Soils, landscapes, and ground-penetrating radar analyses of the Chiefland limestone plain in Levy County, Florida

MISSING IMAGE

Material Information

Title:
Soils, landscapes, and ground-penetrating radar analyses of the Chiefland limestone plain in Levy County, Florida
Physical Description:
xv, 251 leaves : ill., photos. ; 29 cm.
Language:
English
Creator:
Puckett, William Edward, 1958-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 240-250).
Statement of Responsibility:
by William Edward Puckett.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001572708
notis - AHJ6536
oclc - 22830045
System ID:
AA00003756:00001

Full Text











SOILS, LANDSCAPES, AND GROUND-PENETRATING RADAR
ANALYSES OF THE CHIEFLAND LIMESTONE PLAIN
IN LEVY COUNTY, FLORIDA
















BY

WILLIAM EDWARD PUCKETT


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1990































....dedicated to the memory of my father,
George Edward Puckett.













ACKNOWLEDGEMENTS


I would like to thank my major professor Dr. Mary E.

Collins for being a teacher and a friend during my tenure as

a student. I also thank the members of my committee, Dr.

Richard W. Arnold, Dr. Willie G. Harris, Dr. Arthur G.

Hornsby, and Dr. Robert C. Lindquist, for their time and

guidance during my research. Special thanks are extended to

Mr. Gregg W. Schellentrager for his help with the ground-

penetrating radar and his continued support of this

research.

A special thanks goes to the U.S.D.A. Soil Conservation

Service for its support of this research. Special

recognition is extended to G. Wade Hurt, whose patience and

support provided the time needed to finish this manuscript.

I would also like to thank the members of the Levy County

soil survey staff, Mr. Craig A. Ditzler, Mr. Alfred 0.

Jones, Mr. Paul E. Pilney, Mr. Joseph N. Schuster, Mr. James

D. Slabaugh, and Ms. Carol A. Wettstein. Thanks are also

extended to Mr. Fred E. Miller, Mr. William G. Harb, and Ms.

Billie Clark for their contributions during the topographic

survey of the Quincey plot.

I express my sincere gratitude to Mr. Frank Quincey,

whose property was used for the detailed research associated

iii






with this project. Mr. Quincey was patient and supportive

as he witnessed the conversion of his property into Arents.

I thank Ms. Nancy E. Washer, whose friendship and help

during this project will always be remembered. Thanks are

also extended to my friends in the Soil Characterization

Laboratory.

I thank Mr. Larry E. Clark and Mr. Matthew Sauer for

their help with the figures in this dissertation.

My deepest appreciation is extended to my wife

Elizabeth, whose sacrifice, patience, and encouragement kept

the fires burning during this long journey. I also thank my

son Ashton, whose arrival provided the additional motivation

needed to finish this dissertation.













TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS....................................... iii

LIST OF TABLES. ........................................ vii

LIST OF FIGURES......................................... x

ABSTRACT................................................ xiv

CHAPTERS

1 INTRODUCTION AND DESCRIPTION OF STUDY AREA.... 1

Introduction................................... 1
Description of Study Area..................... 2
Location.................................. .. 2
Climate..................................... 2
Drainage................................... 5
Physiography and Geology...................... 6
Structure.......................... ........ 15
Stratigraphy................................ 16
Objectives of Study.......................... 22

2 RADAR AND FLEX-GRID ANALYSIS OF SOILS
ON THE CHIEFLAND LIMESTONE PLAIN.............. 23

Introduction................................... 23
Materials and Methods........................ 26
Location of Study Area...................... 26
Physiography and Geology................... 27
Initial Soils Legend........................ 28
GPR System.................................... 29
Flex-Grid Mapping and GPR Transect
Information....................... ........ 29
Justification and Definition of
Intriplex Map Unit........................ 31
Statistical Analysis....................... 33
Results and Discussion........................ 34
Initial Soils Legend........................ 34
Soil-Landscape Model....................... 36
Grid Mapping................................ 38
GPR Analysis ................................ 42
Final Flex-Grid and Transect Procedure....... 49
v







Intriplex Map Units......................... 54
Conclusions.................................... 62

3 RADAR-GRID CATEGORIES OF SUBSURFACE FEATURES
ON THE CHIEFLAND LIMESTONE PLAIN.............. 68

Introduction................................. 68
Materials and Methods......................... 69
Location and Selection of Study Area........ 69
GPR System................................. 69
Grid Survey and GPR Transects.............. 72
Definition of Radar Class and
Radar-Grid Category....................... 76
Selection of Representative Pedons.......... 77
Laboratory Analysis......................... 77
Statistical Methods......................... 79
Computer-Generated Maps and Diagrams........ 81
Results and Discussion........................ 81
Radar Classes.. ............................ 81
Radar Classes and Soil Characteristics...... 95
Radar Interfaces............................ 113
Radar-Grid Categories....................... 121
Conclusions..... ............................. 146

4 GENETIC PATHWAYS OF SOIL FORMATION ON THE
CHIEFLAND LIMESTONE PLAIN..................... 148

Introduction. ................................ 148
Materials and Methods......................... 152
Location of the Study Area.................. 152
Physiography and Geology.................... 153
Selection of Soils.......................... 154
Laboratory Analysis......................... 154
Results and Discussion....................... 158
Soil Morphology............................. 158
Physical and Chemical Properties............ 169
Soil Mineralogy.................................... 178
Fossil-Limestone Identification............. 186
Insoluble Limestone Residue................ 187
Microscopy Study............................ 189
Uniformity of Parent Materials.............. 193
Parent Material-Time-Soil Sequence Model.... 204
Conclusions ................................. 220

5 GENERAL CONCLUSIONS........................... 222

APPENDIX................................................. 227

LITERATURE CITED........................................ 240

BIOGRAPHICAL SKETCH..................................... 251













LIST OF TABLES


Table Page

1-1 Correlation of terraces and shorelines in Florida
(modified from Healy, 1975)........................ 11

1-2 Geologic formations in Levy County, Florida
(modified from Vernon, 1951)....................... 18

2-1 Classification of series used in the initial soils
legend for the Chiefland Limestone Plain............ 35

2-2 Summary of grid data for 260-ha area on the
Chiefland Limestone Plain.......................... 40

2-3 Summary of ground-penetrating radar transect data
for the 260-ha selected study area on the
Chiefland Limestone Plain.......................... 45

2-4 Comparison of grid data and ground-penetrating
radar transect data for the selected 260-ha study
area on the Chiefland Limestone Plain.............. 47

2-5 Summary of grid data from the four 260-ha areas
located on the Chiefland Limestone Plain........... 48

2-6 Series and classification of soils used in the final
descriptive legend for the Chiefland Limestone
Plain ............. ................................. 51

2-7 Abbreviated field descriptions of soils on the
Chiefland Limestone Plain.......................... 52

2-8 Average composition of map units as determined by
the ground-penetrating radar transect method........ 55

3-1 Radar classes, pedon grid coordinates, and
elevations of representative pedons at the Quincey
plot................................................ 99

3-2 Abbreviated field descriptions of representative
soils at the Quincey plot.......................... 100


vii







3-3 Summary of selected physical and chemical
properties of representative soils in the
Quincey plot..................................... 103

3-4 Selected statistics of surface elevations, first-
radar interface (1RI), second-radar interface
(2RI), and third-radar interface (3RI) for the
437 observation stations on the Quincey plot....... 116

3-5 Criteria for establishing radar-grid categories for
the 437 observation stations on the Quincey plot... 124

3-6 Selected statistics of surface elevations, first-
radar interfaces (1RI), second-radar interfaces
(2RI), and third-radar interfaces (3RI) for each
radar-grid category (RGC) on the Quincey plot...... 129

3-7 Distance statistics for pattern analysis using
nearest-neighbor analysis for each radar-grid
category (RGC) on the Quincey plot................. 134

4-1 Pedon grid-coordinates, soil series, and
classification for the soils sampled at the
Quincey plot........................................ 159

4-2 Proposed discontinuities and sources of
parent materials based on soil morphology for the
soils studied at the Quincey plot.................. 170

4-3 Particle-size data for the soils studied at the
Quincey plot ................................ ...... 171

4-4 Selected chemical data for the soils studied at the
Quincey plot ...................................... 174

4-5 Selected chemical properties by horizon for the
soils studied at the Quincey plot.................. 176

4-6 Qualitative mineralogy and integrated peak
intensities of the coarse clay fraction for the
soils sampled at the Quincey plot.................. 179

4-7 Qualitative mineralogy of the silt fractions for
selected horizons from the Candler and Shadeville
soils at the Quincey plot.......................... 185

4-8 Insoluble residues of the Ocala Group Limestones
and particle-size data of the insoluble residues
for R horizons in the Bushnell and Shadeville soils
at the Quincey plot................................ 188


viii







4-9 Petrographic analysis of the fine-sand fraction
of selected soils and horizons at the Quincey
plot................................................. 194

4-10 Summary of selected stationary components by
horizon for soils at the Quincey plot.............. 200

4-11 Fine, coarse, and total clay by horizon for soils
with Bt horizons at the Quincey plot............... 203

4-12 Proposed parent material-time model for the soils
on the Chiefland Limestone Plain.................... 205

4-13 Summary of lithological discontinuities as
determined using laboratory data for the soils at
the Quincey plot................................... 210













LIST OF FIGURES


Figure Page

1-1 Location of the Quincey Plot and the Chiefland
Limestone Plain in Levy County, Florida............. 3

1-2 Physiographic areas of Levy County, Florida
(modified from Vernon, 1951)....................... 9

1-3 Terraces identified in Levy County, Florida
(modified from Healy, 1975) ....................... 10

1-4 Types of karst in west-central Florida
(modified from Sinclair and Stewart, 1985).......... 14

1-5 Fracture patterns and locations of the Bronson
Graben and Long Pond Fault in Levy County,
Florida (modified from Vernon, 1951)............... 17

2-1 Conceptual soil-landscape model for the
Chiefland Limestone Plain based on the initial
soils legend........................................ 37

2-2 Soil map, ground-penetrating radar (GPR)
transect locations, and topographic map (based
on USGS topoquad) of the 260-ha study area on
the Chiefland Limestone Plain....................... 39

2-3 Ground-penetrating radar graphic profile
illustrating the subsurface variability of the
Bt horizons in soils on the Chiefland Limestone
Plain................. ............................. 44

2-4 Ground-penetrating radar graphic profile of
the Otela-Candler intriplex map unit............... 58

2-5 Ground-penetrating radar graphic profile of
the Shadeville-Otela-Levyville intriplex map
unit............................................... 60

2-6 Ground-penetrating radar graphic profile of
the Jonesville-Otela-Seaboard intriplex map
unit................................................. 64








2-7 Exposed limestone bedrock within the Jonesville-
Otela-Seaboard intriplex map unit illustrating
the subsurface variability of the limestone......... 66

3-1 Ground views of the Quincey Plot................... 71

3-2 Aerial view of the Quincey Plot from the northeast.
Locations of the sampled soils are shown using
grid coordinates................................. 74

3-3 Graphic-radar profile and Otela soil profile at
observation station 0,16 in the Quincey plot....... 83

3-4 Graphic-radar profile and Shadeville soil
profile at observation station 2,4 in the
Quincey plot....................................... 85

