2-3 km
-- Counties
POLK
0 20 40
8'
Kilometers
0
Figure 6.1. Proximity to the nearest sinkhole, indicated by one, two, and three kilometer buffer zones
A
_~ ___ ~ __ ~ ~ ~ ~
0 7777
Nearest sinkhole
I "I 111111e
Neamest Sinkhole
0-5km
5-1km
1 1.5km
POLK
0v
77 -Kilometers
0 20 40
AN
I
Av%~
80
APPENDIX A
GLOSSARY OF ARCMAP TECHNICAL TERMS
Esri's ArcMap 8.3 was used for processing data within this thesis. Some of the
tasks performed are known as Geoprocessing. The following are descriptions of these
procedures:
* Clip: a process where a specified layer, known as the cookie cutter, is used to cut
another layer, know as the input layer. The result is a new layer that has only the
polygons and attributes inside the clip layer, while those outside the clip layer do
not remain.
* Dissolve: the process of combining polygons with identical attributes. The
attribute is specified by the user.
* Geocoding: the process of matching street addresses with geographic coordinates
* Geographic coordinates: a measurement by latitude and longitude on the earth's
surface.
* Join: the process of attaching tabular data to a layer. The fields in the table are
appended to the layer using a common field. Join establishes a one-to-one, one-to
many or many-to-many relationship between map features and table attributes.
* Layer: Geographic information is displayed on a map as layers; each layer
represents a particular type of feature such as streams, lakes, or highways. Layers
are listed in the ArcView table of contents and can be further organized into data
frames. A layer references geographic data stored in a data source, such as
coverage, and defines how to display it.
* Layout: the design or arrangement of elements such as geographic data, elements
like north arrows, legends, scale bars, and text, in a digital map display or printed
map.
* Merge: combining polygons from two or more themes. Attributes with the same
name are retained.
* Projection: a mathematical formula that transforms feature locations from the
earth's curved surface to a map's flat surface. A projected coordinate system
employs a projection to transform locations expressed as latitude and longitude
84
values to x, y coordinates. Projections cause distortions in one or more of these
spatial properties: distance: area, shape, and direction.
S Theme: a category within a layer.
APPENDIX B
HISTORICAL RAINFALL FOR STUDY AREA
Table B.1. Rainfall
Year Hernando Pasco Pinellas Hillsborough Polk
1974 66.83 65.98 67.21 53.11 52.31
1985 48.83 47.96 42.43 44.65 43.29
1994 54.81 51.75 43.03 55.20 56.70
2000 44.55 43.81 39.78 44.31 38.96
Min 44.55 43.81 39.78 44.31 38.96
Max 66.83 65.98 67.21 55.20 56.70
Avg 53.76 52.38 48.11 49.32 47.82
REFERENCES CITED
Arthur, J. D., R. A. Lee, and L. Li. 2001. Lithostratigraphic and Hydrostratigraphic cross
sections through Levy-Marion to Pasco counties, Southwest Florida. Florida
Geological Survey Open File Report no.81. 21 p.
Beck, B., and W. C. Sinclair 1986. Sinkholes in Florida: An Introduction. The Florida
Research Institute Report no.85-86-4. 17 p.
Brooks, H.K. 1981. Geologic Map of Florida, Institute for Food and Agricultural
Sciences, University of Florida, Gainesville, Florida.
Burwell, A.L. and W.L. Wilson. 2001. New sinkhole proximity maps for selected
Counties in Central Florida. Published by Subsurface Evaluations, Inc., Tampa,
Florida. 16 p.
Butler, A.M., B.A. Diskin, K.L. Eastman, D.H. Gatzlaff, R.B. Corbett, C.C. Lilly, and
P.F. Maroney. 1992. Executive Summary in Insurance study of sinkholes. Florida
State University Center for Insurance Research, Tallahassee, Florida. 13 p.
Campbell, Kenneth M. 1984. Geology of Hillsborough County. Florida Geological
Survey Open File Report no.6, Tallahassee, Florida. Florida Geological Survey.
19 p.
Casper, J., B. Ruth, and J. Degner. 1981. A remote sensing evaluation of the potential for
sinkhole occurrence. Remote Sensing Applications Laboratory, Department of
Civil Engineering, University of Florida, Gainesville, Florida. 103 p.
Chen, F.H. 1988. Foundations on Expansive Soils. Developments in Geotechnical
Engineering 54:65-69. New York. Elsevier Scientific Publishing Company.
Compton, J. S., 1997. Origin and paleoceanographic significance of Florida's deposits in
The Geology of Florida edited by A.F. Randazzo and D.S. Jones, 195-216.
Gainesville, Florida. University Press of Florida.
Currey, D.R. 1974. "Continentality of extratropical climates," Annals of the Association
of American Geographers 64:268-280.
Department of Army USA 1983. Foundation in Expansive Soils. Washington, DC, TM
5:818-7.
Dodek, B. 2003. A GIS Analysis of Sinkhole Activity in North-Central Florida: Causes
and Correlations. Master's Thesis University of Florida. Gainesville, Florida.
