|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
INVESTIGATING ARCHAEOLOGICAL SITES, CEMETERIES AND SOILS, WITH
GROUND-PENETRATING RADAR INT FLORIDA
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
O 2007 Christopher Chilton
To my family: Jennifer, Richard, Elizabeth, Cynthia, Richard Jr. and Randall, and all my friends
who have helped through the years.
This research could not have been completed without the advice, guidance and support
from numerous individuals. The integration of multiple disciplines required the assistance of
many friends, colleagues and mentors. Chiefly, I would like to express my gratitude for the vital
support, advice and input from my diverse committee. Dr. Mary Collins was essential to my
understanding and interest in soil science. Her passion for teaching helped create a remarkable
environment for learning and critical thinking. She was an inspiration for all of her students and I
will always appreciate the commitment she had in my education. Dr. James Davidson provided
unique insight and knowledge because of his archaeological expertise. His work on African-
American cemeteries allowed for a critical understanding of much of my research. I was most
fortunate to learn from him not only in the classroom, but in the field as well. Dr. John Schultz
provided valuable counseling and support for my research. His technical and practical
understanding of my work was a tremendous help and allowed me to work through difficult
problems. I was most fortunate to have a committee that provided valuable comments and
patience while writing this thesis.
The Department of Soil and Water Science at the University of Florida has fostered a great
environment for cooperative learning. Never have I seen so many professors and students who
are genuinely interested in each others work and who are willing to assist without reservation.
Dr. Willie Harris has been a wonderful mentor and a remarkable scientist who helped me as well
as many other students throughout his great career. My fellow graduate students who not only
shared an office, but much more, assisted me throughout my research. Dr. Larry 'Rex' Ellis,
Thomas Saunders and Kelly Fischler were instrumental in my graduate school development and
eagerly assisted me whenever needed. I would like to recognize those undergraduate students
who assisted me on many occasions: Victoria Gardner, Aja Stoppe and Leanna Woods. I would
also like to recognize Dr. Rocky Cao, Lisa Stanely and Genero Keehn at the Environmental
Pedology and Land Use Laboratory at the University of Florida for their assistance with my soil
I would like to express my gratitude to the archaeological community that guided me
through many years of support and inspiration. Dr. Kathleen Deagan and Al Woods took me
from the basement of the Florida Museum of Natural History to the field. Their support guided
me to the wonderful people of Southeastern Archaeological Research Inc. The confidence and
opportunity afforded me by Dr. Anne Stokes and James Pochurek allowed for my archaeological
development for which I will always be grateful. Dr. Robert Austin, Dr. Elizabeth Carlson and
Dr. Deborah Mullins have not only had a great impact on my archaeological maturation but have
offered assistance and support for my research.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ...............4.....
LIST OF TABLES ............ ...... .__ ...............9....
LIST OF FIGURES .............. ...............12....
AB S TRAC T ............._. .......... ..............._ 15...
1 INTRODUCTION ................. ...............17.......... ......
Tool s and Method ol ogy of GP R. ............_...... .............. 18..
GPR Antennas ............. .. ... ._ .............. 18....
Radar Propagation and Parameters ............. .. ... ...............21......
Properties and Materials Affecting Radar Propagation ......____ ........._ ..............22
History of GPR .............. .... ...._ .... .......__.. ...........2
Archaeological and Cemetery Application of GPR. ........___....... .._ .........___......25
History and Use of GPR Mapping in Archaeology and Cemeteries .............. ....................26
Soils of Florida ............. ...... ...............29...
Soil Taxonomy .............. ...............29....
Diagnostic Surface Horizons ........._..... ...._... ...............30....
Histic Epipedon ........._.._.. ...._... ...............3 1...
M ollic Epipedon ................. ...............32........ ......
Umbric Epipedon............... ...............32
Ochric Epipedon ........._..... ...._... ...............32.....
Diagnostic Subsurface Horizons .............. ...............32....
Albic Horizon .............. ...............33....
Argillic Horizon............... ...............33
Cambic Horizon............... ...............33
Spodic Horizon ........._..... ...._... ...............34.....
Soil Orders in Florida ........._.._.. ...._... ...............34...
A lfisol s .............. ...............34....
Entisols .............. ...............34....
Hi stosol s .............. ...............3 5....
Incepti sol s............._ ........ ...............35.....
M ollis ol s..............__....._.. ...............36..
Spodosol s............... ...............3
Ulti sol s .....__ ................. .........._.........3
Obj ectives of the Study ..............__.......... ...............37...
2 FIELD SITES .............. ...............38....
Study Site Selection .....__ ................. ........._._.........3
Newberry Cemetery ................. ...............38.................
Description .............. ...............38....
S oil s ................. ...............3.. 9..............
Cultural History ................. ...............41.......... .....
Oakland Cemetery Site ................. ...............43........... ....
Description .............. ...............43....
S oil s ................. ...............43.......... .....
Cultural History ................. ...............45.......... .....
Historic Archaeological Site ................. ...............46................
Description .............. ...............46....
S oil s ................. ...............46.......... .....
Cultural History ................. ...............48.......... ......
Prehistoric Archaeological Site .............. ...............49....
Description .............. ...............49....
S oil s ................. ...............50.......... .....
Cultural History ................. ...............51.......... .....
3 MATERIALS AND METHODS .............. ...............53....
Site Criteria............... ...............53
Soil Sampling............... ...............54
Soil Analyses .............. ...............55....
Physical Analysis............... ...............56
Chemical Analysis............... ...............56
Statistical Analysis............... ...............57
GPR Data Collection .............. ...............57....
GPR Analysis................ .... ............5
GPR Post-Processing Analysis............... ...............59
Study Sites .............. ...............60...
Newberry Cemetery............... ...............60
Oakland Cemetery ................. ...............61.......... .....
Historic Archaeological Site............... ...............63..
Prehistoric Archaeological Site .............. ...............64....
4 RE SULT S .............. ...............65....
Newberry Cemetery ................. ...............65.......... ......
S oil Phy si cal Analy si s ................. ...............65................
Soil Chemical Analysis ................ ...............67..
GPR-Soil Results for Newberry Cemetery ................. ...............70...............
The 500-MHz antenna............... ...............70
The 900-MHz antenna............... ...............73
Oakland Cemetery .............. ...............77....
S oil Phy si cal Analy si s ............__..... ___ ...............77...
Soil Chemical Analysis ............... .. ...............78..
GPR-Soil Results for Oakland Cemetery ................. ...............81......__ ...
The 500-MHz antenna............... ...............8 1
The 900-1VHz antenna............... ...............83
Prehistoric Archeological Site .............. ...............84....
S oil Phy si cal Analy si s ............_...... ._ ............... 5....
Chemical Analysis.................... .... ...................8
GPR-Soil Results for the Prehistoric Archaeological Site .............. .....................8
The 500-1VHz antenna............... ...............89
The 900-1VHz antenna............... ...............93
Historic Archeological Site............... ...............95..
Soil Physical Analysis ............_...... ._ ...............95...
Soil Chemical Analysis .............. .............. ................9
GPR-Soil Results for the Historic Archaeological Site ......____ ..... ...__ ..............99
The 500-1VHz antenna............... ...............99
The 900-1VHz antenna............... ...............102
Summary ............ ..... .._ ...............105...
5 DI SCUS SSION ............ ..... ..__ ............... 106..
Newberry Cemetery ................. ...............107......... ......
Oakland Cemetery ................. ...............110......... ......
Historic Archaeological Site ................. ...............113................
Prehistoric Archaeological Site ................. ...............117...............
Ground-Penetrating Radar Comparisons ................. ...............120................
GPR-Cemetery .............. ...............120....
GPR-Archaeological Sites............... ...............121.
Sum mary ................. ...............122......... ......
6 CONCLUSIONS .............. ...............128....
A SOIL DESCRIPTIONS .............. ...............132....
B PARTICLE-SIZE ANALY SI S ............... ...............14
C SOIL CHEMICAL ANALYSES ................. ...............144...............
D GPR PROFILE S ............ ..... .__ ...............149..
E STATISTICS .............. ...............153....
LIST OF REFERENCES ............ ......__ ...............157...
BIOGRAPHICAL SKETCH ............_...... ...............164...
LIST OF TABLES
1-1 Typical relative dielectric permittivities (RDPs) of common materials. ...........................24
1-2 Diagnostic summary of the four epipedons used to classify soils in Florida. ................... .3 1
1-3 Cation exchange capacities (CEC) of soil materials (Birkeland 1999). ................... .........31
2-1 Graves at the Newberry Cemetery. Preliminary Report, Southeastern Archaeological
Research, Inc. Gainesville, Florida. .............. ...............42....
3-1 Data points for the four corners of the cemetery boundary. The proj section system
used was UTM, NAD 1983, Zone 17 North. ......___. .... ... ._ ...............61
3-2 Data points for the four corners of the cemetery boundary. The proj section system
used was UTM, NAD 1983, Zone 17 North. ......___. .... ... ._ ...............63
3-3 Data points for the four corners of the historic archaeological survey area. The
proj section system used was UTM, NAD 1983, Zone 17 North. .............. ................64
4-1 Selected soil chemical analysis of the Newberry Cemetery natural and anthropogenic
soils along transects 14 and 17. The EC (EC), extractable (Ext.),............... ................6
4-2 Selected values of chemical properties of the Oakland Cemetery natural and
anthropogenic soils along transect 2. The EC (EC), extractable (Ext.), ................... .........81
4-3 Selected values of chemical properties from natural and anthropogenic soils at
prehistoric archaeological site. The EC (EC), extractable (Ext.),............... .................8
5-1 Coordinate positions of the Newberry cemetery of selected transects and location of
graves. ................ ...............110..._ __.......
A-1 Official description of Jonesville soil series in Alachua County, Florida. ................... ...132
A-2 Soil description of Jonesville series at the Newberry gravesite natural soils (NGNS)....133
A-3 Official description of Candler soil series in Orange County, Florida. ........................... 134
A-4 Soil description of Candler series at the Oakland gravesite natural soils (OGNS),
Orange County, Florida. ............. ...............135....
A-5 Official description of Orlando soil series in Lake County, Florida. .............. ..... ........._.136
A-6 Soil description of Orlando series at the prehistoric archaeological natural soils
A-7 Official description of Immokalee soil series in St. Johns County, Florida. ................... 138
A-8 Soil description of Immokalee series at the historic archaeological natural soils
(H AN S). .............. ...............139....
B-1 Soil data from the Newberry Cemetery Anthropogenic Soils (NCAS) and the
Newberry Cemetery Natural Soils (NCNS) ................. ...............140........... ...