3-5 Graphic-radar profile and Pedro soil profile at
observation station 2,21 in the Quincey plot....... 88

3-6 Graphic-radar profile and Bushnell soil profile at
observation station 10,10 in the Quincey plot...... 90

3-7 Graphic-radar profile and pedon 12,0 profile at
observation station 12,0 in the Quincey plot....... 92

3-8 Graphic-radar profile and pedon 14,10 profile at
observation station 14,10 in the Quincey plot...... 94

3-9 Graphic-radar profile and Candler soil profile at
observation station 18,18 in the Quincey plot...... 97

3-10 Computer-generated contour map of surface
elevations and locations of sampled soils
for the Quincey plot (contour interval,
0.2 m).............................................. 98

3-11 X-ray diffraction patterns of Mg-saturated,
glycerol-solvated clay by horizon for pedon
0,16 (Otela soil) at the Quincey plot.............. 105

3-12 X-ray diffraction patterns of Mg-saturated,
glycerol-solvated clay by horizon for pedon
2,21 (Pedro soil) at the Quincey plot.............. 107

3-13 Computer-generated surface-net diagram of
surface elevations above mean sea level (msl) for
the Quincey plot. Sampled pedon locations
are posted above the diagram. Vertical
exaggeration is 10% of horizontal scale............. 115







3-14 Computer-generated net diagram of the first-
radar interface based on elevation above
mean sea level (msl) for the Quincey plot. Sampled
pedon locations are posted above the diagram.
Vertical exaggeration is 10% of horizontal
scale................ .............. ................. 118

3-15 Computer-generated net diagram of the second-
radar interface based on elevation above
mean sea level (msl) for the Quincey plot. Sampled
pedon locations are posted above the diagram.
Vertical exaggeration is 10% of horizontal
scale............. ................................. 120

3-16 Computer-generated net diagram of the third-
radar interface based on elevation above
mean sea level (msl) for the Quincey plot. Sampled
pedon locations are posted above the diagram.
Vertical exaggeration is 10% of horizontal
scale................................ .............. 122

3-17 Plot of the radar-grid categories for the
Quincey plot.................................... .. 128

3-18 Individual plots of radar-grid categories 3,
7, and 8 that have a random spatial arrangement
of points as determined using nearest-neighbor
analysis for soils at the Quincey plot............. 136

3-19 Individual plots of radar-grid categories 4,
5, and 9 that have a clustered spatial
arrangement of points as determined using
nearest-neighbor analysis for soils at the
Quincey plot.................................... .. 138

3-20 Individual plots of radar-grid categories 1 and 2
that have a dispersed spatial arrangement of points
as determined using nearest-neighbor analysis for
soils at the Quincey plot.......................... 140

3-21 Computer-generated contour map of the third-
radar interface based on elevations above mean sea
level for the Quincey plot (contour
interval 0.5 m) ... ................................. 141

3-22 Computer-generated contour map of the first-radar
interface based on elevations above mean sea level
for the Quincey plot (contour interval 0.5 m)...... 143


xii







3-23 Computer-generated contour map of the second-radar
interface based on elevations above mean sea level
for the Quincey plot (contour interval 0.5m)....... 145

4-1 X-ray diffraction patterns of Mg-saturated,
glycerol-solvated coarse clay by horizon for
the Shadeville soil at the Quincey plot............ 182

4-2 X-ray diffraction patterns of Mg- and K-saturated
glycerol-solvated fine clay for the Btl horizon of
the Otela soil at the Quincey plot................. 184

4-3 Thin sections of selected horizons of the
Otela and Bushnell soils at the Quincey plot....... 191

4-4 Dominant vertical sequences of soil strata
occurring on the Chiefland Limestone Plain......... 196

4-5 Generalized landscape evolution model for the
Chiefland Limestone Plain.......................... 206


xiii













Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SOILS, LANDSCAPES, AND GROUND-PENETRATING RADAR
ANALYSES OF THE CHIEFLAND LIMESTONE PLAIN
IN LEVY COUNTY, FLORIDA

By

William Edward Puckett

May 1990


Chair: Mary Elizabeth Collins
Major Department: Soil Science


This research was conducted to determine and evaluate

the soil patterns, soil-landscape relationships, and genesis

of the soils on karst in west-central Florida. The study

was conducted in Levy County, Florida on the Chiefland

Limestone Plain (CLP). The CLP encompassed approximately

28,000 ha of sandy to clayey sediments overlying Eocene-age

limestone. The ground-penetrating radar was used to non-

destructively study the subsurface features of the CLP. Few

relationships existed between the soils and landscapes on

the CLP. The soils occurred in patterns so intricate that

grid mapping had to be used to delineate map units. A new

kind of map unit, the intriplex, was introduced to correctly

define the intricacy and apparent randomness of the soils


xiv






and landscapes on the CLP. A more efficient method of grid

mapping, the flex-grid method, was also introduced and

subsequently used by the U.S.D.A. Soil Conservation Service

in the progressive soil surveys of all the karst areas in

Levy and surrounding counties.

The present CLP surface does not reflect the highly

variable paleo-karst nature of the subsurface. The soils

formed in Eocene-, Miocene-, Pliocene-, Pleistocene-, and

Holocene-aged parent materials that were reworked and

deposited on a paleo-karst surface. The abrupt sandy and

loamy boundaries between the E and Bt horizons, in the

Pleistocene sediments of the CLP soils, were formed by

pedogenic processes. The majority of the soils were not

formed as a residual product of limestone weathering.

Limestone insoluble residues were estimated to accumulate at

a rate of 1.5 cm every 40,000 years. The majority of the

sediments available for soil formation were derived outside

of the CLP area. A flocculation-dispersion mechanism was

purposed to explain the occurrence of abrupt boundaries

between eluvial and illuvial horizons.












CHAPTER 1
INTRODUCTION AND DESCRIPTION OF STUDY AREA


Introduction


The karst landscapes of west-central Florida have been

shaped by sea-level fluctuations and solutional processes.

Numerous cycles of marine deposition and subaerial erosion

have left an intricate stratigraphy of sands, clays, and

limestones. Pedogenic processes in these geologic materials

are affected by the initial composition of these sediments

and karst processes.

Soil formation by the five soil forming factors

(time, parent material, topography, climate, and biota) were

described by Dokuchaev (1883) and Jenny (1941). The varied

effects of topography, climate, and biota through time on

various parent materials are reflected in the multiplicity

of soils found throughout Florida. Cooke (1939) listed four

factors that are responsible for shaping Florida's

landscapes and contributing to the initial composition of

sedimentary units: (a) erosion by running water, (b)

solution of rocks, (c) work of waves, winds, and currents,

and (d) fluctuations of sea level. Karst processes were

considered the dominate geologic process controlling

Florida's topography (Cooke, 1939; Vernon, 1951; White,

1970).









Soil formation on karst is a dynamic process. The

formation of sinkholes and their influence on the soil

system have not been extensively studied. Very little

research has been conducted on soil patterns and soil-

landscape relationships on karst in Florida. Thus, this

dissertation will explore the soils and landscapes on karst

in west-central Florida.


Description of Study Area


Location

The study area was on the Chiefland Limestone Plain

(CLP) in Levy County, Florida (Fig. 1-1). Levy County is

bounded on the northwest by the Suwannee River, to the north

by Gilchrist and Alachua Counties, to the east by Marion

County, to the south by the Withlacoochee River and to the

west and southwest by the Gulf of Mexico. The CLP is

located in the northwestern part of Levy County. The CLP is

divided into northern and southern halves. This study was

confined to the northern half of the CLP. A subarea,

referred to as the Quincey plot (QP), was selected for more

detailed study within the CLP and was located in the

northeastern part of the CLP on the farm of Mr. Frank

Quincey (Fig. 1-1).

Climate

Levy County is characterized by long, warm summers and

mild winters. The average summer temperature was 270C, the



















4--


Limestone Plain


0 25
km


5 /
I
I


Citrus County


Figure 1-1. Location of the Quincey Plot and the Chiefland
Limestone Plain in Levy County, Florida.


.z
e j









average winter temperature was 150C and the average annual

rainfall was 1270 mm from 1841 to 1949 (Vernon, 1951). The

average annual temperature was 20oC and the average annual

rainfall was 1504 mm from 1962 to 1982 (data collected by

the Florida Department of Forestry at the Usher fire tower

which was located 2.5 km south of Chiefland, Florida, on

U.S. Highway 19). Most of the rainfall occurs during

afternoon thunderstorms during the months of June through

October. Levy County is located on the Gulf of Mexico and

is subject to occasional tropical storms and hurricanes.

Extensive flooding can occur along the coastal lowlands

during these events.

Levy County is the northern boundary of the

hyperthermic temperature regime (Soil Survey Staff, 1975) on

the west coast of Florida. The U.S.D.A. Soil Conservation

Service in Florida determined that the boundary between

thermic and hyperthermic soils, for the purpose of making

soil surveys (G.W. Hurt, 1986, personal communication), was

the Suwannee River from the Gulf of Mexico to Gilchrist

County and then east along the Gilchrist-Levy County line.

Brasfield and Carlisle (1976) determined that the boundary

between thermic and hyperthermic soils should be moved south

to the Withlacoochee River between Levy and Citrus Counties.

Vernon (1951) noted that the vegetation was more tropical

and lush in Citrus than Levy County and a difference in

annual air temperature was evident.









The actual boundary between thermic and hyperthermic

soils is not stationary, but probably varies through time

and more than likely is better represented as a transitional

zone stretching from Gilchrist County to Citrus County.

Both thermic and hyperthermic soils were mapped adjacent to

each other on the CLP. Soils that were members of

established series were mapped as hyperthermic and new

series were established as thermic. The new series were

established as thermic so they could be used in the

progressive soil surveys of adjacent northern counties with

similar karst landforms, (i.e., Gilchrist, Dixie and

Lafayette Counties).

Drainage

Overland drainage systems are absent on the CLP.

Subsurface flow is the dominate drainage mechanism in Levy

County (Vernon, 1951). Sandy surfaces and highly porous

limestones contribute to the lack of surface drainage.

The Suwannee and Withlacoochee Rivers have few

tributaries in Levy County and both are entrenched, in

various parts of their courses, into limestone. Numerous

springs along the Suwannee River represent release points

for much of the rainfall on the CLP (Vernon, 1951; Rosenau

et al., 1977). The northern reaches of the Waccasassa River

in Levy County flows across thick, clayey alluvial deposits

and surface drainage is predominate (Vernon, 1951). When

the Waccasassa River floods in this area of Levy County, the









floodwaters flow onto the CLP and slowly disappear into

sinkholes. Further south, the Waccasassa River becomes

entrenched in limestone and few tributaries are present.

Physiography and Geology

Levy County is part of the Coastal Plain Province as

defined by Fenneman (1938) and Thornbury (1965). The

Coastal Plain Province is a subaerial inland extension of

the continental shelf. The Coastal Plain Province is

characterized by sandy to clayey unconsolidated marine and

fluvial sediments. Fennenman (1938) and Thornbury (1965)

suggested that most of central Florida's topography was

controlled, to a large degree, by the presence of soluble

limestone.