79 p.
Fernald, E.A and E.D. Purdum, editors 1998. Water resources atlas of Florida: Institute
of Science and Public Affairs, Tallahassee, Florida. Florida State University,
310 p.
Florida Department of Environmental Protection 2000. Florida Physiographic Provinces.
Florida Department of Environmental Protection ftp site, May 2003.
http://www.dep.state.fl/us/geology/pub.
Florida Department of Environmental Protection 2003. Raster data obtained from FDEP
staff (unpublished data), then it was converted to digital format for this map.
Florida Department of Environmental Protection 2000. Florida Subsurface geology.
Florida Department of Environmental Protection ftp site, May 2003.
http://www.dep.state.fl/us/geology/pub.
Florida Geographic Data Library 1990. Florida county boundaries: http://www.fgdl.org
Florida Geological Survey web site, November 2002. http://www.dep.state.fl/us/geology.
Frank, E.F., B.F. Beck 1990. An analysis of the cause of subsidence damage in the
Dunedin, Florida area. 59 p.
Galloway, D., D.R. Jones, and S.E. Ingebritse 1999. Sinkholes, West-Central Florida.
Excerpt from Circular 1182:121-140. United States Geological Survey.
Green, R., J.D Arthur, and D. DeWitt 1995. Lithostratigraphic and Hydrostratigraphic
cross sections through Pinellas and Hillsborough counties, Southwest Florida.
Open File Report no.61, 1-26. Florida Geological Survey, Tallahassee, Florida.
Healy, H.G. 1975. Terraces and shorelines of Florida. Florida Bureau of Geology Map
Series no.71.
Heatherington, A.L. and P.A. Mueller 1997. Geochemistry and origin of Florida crustal
basement terranes, in The Geology of Florida edited by A.F. Randazzo and D.S.
Jones, 27-37. University Press of Florida, Gainesville, Florida.
Henry, J.A., K.M Portier, and J. Coyne 1994. The Climate and Weather of Florida.
Sarasota, Florida. Pineapple Press, Inc. 249 p.
Hine, A.C. 1997. Structural and paleoceanographic evolution of the margins of the
Florida platform, in The Geology of Florida edited by A.F. Randazzo and D.S.
Jones, 169-194. University Press of Florida, Gainesville, Florida.
Hyde, L.W. 1975. Principal Aquifers in Florida, Map Series No. 16, United States
Geologic Survey and Bureau of Geology, Tallahassee, Florida.
Lane, E. 1986. Karst in Florida, Special publication # 29, Florida Geological Survey,
Tallahassee, Florida. 100 p.
Miller, J.A. 1986. Hydrogeologic Framework of the Floridan aquifer system in Florida
an in parts of Georgia, South Carolina, and Alabama, United States, professional
paper 1403-B. Geological Survey, Department of the Interior, Washington D.C.-
91 p
- 1997. Hydrogeology of Florida, in The Geology of Florida edited by A.F. Randazzo
and D.S. Jones, 69-88. University Press of Florida, Gainesville, Florida.
National Oceanic and Atmospheric Administration, National Weather Service 2003
http://www.srh.noaa.gov/tbw/climate, May 5, 2003.
Nelson, J.D. and D.J. Millerl992. Expansive Soils: problems and practice in foundation
and pavement engineering, New York. John Wiley and Sons, Inc. 43 p.
Pirkle, E.C., W.H. Yoho, and C.W. Hendry, Jr. 1970. Ancient Sealevel Stands in Florida
Bulletin no.52, 61 p. Published by the Bureau of Geology Tallahassee, Florida.
61 p.
Randazzo, A.F. 1997. The sedimentary platform of Florida: Mesozic to Cenozoic, in The
Geology of Florida edited by A.F. Randazzo and D.S. Jones, 39-56. Gainesville,
Florida. University Press of Florida.
- and D.L.Smith 2003. Subsidence-induced foundation failures in Florida's karst terrain.
Sinkholes and the engineering and environmental impacts of karst. Proceedings of
the ninth multidisciplinary conference, Huntsville, Alabama, 82-94. Edited by
Barry F. Beck, American Society of Civil Engineers.
Schmidt, W., 1997. Geomorphology and physiography of Florida, in The Geology of
Florida edited by A.F. Randazzo and D.S. Jones, 1-12. Gainesville, Florida,
University Press of Florida.
- and T.M. Scott, 1984. Florida karst-Its relationship to geologic structure and
stratigraphy. Proceedings of the First Multidisciplinary Conference on Sinkholes,
11-16.
Scott,T. M. 1992. A Geolocgical Overview of Florida. Florida Geological Survey Open
File Report no.50, Tallahassee, Florida. 77 p.
Scott, T.M. 1997. Miocene to Holocene history of Florida, in The Geology of Florida
edited by A.F. Randazzo and D.S. Jones. University Press of Florida, Gainesville,
Florida. Structural features of the Florida peninsula p 58.
- 1997. Miocene to Holocene history of Florida, in The Geology of Florida edited by
A.F. Randazzo and D.S. Jones, 57-67. University Press of Florida, Gainesville,
Florida.