B-2 Soil data from the Oakland Cemetery Anthropogenic Soils (OCAS) and the Oakland
Cemetery Natural Soils (OCNS)............... ...............141
B-3 Soil data from the Prehistoric Archaeological Anthropogenic Soils (PAAS) and the
Prehistoric Archaeological Natural Soils (PANS) ................. ................ ......... .142
B-4 Soil data from the Historic Archaeological Anthropogenic Soils (HAAS) and the
Historic Archaeological Natural Soils (HANS) ................. .............. ......... .....143
C-1 Soil data for the Newberry Cemetery Anthropogenic Soils (NCAS) from selected
transects. Extractable (Ext.). NCAS Transect 22 15N/21E ........._._..... ....._.._. .......144
C-2 Soil data for the Newberry Cemetery Anthropogenic Soils (NCAS) from selected
transects. Extractable (Ext.). NCAS Transect 14 5N/13E ................. ......................144
C-3 Soil data for the Newberry Cemetery Anthropogenic Soils (NCAS) from selected
transects. Extractable (Ext.). NCAS Transect 10 8N/9E ................. ........___..........144
C-4 Soil data for the Newberry Cemetery Anthropogenic Soils (NCAS) from selected
transects. Extractable (Ext.). NCAS Transect 14 9N/13E ................. ......................145
C-5 Soil data for the Oakland Cemetery Anthropogenic Soils (OCAS) from selected
transects. Extractable (Ext.). OCAS Transect 9 40N/10E ................. ............ .........145
C-6 Soil data for the Oakland Cemetery Anthropogenic Soils (OCAS) from selected
transects. Extractable (Ext.). OCAS Transect 22 40N/54W ................. .....................145
C-7 Soil data for the Oakland Cemetery Anthropogenic Soils (OCAS) from selected
transects. Extractable (Ext.). OCAS Transect 22 14N/52W ................. .....................146
C-8 Soil data for the Prehistoric Archaeological Anthropogenic Soils (PAAS) from
selected transects. Extractable (Ext.). PAAS Transect 2 2N/1E............... .................14
C-9 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from
selected transects. Extractable (Ext.). HAAS Test Unit 17, Transect 1 52N/44E ...........146
C-10 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 17, Transect 1 50N/44E ................... ......146
C-11 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 17, Transect 3 52N/46E ................... ......147
C-12 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 17, Transect 3 51N/46E ................... ......147
C-13 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 17, Transect 3 50N/46E ................... ......147
C-14 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 18, Transect 8 50N/40E ................... ......147
C-15 Soil data for the Historic Archaeological Anthropogenic Soils (HAAS) from selected
transects. Extractable (Ext.). HAAS Test Unit 18, Transect 8 52N/40E ................... ......147
C-16 Soil data for the Historic Archaeological Natural Soils (HANS) from selected
transects. Extractable (Ext.). HANS, Transect 12 48N/30E ................. .....................148
C-17 Soil data for the Historic Archaeological Natural Soils (HANS) from selected
transects. Extractable (Ext.). HANS, Transect 14 40N/20E ........._.._.... ......_.._. ......148
E-1 Results of t-tests for Newberry Cemetery natural soils (NCNS) and Newberry
Cemetery .........._ _.......__ ...............153...
E-2 Results of t-tests for Oakland Cemetery natural soils (OCNS) and Oakland Cemetery .154
E-3 Results of t-tests for Prehistoric archaeological natural soils (PANS) and Prehistoric
archaeological anthropogenic soils (PAAS) samples ................. .......... ...............155
E-4 Results of t-tests for Historic archaeological soils (HANS) and Historic
archaeological anthropogenic soil (HAAS) samples. ............. ...............156....
LIST OF FIGURES
1-1 Scale of relative depth of penetration to resolution for ground-penetrating radar
1-2 A schematic of reflected and scattered electromagnetic waves as they are transmitted
through differing soil properties.. ............ ...............21.....
2-1 Location of research sites in Alachua County, Lake County, St. Johns County, and
Orange County, Florida. ............. ...............39.....
2-2 Soil survey map showing soil series and complex at the Newberry cemetery .. ............... .40
2-3 Soil depth to limerock variability of an Alfisol in Alachua County, Florida. ...................41
2-4 Soil survey map showing the Candler-Apopka soil complex at the Oakland cemetery....44
2-5 Photograph of an Entisol profile in Florida. ............. ...............45.....
2-6 Soil survey map showing soil series and complex at the historic archaeological site.......47
2-7 Photograph of an Immokalee soil series profile. ............. ...............48.....
2-8 Photograph of an excavation at the prehistoric archaeological site along the
southwest shore of Lake Apopka, Montverde, Florida ......... ................. ...............50
2-9 Aerial photograph (1999) of the prehistoric archeological site at Monteverde
showing shovel tests and test unit location (Austin 2006)............... ...............51.
3-1 Aerial photograph of Oakland Cemetery from 1947. Site boundary is highlighted in
red. ............. ...............62.....
4-1 Depth distribution of total clay content of natural transectt 17) and anthropogenic
transectt 14) soils at the Newberry cemetery............... ...............66
4-2 Depth distribution of Transect 14 pH values and Transect 17 pH values at the
Newberry cemetery ................. ...............67.......... ......
4-3 Ground-penetrating radar profiles of the Arenic Hapludalf using a 500-MHz antenna
along transect 17 .............. ...............72....
4-4 Ground-penetrating radar profiles of anthropogenic altered Arenic Hapludalf ................73
4-5 Ground-penetrating radar profile of anthropogenic altered Arenic Hapludalf .................74
4-6 Ground-penetrating radar profile of natural Arenic Hapludalf using the 900-MHz
antenna along transect 17............... ...............75...
4-7 Ground-penetrating radar profile of anthropogenic Arenic Hapludalf using a 900-
MHz antenna along transect 14............... ...............76...
4-8 Depth distribution of pH values for Transect 2 at the Oakland cemetery. ......................79
4-9 Ground-penetrating radar profile using the 500-MHz antenna with two suspect
graves from a Lamellic Quartzipsamment. ................ ....................................83
4-10 Ground-penetrating radar profile using the 900-MHz antenna of two suspect graves
in a Lamellic Quartzipsamment ................. ...............85................
4-11 Depth distribution of pH values for transects at the prehistoric archeological site. ..........87
4-12 Ground-penetrating radar profiles of a natural Psammentic Dsystrudepts using the
500-M Hz antenna. ........... ..... .._ ...............91....
4-13 Ground-penetrating radar profiles of an anthropogenic Psammentic Dsystrudepts
using the 500-MHz antenna. ........... ..... .._ ...............92...
4-15 Ground-penetrating radar profiles of an anthropogenic Psammentic Dsystrudepts
using the 900-MHz antenna. .............. ...............94....
4-16 Depth distribution of selected pH values for transects at historic archaeological site.
Historic site data did not exceed 0.50 m. ............. ...............97.....
4-17 Ground penetrating radar profiles of a natural Arenic Alaquodos using the 500 MHz
antenna.. ............ ...........10
4-18 Ground-penetrating radar profiles of an anthropogenic Arenic Alaquods using the
500-M Hz antenna. ........... ..... .._ ...............102...
4-19 Ground-penetrating radar profiles of a PANS using the 900-MHz antenna. A)
Background filter is applied and antenna noise is removed. ............. .....................0
4-20 Ground-penetrating radar profiles of a historic archaeological anthropogenic soil
using the 900-MHz antenna. ........... ..... .._ ...............104..
5-1 Example of a root ball inversion caused by severe weather in Newberry cemetery
area. ........... ..... ._ ...............109...
5-2 GPR profile at Newberry cemetery of transect 22 using 500-MHz antenna. Grave is
outlined demonstrating how a vertical feature produces scattering. .............. ..............110
5-3 Photograph of excavation unit from the historical archaeological site. Wall footer/
trench is located in southwest corner of unit. ................ .......... ............. .....15
5-4 Ground-penetrating radar profile (GPR) of the prehistoric archaeological natural soil
(PANS) using the 900-MHz antenna. ....__ ......_____ ......___ ............1
5-5 Photograph of soil in abandoned excavation unit from the prehistoric archaeological
site.. ............ ..........11
5-6 Ground-penetrating radar profile using 500-MHz antenna from the Newberry
cemetery anthropogenic soils (NCAS). ............. ...............124....
D-1 Ground-penetrating radar (GPR) transect 22 of the Newberry Cemetery
Anthropogenic Soils (NCAS) using 500-MHz antenna .........__ ...... ...__ ............149
D-2 Ground-penetrating radar (GPR) transect 22 of the Newberry Cemetery
Anthropogenic Soils (NCAS) using 500-MHz antenna .........__ ...... ...__ ............149
D-3 Ground-penetrating radar (GPR) transect 4 of the Oakland Cemetery anthropogenic
Soils (OCAS) using 500-MHz antenna. ....__ ......_____ .......__ ...........15
D-4 Ground-penetrating radar (GPR) transect 4 of the Oakland Cemetery Anthropogenic
Soils (OCAS) using 500-MHz antenna. ....__ ......_____ .......__ ...........15
D-5 Ground-penetrating radar (GPR) transect 4 of the Oakland Cemetery Anthropogenic
Soils (OCAS) using 500-MHz antenna. ....__ ......_____ .......__ ...........15
D-6 Ground-penetrating radar (GPR) transect 2 of the Prehistoric Archaeological
Anthropogenic Soils (PAAS) using 500-MHz antenna...................... ...............15
D-7 Ground-penetrating radar (GPR) transect 2 of the Prehistoric Archaeological
Anthropogenic Soils (PAAS) using 500-MHz antenna .....................151
D-8 Ground-penetrating radar (GPR) transect 2 of the Prehistoric
ArchaeologicalAnthropogenic Soils (PAAS). ............. ...............15 1....
D-9 Ground-penetrating radar (GPR) transect 3 of the Historic
ArchaeologicalAnthropogenic Soils (HAAS) using 500-MHz antenna. .......................152
D-10 Ground-penetrating radar (GPR) transect 3 of the Historic Archaeological
Anthropogenic Soils (HAAS) using 500-MHz antenna. ......._____ .... ..__ ............152
D-11 Ground-penetrating radar (GPR) transect 3 of the Historic Archaeological
Anthropogenic Soils (HAAS) using 500-MHz antenna. ............. ....................15
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
INVESTIGATING ARCHAEOLOGICAL SITES, CEMETERIES AND SOILS WITH
GROUND-PENETRATING RADAR INT FLORIDA
Chair: Mary E. Collins
Major: Soil and Water Science
Ground-penetrating radar (GPR) is an efficient and nondestructive instrument for
collecting information below the ground surface relating to potential archaeological sites and
cemeteries. This research assessed the application of GPR technologies for archaeological
investigations with multiple soil orders of Florida. By differentiating the archaeological anomaly
from natural pedogenic properties, a qualitative interpretation of a site can determined. Four sites
were selected based upon archaeological context and soil properties in Florida. The
anthropogenic sites included two cemeteries; the Newberry cemetery located in Alachua County
and the Old African American Cemetery in Orange County. The one historic archaeological site
is located in St. Johns County and a prehistoric archaeological site is in Lake County. Each site
had a different soil order commonly encountered in Florida (Alfisols, Entisols, Inceptisols and
Spodosols). Natural and anthropogenic soils were surveyed using 900-and 500-MHz antennae.
Soil samples were collected along GPR surveyed transects and analyzed for physical and
chemical properties including texture, pH, and electrical conductivity. Post-processing of the
GPR data was conducted to augment profile interpretation and assess the differences in imagery
of the 900- and 500-MHz antennae.