The Florida peninsula is divided into three

physiographic zones (White, 1970). Levy County is located

in the mid-peninsular physiographic zone. The mid-

peninsular zone extends from approximately Lake City,

Florida, to the northern end of Lake Okeechobee. The mid-

peninsular zone is characterized by discontinuous parallel

ridges separated by broad valleys. Geomorphic subzones are

differentiated within the mid-peninsular zone based on

elevations. In Levy County, two geomorphic subzones are

recognized; the Central Highlands and the Gulf Coastal

Lowlands (White, 1970). The Central Highlands are

restricted to eastern Levy County. The Gulf Coastal

Lowlands occupy the western part at elevations below 33 m.









The Gulf Coastal Lowlands consist of nearly level to gently

sloping plains with limestone at or near the surface.

Vernon (1951) subdivided Levy County into two major

physiographic areas: the Terraced Coastal Lowlands and the

River Valley Lowlands (Fig. 1-2). The Terraced Coastal

Lowlands include four coastwise terraces: (a) the Coharie at

66 m, (b) the Okefenokee at 45 m, (c) the Wicomico at 30 m

and (d) the Pamlico at 7.5 m. In addition, Healy (1975)

added the Silver Bluff terrace at approximately 3 m (Fig.

1-3). Correlations of Florida terraces by several authors

are presented in Table 1-1.

The River Valley Lowlands include the alluvium of the

Wicomico and Pamlico terraces and Holocene floodplains of

the Suwannee and Waccasassa Rivers. Vernon (1951) and White

(1970) postulated that the Suwannee River once flowed along

the eastern edge of the CLP through the Waccasassa River

valley. Vernon (1951) suggested that the Suwannee River

might have migrated from the Waccasassa River valley, across

the CLP into its present valley. However, Puri et al.

(1967) described a topographic high in Gilchrist County

which they believed would have kept the Suwannee River from

occupying the Waccasassa River Valley.

Several geomorphic subdivisions are identified within

the Terraced Coastal Lowlands and River Valley Lowlands

(Fig. 1-2). The CLP was one of the geomorphic surfaces

defined by Vernon (1951).


















Figure 1-2. Physiographic areas of Levy County, Florida
(modified from Vernon, 1951).

LeQend

TERRACED COASTAL LOWLANDS

Coharie-Okefenokee Terrace
MBrooksville ridge

Wicomico Terrace
(2Sand belt and coast line
EChiefland limestone plain

Pamlico Terrace
[BSand belt and coast line
ELimestone shelf and hammocks
RCoastal marst belt

RIVER VALLEY LOWLANDS

Suwannee River Valley Lowland
HFlood plain
Sand bars of Pamilco formation
OSand bars of Wicomico formation

Waccasassa River Valley Lowland
iDelta plain of Pamilco formation
5Alluvium of Wicomico formation
RWilliston limestone plain





Puckett, William
1990 Soils, Landscapes, and Ground=Penetrating Radar
Analyses of the Chiefland Limestone Plain in Levy County,
Florida. Thesis (Ph.D.), University of Florida.











Gilchrist


0












Gulf of Mexico 1 +
+ +












0 17
ICitrus County
km




Puckett, William
1990 Soils, Landscapes, and Ground=Penetrating Radar
Anial,:-,es of the Chiefland Limestone Plain in Levy County,
Florida. Thesis (Ph.D.), University of Florida.





















0 25
-km


Silver Bluff terrace (< 3 m).

Pamlico terrace (3 to 8 m).

Penholoway terrace (14 to 23 m).


Wicomico terrace (26 to 33 m).


Coharie-Okefenokee terrace (> 33 m).


Figure 1-3.


Terraces identified in Levy County, Florida
(modified from Healy, 1975).


Puckett, William
1990 Soils, Landscapes, and Ground=Penetrating Radar
Analyses of the Chiefland Limestone Plain in Levy County,
Florida. Thesis (Ph.D.), University of Florida.


*n
I.


Li









Table 1-1.


Correlation of terraces and shorelines in
Florida (modified from Healy, 1975).


Cooke MacNeil Vernon Bermes and
others
---------------------- ------------- -----------
1939,1945 1959 1951 1963
(Statewide) (Statewide) (Citrus and (Flagler,
Levy Putnum, and
Counties) St.Johns
Counties)


Hazlehurst
72-90 m


Coharie
57-72 m
(marine)

Sunderland
33-57 m
(marine)

Wicomico
23-33 m
(marine)

Penholoway
14-23 m
(marine)


Talbot
8-14 m
(marine)

Pamlico
2-8 m
(marine)


Silver Bluff
0-3 m
(marine)


High Pliocene
50-93 m
(subaerial)


Coharie
73 m
(marine)


Okefenokee
50 m
(marine)

Wicomico
33 m
(marine)


Pamlico
2-12 m
(marine)


Okefenokee
50 m
(marine)

Wicomico
33 m
(marine)


Pamlico
2 m
(marine)


Silver Bluff
2-3 m
(marine)


Coharie
57-72 m
(marine)

Sunderland
33-57 m
(marine)

Wicomico
23-33 m
(marine)

Penholoway
14-23 m
(marine)

Talbot
8-14 m
(marine)

Pamlico
3-8 m
(marine)

Silver Bluff
0-3 m
(marine)









The CLP is a broad, relatively flat surface of

unconsolidated marine sediments overlying the Eocene-age

Ocala Group limestones. The surficial sediments of the CLP

were associated with the Pleistocene-age Wicomico terrace

which was correlated with the Sangamon interglacial period

(Cooke, 1939; Vernon, 1951).

Vernon (1951) listed eight agents responsible for the

present landscape development in Levy County: (a) warm,

humid climate, (b) high annual rainfall, (c) bedrock

composed of soluble carbonates, (d) low surface elevations,

(e) flat to gently dipping porous rock covered by sand and

phosphatic beds, (f) high concentrations of phosphoric,

humic, and carbonic acid waters, (g) fracturing along the

Ocala Uplift and (h) groundwater. Cooke (1939) listed

waves, winds, currents, and the fluctuations of sea level as

the primary agents responsible for regional landforms. The

agents listed above by Vernon (1951) are also necessary for

the development of karst landscapes (Sweeting, 1973; Bloom,

1978).

Karst is a type of topography that is characterized by

closed depressions or sinkholes, and is formed by

underground solution and diversion of surface waters to

underground routes (Soil Survey Staff, 1983). The

disappearance of surface waters and its circulation

underground is the essence of the karst process as described

by Sweeting (1973).







13

The type of karst on the CLP is normally referred to as

"covered karst". Covered karst is where the limestone is

covered by soils (Bloom, 1978). Sinclair and Stewart (1985)

referred to the CLP karst as "bare or thinly covered

limestones" (Fig. 1-4). They defined bare or thinly covered

limestones as being covered by materials ranging in

thickness from less than 0.3 to about 8 m.

Karst features are common on the CLP. Bowl-shaped

depressions and small hills make up the dominate landforms

on the CLP. A majority of the depressions are related to

karst activity as interpreted from ground-penetrating radar

data (G.W. Schellentrager, 1987, personal communication).

Most of the bowl-shaped depressions on the CLP would

classify as solution sinkholes (Sinclair and Stewart, 1985).

Solution sinkholes are formed when lateral variations in

solution are minimum.

The depressions vary in diameter from less than 10 m to

more than 100 m. Depression sideslopes range from 3 to 12

percent. Minor features on the CLP are the vertically and

horizontally oriented solution channels, limestone outcrops,

and drowned caves. Elevations range from 8 to 18 m above

mean sea level in a SW to NE trend across the CLP.

Two additional Pleistocene terraces were mentioned by

Vernon (1951) that may have influenced the topography of the

CLP. These were the Talbot at 7 to 14 m and the Penholoway

at 14 to 24 m. Vernon (1951) believed that karst activity























km 0


D_ Area 1. Bare or thinly covered limestone.

Area 2. Cover is 10 to 66 meters thick (permeable sands).

Area 3. Cover is 10 to 66 meters thick (cohesive clays).

S Area 4. Cover is more than 66 meters thick
(cohesive sediments).

Figure 1-4. Types of karst in west-central Florida
(modified from Sinclair and Stewart, 1985).








removed any evidence of these terraces on the CLP. Opdyke

et al. (1984) demonstrated how the solution of limestone

resulted in isostatic adjustments and subsequent uplift of

beach ridges along Florida's Trail ridge. The result of the

isostatic adjustments was elevational differences on similar

marine terraces. Difficulties arise in assigning

Pleistocene shorelines on karst terrains using elevational

data because of solutional activity (R.C. Lindquist, 1987,

personal communication). Based on elevational and terrain

analysis in Alachua, Gilchrist, and Levy Counties, the CLP

would be considered part of the Penholoway shoreline (R.C.

Lindquist, 1987, personal communication).

Structure

The major structural feature within Levy County is the

Ocala Uplift, Fig. 1-1. The Ocala Uplift represents a

gentle uplift developed in Tertiary sediments (Vernon,

1951). Cooke (1945) and Vernon (1951) believed that the

Ocala Uplift started during Late Eocene to Early Miocene.

Pirkle (1956) agreed with Cooke (1945) and Vernon (1951),

but Pirkle (1956) believed that the uplift continued during

later-Tertiary times. The Ocala Uplift is 434 km long and

113 km wide with the crest of the uplift located in Levy

County. The trend of the Ocala Uplift is NW to SE across

the central part of the state.

Vernon (1951) was the first to recognize structural

faults in Florida. Regional fracture patterns occurred









parallel and perpendicular to the Ocala Uplift (Fig. 1-5).

Extensive faulting has been observed in western Alachua

County on the Newberry Limestone Plain (Williams et al.,

1977). The linear occurrence of sinkholes and caves was

attributed to solutional processes occurring along fracture

zones. Other major structures recognized by Vernon (1951)

included the Bronson graben and Long Pond fault, Fig. 1-5.

The Bronson graben was buried by alluvial sediments and may

have influenced the course of the Waccasassa River. The

Long Pond fault was believed to be responsible for the

separation of the CLP into two sections.

Stratigraphy

The oldest rocks identified in Levy County are the Avon

Park Limestones of Middle-Eocene age. The Avon Park

Limestones are exposed at the surface in southwestern Levy

County. Eocene-age limestones form the basic stratigraphic

units from which karst has formed in Levy County. The CLP

is underlain by Upper-Eocene limestones and younger

sediments. Therefore, the discussion in this text will be

confined to the Upper-Eocene and younger sediments.

Undifferentiated Miocene-, Pliocene-, Pleistocene-, and

Holocene-age sands and clays form most of the surficial

sediments. The geologic history modified from Vernon (1951)

is presented in Table 1-2.







Gilchrist


Citrus County
LEGEND
1= Bronson graben.
2= Long pond fault.
3= Chiefland limestone plain.
\= Fracture or fault.


Figure 1-5. Fracture patterns and locations of the Bronson
Graben and Long Pond Fault in Levy County,
Florida (modified from Vernon, 1951).










Table 1-2. Geologic formations in Levy County, Florida
(modified from Vernon, 1951).

Era Period Epoch Formation


Cenozoic Quaternary


Tertiary


Holocene


Pleistocene


Miocene


sand, peat and clay,
unnamed.