- 2001. Text to accompany the Geologic map of Florida, Florida Geological Survey
Open File Report 80. Tallahassee, Florida.
Sinclair, W.C. 1982. Sinkhole development resulting from ground-water withdrawal in
the Tampa Area, Florida: U.S. Geological Survey Water Resources Division. 19 p.
-, J.W. Stewart, R.L. Knutilla, A.E. Kilboy, and R.L. Miller 1985. Types, features, and
occurrence of sinkholes in the karst of west-central Florida: U.S. Geological
Survey Water Resources Investigations.81-50. 19 p.
and J.W. Stewart. 1985. Sinkhole type development and distribution in Florida:
Florida Geological Survey Map Series 110. Tallahassee, Florida. Florida
Geological Survey.
Smith, D.L. and K.M. Lord. 1997. A Techtonic Evolution and Geophysics of the Florida
Basement. In The Geology of Florida edited by A.F. Randazzo and D.S. Jones,
13-26. Gainesville, Florida. University Press of Florida.
Southeastern Geological Society Ad Hoc Committee on Florida Hydrostratigraphic Unit
Definition 1986. Hydrogeological units of Florida: Florida Geological Survey
Special Publication 28. Tallahassee, Florida. 8 p.
Southwest Florida Water Management District 2002. Recharge and Discharge areas of
the Floridan aquifer system. http://www.swfwmd.state.fl.us/data/gis, November
20, 2003.
-2002. Map of five-county study area.
http://www.swfwmd.state.fl.us/data/gis/library/political.htm. May 2003.
-2003. Potentiometric base maps and potentiometric surface obtained separately from
http://www.swfwmd.state.fl.us/data/data/dataonline.htm. May 2003.
-2003. Well pumpage and locations were supplied in Excel format by SWFWMD
employee via email June 2003.
Spechler, R.M. and D.M. Schiffer 1995. United States Geological Survey fact sheet
FS-151-95, 1 sheet.
Spencer, S.M. and Ed Lane 1995. Florida Sinkhole Index. Open File Report no.58,
Talahassee, Florida. Florida Geological Survey. 18 p.
Sridharan, A. and K Prakash 2000. Shrinkage limit of soil mixtures. ASTM Geotechnical
Testing Journal 23:3-8.
Teleatlas: http://www.na.Teleatlas.com, November 3, 2002
Tihansky, A.B. United States Geological Survey. 1999. Human-induced sinkhole
development in the study area, Pasco and Hernando counties. Sinkholes, West-
Central Florida, A Link Between Surface Water and Ground Water, an excerpt
from Galloway, Devin, D.R. Jones, and S.E. Ingebritsen, Circular 1182, p 138.
-,United States Geological Survey. 1999. Sinkholes, West-Central Florida, A Link
Between Surface Water and Ground Water, an excerpt from Galloway, Devin, D.R.
Jones, and S.E. Ingebritsen, Circular 1182, p 127.
-,United States Geological Survey. 1999. Sinkholes, West-Central Florida, A Link
Between Surface Water and Ground Water, an excerpt from Galloway, Devin, D.R.
Jones, and S.E. Ingebritsen, Circular 1182, 121-139.
United States Census Bureau 2000, Population density by U.S. Census block groups:
http://www.census.gov/geo/www/cob/bg2000.html
United States.Geological Survey 1990. Confining units of the Floridan aquifer system.
(modified from Miler). Ground Water Atlas of the United States: Alabama,
Florida, Georgia, and South Carolina, publication HA 730-G
- 1990. Ground Water Atlas of the United States: Alabama, Florida, Georgia, and
South Carolina, publication HA 730-G. Online version;
http://capp.water.usgs.gov/gwa, November 18, 2003.
United States Geological Survey 1990. National Water Summary 1987: hydrologic
events and water supply and use. U.S. Geological Survey Water-Supply Paper
no. 23, 550-553.
Upchurch, S.B. and A.F. Randazzo. 1997. Environmental geology of Florida. In
Geology of Florida, edited by A.F. Randazzo and D.S. Jones, 217-48. Gainesville,
Florida. University Press of Florida.
Walker, K.R., G. Shanmugam, and S.C. Ruppel 1983. A model for carbonate to
terrigenous plastic sequences. Bulletin of the Geological Society of America
no. 94, 700-712.
Wetterhall, W.S. 1964. Geohydrologic Reconnaissance of Pasco and Southern Hernando
Counties, Florida. Report of Investigations No. 34, Tallahassee, Florida. Florida
Geological Survey. 12 p.
White, W.A., 1970. The Geomorphology of the Florida Peninsula:Bureau Geology.
Florida Department of Natural Resources, Bulletin no.51, 69 p.
91
Zisman, E.D. 2003. Guilty until proven innocent Sinkhole definition & identifying
features. Sinkholes and the engineering and environmental impacts of karst.
Proceedings of the ninth multidisciplinary conference, Huntsville, Alabama.
Edited by Barry F. Beck, American Society of Civil Engineers, 124-129.