Results of the GPR and soil analyses revealed that each was dependent upon the
anthropogenic context. Diagnostic horizons had a greater influence than soil order classification
on the successful application of GPR on archaeological sites and cemeteries. Soil physical
properties, specifically texture, had a greater overall impact than soil chemical properties on the
radar propagation in every site. The 500-MHz antenna was preferred over the 900-MHz antenna
for general performance in all four research sites, regardless of soil order. 500-MHz antenna
preference for the GPR was based on depth of radar wave penetration and the general
background noise associated with the higher frequency antenna. Though diagnostic surface
horizons and soil texture influenced GPR interpretation, this study showed that GPR is
applicable in a wide variety of soil and archaeological contexts in Florida.
Ground-penetrating radar (GPR) has become a comprehensive and reliable remote sensing
tool for gathering and processing geophysical information. GPR is a method that provides rapid
and non-invasive identification of subsurface features and materials. Recent technological
innovations, such as real time video display with post processing software on the control unit,
have improved the effectiveness of data acquisition and processing along with increasing the use
and application of this tool. The technological advances in GPR in the past decades have allowed
its use today to become common and accepted procedures in many disciplines.
Many land-use activities have the potential for the application of GPR. Numerous
scientific and commercial fields such as: geology, mining, engineering, construction, agriculture,
environmental science and archaeology are turning to real-time and nondestructive data
acquisition tool. Also, because of the easy and rapid use of this instrument, there has been an
increase in the utilization for surveying, mapping, mitigating and delineating subsurface features.
Soil scientists and geologists have used GPR to map subsurface horizons, delineate karst
features, detect sinkhole formation, and differentiate bedrock units for industrial minerals
(Doolittle and Collins 1998, Collins 1990, Puckett 1990, and Dekeyser 2005). The military is
able to save lives by detecting buried land mines and unexploded ordinance (Hyde 1997).
Engineers have applied this efficient technology to high-speed rail trackbed investigations
associated with high-speed trains (Doshi and Al-Nuaimy 2006). GPR was used in May 2005 for
a search mission after a mountain accident in the Chilean Andes involving the death of 45
military troops (Casassa 2006). The success of GPR in the past has created strategies for the
future; scientists are planning to conduct geophysical surveys on Mars (Olhoeft 1998).
Tools and Methodology of GPR
GPR is a method that is able to produce large amounts of high resolution, multi-
dimensional information through the transmission of electromagnetic waves. This tool actively
produces multiple pulses of energy derived from a control unit. Fiber-optic cables transmit a
digital signal to and from the antennas; the raw data derived from the generated electrical pulses
are received and displayed in a video or printed format (Conyers and Goodman 1997, Davis and
Annan 1989, and Schultz 2003). The transmission of electromagnetic radar pulses travel through
the earth's surface and any of these pulses which are reflected will be received and recorded.
Energy is transmitted and energy is received through antennas; the frequencies of these antennas
range from a bandwidth of about 10 megahertz (1VHz) to 1500 1VHz.
Transmitting and receiving antennas used for GPR are configured in either a monostatic or
bistatic mode. The monostatic mode is where the transmitting and receiving antennas is at a zero-
offset, there is no separation from the point of transmission to the point of reception. This allows
for a self-contained unit with less variability. On the contrary, the bistatic mode has two separate
points for transmission and reception of radar energy. Separation of these independent points can
be manipulated in the field to determine radar wave velocity. The recorded time of transmission
from one antenna to another through a known distance of material can be measured and thereby
calculate the radar wave velocity (Conyers and Goodman 1997). Configuration choices for GPR
antennas may be based on the type of field questions asked.
Different GPR applications demand different antennas. The choice for standard
transmitting antennas is generally straightforward, the dipole or bowtie traveling-wave antenna
have been primarily used for the impulse-based radar system (Daniels 1996). The transmitting
antenna emits an elliptical cone of radar energy with the apex of the cone located in the center of
the antenna (Davis and Annan 1989, Schultz 2003). The elongated shape of the elliptical is
usually parallel to the direction of the antenna movement along the ground (Conyers 2004,
Conyers and Goodman 1997). This shape of transmitted radar energy plays a role in survey
parameters. The size and shape of the transmission cone is primarily dependent upon the
physical properties of the soil, thereby dictating the spatial parameters of collecting data in the
Some antennas have an added feature that allow for a reference mark to be employed
during a survey. This mark is activated when a trigger switch is pushed on the antenna handle by
the operator, allowing for a spatially referenced point along a GPR transect. Marked points on
the radar transect are coordinated with field grid markers that can be located using a global
positioning system (GPS) to create a three-dimensional reference points (Conyers 2004, Conyers
and Goodman 1997, Tischler 2003).
Techniques for GPR vary widely, each according to their application. There are restrictions
to the range of the operating system. Selections of antennas are based on the depth and resolution
for area of study (Davis and Annan 1989). Figure 1-1 illustrates the commercial range of antenna
frequencies and properties. Higher frequency antennas (e.g. 900 MHz) offer a greater ability to
differentiate features of varying size from a surrounding matrix, but are limited in depth of
penetration. Conversely, the lower frequency antennas (e.g. 120 MHz) allow for greater depth at
which a GPR signal will be reflected back to the surface, but provide a lower range of resolution.
Any radar frequency used will be determined upon not only by the depth of study and the
subsurface features of that study, but upon the inherent properties of site area.
Antenna Frequency Properties Blu' = Pcso~lullan
Red = Depin
Figure 1-1. Scale of relative depth of penetration to resolution for ground-penetrating radar
antennas. Resolution capability is inversely proportional to depth of penetration.
(From Tischler, 2003. Integrating Ground-Penetrating Radar, Geographic Information
Systems and Global Positioning Systems for Three- Dimensional Soil Modeling).
The type of antenna frequency used is limited by the properties of the material that is
surveyed. A receiver bandwidth in excess of 500 MHz, and typically 1 GHz, is required to
provide a typical resolution of between 5 and 20 cm, depending on the relative properties of the
soil (Daniels 1996). Antennas used in most archaeological contexts are the 500-MHz and 900-
MHz antennas; while lower frequency antennas are often used to interpret the geomorphology of
many prehistoric sites. Equipment choice can be a compromise between resolution, scope of
signal processing, and the ability to effectively offset the negligible properties of subsurface
material. Frequency selection is dependent upon site parameters and survey obj ectives. Annan
and Cosway (1994) listed a number of factors controlling frequency selection: (i) electrical
properties of the host environment, (ii) depth resolution, (iii) clutter dimension, (iv) depth of
exploration and target size, (v) site access and (vi) any external interference.
Radar Propagation and Parameters
The effectiveness of the radar is dependent upon the ground in which it penetrates;
specifically the properties beneath the surface. GPR has been used to investigate ice, water, rock,
pavement, soil and many other natural and human environments (Olhoeft 1996). The success of
these numerous applications has all been based on the ability of the transmission of
BE DROC K
Figure 1-2. A schematic of reflected and scattered electromagnetic waves as they are transmitted
through differing soil properties. Antenna displayed is a monostatic system with both
a transmitter (Tx) and receiver (Rx). (From Schultz, 2003. Detecting Buried Remains
in Florida Using Ground-Penetrating Radar. Doctoral Thesis in Forensic
Anthropology, University of Florida, p.28, Figure 4).
The physical properties that affect the radar waves as they pass through a medium are electrical
conductivity and magnetic permeability (Annan et al. 1975). Radar pulses generate a wavefront
which propagates downward and is partially reflected by any change in the electrical bulk
properties of the soil (Figure 1-2) (Vaughan 1986), GPR measures the changes in these
properties. Any of the electromagnetic energy that is reflected because of changes in these
properties is amplified and recorded through the receiving antenna.
Use of GPR requires the system operator to accurately understand the constraints of the
study area. Two factors must be determined prior to the operation of the equipment: the dielectric
constant and the range. Accurate configuration of the system will determine the proper depth of
signal penetration and resolution of that signal. Both range and dielectric constant are system
parameters that are set individually, but are dependent upon one another. An accurate range
cannot be set without a general knowledge of the subsurface dielectric constant.
The range is a parameter that has a time value measured in nanoseconds (ns.) A range
value is entered into the GPR control unit in order to determine how long to record received
signals after a radar pulse has been transmitted (Geophysical Survey Systems Inc. 1994). Any
portion of the radar pulse which can penetrate into the subsurface material, reflect from a
boundary or obj ect and return to the antenna is recorded and displayed in real-time. As time is
increased for the range setting, radar waves will penetrate the subsurface further in depth, yet
resolution and reflected data decrease with depth. There are finite restrictions to range settings
and the distance for which radar waves will travel. The time (ns) for a pulse to reach an obj ect or
boundary is dependent upon the material through which it passes. Restrictions to the range are
reliant upon frequency of the antenna and the dielectric constant.
Properties and Materials Affecting Radar Propagation
Electromagnetic pulses that are transmitted from the GPR antenna pass through space. If
this space is in a vacuum, then an electromagnetic pulse will travel at 0.2999 meters per
nanosecond. However, GPR is used over the ground and the materials located beneath the
surface range from soil, rock, ice, water and air to anthropogenic remains. Materials in the earth
have a range of dielectric properties which allow the passage of electromagnetic energy. The
more electrical conductivity a given material has the less dielectric it is; therefore for maximum
radar-energy penetration, a highly dielectric medium with a low electrical conductivity is
required (Conyers and Goodman 1997).
The dielectric constant is a parameter required in the calibration for the maximum effective
depth of the GPR energy waves. This constant, also known as the relative dielectric permittivity
(RDP), is measured against the heterogeneity of an obj ect or earth' s natural subsurface system.
The RDP is the capacity of a material to store, and then allow the passage of electromagnetic
energy (von Hippel 1954). Dielectric characteristics of materials vary greatly. Calculation and
measurement of RDP is based on of the velocity of an electromagnetic pulse in a vacuum. When
a pulse travels in a vacuum at 0.2998 meters per nanosecond then the RDP is 1 (Conyers and
Goodman 1997). Materials with a low dielectric constant, allow the passage of electromagnetic
waves with little or no dissipation, while higher conductivity will create greater dissipation of
radar waves or attenuation.
Soil and consolidated earthen material are both filters and boundaries to radar waves,
depending upon the dielectric properties of the medium. Dielectric constant values range from air
(low), to sea water (high) and vary according to the properties associated with subsurface
materials (Table 1-1), (Davis and Annan 1989, Geophysical Survey Systems Inc. 1994).
Materials composed primarily of air are less electrically conductive and are termed to be less
dielectric. Conversely, the RDP of water is of the order of 80, therefore even small amounts of
moisture will cause a significant increase of the relative permittivity of material (Daniels 1996).
Table 1-1. Typical relative dielectric permittivities (RDPs) of common materials.
Dysandy coastal land 10
Average organic-rich surface soil 12
Marsh or forested land 12
Organic-rich agricultural land 15
Saturated sand 20-30
Fresh water 80
Sea water 81-88
Modified from Davis and Annan (1989) and Geophysical Survey Systems, Inc. (1987).