Lake Flirt Marl of
late Wisconsin
glacial stage.


Pamilco formation of
Peorian age.
Wicomico formation
of Sangamon age.
Okefenokee formation
of Yarmouth age.
Coharie formation
of Aftonian age.


Alachua formation
(terrestrial)
_a ??? ???--???
Hawthorn formation
(marine)


Oligocene Absent in Levy
County


Eocene


Ocala
Group
Limestones


L _______________________________1_ ___ ^___ __ I-___


Crystal
River
formation
Williston
formation
Inglis
formation


I i








Eocene series

The term "Ocala" limestone was first used by Dall and

Harris (1892) after examining an exposure of limestone near

Ocala, Florida. The Ocala limestones overlie the Avon Park

limestones in all parts of Levy County except in the

southwestern corner of Levy County (Vernon, 1951). Vernon

(1951) described two formations within the Ocala limestones:

(a) the Ocala limestone-restricted and (b) the underlying

Moodys Branch Formation. Two members were recognized within

the Moodys Branch Formation, the Williston and underlying

Inglis. Puri (1953) raised the Williston and Inglis members

of the Moodys Branch Formation to "Formational status". He

also renamed the Ocala limestone-restricted, the Crystal

River Formation. The Ocala limestones are now known

collectively as the Ocala Group Limestones. The Ocala Group

Limestones, therefore, consist of three formations; the

Inglis Formation, the Williston Formation, and the Crystal

River Formation. The Crystal River and Williston Formations

are the major limestone components of the CLP (Vernon,

1951).

The Ocala Group Limestones are relatively pure (> 98%)

calcium carbonate and solutional features are common

(Vernon, 1951; Williams et al., 1977). Groundwater freely

circulates through this porous limestone which forms the

basic hydrologic unit of the Floridan aquifer (Spangler,

1982).









Oligocene series

The Suwannee limestone which represents the Oligocene

series was not recognized in Levy County (Vernon, 1951;

Crane, 1983). Vernon (1951) proposed that the Suwannee

limestones were removed from Levy County by severe erosion.

Miocene to Pleistocene series

The Ocala Group Limestones, in Levy County, are

commonly covered by Pleistocene sands or by erosional

remnants of the Hawthorn and/or Alachua Formations (Cooke,

1945).

The Hawthorn Formation is predominately associated with

the Miocene (Vernon, 1951; Pirkle, 1956; Scott, 1983). The

Hawthorn Formation consists of widely varying mixtures of

clay, quartz sand, carbonates and phosphates (Scott, 1983).

Scott (1983) applied the term F.U.B.A.R. (Fouled Up Beyond

All Recognition) to the sediments of the Hawthorn Formation.

Variability in the Hawthorn sediments is the rule rather

than the exception. The Hawthorn Formation is present in

Alachua and Marion Counties (Pirkle, 1956; Williams et al.,

1977), but was not recognized in Levy County by Vernon

(1951). Cooke (1945) suggested that all of Florida was

covered by the Hawthorn sea during Miocene times, but Vernon

(1951) believed that Levy County was terrestrial during most

of the Miocene due to the Ocala Uplift. Pirkle et al.

(1965) reported that the Hawthorn Formation was present

throughout most of Florida, except along the crest of the






21

Ocala Uplift. Erosional remnants of the Hawthorn Formation

were probably present over the crest area of the Ocala

Uplift in Levy County. Several coral heads of Siderastraea

siderea, were discovered in a sinkhole near Chiefland,

Florida. This species of coral has been associated with the

presence of the Hawthorn sea in western Alachua County

(Pirkle, 1956; Williams et al., 1977). Transportation of

these coral heads by former ocean currents may have been

responsible for the location of the coral heads on the CLP.

The Alachua Formation is composed of clays, sands,

rubble of phosphatic rock, silicified wood, and limestone

rubble (Vernon, 1951). The Alachua Formation was identified

in eastern Levy County (Vernon, 1951) and on a small area of

the CLP (Vernon, 1951; Crane, 1983). Cooke (1945)

considered the Alachua Formation as the collapsed and

compacted residue of the Hawthorn Formation. Vernon (1951)

correlated the Alachua Formation as a contemporaneous

terrestrial component associated with the Hawthorn

Formation. Pirkle (1956) considered that the Alachua

Formation consisted of terrestrial deposits which

accumulated at intervals during Early Miocene, Pliocene, and

Pleistocene times. Pirkle (1956) considered the Hawthorn

Formation the source of phosphorus in the Alachua Formation.









Williams et al. (1977), considered the Alachua Formation a

terrestrial residual of the Hawthorn Formation.


Objectives of Study


The objectives of this research were; (a) to map,

characterize, and classify the soils on the CLP as part of

the progressive soil survey in Levy County, (b) combine GPR

technology and grid mapping to better determine map unit

type and composition, and (c) provide information on soil-

landscape relationships (Chapter 2); (d) develop

"relationships" between the graphic GPR profiles and the

soils on the CLP, and (e) qualify the spatial

characteristics of the soils identified using the GPR

(Chapter 3); (f) determine the origins of the soils on the

CLP, (g) propose a parent material-time sequence model, and

(h) propose genetic pathways for soil genesis on the CLP

(Chapter 4).











CHAPTER 2
RADAR AND FLEX-GRID ANALYSIS OF SOILS ON
THE CHIEFLAND LIMESTONE PLAIN


Introduction


The landscapes of west-central Florida have been shaped

by sea-level fluctuations and karst processes (Cooke, 1939).

The CLP located in Levy County, Florida, exhibits

characteristics of these processes (Vernon, 1951).

Karstsification has created an irregular bedrock subsurface

while Pleistocene sea-level changes have deposited and

reworked the surficial sediments. The resultant surface is

nearly level to gently rolling with numerous sinkholes.

Depth to limestone and diagnostic subsurface horizons

exhibit extreme variations within short horizontal

distances. Relationships between soils and landscapes on

this karst terrain are not well understood and delineation

of soil map units based on soil-landscape relationships is

difficult.

Variation between soil occurrence and landscape

position in any terrain can be classified as either

systematic or random (Wilding and Drees, 1983). A

systematic relationship occurs when soils can be predicted

based on geomorphic-relationships. A random relationship









would exist when such a prediction about soils and

landscapes is not possible.

Soil scientists normally rely on systematic variations

in soils and landscapes to design and delineate map units

that describe how the soils and landscapes are related. The

type of map unit (i.e., complex, consociation, etc. (Soil

Survey Staff, 1983)) depends on the scale of mapping and its

subsequent use. After the map units are delineated, a

percentage of the map units should be transected to

determine the percentage of component soils. Component

soils are defined in this text as the soils identified in

the map unit.

Johnson (1961) described two transect methods, line-

and point-intercept, for estimating the composition of map

units. These methods are based on whether map unit

boundaries and/or component soil boundaries can be

identified within a landscape.

The line-intercept method is used when component soil

boundaries are easily observed within map unit boundaries.

A soil scientist can walk across the landscape and estimate

the composition of the map unit based on changes of

geomorphic features.

The point-intercept method is used when map unit

boundaries can be delineated within a landscape but

component soil boundaries are not apparent within the map

unit delineation. The point-intercept transect is conducted






25
with observation points at pre-selected intervals within the

map unit. The point-intercept method has been used

successfully in various terrains (Powell and Springer, 1965;

Steers and Hajek, 1979; and Schellentrager et al., 1988).

Lack of soil and landscape relationships in karst

terrains are normally expected (Sweeting, 1973). But, when

soils and landscapes vary independently of each other,

(i.e., karst terrain in Florida), map unit boundaries and,

therefore, component soils are not easily determined. Soil

survey techniques in areas where soil and landscape

relationships are not apparent normally depend on labor

intensive "grid-mapping" procedures. Grid-mapping involves

observing soils at predetermined intervals along parallel

traverse lines and drawing boundaries between observation

points in order to separate map units. Transects may or may

not have to be conducted to supplement the information

obtained using the grid method of mapping.

Ground-penetrating radar (GPR) can be used to

supplement grid mapping in terrains where soils and

landscapes vary independently. Ground-penetrating radar has

the advantage of providing a continuous graphical display of

diagnostic soil horizons below the surface (Johnson, et al.,

1980; Doolittle, 1982; Collins et al., 1986; Collins and

Doolittle, 1987; Schellentrager et al., 1988) and,

therefore, does not require the assumption of soil

homogeneity between observation points.









Ground-penetrating radar has been used to investigate

soil variability and composition in several physiographic

areas in Florida (Shih and Doolittle, 1984; Shih et al.,

1985; Collins et al., 1986; Collins and Doolittle, 1987;

Schellentrager et al., 1988). None of these studies

concentrated on mapping and transecting procedures in karst

areas; nor did they evaluate the effectiveness of

incorporating GPR and grid mapping techniques into an on-

going soil survey. The objectives were to (a) combine GPR

technology and grid mapping to better determine map unit

type and composition and (b) provide information on soil-

landscape relationships. Arnold (1988) considered the study

of soil-landscape relationships to be more intensive during

progressive soil surveys than most other types of soil-

landscape research. Because of this, the study was

conducted in conjunction with the on-going soil survey in

Levy County, Florida.


Material and Methods


Location of Study Area

Levy County is located on the west coast of peninsula

Florida (Fig. 1-1). The mean annual temperature is 21C and

the average annual rainfall is 1387 mm. The study area was

located on the CLP. Elevations range from 8 to 20 m, above

mean sea level, in a SW to NE trend across the limestone

plain. The CLP encompassed approximately 28,000 ha.








Physiography and Geology

Levy County is part of the Coastal Plain Province as

defined by Fenneman (1938) and Thornbury (1965). The

Coastal Plain Province is characterized by sandy to clayey

unconsolidated marine and fluvial sediments. Vernon (1951)

subdivided Levy County into two major physiographic areas;

the Terraced Coastal Lowlands and the River Valley Lowlands.

The CLP is part of the Terraced Coastal Lowlands (Fig. 1-2).

The Terraced Coastal Lowlands include four coastwise

terraces; (a) the Coharie at 66 m, (b) the Okefenokee at

45 m, (c) the Wicomico at 30 m, and (d) the Pamlico at

7.5 m (Fig. 1-3).

The CLP is a broad, relatively flat, surface of

unconsolidated marine sediments overlying the Eocene-age

Ocala Group limestones. The CLP was described by Vernon

(1951) to be a submarine limestone shelf formed as part of

the Wicomico shoreline. However, evidence from Alachua,

Gilchrist, and Levy Counties suggested that the CLP should

be considered part of the Penholoway shoreline (R.C.

Lindquist, 1987, personal communication). The Wicomico and

Penholoway terraces were both correlated with the Sangamon

interglacial period (Cooke, 1939 and Vernon, 1951).

The type of karst on the CLP is normally referred to as

"bare or thinly covered limestone", Fig. 1-4, (Sinclair and

Stewart, 1985). They defined bare or thinly covered







28

limestone as being covered by sediments ranging in thickness

from less than 0.3 to about 8 m.

Karst features are common on the CLP. Bowl-shaped

depressions and small hills create the dominant landforms.

Most of the bowl-shaped depressions are classified as

solution sinkholes (Sinclair and Stewart, 1985). Solution

sinkholes are formed when lateral variations in solution are

minimum. The depressions caused by the sinkholes vary in

diameter from less than 10 m to more than 100 m. Slopes

into these depressions ranged from 3 to 12 percent. Slopes

of the small hills ranged from 4 to 6 percent. Minor

features of the CLP are vertically and horizontally oriented

solution channels, limestone outcrops, and drowned caves.

Undifferentiated Miocene-, Pliocene-, Pleistocene-, and

Holocene-age sands and clays blanket the Ocala Group

limestones.

Initial Soils Legend

A preliminary field study was conducted to evaluate the

soil and landscape relationships on the CLP. Soil surveys

from Alachua (Thomas et al., 1985) and Marion (Thomas et

al., 1979) Counties, which have similar karst areas, were

used to provide initial soil and landscape data.

Reconnaissance soil mapping across the CLP was used to

gather soil and landscape information. The minimum-size

delineation for the survey was established at 2-ha based on

the 1:24000 mapping scale.








GPR System

The GPR system used was the Subsurface Interface Radar

(SIR) System-8 (The trade name has been used to provide

specific information. Its mention does not constitute

endorsement). The 120-MHz antenna was towed at a speed of

4.0 to 5.5 km/hr. The scanning time of the GPR was 70 ns,

with a scanning rate of 25.6 scans/s.

Flex-Grid Mapping and GPR Transect Information

A flex-grid system of mapping was developed where grid-

interval varied according to soil complexity. The grid-

interval of the flex-grid survey was not fixed but was

allowed to vary according to the complexity of the map unit.

Observation intervals within the flex-grid, ranged from 30

to 200 m.

The flex-grid method was devised after the traditional

grid-method (fixed interval) was unsuccessful in correctly

identifying the complexity of soil map units. Ground-

penetrating radar data confirmed the failings of the

traditional grid-method to identify map unit complexity.

Traditional grid-methods would have been successful with the

use of GPR transects, but because the grid-interval was

fixed, it was more laborious than the flex-grid system. The

primary goal of the flex-grid survey was to locate map unit

boundaries. The GPR was used to determine the periodicity

of the soils or composition. Some soil scientists (Wilding

and Drees, 1983) have expressed concern that grid surveys







30

may not properly delineate map units due the coincidence of

the grid interval with the periodicity of the soils. Since

the grid interval of the flex-grid survey was not constant,

errors associated with soil periodicity were reduced. The

flex-grid system was used to locate areas of modal soil

components and the GPR was used to determine the complexity

of the soils within these areas.

The GPR graphic profiles were used to check map unit

boundaries, determine the linear regularity of diagnostic

subsurface features, and determine map unit composition.

Criteria to separate map units were soil drainage, depth to

Bt horizon, and depth to limestone. These soil

characteristics were determined to have a major influence on

the soil use and interpretations for the survey area.

Each GPR transect was 305 m long; observational markers

were placed at 10 equally-spaced points. One to three soil

borings were made along each transect in order to collect

field notes (i.e., diagnostic horizons, colors, textures,

drainage, and depth to limestone) needed to scale the radar

imagery. The difference between the scaled radar images and

actual depths to subsurface features as determined by soil

borings ranged from 3 to 20 cm. This range is in agreement

with findings by Johnson et el. (1980), Shih et al. (1985),

and Zobeck et al. (1985).








Justification and Definition of Intriplex Map Unit

A map unit is "a collection of areas defined and named

the same in terms of their soil components or miscellaneous

areas or both" (Soil Survey Staff, 1981). Map units are

designed to meet certain taxonomic and interpretative

criteria. Kinds of map units (i.e., consociation, complex,

association, and undifferentiated groups) used in a soil

survey depends on the soil patterns and the purpose and

proposed intensity of the survey. The mapping scale of the

soil survey is a primary factor in determining the kind of

map unit selected.

Properly designed map units convey characteristics

about the soils and landscapes to the soil survey user.

Soil areas delineated as a complex map unit indicate that

two or more soil components occur in a regularly repeating

pattern so intricate that the soil components cannot be

separated at the selected scale of mapping (Soil Survey

Staff, 1983). An undifferentiated group is used when two or

more soil components do not occur in a regularly repeating

pattern, but because of some common limitation (i.e.,

steepness, stoniness, or flooding) which limits use and

management, the soils are combined (Soil Survey Staff,

1983). The major soil components of an undifferentiated

group are generally large enough to be separated at the

scale of mapping. However, when soils and landscapes do not

occur in a regularly repeating pattern and delineations are







32

not limited by a common feature, a gap exists in the kind of

map units available for interpreting soil surveys. To refer

to such soil areas as complexes, no doubt conveys an idea of

"intricacy" of soils to the layperson, but to a soil

scientist it probably conveys the concepts defined in the

National Soils Handbook (Soil Survey Staff, 1983).

In the karst regions of west-central Florida, another

kind of map unit is purposed to bridge the gap between

complexes and undifferentiated groups. The new kind of map

unit is termed "Intriplex", a combination of the word

"intricate" and "complex".

Intriplex map units are similar to complexes except for

the requirement that the soil and landscapes occur in a

regular repeating pattern. Soil intriplexes are map units

that consist of areas of two or more kinds of soils (taxa)

or miscellaneous areas that are so intricately associated

that they appear to occur at random. Individual components

would be difficult to delineate even with intensive onsite

investigations. The proportions of the major components

vary from one delineation to another but each occupies a

significant part of the delineation. Grid mapping is

normally required to delineate components of intriplexes.

Transects are conducted within selected delineations to

determine the composition of component soils. No one soil

dissimilar to the named components is to exceed 10% of the

whole and the aggregate of dissimilar soils can not be more








than 25%. Intriplexes are usually named using soil series

names, although other taxonomic class names and names of

miscellaneous areas can be used. Interpretations for some

uses may be made for the intriplex as a whole, determined by

the overriding limitation of any one component. For some

other land uses, each component of the intriplex is

interpreted separately. This definition is based on and

modified from the definition of a complex in the National

Soils Handbook (Soil Survey Staff, 1983).

Statistical Analysis

The mean was calculated for the component soils and the

confidence interval and coefficient of variation were

calculated for each map unit according to the method

outlined by Arnold (1979, 1980). The data were not tested

for normality due to the assumption that a binomial

distribution tends toward normality as the sample size is

increased (Snedecor and Cochran, 1967). The calculated

statistics are based on a binomial distribution, where the

named and similar soils make up one population and the

dissimilar soils make up another population. Bigler and

Liudahl (1984) tested this method of determining map unit

composition in North Dakota and concluded that it was rapid

and reliable. The method was also used in Florida by

Schellentrager et al., (1988).









Results and Discussion

Initial Soils LeQend

The initial soils legend which was established based on

soil surveys from neighboring counties and reconnaissance

mapping within the CLP is presented in Table 2-1. The

classification of these soils is also presented in

Table 2-1. The Arrendondo, Cadillac, and Millhopper soils

are defined as having A and E horizons greater than 1 m in

combined thickness. Arrendondo and Millhopper soils are

both Ultisols that normally occur in non-karst landscapes.

Arrendondo soils are well drained and have coated (Soil

Management Support Services, 1987) sands between a depth of

25 and 100 cm. Millhopper soils are moderately-well drained

and are uncoated (Soil Management Support Services, 1987).

Cadillac soils are well drained and are formed on karst

landscapes. Cadillac soils are Alfisols with 7.5YR colors

throughout the Bt horizon. Arrendondo, Cadillac, and

Millhopper soils are mapped within the same landscapes in

Alachua and Marion counties.

Candler and Tavares soils both have sandy profiles and

are uncoated. Candler soils are excessively-well drained

and have lamella normally between a depth of 150 to 200 cm.

Tavares soils are moderately-well drained and lack lamella.

Hague, Jonesville, and Pedro soils have Bt horizons at

depths less than 100 cm and are well drained. Hague soils








Table 2-1. Classification of series used in the initial
soils legend for the Chiefland Limestone Plain.

Series Family Classification

Arrendondo Loamy, siliceous, hyperthermic
Grossarenic Paleudults

Cadillac Loamy, siliceous, hyperthermic
Grossarenic Paleudalfs

Candler Hyperthermic, uncoated Typic
Quartzipsamments

Hague Loamy, siliceous, hyperthermic
Arenic Hapludalfs

Jonesville Loamy, siliceous, hyperthermic
Arenic Hapludalfs

Millhopper Loamy, siliceous, hyperthermic
Grossarenic Paleudults

Pedro Fine-loamy, siliceous, hyperthermic
shallow Typic Hapludalfs

Tavares Hyperthermic, uncoated Typic
Quartzipsamments







36
are defined as having a Bt horizon between depths of 50 and

100 cm, hue of 7.5YR somewhere within the Bt horizon, and a

decrease in clay content with depth. Hague soils occur in

both karst and non-karst landscapes. Jonesville soils have

a Bt horizon between 50 and 100 cm overlying limestone. The

Pedro series was established as having a cyclic,

discontinuous Bt horizon over limestone that occurred at

depths less than 50 cm. Hague soils are mapped normally as

consociations. Generally, Jonesville and Pedro soils are

mapped as complexes.

Soil-Landscape Model

A soil-landscape model for mapping the CLP was

conceptualized based on the preliminary field study using

the initial soils legend (Fig. 2-1). The Typic

Quartzipsamments (Candler and Tavares soils) were

conceptualized to occur predominately on elevated positions

where eolian sands had accumulated and in depressional areas

where sands had collected due to subsurface collapse. The

Grossarenic Paleudults and Paleudalfs (Arrendondo, Cadillac,

and Millhopper soils) occurred on the broad level areas

where slopes ranged from 1 to 5 percent. The matrix of the

CLP was composed of this type of terrain. The Paleudults

developed in leached parent materials which were not

influenced by the deeper underlying limestone. The

Hapludalfs (Hague, Jonesville, and Pedro soils) occurred

























D A and E Horizons
Bt Horizons
Limestone


Figure 2-1. Conceptual soil-landscape model for the
Chiefland Limestone Plain based on the initial
soils legend. (l=Typic Quartzipsamments;
2=Typic, Lithic, and Arenic Hapludalfs and
Lithic Quartzipsamments; and 3=Arenic
Hapludalfs).







38

where limestone was nearer to the surface, i.e., such as the

rim areas around sinkholes.

The initial mapping model separated the Typic

Quartzipsamments, Grossarenic Paleudults, and Grossarenic

Paleudalfs as consociations (Soil Survey Staff, 1983) and

the Arenic and Typic Hapludalfs as complexes (Soil Survey

Staff, 1983).

Field mapping was initiated based on the initial soils

legend and the conceptual soil-landscape model. An

immediate conclusion was that the soils and landscapes of

the CLP were not occurring systematically. The landscapes

could be delineated but the soils could not be predicted

based on landscape position. Therefore, a grid mapping

procedure was used to provide additional data in order to

delineate the boundaries between soil map units across

landscapes.

Grid Mapping

A soil map based on the initial soils legend and grid

mapping is presented in Fig. 2-2b. Elevations in this

selected 260-ha study area (section 21, R15E, T11S, Trenton

USGS quadrangle) ranged from 12 to 20 m (Fig. 2-2a). One

hundred and one borings were made in this study area and map

unit boundaries were delineated based on the grid borings.