History of GPR
GPR has a history that was a precursor to its own rediscovery. The first practical
application of GPR was in 1929 by W. Stern for determining the depth of an Austrian glacier
(Olhoeft 1996). The United States Air Force developed better radar systems in the late 1950's
when several pilots crashed on icy runways because conventional radar was transmitting through
the ice and reading inaccurate altitudes. In 1960, John Cook made the first proposal for the
detection of subsurface reflections using radar in his article "Proposed monocycle-pulse, VHF
radar for airborne ice and snow measurements." From this work, the development of radar
systems to detect subsurface reflections continued. The subsurface radar system from the late
1920's had not changed all that much when used for the Apollo 17 lunar mission in 1972. The
need for alternative applications by the U.S. military during the Vietnam War created an increase
in GPR development and advancement, hence the redundant use of acronyms. Radar is an
acronym for Radio Detection and Ranging.
Commercial availability for GPR technology was introduced in 1972 by Geophysical
Survey Systems Inc. (GSSI), ushering in a myriad of applications and extensive research to
universities and private companies alike. GPR development increased during the 1970's; Moffatt
and Puskar (1976) improved the target-to-clutter ratio of antennas and thereby increased
subsurface reflection accuracy. With increased antenna capability, new methods were able to
estimate the location of mines, faults and underground tunnels (Cook 1975, Moffatt and Puskar
1976). GPR technology in the early 1980's was used as a pedologic tool (Collins 1992, Doolittle
1982) and to investigate the depth of groundwater (van Overmeeren 1994). GPR has been used
to map bedrock surfaces, detect sinkholes and delineate karst features (Collins et al., 1990,
Doolittle and Collins 1998, Puckett 1990). As advancements in technologies emerged during the
1980's, methods of processing and analyzing GPR data became more widespread and available
to a variety of scientific Helds. This increase in available GPR technology created many diverse
Archaeological and Cemetery Application of GPR
The Hield of archaeology has long relied upon the use of multiple scientific disciplines,
ranging from geology to zoology. Techniques and tools from these other areas of studies have
worked well in shaping the way archaeologists conduct their research. In the past three decades,
the use of GPR has provided a means for analyzing an archaeological site in a nondestructive
way. GPR is a tool that increases the spatial extent of a study area, while reducing the amount of
time to conduct a preliminary survey. Data is collected in real-time and conveyed to the operator
allowing for rapid reconnaissance.
Archaeologists often destroy the area they study through excavation. GPR nondestructively
records and saves all data that are collected, thereby allowing decisions to be made prior to
disturbing a site. The capability of GPR to estimate depth and shape of buried obj ects make its
use suitable to grave detection (Bevan 1991). Varying site characteristics and context are also
amenable to the use of radar. To determine whether artifacts were present in a site often required
a great deal of exploratory digging, thereby greatly increasing time and expense (Vaughan
1986). Field season at an archaeological site is usually short and successes are often dependent
upon many variables including equipment, personnel and weather. Surveys conducted with GPR
may narrow the scope of interest and reduce the labor-intensive process associated with
History and Use of GPR Mapping in Archaeology and Cemeteries
The increased success of GPR in archaeology has been related to the various circumstances
in which the technology is used. Advances in techniques have created greater interpretation of
data. Until recently, GPR was simply used to identify subsurface "anomalies" that may not have
represented archaeological features (Conyers and Cameron 1998). Fast and accurate
archaeological surveying of large areas using GPR has even created new word usage.
Geophysical 'prospection' is now considered a standard method for detecting buried
archaeological features and structures (Leckebusch 2003). No matter the terminology, GPR has
become recognized as an important tool for numerous archaeological applications.
The archaeological applications of GPR range from spatial parameters to temporal context.
Spatially, surveys are able to locate sites and delineate anthropogenic boundaries, while
superposition can be conveyed through stratigraphic sequencing and the mapping of features
(Dalan and Bevan 2002, Conyers 1995, Conyers and Goodman 1997, Sassaman et al. 2003,
Schultz 2003). The absence or presence of soil heterogeneity may also indicate an anthropogenic
site. This is contextual and site specific. A GPR survey of a possible prehistoric mound from the
Effigy Mounds National Monument in Iowa revealed symmetrical curving strata more consistent
with a prehistoric cultural mound than caused by natural phenomena (Whittaker and Storey,
2005). Reinterpretation and nondestructive analyses of the mounds has continued for the past 40
years. Whittaker and Storey (2005) affirm the potential of GPR for archaeological preservation
with ongoing research at Effigy Mounds National Monument. Sufficient knowledge of local
landforms and processes may allow a determination of earthen features (Dalan and Bevan,
2002). When taken in context to surrounding geomorphology, sequencing and uniformity in
radar profiles may suggest possible human augmentation.
GPR use can incorporate surface anomalies with subsurface uniformity. Discontinuities
between cultural features with the natural soil matrix extend spatial parameters for delineating
the lateral extent and thickness of shell middens (Chadwick and Madsen 2000). Near surface
lithological features create discrete anomalous characteristics. Such highlighted features may
expose potential voids beneath such as gravestone chambers (Lorenzo and Arias 2005). GPR in
archaeology not only highlights discontinuities within the soil, but reveals materials and obj ects
anomalous to a known area.
The successful use of GPR for locating historic graves has ranged widely, often with
varying results in the same site area. Intrasite soil conditions and burial practices may vary,
enabling clear and distinct reflections or attenuation. Even with mixed success, GPR is
considered the most reliable geophysical tool for locating graves. A model study by Bevan
(1991) was conducted at nine different sites in the United States that used both GPR. One study
site, the Poor Farm Cemetery in Rockville, Maryland, had no marked graves. Radar reflections at
this site exhibited grave features; however, excavation showed the reflection to be a natural
change in the soil. The study Bevan (1991) conducted at the different sites illustrated the varying
degree of success with GPR on cemetery graves.
Controlled archaeological and forensic tests have been conducted to establish the efficacy
of GPR. Schultz (2003) tested the applicability of GPR to detect buried bodies in two different
Florida soils. Pig cadavers, with varying weights and sizes, were selected as surrogates for
human bodies and buried at two depths. Factors that controlled decomposition were monitored
and assessed with the GPR. Schultz (2003) concluded that depth and time were the most
significant factors in the decomposition of the cadavers.
Another controlled test conducted with GPR was a simulated archaeological site in Illinois.
This study replicated features present in some North American archaeological sites under a
controlled environment (Hildebrand et al. 2002). Features included pig and dog burials in
mounds, isolated pits, hearths, artifact clusters and structures. In addition to the GPR, seismic
reflection imaging (SRI) was used. This allowed the authors to compare geophysical methods
based on reflection imaging in anthropogenic soil conditions. Both methods worked well in the
study with the GPR providing a more accurate depth range to target (Hildebrand et al. 2002).
A combination of pedological insight with the use of GPR allows for comprehensive site
analysis. Distinct change in soil structure or bulk density may enable a GPR operator to
distinguish cultural augmentation from natural occurrences. Surficial features, such as circular
trample zones associated with animal feeding areas, were identified with time-slice depth maps
created by processed radar data (Nishimura and Goodman 2000). A 300-MHz antenna was
chosen for a Roman period site in Wroxeter, England as a consequence of anticipated depth of
stratigraphy. This approach in data acquisition increases opportunity in locating site specific
feature characteristics. Prior knowledge of a site's archaeological component, along with
physical soil properties, such as soil texture, can dictate proper antenna use. Another Roman
period building in the United Kingdom was surveyed using GPR; a 450-1VHz antenna was
chosen over a lower frequency due to the potential signal attenuation in clay-rich soils at great
depths (Linford and Linford 2004). These are two examples of balancing depth of penetration
with the potential of known feature resolution in similar archaeological site parameters.
Soils of Florida
To understand the limitations and efficacy of GPR in Florida, characteristics and properties
of the soil will be discussed. Soil "is a collection of natural bodies on the earth' s surface, which
may be modified or created by human activity, containing living matter and capable of
supporting plants" (Collins 1992). All soils are the product of the environment factors in which
they are associated. The Hyve soil forming factors (time, parent material, topography, climate and
biota) were described first by Dokuchaev (1883) and later by Jenny (1941). Distinct soil forming
factors associated with Florida are reflected in the dynamic and varied soil properties. Resulting
pedogenic processes are the basis for describing and classifying soils.
Soils in the United States are classified according to Soil Taxonomy (Soil Survey Staff,
1999). Specific information is based on this system of soil classification will be from the Soil
Survey Staff (1999), unless cited otherwise. Soil Taxonomy is based on measurable or
observable properties of soils as described in the field and laboratory. Classification indicators
include diagnostic subsurface and surface (epipedon) horizons based on physical, mineralogical,
morphological, and chemical properties. Soil indicators are referenced and classified to a
maximum depth of two meters, while diagnostic features may be restricted to or exceed this
Soil Taxonomy is hierarchical with the following categories from the highest to lowest: (i)
Order, (ii) Suborder, (iii) Great Group, (iv) Group, (v) Family and (vi) Series. Soil Orders are the
highest and most inclusive category, while the soil Series is the most comprehensive category.
The state of Florida has seven of the twelve soil Orders: Alfisols, Entisols, Histosols, Inceptisols,
Mollisols, Spodosols and Ultisols (Collins 1997). Soil Order classification is based on the
presence or absence of maj or diagnostic surface or subsurface horizons.
Diagnostic Surface Horizons
All diagnostic surface horizons are termed epipedons, based on the Greek words epi,
meaning over or upon, and pedon, soil. Epipedons that exist in Florida are either organic (histic,
folistic) or mineral (mollic, ochric, and umbric). A summary table of the Hyve epipedons in
Florida is listed in Table 1-2, with the folistic epipedon not included due to the very specific
nature of its classification and limited coverage. Diagnostic surface horizons that have been
altered through anthropogenic influences could be considered anthropic epipedons.
Understanding diagnostic surface horizons is vital in determining successful GPR application.
Epipedon determination may also establish soil Order classification.
Epipedons range in physical, chemical and morphological characteristics. Physical
parameters include mineral and organic soils, while morphological include soil thickness,
structure, and color. Chemical diagnostic properties associated with epipedons include base
saturation (B S) percentage including cation exchange capacity (CEC), listed in Table 1-3.
Particles <0.002 mm in size in Florida have a net negative charge. Positively charged cations are
attracted to these negatively charged surfaces. Most of the cations attracted to the sites are
exchangeable. These soil chemical properties have a tremendous impact on the performance of
the GPR. According to Daniels (2004) soils with a high CEC are more attenuating to GPR than
soils with an equivalent percentage of low CEC. This is similar for soils with elevated levels of
BS, more cation activity in the soil will increase attenuation in the GPR.
Table 1-2. Diagnostic summer of the four eidons used to classic soils in Florida.
Epipedon Base Typical Thickness Organic Horizon General
Saturation Color Content Description
Hi stic Low to 10YR 2/1, > 20 cm > 12% Organic Saturated >
high. 0- 2/2 30 days a
Mollic > 50% 10YR 3/3 > 25 cm > 0.6 to Mineral > 50% B.S.
or less 12%
Ochric 0-100% 10YR 4/4 < 25 cm < 0.6% Mineral Fails to
or more meet
Umbric < 50% 10YR 3/3 > 25 cm > 0.6 to Mineral Similar to
or less 12% Mollic but
Table 1-3. Cation exchange caacities (CEC) of soil materials (Birkeland 1999).