A summary of the grid data is provided in Table 2-2.

Cadillac variant was the dominate soil in map unit 12

according to the grid data. The Cadillac soils were














16




a.








12


ci 15

/ 12

14 15
b.

0 .6
i------i~
km
Figure 2-2. Soil map, ground-penetrating radar (GPR)
transect locations, and topographic map (based
on USGS topoquad) of the 260-ha study area on
the Chiefland Limestone Plain.
a) Topographic map of the 260-ha study area
based on the USGS topoquad (1.5 m contour).
b) Soil map (12-Cadillac variant soils;
14-Hague variant and Typic Hapludalfs; and
15-Candler soils) and GPR transect
locations indicated by letters.









Table 2-2. Summary of grid data for 260-ha area on the
Chiefland Limestone Plain.

Dominant soils and similar soils

Map Cadillac Hague Variant and
Unit Variant Candler Typic Hapludalfs

-------------no. of observations--------------
12 54 6 0
*(90) (10)

14 1 0 14
(7) (93)

15 3 23 0
(12) (88)


*Percent of named plus similar soils.






41

determined to be variants due to drainage and color. Depth

to limestone in this unit was generally below 200 cm.

Arrendondo and Millhopper soils were not identified in map

unit 12. Cadillac Variant soils occurred on all landscape

positions but were generally located on the nearly level

landscapes.

Map unit 14 was mapped as a complex of Hague Variant

and unnamed Typic Hapludalfs. Hague was a variant due to

high clay contents in the Bt horizon and the occurrence of

limestone within 200 cm. The Typic Hapludalfs had

characteristics similar to the Hague Variant except that the

Bt horizons were within 50 cm of the soil surface. The

Typic Hapludalfs and Hague Variant soils occurred adjacent

and in depressions on the landscape and also on the nearly

level areas where map unit 12 normally occurred. Jonesville

and Pedro soils were of minor extend in this map unit.

Candler and similar soils made up 88 percent of map

unit 15 and were represented as a consociation based on the

grid data. The map unit delineated as Candler soils

contained most of the isolated "hills" that occurred on the

nearly level topography (Figs. 2-2a and 2-2b). These hills

could have been the product of eolian processes. Tavares

soil did not occur in the study area.

The conceptual soil-landscape model was correct in some

cases but it could not be applied in all cases. One reason

the model could not be applied directly might be due to the







42

nearly level topography and variable subsurface features of

the CLP.

GPR Analysis

As a final check, 5 GPR transects were conducted within

the map units that had been delineated in the 260-ha study

area. Ground-penetrating radar transect locations are given

in Fig. 2-2b. In these areas, GPR investigations

illustrated that the subsurface variability of the Bt

horizons was much greater than had been observed by the grid

mapping procedure (Fig. 2-3). Areas where the proposed

consociations were expected to occur were complexes.

Transects A, B and C were conducted within the proposed

Candler map unit (Table 2-3). Transect A was conducted

across the footslope, sideslope, shoulder slope, and summit

of a small hill with elevations ranging from 16 to 18 m as

determined from the USGS topographic map. Transect B was

conducted parallel to the summit of a narrow hill at an

approximate elevation of 20 m. Transect C was across a

nearly level area of the Candler map unit. In contrast to

the grid results, the average of the three transects was 50

percent Cadillac Variant and similar soils and 47 percent

Candler and similar soils.

Transect D was conducted across a depressional area

within the Hague map unit. The results were 50 percent

Cadillac Variant and similar soils and 50 percent Hague

Variant, Typic Hapludalfs, and similar soils. According to
























r.0


4)
4-)0


44 -H
0*r-4 V.
r4 -rI
04 -H r0

(d 4-4
.1-i I-I..

u 9)
-A



0
04io





9 0
-H 0 UM
04-J 4
$.4 -.4
WCU)



04j 0





0r4 0
O) -'4












C14






















- ji









Table 2-3. Summary of ground-penetrating radar transect
data for the 260-ha selected study area on the
Chiefland Limestone Plain.

Dominant soils and similar soils

Hague Variant
# GPR Cadillac and
Transect Variant Candler Typic Hapludalfs

--------------- %----------------__-
A 60 40 0

B 30 70 0

C 60 30 10

D 50 0 50

E 50 50 0


# Ten observations stations per transect.







46

the transect information, a Grossarenic soil component would

need to be added to map unit 14.

Transect E was conducted across a depressional area

within the Cadillac map unit. Fifty percent of the transect

was Cadillac Variant and similar soils and 50 percent was

Candler and similar soils.

Based on the data presented for the 260-ha area plus

GPR transect information from other areas, map units 15 and

12 were combined due to the inconsistency with which they

could be separated. Another interesting result was the

comparison of the grid data for the 260-ha area and the GPR

transect data (Table 2-4). The grid data supported map

units 12 and 15 as consociations and map unit 14 as a

complex. But the GPR data suggested that map units 12 and

15 be combined and map unit 14 remain a complex but with an

additional Grossarenic soil component. The grid results

were based on 101 observations and the GPR results were

based on 50 observations for this area. The average grid

density was 1 observation for every 2.4-ha. Data from the

GPR transects were taken every 30.5 m along a 305 m

transect.

The grid data covered the entire area but spacing

probably was too wide to accurately reflect subsoil

variations. Three additional 260-ha grid plots were sampled

for grid analysis. A summary of the grid data for the three

plots plus the previous plot is given in Table 2-5, (map








Table 2-4. Comparison of grid data and ground-penetrating
radar transect data for the selected 260-ha
study area on the Chiefland Limestone Plain.

Dominant soils and similar soils

Map Cadillac Hague Variant and
Unit Variant Candler Typic Hapludalfs

Grid GPR Grid GPR Grid GPR

------------no. of observations-------------

12 54 5 6 5 0 0
*(90) (50) (10) (50)

14 1 5 0 0 14 5
(7) (50) (93) (50)
15 3 15 23 14 0 1
(12) (50) (88) (47) (3)


Total 58 25 29 19 14 6
(57) (50) (29) (38) (14) (12)

* Percent of named plus similar soils.









Table 2-5. Summary of grid data from the four 260-ha areas
located on the Chiefland Limestone Plain.

Dominant soils and similar soils

# Map Cadillac Hague Variant and
Unit Variant Candler Typic Hapludalfs

-----------no. of observations------------

12 184 77 5
*(68) (28) (4)
14 20 2 92
(17) (2) (81)

# Map units 12 and 15 where combined based on GPR transects.
* Percent of named plus similar soils.








units 12 and 15 were combined based on earlier results).

After summarizing the data from the four plots, the

percentage of Cadillac Variant decreased from 90 percent for

the initial plot to 68 percent for all 4 plots. Candler and

similar soils increased from 10 percent to 28 percent for

map unit 12. Arenic and Typic Hapludalfs increased from 50

to 81 percent with Grossarenic soils decreasing from 50 to

17 percent for map unit 14. Map unit 12 would now be a

complex based on grid data. The grid data for the 4 plots,

which covered over 1040 ha, were similar to the data values

determined using GPR (Table 2-8). Even though 61 GPR

transects were conducted throughout the CLP, the combined

area investigated by the GPR was 2-ha.

Final Flex-Grid and Transect Procedure

On the basis of GPR interpretations, the final

procedure combined the flex-grid mapping system and GPR

transects. The flex-grid was used to delineate the map unit

boundaries and the GPR was used to verify the map unit

components. The grid interval varied according to soil

complexity as determined by the initial GPR investigation.

In areas where the depth to the Bt horizon was greater than

1 m, the flex-grid interval was approximately 200 m, but in

areas where the depth to the Bt horizon was less than 1 m,

the flex-grid interval varied from 30 to 100 m.

New soil series were proposed to replace the variants.

Classification of the soils identified in the final survey









is given in Table 2-6. Brief field descriptions of each

soil are presented in Table 2-7.

Sixty-one GPR transects were conducted across the CLP

during the flex-grid mapping procedure. Transects were

located within map units as randomly as possible. Landowner

permission and radar accessibility were the major limiting

factors in locating transects.

The radar images of diagnostic soil horizons were

traced laterally on the GPR graphic profiles for each

transect. The percent of each soil, as represented on the

GPR graphic profile, could not be estimated by measuring

it's length on the graphic profile due to the variations in

the horizontal speed of the radar. Each observation point

on the GPR profile, which was related to known points on the

ground, was assigned a soil series name. For example, if

the Bt horizon was within 1 m of the surface and limestone

was > 1.0 m deep, Shadeville fine sand was noted for that

observation point. Soils with similar interpretative

characteristics were grouped for each transect (Soil Survey

Staff, 1983). Similar and dissimilar soils were

distinguished by drainage and depths to Bt horizons and

limestone. Similar soils normally differed from named soils

due to slight color or texture variations which were

outside the range of the named soil series. The map units

were thus defined based on the interpretative

characteristics of the soil population (Edmonds et al.,








Table 2-6. Series and classification of soils used in
the final descriptive legend for the Chiefland
Limestone Plain.

Series Family Classification

Candler Hyperthermic, uncoated Typic
Quartzipsamments

Shadeville Loamy, siliceous, thermic Arenic Hapludalfs


Jonesville Loamy, siliceous, hyperthermic Arenic
Hapludalfs

Levyville Fine-loamy, siliceous, thermic Arenic
Hapludalfs

Otela Sandy, siliceous, thermic Arenic
Hapludalfs +

Tavares Hyperthermic, uncoated Typic
Quartzipsamments

Seaboard Hyperthermic, uncoated Lithic
Quartzipsamments

* These series have been proposed.
+ Classification of soil would be Grossarenic if allowed by
Soil Taxonomy (Soil Survey Staff, 1975).









Table 2-7. Abbreviated field descriptions of soils on the
Chiefland Limestone Plain.

Horizon Depth Color Texture Structure+ Consistence+
(cm) (moist) (moist)

------------------------tela-------------------------


0-20
20-54
54-81
81-127
127-155
155-173

173-203


10YR 4/2 fs
10YR 5/3 fs
10YR 8/3 fs
10YR 8/1 fs
10YR 6/6 fsl
10YR 6/6 scl
*c2d 10YR 7/2
10YR 7/2 scl


Ifgr
sgl
sgl
sgl
2msbk
2fsbk

massive


---------------------------- Candler-------------------------


0-35
35-75
75-152
152-200


10YR
10YR
10YR
10YR
10YR


5/2
6/3
7/3
8/2
5/6


fs
fs
fs
fs (E)
Ifs (Bt)


sgl
sgl
sgl
sgl


----------------------------Tavares-------------------------


0-18 10YR 3/2
18-61 10YR 5/3
61-104 10YR 5/3
*cld 10YR 5/8
104-147 10YR 6/3
*c2d 10YR 5/8
147-200 10YR 8/2
*fld 10YR 6/6


sgl
sgl
sgl

sgl

sgl


--------------------------Jonesville------------------------


0-13
13-35
35-68
68-88


10YR
10YR
10YR
10YR
10YR


5/1
6/2
7/1
6/6
8/1


scl
limestone


sgl
sgl
sgl
Ifsbk


-------------------------Levyville------------------------


0-18
18-43
43-79
79-112
112-203


10YR 3/3
10YR 6/6
7.5YR 5/8
10YR 5/8
10YR 6/4


fs
s
fsl
fsl
fs


Ifgr
sgl
Imsbk
Imsbk
sgl


vfr

fr
fr
fr


Ap
AE
El
E2
Btl
Bt2

2Cg


Ap
El
E2
E/Bt


Ap
E
Btl
Bt2
C








Table 2-7--Continued.