Soil Material Net Negative Charge (cmol c/k)
Humus (Organic) 100-550
The histic epipedon (from Greek histos, meaning tissue) is a horizon 20 cm or more thick
of organic soil material consisting of either muck, peat, hemic or a combination. This organic
soil horizon overlies a mineral subsurface horizon and has a very low bulk density. This
epipedon is saturated with water at some period (30 days or more) of the year and are common in
very poorly drained areas in Florida. Histic epipedons have high electric conductivity, high BS
and very high CEC, (Table 1-2).
A mollic epipedon is a mineral surface horizon with high accumulated organic matter
content (from .6 to 12 % organic carbon throughout), allowing for a dark brown (10 YR 3/2) to
black moist color (10 YR 2/1). Other diagnostic properties associated with a mollic epipedon are
it must exceed 25 cm in thickness and have a B.S. greater than 50%. Mollic epipedons have a
high electric conductivity and a high CEC.
Morphologically, the umbric epipedon appears to have the same general characteristics as
a mollic epipedon; the only difference is the B.S. is lower than 50%. Electric conductivity and
CEC are lower than the mollic epipedon.
The most common mineral epipedon in soils is the ochric (Collins 1992). The
morphological description of an ochric epipedon is a default category by not meeting the
classification of any other epipedons. This surface horizon does not have the diagnostic
properties associated with other epipedons; it is low in organic matter and not dark or thick
enough to be a mollic or umbric epipedon. Electric conductivity and CEC have a wide range of
values in an ochric epipedon and often tend to be lower than most epipedons.
Diagnostic Subsurface Horizons
Diagnostic subsurface horizons provide morphological, mineralogical and chemical
characteristics that enable the soil to be classified. Many of the diagnostic subsurface horizons
are defined by clay mineralogy and chemical components. Only 5 of the more than 18 subsurface
horizons (Soil Survey Staff, 1999) used to classify soils apply to this research and will be
considered. All 5 are located in Florida and pertinent to this study: albic, argillic, cambic, kandic,
An albic horizon is defined by a light-colored horizon of eluviation that is low in, organic
matter, clay minerals and iron and aluminum oxides. This horizon is generally associated above
subsurface illuviated horizons in Florida: spodic, argillic, or kandic (Collins 1997). Albic
horizons are predominately made up of white to light gray sand (ex: 10YR 7/1-8/1) particles in
Florida with low electrical conductivity and low CEC.
An argillic horizon is a subsurface illuviation of high-activity silicate clays. The
accumulation of layer-lattice silicate clays in the argillic horizon is due to physical translocation
and chemical transformation soil processes. Concentrations of clay translocated from upper
horizons are called argillans. These interstitial clay skins often coat and form on the surface of
peds and within soil pores between sand and silt particles; it is often the amount of clay that
determines an argillic horizon. An argillic horizon is an illuvial horizon > 15 cm thick in a soil
with 15 to 40% clay. Argillic horizons have medium values of electrical conductivity and a wide
range of CEC.
A subsurface soil horizon that is weakly developed with some color, texture and structure
change is classified as a cambic horizon. The cambic horizon has very little or no significant
accumulation and is characterized by the alteration or minor accumulation of mineral material.
This diagnostic subsurface horizon is weakly to moderately weathered and relates a recent soil
genesis. Cambic horizons have moderate to high electrical conductivities and low to moderate
range of CEC.
An illuvial horizon having an accumulation of colloidal organic matter, Al, and/or Fe
oxides is classified as a spodic horizon. Spodic horizons are highly illuviated subsurface soil
horizons commonly located in Florida' s pine flatwood forests. The black or reddish organic
illuvial horizon is very acidic. Often spodic horizons in Florida underlay highly eluviated
subsurface horizons or mineral and organic epipedons. The amorphous accumulation of organic
matter and oxides in a spodic horizon will increase the bulk density and lower the pH in relation
to surrounding horizons.
Soil Orders in Florida
According to soil taxonomy there are currently 12 soil orders assigned to soils throughout
the world. Soil taxonomy is the classification scheme in use in the United States, and is based
mainly on observable properties (Birkeland 1999). A wide range of soil orders (7) are present
throughout Florida: (i) Alfisols, (ii) Entisols, (iii) Histosols, (iv) Inceptisols, (v) Mollisols, (vi)
Spodosols, (vii) and Ultisols.
Alfisols are classified by the subsurface diagnostic argillic horizon. An argillic horizon
and also a thin gray to brown ochric epipedon are commonly associated with this soil Order.
Alfisols in Florida are characterized by a subsurface accumulation of silicate clays (argillic
horizon) and have a moderate to high level of bases (> 35%). Alfisols are moderately leached
soils that are commonly associated with mixed hardwood forests in well drained to poorly
This Order is a very common and morphologically diverse group of soils in Florida.
Entisols are recently developed or show little if any pedogenic development. Parent materials of
this soil Order range from to a marine, aeolian, or alluvium deposition in Florida. Diagnostic
horizons are limited and may include ochric epipedons and albic horizons, yet argillic or spodic
horizons may be present below a depth of 2 m (Collins, 1997, W.G. Harris, personal
communication, 2006, Schultz, 2003). Entisols are primarily quartz sand and may range from
well to poorly drained soils; the CEC and electric conductivity range of these soils are widely
varied as well.
Histosols are comprised of organic soil material. Histosols have soil organic material in
half or more of the upper 80 cm, or in two-thirds of a soil overlying shallow rock. Due to the
high organic content, these soils are generally black to dark brown in color and have high water-
holding capacities on a mass basis. Not all wetlands in Florida contain Histosols, however all
Histosols (except Folists) occur in wetland environments. Common names for Histosols are peat
(slightly to moderately decomposed) or muck (highly decomposed). Generally, Histosols have
very high amounts of CEC.
Inceptisols are the 'inception' of soil development with few diagnostic features. Unlike
Entisols, Inceptisols show more significant soil development in both surface and subsurface
horizons. Umbric and ochric epipedons are very common in this soil Order, yet may have a
mollic or histic surface as well. Subsurface horizons, such as albic or cambic horizons are
represented with the establishment of a structural, textural or color change. Soil properties
associated with Inceptisols range from textures of a sandy loam to fine sand, and moderate to
high CEC in the clay fraction of recently developed illuvial horizons (Collins 1997).
Mollisols are mineral soils which have an accumulation of high organic matter, with a
B.S. greater than 50%. All Mollisols have a dark, humus-rich mollic epipedon 18-25 cm in
thickness (10 cm over bedrock); however a histic epipedon may overlay a mollic. In Florida,
Mollisols have argillic, albic or cambic diagnostic subsurface horizons. A maj ority of Mollisols
are associated with grassland or prairie environments; however, in Florida this soil Order is
primarily poorly to very poorly drained conditions (Collins 1997). High ranges of CEC and low
ranges of electric conductivity occur due to the high organic matter and B.S. inherent to
Spodosols are mineral soils with a spodic horizon within 2 m of soil surface. Spodic
horizons often underlay very leached eluvial albic horizons, but may have argillic or kandic
horizons as well. Epipedons associated with Spodosols range from ochric to a thin histic. These
soils are often sandy, acidic forest soils with low bases and high amounts of Al. Most Spodosols
occur in poorly drained pine flatwood regions throughout Florida (Collins 1997). Spodosols have
low CEC ranges, while electric conductivity may be quite high down to the spodic horizon.
Ultisols are highly weathered soils with silicate clays and low bases. Intense leaching
from upper horizons allow for the translocation of clay minerals into argillic or kandic
subsurface horizons. Prolonged leaching of Ultisols is characterized by a low amount of bases
(<3 5%), relatively acidic silicate clays, and Fe and Al oxides. Ultisols may have either an ochric
or umbric epipedon. This soil is commonly associated with stable landscapes with mature forest
vegetation. Ultisols are morphologically similar to Alfisols, but are more highly weathered and
acidic. In Florida, Ultisols range from well-drained to poorly drained soils (Collins 1997).
Ultisols have low CEC ranges, with moderate electric conductivity.
Objectives of the Study
The general obj ective of this research was to demonstrate the efficacy of GPR on
anthropogenic sites with selected Florida soils Orders. The suitability of GPR for archaeological
and gravesite detection is not always applicable to pedogenic models. The purpose of this
research is to assist archaeologists and GPR operators to better understand the context of
anthropogenic influences on natural soils.
* Conduct GPR surveys to obtain information about the selected archaeological sites and
cemeteries in Florida
* Establish the extent of archaeological and cemetery soil properties in producing distinctive
* Document resolution differences between 900-MHz and 500-MHz GPR antenna usage for
obtaining anthropogenic targets in study areas.
* Determine the suitability of GPR on various Florida soil Orders in an archaeological and
gravesite context by demonstrating the influences of soil physical and chemical properties.
Study Site Selection
Four sites were studied to evaluate the application and feasibility for the use of GPR in an
archaeological and cemetery context in the North-Central and Central Florida (Figure 2-1). Field
sites were based on archaeological context and soil variability. Archaeological sites included a
prehistoric habitation site in Lake County on the shore of Lake Apopka and a historic structure
site in St. Johns County, north of St. Augustine. The two historic cemeteries date from the late-
nineteenth century in Orange County to the early and mid twentieth century in Alachua County.
Both the Orange and Alachua County sites are associated with African-American cemeteries.
These study areas were based on four different soil Orders. In Florida soils there are
contrasting subsurface features, i.e. argillic horizons, spodic horizons, layers (organic); lithic
(bedrock) contacts; wetting fronts; and lamellae that are easily exhibited by GPR (Collins 1992).
The properties from a maj ority of Florida soils are conducive to the use of GPR (Collins et al.
1996, Doolittle et al. 2002). The four field study sites selected will elucidate potential concerns
for the use of GPR when applied with soil properties in an anthropogenic context.
The Newberry cemetery (Davis Cemetery 8AL4992) is located in Section 34 of Township
9, Range 17 East in western Alachua County, Florida. This site lies east of the Brooksville range
on the Wicomico terrace; underlain by alternately thick layers of Plio-Pleistocene material
overlying loamy marine sediments and limestone bedrock (Thomas et al. 1985). Part of the
Alachua Formation, the Newberry Cemetery site area is in the southwestern part of Alachua
County, where it forms low and rolling hills (Thomas et al. 1985). The substrate materials
FI 16~01 12 1240Kilorreten1
represent the residuum of post-Eocene formations (Schmidt 1997). This area has been
extensively mined for rock. This formation varies in depth from 8-12 meters (Thomas et al.
1985). Naturally occurring vegetation in this area includes: live oak (Quercus virginiana), laurel
oaks (Quercus laurifolia) and water oaks (Quercus nigra); longleaf pine (Pinus palustris) and saw
palmetto (Serenoa repens).
Figure 2-1. Location of research sites in Alachua County, Lake County, St. Johns County, and
Orange County, Florida.
The soils at the Davis cemetery are well-developed Alfisols (Thomas et al. 1985). They
are mapped as Pedro- Jonesville complex, with varying depths of argillic horizon; which are so
intermixed, that they cannot be separated at the scale of mapping (Figure 2-2). Mapped areas of
the complex range from 10 to 50 hectares. Pedro soils are fine-loamy, siliceous,
Figure 2-2. Soil survey map showing soil series and complex at the Newberry cemetery. The
map was created in ArcGIS 9.1 using data derived from the Alachua County Soil
Survey (Thomas BP, et al. 1985).
hyperthermic, shallow Typic Hapludalfs. The thickness of the solum and the depth to limestone
range from 15-45 cm. The other soil in this complex is the Jonesville series; loamy, siliceous,
hyperthermic Arenic Hapludalfs. Mapped as a complex, the maj or difference between the
Jonesville soil as compared to the Pedro soil is the depth to limestone. Both soils have an argillic
horizon which is from 32-72 cm in depth. Depths to limestone and the argillic horizon greatly
vary within the site. An example of depth to lime rock variability for Alachua County soils is in
Figure 2-3. Overall, this complex has severe limitations for cemetery use.
Figure 2-3. Soil depth to limerock variability of an Alfisol in Alachua County, Florida.
Three headstones were identified in a heavily wooded area north of Newberry, Florida
along transects 14, 17 and 21 in Figure 2-2. In addition to the marked graves, four suspect graves
were located in immediate vicinity; two linear depression and two uninscribed marble footstones.
All three marked graves were inscribed with name, date of birth and date of death. The graves
were identified bearing the names Catherine Alridge, Willies Davis, and W.J. Long. The earliest
recorded date of death was Willie Davis on February 14, 1892. Catherine Alridge was the most
recent grave, dated to June 4, 1918. All marked graves and linear depressions were in an east to
west alignment, denoting a Western Christian burial pattern. Gravestone type and dates of the
four marked graves are in Table 2-1.
Table 2-1. Graves at the Newberry Cemetery. Preliminary Report, Southeastern Archaeological
Research, Inc. Gainesville, Florida.
Name Date of Birth Date of Death Gravestone Material
Catherine 1832 June 4, 1918 Rounded Marble
Willie Davis May 14, 1818 February 14, Shouldered Marble
W.J. Long May 1867 August 14, Rounded Marble
Historic research conducted by Southeastern Archaeological Research, Inc. (SEARCH) in
the area was inconclusive as to the original name of the cemetery. No known churches or
homesteads with common last names were connected with the Davis Cemetery. No information
could be found on Catherine Alridge or Willie Davis, while only one report of W.J. Long (black
male) places him in the Newberry area during the year of 1916 (SEARCH 2005). Historic deed
research of the Davis Cemetery area indicated that in the first decade of the 20th century the
property was owned, but not developed, by the Florida Land Development Company and was
sold on July 14th 1910 to the Mutual Mining Company of Marion County (SEARCH 2005).
Located in a mixed hardwood hammock, this site area has a high canopy with little under
story growth. North of the graves are three small cedar trees. Cedar trees were typically planted
near the periphery of cemetery boundaries during the late 19th and 20th century in Florida. The
cemetery boundary is identified to be no larger than 32 m east to west and 25 m north to south.
During the time of the research at this site area a 27 m preservation buffer zone was put in place.
No protective fences or markers are located around the heavily wooded area. A town facility is
adj acent to the Davis Cemetery. As of the fall of 2006, this cemetery does not meet the criteria
for eligible inclusion in the National Register of Historic Places (NRHP).
Oakland Cemetery Site
Located in Orange County, Oakland Cemetery (OR9576), also known as the Old African
American Cemetery, is situated in Section 20 of Township 22, Range 27 East. This site lies south
of Lake Apopka in the southern end of the Central Valley. The Lake Apopka area is underlain by
undifferentiated Plio-Pleistocene sediments and sedimentary deposits of the Hawthorn Group
overly the upper Eocene limestone (Doolittle and Schellentrager 1989, Schmidt 1997, White
1970). The geomorphology of the site area is at the interface of the Central Valley and Lake
Wales ridges, part of the Wicomico Terrace (Doolittle and Schellentrager 1989, White 1970).
Topographic variability, associated with the Central Valley and Lake Wales ridge, range from a
5-12 % slope in the site area (Figure 2-3). Natural vegetation is predominately live oak (Quercus
virginiana), bluej ack oak (Quercus incana), turkey oak (Quercus laevis), and sand pine (Pinus
clausa). Surrounding site area has been used for citrus crops during the past century.
Oakland Cemetery is predominately comprised of Entisols and Ultisols, and is mapped as
Candler-Apopka fine sands with 5-12% slopes and adj acent, depressional Basinger fine sands
(Doolittle and Schellentrager 1989). The Candler-Apopka complex is an excessively drained
soil. This soil complex is associated with uplands and low ridges (Figure 2-4). Candler soils are
classified as hyperthermic, uncoated Lamellic Quartzipsamments; they form in thick deposits of
aeolian or marine sand. Similar in pattern and proportion to the Candler, the Apopka soil series is
a loamy, siliceous, hyperthermic Grossarenic Paleudult. This soil complex is widespread
throughout the site area. The individual soil series cannot be differentiated on a small scale and
are therefore mapped together as a complex.
Figure 2-4. Soil survey map showing the Candler-Apopka soil complex at the Oakland
cemetery. The map was created with ArcGIS 9.1 using data derived from the Orange
County Soil Survey (Doolittle and Schellentrager 1989).
Apopka consists of well drained soils that formed in sandy and loamy marine sediment.
Both the Apopka and Candler soil series are morphologically very similar in the upper meter of
each horizon (see Discussion). The Basinger series is classified as a siliceous, hyperthermic
Spodic Psammaquents; which form in sandy marine sediment. These poorly drained soils are in
shallow depressions and along the rim of larger depressions. An example of a Florida Entisol soil
is in Figure 2-5. The homogenous matrix of the C horizon is evident in the uniform subsurface
~w9r r- r~Yc-f
.~ .. ~ =IFi'
Figure 2-5. Photograph of an Entisol profile in Florida.
The Old African American Cemetery is located north of the State Road 50 and Florida
Turnpike intersection. Along the south border of the cemetery is the remnant of the Tavares and
Gulf Railroad (1891-1971) roadbed. No longer an active cemetery, approximate dates of early
graves are from the early 1890's. The last known graves are from as late as 1952. The cemetery
was not affiliated with any religious organizations and African Americans from the surrounding
communities were also buried in the cemetery (SEARCH 2005). Both the citrus crops and the
railroad line facilitated an influx of migrant workers in the area. Many of the people buried in the
cemetery were citrus farmers and service workers. Also, if any of the migrant workers or
possible transients died while in the area, they were buried in the cemetery as well (Deacon
Moore 2006: personal communication). Census searches revealed most of those buried in the
cemetery are originally from North or South Carolina and Georgia (SEARCH 2005, Florida
Census 1920). African American burial practices associated with this site had a significant
relationship with results of the GPR survey and will be discussed in Chapter 5.
Historic Archaeological Site
This British Period occupation site area is located west of the Tolomato River and north
of Marshall Creek in St. Johns County, Florida. The site (8SJ3 149) lies within the Atlantic
Coastal Lowlands and is less than one-half mile west of the Atlantic Ocean on the Pamlico
terrace. This area was formed as barrier islands and lagoons from the Plio-Pleistocene to recent
times (Davis 1997). Silver Bluffs and Pamlico terraces occur in this intercoastal area. Physical
environment is dominated by pine flatwoods, hardwood hammocks and salt marshes. Hammocks
within the Marshall Creek property are associated with the bottomland hardwood wetlands that
border the coastal salt marshes.
The soils mapped in the site area are weakly to moderately developed Spodosols with
associated Entisols; and are classified as either Myakka or Immokalee (Spodosols) soils (Readle
1983). The Myakka-Immokalee complex is a nearly level, poorly drained sandy soils located on
flat marine terraces in broad flatwood areas (Figure 2-6). Most of the soils in this complex have
loamy to sandy subsoils that are stained with dark organic accumulations. The Myakka series is
Florida' s state soil and is classified as a sandy, siliceous, hyperthermic Aeric Alaquod.
Immokalee soils are classified as sandy, siliceous, hyperthermic Arenic Alaquods (Figure 2-7).
Figure 2-6. Soil survey map showing soil series and complex at the historic archaeological site.
The map was created with ArcGIS 9.1 using data derived from the St. Johns County
Soil Survey (Readle 1983).
Figure 2-7. Photograph of an Immokalee soil series profile.
The Tolomato River and Marshall Creek areas have varied historic and prehistoric
archaeological components. Multiple phases of the St. Johns culture occupied northeast Florida
from around 500 B.C. and lasted until shortly after permanent historical settlements occurred in
1565 A.D. (Milanich 1994). A European power struggle of French and Spanish interests
occurred in the St. Johns River area during the mid-16th century (Contact Period: from 1565
A.D.). In 1565 the Spanish, led by Pedro Menendez de Aviles dispatched the French and
established St. Augustine. Chosen for its strategic location, St. Augustine existed as a military
outpost and a base for missionaries for the area (Deagan 1983). The First Spanish Period (1565-
1763 A.D.) in northeast Florida was a rural mix of Franciscan missions, Indian villages, Spanish
cattle ranches and military fortifications (Bond et al. 1990, SEARCH, 2005). During the British
Period (1763-1784 A.D.), rural construction in northeast Florida increased; plantations, industrial
construction, trading posts, slave cabins and plantation houses developed throughout the region
(Bond et al. 1990, SEARCH, 2005).
Numerous excavations were conducted in the Marshall Creek (8SJ3149) area over the last
decade. Extensive amounts of information have been collected on both prehistoric and historic
archaeological components from the site area. Archaeology on 8SJ3149 was conducted in order
to mitigate development in the vicinity. Multiple scientific disciplines were employed to
facilitate the survey and excavation of this site.
Prehistoric Archaeological Site
This prehistoric archaeological site is known as the Montverde site (8LA243), located in the
town of Montverde in Lake County, Florida. The site is situated in an agricultural area on the
west side of Lake Apopka (Figure 2-8). The site 8LA243 is bordered to the east and north by the
lake, and to the northwest by a freshwater marsh. The geomorphology of the area includes Plio-
Pleistocene sediments overlying Hawthorn deposits and upper Eocene limestone (Doolittle and
Schellentrager 1989, Schmidt 1997, White 1970). Lake Apopka lies in the Central Valley,
associated with the Lake Wales and Winter Haven Ridge along the Wicomico terrace (White
1970). The well-drained soils and proximity to Lake Apopka (Figure 2-9) was conducive to long
term citrus cultivation at the site area. Natural vegetation includes live oak (Quercus virginiana),
turkey oak (Quercus laevis), longleaf pine (Pinus palustris), slash pine (Pinus elliottii), and saw
palmetto (Serenoa repens).
Figure 2-8. Photograph of an excavation at the prehistoric archaeological site along the
southwest shore of Lake Apopka, Montverde, Florida.
Two soils are mapped with the site area; a well-drained Orlando series and a very well-
drained Entisols classified as the Lake soil series (Furman et al. 1975). Predominately, the site
associated with the Orlando series (Figure 2-6). This series is classified as a sandy siliceous,
hyperthermic Humic Psammentic Dystrudepts. The Orlando series is a nearly level to gently
sloping (0-5%), well-drained sandy soils in association with upland ridges. This Inceptisol
formed in sandy marine sediments and has a thick umbric epipedon. The other soil in the
research site is the Lake series; classified as a hyperthermic, coated Typic Quartzipsamment.