Depth Color
(cm) (moist)


Texture


Structure+ Consistence+
(moist)


--------------------------Shadeville-----------------------


0-28
28-68
68-89
89-152
152-162


10YR 3/2
10YR 4/6
10YR 7/6
7.5YR 5/8
10YR 7/2
10YR 8/1


s
fs
fs
fsl
fsl


Ifgr
sgl
sgl
2msbk
massive


vfr


fr
fr


----------------------------Seaboard-------------------------


0-18 10YR 4/2
18-45 10YR 8/3
10YR 8/1


sgl
sgl


+ l=weak; 2=moderate; f=fine; m=medium; sgl=single grain
loose; gr=granular; sbk=subangular blocky; vfr=very
friable; fr=friable. Coded according to Appendix I, Soil
Taxonomy (Soil Survey Staff, 1975).
* Mottles.


Horizon


Ap
El
E2
Bt
2Cg
3R









1982; Edmonds et al., 1985a, 1985b; Edmonds and Lentner,

1986) rather than strict adherence to taxonomic

classification.

Intriplex Map Units

As the soil survey of the CLP progressed, complexes

were found to be inadequate in describing soil and landscape

patterns. Complexes inferred a certain knowledge about soil

and landscape patterns. Variations in depths of Bt and R

horizons were not predictable based on interpretation of the

present landscape surface on the CLP. Therefore, to

delineate map units as complexes would convey more knowledge

about the soils than is presently known. Soil scientists

should not only be concerned with the accuracy of soil

boundaries, but also with the design and validity of the map

units as measured against the standards established for

nomenclature. The "intriplex" map unit was defined to more

accurately convey a concept about soil and landscape

patterns.

The result of the flex-grid survey and GPR procedure

was the identification of four map unit intriplexes (Table

2-8) within the CLP. Representative GPR graphic profiles

for the map unit intriplexes are given in Figs. 2-4 to 2-6.

A confidence level of 95% was used for the Otela-

Candler, the Otela-Tavares, and the Shadeville-Otela-

Levyville intriplexes. Confidence level for the Jonesville-

Otela-Seaboard intriplex was at the 80% level. This latter








Table 2-8.


Average composition of map units as determined
by the ground-penetrating radar transect method.


Map Unit
Name
(intriplex)


Otela-
Candler


Soil


Named ++
and
Similar
(%)


No. of *
Transects


Otela
(Bt>lm)


Candler
lamellaee)


Otela
(Bt>lm)

Tavares
(MWD) +


Shadeville-
Otela-
Levyville


Shadeville
(Bt (R>1.5m)


Otela
(Bt>lm)


Jonesville-
Otela-
Seaboard


Levyville
(Bt<0.5m)

Jonesville
(Bt (R

88-76


80 15


Otela
(Bt>lm)


Seaboard 21
(R<0.5m)

++ "Dissimilar soils make up the remaining percentage to 100.
+ Moderately well drained.
* Each transect had 10 equally spaced observation points.
# C.I. is confidence interval. C.L. is confidence level. C.V. is
coefficient of variation.


#
C.I.
(%)


C.L.
(%)


Otela-
Tavares


#
C.V.
(%)


97-84


99-78


95-76







56

intriplex only had 8 transects, which accounts, in part, for

its lower confidence level.

Otela-Candler and Otela-Tavares intriplexes

The Otela-Candler intriplex, which is illustrated in

Fig. 2-4, was estimated to be 57% Otela and similar soils

and 33% Candler and similar soils. Twenty-one GPR transects

were conducted in this unit. The confidence interval for

the named plus similar soils was 84 to 97% (Table 2-8).

Otela-Tavares intriplex was similar to the Otela-

Candler intriplex except that the Tavares soil is

moderately-well drained. Candler soils are excessively

drained. The Otela series plus similar soils were 47% of

the mapped area and the Tavares series and similar soils

were 42%. The named plus similar soils made up 78 to 99% of

the Otela-Tavares map unit.

Generally, the observational density of the flex-grid

for these map units was 1 grid intersect per 3-ha. The

Otela-Tavares intriplex occurred between the 3.0 to 7.5 m

elevation range while the Otela-Candler intriplex occurred

from the 7.5 to 15 m range. The subsurface features of

these soils were highly irregular but the topography of the

surface was nearly level.

Shadeville-Otela-Levyville intriplex

Shadeville-Otela-Levyville intriplex was composed

mostly of well-drained Hapludalfs which varied in depths to

the Bt horizon and to limestone (Fig. 2-5). The Shadeville

























(4.4
0

1-4
*ri










-4
0CM

0 -0






0 C
4-4
(0



C4
.X


0 *H




CV












*a
^ X!











*i-l








58



















* aI





am I




Id




* 1






-I


"U
q44









0 ~CM

























-40


4)x
-r4-

o -dq

134 4J








w>

0-I




4)1





r. r) "4
Ow










S4J :
C14
UI'











0
3.4
frA







60

















I s
---

























. -- -.C'








series and similar soils made up 43% of this map unit. In

Fig. 2-5, the area where the Bt horizon was greater than 1 m

below the soil surface was identified as a collapsed feature

and was mapped as the Otela series. The Otela series

accounts for 26% of this map unit. The Levyville series and

similar Typic Hapludalfs accounted for 17% of the unit. The

flex-grid density within this unit was 1 to 3 grid

intersects per ha. More grid intersects were needed to

delineate the boundaries of map unit 14 than for the deeper

sola map units.

In some areas, the Bt horizons were well developed and

produced a broad, dark, band on the GPR profile (Fig. 2-5).

However, the interfaces between the Bt and R horizons were

not as easily distinguished due to the similar dielectric

properties of the Bt horizon and the limestone. Limestone

was normally detected by the presence of hyperbolic images

on the radar graphic profiles. Generally, cavities are

represented as hyperbolic images on GPR graphic profiles

(Kuhns, 1982; Ballard, 1983). The cavities in Fig. 2-5, are

adjacent to the word "limestone". Cavities in the limestone

are normally filled with air and/or water. The dielectric

properties of air and water contrast strongly with that of

limestone (Ulriksen, 1982) and cause the radar waves to be

reflected strongly.









Jonesville-Otela-Seaboard intriplex

The Jonesville-Otela-Seaboard intriplex consisted of

dominantly moderately deep (R horizon < 1 m) and shallow (R

horizon < 0.5 m) well-drained soils (Fig. 2-6). The

Jonesville series and similar soils made up 38% of this

unit. Ground-penetrating radar signals associated with the

Seaboard series (R horizon <0.5 m) were in the central

portion of Fig. 2-6. The large pinnacle of limestone, in

the center of Fig. 2-6, has an irregular pitted surface

which accounts for the intricate soil patterns in this map

unit. The Otela soils are on the left side of the central

limestone feature. The solution features on each side of

the limestone pinnacle were believed to have formed along

fracture planes in the limestone. These solution features

were not well expressed at the surface. The subsurface

variability of the Jonesville-Otela-Seaboard intriplex was

easily observed where the soil cover had been removed

exposing the underlying bedrock (Fig. 2-7). Thin,

discontinuous Bt horizons mantled the limestone in some

areas. The Seaboard soils and similar soils occurred where

the Bt horizon was absent. The flex-grid density for this

map unit varied from 1 to 3 observations per ha.


Conclusions


A survey method for mapping soils on karst using flex-

grid techniques and GPR proved to be an effective




























0 (
ola
x
Ir4X
06- 04
cf-4


M4.H






*r4 p
4J







00


M -4

.0

4 0

11 4) i
ro rn-
0

.-I
.rq





I C
10 0
4-)
0 4 -4
4 r




Ql4
0L
tp
.r4

P4Qc









1
U
I
a
I
E
I
U
I
U
I
E

I
5
I
I
U
I

a
I
U
I


E
I
U
I
I
3
I
!
3
U



























Figure 2-7. Exposed limestone bedrock within the
Jonesville-Otela-Seaboard intriplex map unit
illustrating the subsurface variability of the
limestone.



























.*I' A

L

r. *0 -

a .--O








alternative where soils and landscapes were independent of

each other. In the karst areas of west-central Florida,

where subsurface soil features often change rapidly and

dramatically over short horizontal distances with no surface

expression, the GPR was ideal for studying the magnitude,

extent, and variability of this change. The radar provided

a continuous subsurface profile where landscape features

could not be used to separate soils. The GPR also allowed a

larger sample of the soil population to be examined, which

increases the accuracy of the map unit descriptions in the

soil survey report. The data presented in Table 2-8 will be

published in the "Soil Survey Report of Levy County,

Florida".

The intriplex map unit was also introduced as an

alternative to the complex map unit in areas where soils and

landscapes appeared to occur randomly. The intriplex map

unit was proposed to more accurately reflect the state of

soils and landscapes on karst in west-central Florida.













CHAPTER 3
RADAR-GRID CATEGORIES OF SUBSURFACE FEATURES
ON THE CHIEFLAND LIMESTONE PLAIN


Introduction


The soils and landscapes of the CLP appeared to vary

independently at a scale of 1:24000 (Chapter 2). The soils

on the CLP have been strongly influenced by sea level

fluctuations and karst processes. Identification of

subsurface features on the CLP and mapping their lateral and

vertical variability by normal soil survey procedures were

not effective.

The ground-penetrating radar (GPR) has been used for

determining soil components in map units (Johnson et al.,

1980; Doolittle, 1982; Collins et al., 1986; Schellentrager

et al., 1988), determining soil microvariability (Collins

and Doolittle, 1987; Collins et al., 1989; Rebertus et al.,

1989), determining soil thickness (Shih and Doolittle, 1984;

Olson and Doolittle, 1985; Collins et al., 1986), and

detection of subsurface cavities and geologic materials

(Kuhns, 1982; Ballard, 1983; Hearns, 1987). Asmussen et al.

(1986) used the GPR to study the subsurface dimensions of

the Hawthorn Formation and to develop radar signatures for

dominant Coastal Plain soils in south Georgia.








The objectives of this chapter were to (a) develop

relationships between the graphic GPR profiles and the soils

of the CLP, and (b) to qualify the spatial characteristics

of the soils identified using GPR.


Materials and Methods


Location and Selection of Study Area

The 4-ha study area shown in Fig. 3-la and 3-1b,

referred to as the Quincey plot (QP), was located on the

CLP, in Levy County (Fig. 1-1). The QP was selected after

analysis of information obtained from the 61 GPR transects

that had been conducted on the CLP (Chapter 2). Also, field

observations indicated that the QP was composed of soils and

karst features that were representative of the CLP.

Elevations on the QP range from 12 to 17 m above mean sea

level.