Figure 2-9. Aerial photograph (1999) of the prehistoric archeological site at Monteverde
showing shovel tests and test unit location (Austin 2006). Shovel tests and excavation
units were conducted prior to survey by Southeastern Archaeological Research Inc.
Soil series information derived from Lake County Soil Survey (Furman et al. 1975).
The Orlando and Lake soil soils make up 93 to 99 percent of mapped areas with
marine sediment on upland slopes of 0-5% (Furman et al. 1975).
Previous excavations in the Montverde site area have indicated the presence of two
archaeological components. Upper levels of the excavation units contained St. Johns Check
Stamp pottery, indicating a St. Johns II occupation site; while the lower levels appear to be a St.
Johns I period, dominated by St. Johns Plain ceramics (Austin 2006). The organic midden base is
an early St. Johns I period with related features dating from as early as 1600 B.C.; while dates
within the midden go up to 400 B.C.(R. Austin, personal communication, 2006). Two cultural
features were identified in an excavation unit as circular, basin-shaped pits with fragments of
burned wood in the sandy fill (Austin 2006). The burned wood associated with pit features
allowed for accurate 14 C dating.
MATERIALS AND METHODS
Each field site area was selected on the key variables of temporal component, spatial
extent, pedological criteria and availability for research. Every site was to be represented by a
soil Order in Central Florida and not to exceed a survey size of 2 hectares. Selected sites were to
reflect field conditions encountered during standard operating archaeological surveys and
excavations. Every field site contained an element of bioturbation and anthropogenic influence
of varying degrees. A common element in all four research areas were the materials used and the
methods employed. Each individual site will include: archaeological field work, GPR survey,
soil sampling and description. Equipment and software used and soil analyses conducted were
identical for all four sites.
Equipment and Software: All four field sites were surveyed with the same GPR system:
SIR-2000 control unit with 500 MHz and 900 MHz ground-coupled monostatic, dipole antennas
with control unit (Geographical Survey Systems, Incorporated, North Salem, New Hampshire).
Both antennas have marker switches located on the handle and manually controlled by the
antenna operator. The control unit is a digital pulse system, with internal hard drive, keypad and
color screen. GPR files may be viewed real-time in the field or data may be downloaded to an
external computer. Each antenna is connected to the control unit by a 60-meter cable. A twelve-
volt battery powers the GPR unit.
Multiple software packages were used in the field and laboratory for GPR, Global
Positioning System (GPS) and Geographic Information System (GIS) data calculation. For all
four sites, GPR files were downloaded from the field to the computer for post-processing
analyses with Data Transfer Utility for Windows 95 and Windows NT, proprietary software of
GSSI (Geophysical Survey Systems Inc.). Post-processing analyses for the manipulation of GPR
data of all sites were conducted using the same computer software package (ReflexW 2.5X,
Sandmeier Scientifie Software, Karlsruhe, Germany). The primary GIS software package used
was ArcGIS 9.1 (Environmental Research Institute, Redlands, CA.). Two different GPS devices
were used in the acquisition of spatially referenced data, TOPCOM 3C total station with
TOPCOM SC data collector. The Garmin etrex VistaThl handheld GPS receiver was also used.
Soil samples were collected on site specific criteria. Collection of anthropogenic soils
samples was uniform in depth and quantity within each site area. Physical data acquisition and
soil sampling was done with a bucket auger (8 cm diameter and 17 cm in length). Archaeological
Hield sites incorporated both bucket auger and hand trowel during excavation. Auger samples
were taken with a bucket auger, which was found to pack and hold a greater range of soil
textures encountered from all four site areas. Soils were described and sampled up to a depth of 2
meters for every site. Anthropogenic soils were described on a site specific basis, with no
maximum depth requirement.
Soil samples of approximately 400-500 g each were collected in the Hield and placed in
labeled plastic bags. Soil descriptions were recorded and sample bags were returned to the
University of Florida Environmental Pedology and Land Use Laboratory. In the pedology
laboratory, soil color was determined by visual comparison on moist samples to a Munsell@ Soil
Color Book (GretagMacbeth, 2000). Morphological descriptions from the field and from NRCS
soil survey data were augmented with laboratory data and are listed in Appendix A. Appendices
B and C list the physical and chemical data from each site.
Two types of sampling increments were used for anthropogenic and pedogenic soils:
1. Natural soil samples were taken from each designated genetic horizon revealed in auger
samples for morphological descriptions and lab analyzed in the laboratory.
2. Anthropogenic soils sampled were taken from 20-30 cm interval (depth) samples for
physical and chemical analysis from surface to appropriate depth.
Procedural analyses for all soil samples of the four site areas were identical. Processing of
soil samples was done at the same time, using the same equipment and environmental conditions.
Samples were air-dried, screened through a 2-mm sieve, and stored in Analytical Research
Laboratory (ARL) approved soil sample bags. Each of the 232 collected soil samples was
assigned an individual catalogue number for use in laboratory testing and curation. Site specific
identification numbers (e.g., 8LA243: 8 = state of FL, LA = Lake County, 243 = site #), included
the provenience descriptions.
Soil samples were subj ect to physical and chemical analyses. Specific analyses included:
particle size, pH, organic carbon (OC) content, electrical conductivity (EC), extractable
aluminum (Al), extractable iron (Fe), extractable phosphorus (P), extractable magnesium (Mg),
extractable sodium (Na), extractable calcium (Ca) and extractable potassium (K) contents.
Particle-size analysis was used to determine the percentage of sand, silt and clay for textural
classification and subsequent sand fractionation. Soil pH is known as the "master variable"; this
test is done in order to understand the chemical range and parameters within the soil. The amount
of OC was determined to differentiate between a mineral soil (<12%) and an organic soil. The
percentage of O C may also greatly influence the taxonomic classification of a soil, chemical
properties and indicate the presence archaeological middens.
Amounts of extractables P, Ca, Na, Mg, K, Fe, and Al were analyzed to determine the
extent of anthropogenic influences on a pedologic landscape using the GPR Archaeologists have
used soil chemical analyses as both a temporal and spatial indicator of site habitation. Soil
chemical testing of the four bases (Na, Ca, K, and Mg) along with the two acid cations (Al, Fe)
will allow for an increased understanding of natural and anthropogenic site context. Extractables
P and Ca were analyzed to determine if bone diagenesis occurred in the site areas; both P and Ca
are the two most common minerals in bone (Schultz 2003). Distribution of extractable P is a
good indicator of anthropogenic activity, because mean extractable P is notably lower in
naturally occurring soils.
All particle-size analysis was conducted at the Environmental Pedology and Land Use
Laboratory, Soil and Water Science Department, at the University of Florida. Particle-size
distribution analysis was conducted using the pipette method (Gee and Bauder, 1986). Soil
samples of 50 g were treated with hydrogen peroxide and heated on hot plates to oxidize organic
material. The samples were then dispersed with sodium hexametaphosphate in de-ionized water.
Dispersed samples were immersed in a 200 C water bath and pipetted to retrieve the clay
fraction. The clay fractions retrieved were dried and weighed. Sand fractions were obtained
from the original sample and removed of remaining silt and clay, then oven-dried and shaken
through nested sieves. Silt content was derived by the subtraction of sand and clay weight from
the total sample weight.
Organic carbon content, pH, EC, extractables P, Fe, and Al were all analyzed at the
Analytical Research Laboratory, Institute of Food and Agricultural Sciences, Soil and Water
Science Department, at the University of Florida. OC content was determined by a modified
Walkley- Black method of dichromate sulfuric acid (Soil Survey Staff, 1996). The calculation of
pH was done by 1:2 water to soil solution by glass electrode of pH meter and deionized water.
EC was determined by 2:1 deionized water to soil solution using a glass electrode. This
procedure was done because EC is directly related to the amount and distribution of ionic
charges in the soil (McNeill, 1980).
The procedure to determine extractables P, Al, and Fe was using a Mehlich 1 extraction
(Mehlich, 1953). This procedure entails 5 grams of soil sample to be added to 20 ml Mehlich 1
solution (0.05 M HCL and 0.0125 M H2SO4). The solution and sample are shaken for 5 minutes
and filtered through Whatman # 42 paper. The extract is then analyzed by an inductively coupled
plasma-atomic emission spectroscopy (ICP-AES) (EPA Method 200.7, Analytical Research
Laboratory 2006). Mehlich 1 was also used for the analyses of extractables Ca, K, Mg, and Na;
the extract is analyzed by atomic adsorption spectroscopy (AAS).
Statistical analyses of the data was based on a standard two-tailed, paired t-Test. Analyses
of the soil data with a t-Test was used to determine whether two soil samples were likely to have
come from the same two populations that have the same mean. The dependent variables are
based on the pedogenic soils from the site. Anthropogenic soil samples are the independent
variable, measured with the dependent pedogenic soil samples. Data from physical and chemical
soil properties sampled from the site associated with GPR transects. The two-tailed t-Test was
performed to determine if soils from a potential anthropogenic site corresponds to the probability
of the known values of the pedogenic site. The statistical program used to perform the standard
two-tailed, paired t-Test was the Microsoft Office Excel 2003, statistical functions.
GPR Data Collection
Each research site was cleared of underbrush prior to GPR survey. Cleared transect paths
allow for an uninterrupted flow of data collection; thereby minimizing post-processing
corrections. Datum points were acquired for every site by a TOPCOM 3C total station and
TOPCOM SC data collector. Reference points were recorded from the datum points located in
the southwest corner of each site. A hand held GPS receiver was used to correlate spatial
coordinates of transects emanating away from the datum points. The GPS handheld receiver was
used along transects in all four field sites. Signal error due to tree canopy obstruction was
minimized by the GPS receiver being placed over known coordinates and physically rectified by
GPR transects for each site were collected in single line grid formation emanating from
datum point coordinates. A 100-meter pull tape laid out each transect in a north-south direction.
Direction and distance were corrected with the use of GPS and hand held SU7UNTOTM Sighted
compass. Pin flags were placed every one to two meters along each transect. Transects were
spaced in one-meter interval grid patterns. A one meter distance between transects was necessary
for complete site coverage with GPR. An antenna was walked along each transect and referenced
with corresponding intermediary trigger-switch marks associated with measured pin flag points.
The antenna was pulled along the same side of the line for consistency. All sites received the
same spatial reference methodology.
Each site contained different and varying soil properties, and thereby required different
range settings and RDP calibrations to determine time-depth conversions. An accurate
conversion of two-way travel time to depth is necessary before the interpretation of data can
commence. Time-depth calculation was determined by auguring to a known depth just off-site.
At one and two meter depths, water was added to the base of the auger hole and immediately
geo-referenced with the GPR with both 900-MHz and 500-MHz antennas. Ranges and dielectric
constants were established and set on each individual site location and date from this information
and referenced with known RDP ranges.
Analysis of GPR data often is conducted in a two-step process. The GPR technology used
in this research allowed for real-time data viewing in the field. This enabled the operator to take
corrective measures in equipment calibration and survey technique. In the field, GPR data files
were viewed using a Color Table of 2 and Color Xform of 2. These visual screen settings are
qualitative and subj ect to viewer preference for optimum data interpretation. Field interpretation
of visual GPR data was considered the preliminary step for overall data acquisition. Often times
transects were discounted and field data were reacquired based upon preliminary field
observations. Raw field data contain extraneous background noise or reflections that make
definitive interpretations difficult. Horizontal scale and vertical scales are often not coordinated
effectively in non-processed data.