GPR System

The GPR is an impulse radar system designed for shallow

subsurface investigations. The GPR generates an

electromagnetic wavefront which is propagated into the

ground through an antenna that is electromagnetically

coupled to the ground. The generated pulses are composed of

a number of frequencies that are distributed around the

central frequency of the antenna (Olson and Doolittle,

1985). When a pulse strikes an interface of differing

electrical properties, a part of the pulse's energy is

























Figure 3-1. Ground views of the Quincey Plot.
a) Looking from the southwest to the
northeast from observation station 0,0.
Wooden stakes indicate locations of
each observation station (interval
between stakes was 10 m).
b) Looking from the east to the west-
northwest from observation station 18,7.












































'. I I I
.1 ,^M
}'r









reflected back to the antenna. The reflected pulses are

processed and displayed on a graphic recorder in various

shades of gray. Strong reflections are displayed by the

graphic recorder as black with intermediate reflections

recorded in shades of gray. Radar pulses that are

attenuated and not returned to the antenna are recorded as

white.

The probing depth of the GPR is largely determined by

the electrical properties of the soil. Soil properties that

influence the electrical conductance of the soil are water

content, concentration of salts in solution, adsorbed ions

on clay particles, and amount and type of clays (Johnson et

al., 1980; Shih and Doolittle, 1984; and Olson and

Doolittle, 1985).

The GPR system used was the Subsurface Interface Radar

(SIR) System-8 (The trade name has been used to provide

specific information. Its mention does not constitute

endorsement). A 120-MHz antenna was used at a speed of 4.0

to 5.5 km/hr. The scanning time of the GPR was 70 ns, with

a scanning rate of 25.6 scans/s.

Grid Survey and GPR Transects

A rectangular grid, using 10 m spacings, was

established on the QP. The grid was 230 m in the north-

south direction and 190 m in the east-west direction (Fig.

3-2). Surface elevations were recorded at each grid



























0)

$4
H

0) -r
to


.00)
4J t
010
~44

1tow
P-4w
4-) 4-)
014) 41
(1)

0 0-H


0 0

4-) (
UV~



44) 0


0) .
r.l

M 0


r-4F

-H 4.) -H
5ww





0) 0.4
0) L

-H r.
V:4








74















: 1


i.






75

interval (observation stations) using a transit. Elevations

were tied to a USGS benchmark located near the QP.

The first transect line (0,0 to 0,22) was transected

north to south using a 120-MHz, 300-MHz and 500-MHz antenna.

This was done in order to determine which antenna or

antennae would be best utilized for the investigation. The

120-MHz antenna was selected due to better resolution

between soil depths of 34 and 285 cm.

Nineteen GPR transects were conducted in a north-south

direction on the QP. Thirty-five soil borings were made in

order to collect field notes (i.e., diagnostic horizons,

colors, textures, drainage and depth to limestone) needed to

scale the radar imagery. Soil data were collected to a

depth of 2 m or to limestone, whichever was shallower. The

GPR graphic profiles were scaled by measuring the distance

from the ground surface pulse to the first interface. The

GPR data were regressed against measured depths at 29 of the

35 observation stations (R-squared=0.94). Six of the

observation stations were excluded due to poor resolution of

the 120-MHz antenna at shallow depths (0.5 m) or lack of

depth measurements below 2 m. The minimum resolution of the

120-MHz antenna was 34 cm. The 34 cm depth was in agreement

with the findings of Rebertus et al. (1989), who reported

that strong surface and near surface reflections made

interpretations of subsurface interfaces less accurate







76

within depths of 25 to 35 cm than at deeper depths using the

120-MHz antenna.

Temporal effects on GPR graphic profiles are not known.

Therefore, the results reported here are relative to the

field conditions the day the transects were conducted. The

GPR transects for the QP were conducted during a period of

drought to minimized the effects of moisture on the GPR

graphic profiles. Perched or apparent water tables produce

strong interfaces on the graphic profiles and, therefore,

confound the process of horizon identification.

Definition of Radar Class and Radar-Grid Category

Radar class

A radar class (RC) is defined as "a unique GPR graphic

profile that is representative of a specific arrangement of

soil horizons or geologic strata". A radar class represents

a subdivision of a GPR graphical profile into its most basic

recognizable graphic component.

Radar-grid category

A radar-grid category (RGC) is defined as "a collection

of similar RCs that have spatial distribution and are

representative of a defined soil component or components".

Soil components were defined in Chapter 2. A RGC is a

grouping of RCs into a useful and meaning spatial

arrangement. The characteristics of a RGC are defined by

the graphic properties of the RC and the range of








characteristics that define the basic soil component or

components.

Selection of Representative Pedons

The 19 GPR graphic profiles were carefully studied and

seven RCs were chosen as unique and representative of all

radar images in the QP. The seven RCs represented seven

different graphical image profiles. After the RCs were

identified and defined, soils were selected and sampled

within each RC to establish qualitative relationships

between the graphical radar profiles of each RC and soil

characteristics.

Laboratory Analysis

The seven selected pedons were described and sampled by

horizon. Bulk soil samples were air dried and sieved to

remove particles larger than 2 mm. A subsample was ground

in a ball mill to pass a 250-um sieve. The subsamples were

used to determine total phosphorus (TP) contents.

Particle-size analysis was performed using the pipette

method (Day, 1965). Total phosphorus content was determined

using the alkaline-oxidation method developed by Dick and

Tabatabai (1977) as modified by Walter (Collins, 1977).

Cation exchange capacity (CEC) was calculated by

summation of the extractable bases (Na, K, Mg and Ca), using

as an extractant 1 M NH4OAc pH 7.0 (Soil Survey Staff,

1984); and extractable acidity, using 0.25 M BaCl2 TEA pH







78
8.2 (Soil Survey Staff, 1984). Electrical conductivity (EC)

measurements were made using a 1:1 ratio of soil to water.

The A and E horizon samples were treated with NaOCl to

remove organic matter in preparation for mineralogical

analysis. Samples were pretreated with Na citrate-

dithionate-bicarbonate to remove oxide coatings (Mehra and

Jackson, 1960). Organic matter and iron were removed due to

the confounding effects of these cementing agents on

dispersion and the proper orientation of clay on the ceramic

tile.

Sand was separated by wet sieving. Silt and clay were

dispersed by adjustment of the suspension pH to 10 with

Na2CO3 and then separated by centrifugation and decantation.

Oriented mounts of clay fractions were prepared by

depositing approximately 250 mg from the suspension onto a

ceramic tile under suction, saturating with Mg or K, washing

free of salts using distilled water, and adding glycerol to

the Mg-saturated samples.

Clay minerals were identified from x-ray diffraction

(XRD) patterns of Mg-saturated clay (air dry and 1100C) and

air dry, 1000, 3000 and 550C heated, K-saturated clay. X-

ray diffraction analysis was conducted using a Nicolet 12

computer-controlled system. Samples were scanned at 20 28

per minute using CuK-alpha radiation. Minerals were

identified from XRD patterns using differentiating criteria

outlined by Whittig and Allardice (1986).








Statistical Methods

The UNIVARIATE procedure (SAS Institute, 1988) was used

to calculate the means (X), medians (Md), modes (Mo),

coefficients of variation (CV), and standard error of the

mean (SE). The NORMAL option was used to compute the

probabilities (p-values) associated with the Shapiro-Wilk W-

statistic for testing for normal distributions.

The nearest-neighbor analysis method developed by Clark

and Evans (1954) was used to determine the pattern

characteristics of the soils in the QP. In this analysis,

the observed mean nearest-neighbor distance for a set of

points was compared to the theoretical mean distance for a

(a) clustered, (b) random, and (c) dispersed arrangement.

The expected mean nearest-neighbor distance for a random

arrangement of points is given by

Dran=l/(2*p) [1]

where Dran is the expected mean nearest-neighbor distance

for a random arrangement and p is the density of points per

unit area (number of observations divided by the area).

The expected mean nearest-neighbor distance for a

dispersed arrangement of points is given by

Ddis=1.07/p [2]

where Ddis is the expected mean nearest-neighbor distance

for a dispersed arrangement. The calculation of Ddis is

based on a hexagonal arrangement of points, which gives the

maximum distance that could exist between points.









The observed mean nearest-neighbor distance for an

arrangement of points is given by

Dobs=d/n [3]

where Dobs is the observed mean nearest-neighbor distance, d

is the sum of the measured nearest-neighbor distances for

the observed points, and n is the number of points.

The test statistic or nearest-neighbor index is given

by

R=Dobs/Dran [4]

where R is the nearest-neighbor index. The value of R can

be compared with standard table values (Ebdon, 1983) to test

its significance. The value of R is equal to 0 if the

arrangement of points is clustered. If the arrangement of

points is random then the value of R is equal to 1. For a

completely dispersed arrangement of points, the value of R

is equal to 2.15.

The hypothesis tested for the QP was

Ho: The observed arrangement is the result of
points being located at random in the QP.
Ha: If R < tested R then points are clustered.
If R > tested R then points are dispersed.

Relationships between the soils and landscapes on the

CLP were not apparent at the scale of 1:24000 (Chapter 2).

The nearest-neighbor analysis was intended to test the

hypothesis that the soils in the QP occurred at random and

exhibited no apparent pattern. Burrough (1983) suggested

that systematic and random variations were entirely scale







81
dependent because increasing the scale of observation almost

always reveals structure in the random variations.

Computer-Generated Maps and Diagrams

Computer-generated contour maps and surface-net

diagrams were created using the SURFER software programs.

The Inverse Distance method with a weighting power of 2 was

used to create the surface-net diagrams.


Results and Discussions


Radar Classes

The seven RCs identified from the GPR graphic profiles

will be identified by letters. Radar class "A" was

identified by the occurrence of two radar interfaces below 1

m. The first radar interface was a continuous broad, black

band and the second a discontinuous grayish-black band (Fig.

3-3a). The wide gray band below the surface pulse was

smooth and continuous which indicated that interference was

not occurring between the surface pulse and the first

subsurface interface. This means that the distance between

the soil surface and the first soil horizon was greater than

one-half of the wavelength of electromagnetic pulse.

Radar class "B" was characterized by a broad,

continuous black band above 1 m (Fig. 3-4a). The hyperbolic

reflections below 1 m and to the right of center were

considered to be associated with the presence of limestone.

Cavities or cracks in the limestone produce characteristic




























H*
1-4


Q4*H m0 P
HI CC

OW Of O *H
U00 0O F
44 ) to -H
r- C o II C
+J 1
0 ) p 0 0)
to ao a
co o') -


S i --Q )
4O k-) n
C *H-O
00 40 0 V1
C*04 0) )
r-,-I o0 -4 -l

4 0 a 4 o
o io
0U) 4C C
Q ( *H 0


0C 1 0 3


0 0 4( m 0
(0 0 0 g
kC ,Q W 4
I H- II r- 0A
u 0) a0p AS a4 (a


0 UA to *4 0






14U




.r,
44











































-UC


" mm"


Alp8































0 4J U)


'-4 -W -)
> ,

-4.,f w 0)



0) 0 -44




m0 0
0) tv W
00- Q -I b





.- o 10 -rq


0~4 ~4) 0

o004.) k 04w
ra 0 1
04Jr- UV
-4 0 $4 CO


00*-4 0
0. 0~4 ofMIQc









J.z4
kdr

















w.

1


P ~---