GPR Post-Processing Analysis
Post-processing is an important step in GPR analysis and interpretation. Qualitative
assessments are greatly enhanced with the manipulation of GPR data; often images and
anomalies are not clear to the operator until the data has been processed. Interpretive strategies
and procedures for the processing of GPR data are numerous. There are fundamental procedures
in the rectification and manipulation of data sets. Spatial adjustment, both vertically and
horizontally, along with background filtering are primary procedures often employed by the
operator. Processing of GPR profiles required a minimum 4 different procedures in order to
interpret the raw data.
Horizontal adjustment was the first procedure for every GPR file. Linear correction is
necessary because the GPR antenna is not pulled across a transect at a constant speed.
Intermediary marks were created at prescribed intervals along each transect during the collection
of data. The known distances between each interval allows the software to the correct distances
between each mark. This step was a trace interpolation/resorting procedure. Vertical correction
was the next step taken for proper depth. GPR antennas are housed in a fiberglass box and
suspended above the ground surface. The actual distance between the antenna and the ground is
seen as an 'antenna ringing' noise that can be corrected. This 'ringing' is seen as a uniform
reflection image and is included in the original data as part of the profile depth. Elimination of
this noise will set the time transmission depth at 0 ns, as 0 meters; this is done through gain
After vertical and horizontal adjustments are made, filtering the files of horizontal banding
is done. This process allows for enhanced viewing of anomalies and the removal of unwanted
clutter. Radar attenuation is filtered out, leaving continuous pedogenic horizons, planar and
point-source anomalies. This procedure is a combination of a running and background removal
The Alfisols at the Newberry Cemetery are not considered highly suitable soil orders in
Florida for GPR application (Collins et al. 1996). Thus, GPR use for this site is not ideal from a
pedogenic context. However, this site was selected for availability, pedogenic diversity from
other selected soil orders and for the community need. The area is heavily wooded and
surrounded by planted pine. Cemeteries are confined to area no greater than 500 sq m; coupled
with a strict temporal range of less than 30 years. There are numerous occurrences of
bioturbation throughout the site that include animal burrows (krotovinas), tree stumps, root ball
inversion and roots. These subsurface anomalies are detected by the GPR and have varying
influences on both the 900-MHz and 500-MHz antennas.
A total of 24 transects were conducted in the cemetery every meter. Each transect was 20
m in length with marker flags placed every two meters for reference points. The general
cemetery area surveyed and staked at 27 m east to west and 21 m north to south by a previously
contracted survey team (Table 3-1). GPR surveys and soil sample collections were conducted
several times over a five-month period (October of 2005 to February of 2006).
Table 3-1. Data points for the four corners of the cemetery boundary. The proj section system used
was UTM, NAD 1983, Zone 17 North.
Northwest Comer: 17R 0344765 W Northeast Corner: 17R 0344786 W
3283016 N 3283015 N
Southwest Comer: 17R 034464 W Southeast Comer: 17R 0344886 W
3282990 N 3282989 N
The selection of Oakland Cemetery as a research site was a combination of soil
suitability, anthropogenic criteria and the need for mitigation before future development in the
area. The presence of unmarked graves outside the cemetery boundaries was not known by the
town residents or the local, affiliated church. A survey was conducted in order to determine if
any unmarked graves existed outside the historic boundaries of the Oakland Cemetery. This was
achieved through a combination of archived aerial photographs (Figure 3-3), ethnographic
accounts and a GPR survey.
On the eastern shoulder of the site area is a transition from the Lamellic Quartzipsamment
(Candler) down slope to the Grossarenic Paleudult (Apopka). Both of these series are ideal soils
orders in Florida for GPR applications in a pedogenic context (Collins et al. 1996). Typical
vegetation of these soils is sand pine (Pinus clausa), slash pine (Pinus elliottii), chapman oak
(Quercus chapmanii), live oak (Quercus virginiana, and saw palmetto (Serenoa repens). The area
is subj ect to krotovinas. At the base of the slope is a Spodic Psammaquent (Basinger), which is
seasonally flooded for extended periods of time. Naturally occurring vegetation includes
pondcypress (Taxodium ascendens), sweetgum (Liquidambar styraciflua) and scattered pond
pine (Pinus serotina).
A P sre wscndcedaog h isoiclbunais fth aladCmeey
Tranect emnatng utwrd rmkongae er opee nagi pten ahtasc
was ~ ~ seaae ya ee n efrnepitswr lcd vr eer.Telre cl fti
survey as augmnted b a numbr of vlunerwoweeal oplc i-lgso upce
anomaies drin th aacleto.Attlo upce nakdgae eelctdi h
course of three days of surveying. The site was staked with survey markers (Table 3-2) in all four
corners prior to the GPR survey.
Table 3-2. Data points for the four corners of the cemetery boundary. The proj section system used
was UTM, NAD 1983, Zone 17 North.
Northwest Corner: 17R 0437596 W Northeast Corner: 17R 0437696 W
3158418 N 3158417 N
Southwest Corner: 17R 0437596 W Southeast Corner: 17R 0437696 W
3158285 N 3158347 N
Historic Archaeological Site
Research interest in the Marshall Creek site (8SJ3149) was based primarily on the historic
occupation component. The field site in St. Johns County contained Spodosols and peripheral
Entisols, which are two of the best soils orders in Florida for GPR applications in a pedogenic
context (Collins et al. 1996). The area is heavily wooded with numerous occurrences of
bioturbation that throughout the site that include; animal burrows (krotovinas), tree stumps and
roots. Large areas of underbrush were cleared to prepare the site and restrict the unwanted
influences of the GPR survey. Both the 900- MHz and 500-MHz antennas were used and had
varying influences on subsurface detection.
The research site at Marshall Creek was a collaborative effort with SEARCH.
Archaeological excavations were conducted in the area over a six-year period. A small scale
GPR survey was conducted in the August 2005 in a known multi-component site area. GPR data
collection was confined to 100 m2 area adj acent to a block excavation. The excavation units were
placed near previously tested units which contained abundant British period materials (Carlson
2005). The coordinates of site area is shown Table 3-3.
A total of 16 transects were placed in a grid pattern and ran in both north-south and east-
west directions. This was conducted to achieve a greater spatial context of any possible historic
features. Transects were 3 to 5 meters in length with half-meter intervals. Interval distance was
reduced due to the depth of potential targets and smaller elliptical radiation footprint. GPR
surveys and soil sample collections were conducted two times over a period of one month.
Table 3-3. Data points for the four corners of the historic archaeological survey area. The
proj section system used was UTM, NAD 1983, Zone 17 North.
Northwest Corner: 17R 0463935 Northeast Corner: 17R 0463926 W
3322503 3322502 N
Southwest Corner: 17R 0463933 Southeast Corner: 17R 0463925
Prehistoric Archaeological Site
Preliminary archaeological work at the Montverde site (8LA243) was conducted in 2004
due to possible future development. Independent to the research work, a contract archaeological
firm (SEARCH) began testing with a subsurface survey of 87 shovel tests throughout the proj ect
area. Prehistoric artifacts were recovered from 50 shovel tests; the center and the southwest
corner of the grove were the only areas void of cultural material (Austin 2005). Excavations units
were then placed in areas of high artifact concentration, along the eastern proj ect area. The units
were situated between the citrus grove edges and the lake shore. The site area extends about 25
hectares across an active orange grove. The site area has a marked slope (>5%) from west at 36
m to the lake shore at 21 m above mean sea level.
The scope of the GPR research site area was confined to potential archaeological
concentrations. A GPR survey was conducted from the block excavation areas between previous
T.U. 1 and T.U. 2. The survey began from an Orlando series (Inceptisol) and continued upslope
in to the Lake series (Entisols). A total of 24 transects were conducted in an east to west
direction in order to determine the transition of the umbric epipedon to an ochric epipedon.
Transects were one meter apart and restricted in grid size due to presence of orange trees, and
were referenced off existing excavation unit datum points and handheld GPS receivers.
Results of interpreting GPR profiles and soil analyses are separated into sections covering
the four research sites. Each section will include: (i) physical and (ii) chemical analyses of
anthropogenic and natural soils, (iii) GPR profiles with the 500-MHz antenna and (iv) 900-MHz
antenna of anthropogenic and natural soils. Comparison of all four sites will be covered in the
discussion chapter. A summary of morphological pedon descriptions and selected physical and
chemical properties are given in Appendices A and C. Additional GPR profiles suing the 500-
MHz and 900-MHz antennas are given in Appendix D.
Located in western Alachua County, Florida, the Newberry cemetery was selected for this
study based on soil Order (Alfisols) and on anthropogenic context. Research was conducted to
see the effect of the soil on GPR to detect unmarked graves. In order to asses the suitability of
GPR with the soil properties, two different frequency antennas (900-MHz and 500-MHz) were
compared during the survey. The soils from the site area are mapped as well developed Alfisols
(Thomas et al. 1985). These soils are classified as a Pedro- Jonesville complex, with varying
depths of argillic horizon; which are so intermixed, that they cannot be separated at the scale of
mapping (Thomas et al. 1985). The marked graves are from the early twentieth century; several
unmarked graves are located within proximity of only a few meters. This site addressed several
obj ectives of the study: the application of GPR to verify the existence of a clandestine grave
while meeting the criteria of a specific Florida soil Order.
Soil Physical Analysis
Particle-size distribution of native soils at this site indicated soil texture ranging from fine
sand to clay textures. The Newberry cemetery natural soils (NCNS), native Arenic Hapludalfs,
were approximately 91% medium to very fine sand throughout the A and E horizons, with less
than 4% silt or clay. The Ap horizon had a silt content of 3.5 % with a decrease to 2.3% in the E
horizon. The amount of clay increased from < 4% in the Ap and the E horizons to >21% in the
Bt horizon. The silt content increased to < 6% in the Bt horizon for the native soils.
The particle-size distribution of the Newberry cemetery anthropogenic soils (NCAS)
ranged from medium grain sand to a clay fraction. The mixed horizons of the NCAS varied
widely in particle-size distribution. Soils from suspected surveyed graves were ranged
approximately from 86% to 91 % medium to very fine sand in the Ap horizon, with ranges of
4%-6% silt and 3%-6% clay. Soils from the altered El horizon (45-60 cm) had a backfill
scattering that ranged from 11.6% to 42.2% clay. (Figure 4-1).
30 o-% Clay
[L15 -m-% Clay
10 V (pedogenic)
0-15 (A) 45-60 (E1) 83-98(E2) 130-142(Bt)
Figure 4-1. Depth distribution of total clay content of natural transectt 17) and anthropogenic
transectt 14) soils at the Newberry cemetery.
Medium to very fine sand ranged from 64.2% to 85.1% and the silt content was 5%-6% in the
soil from the anthropogenic altered E horizon. Clay content increased in the associated Bt
horizon in the anthropogenic soil to 41.2%. When the percentage of clay is compared with depth,
clay increased in the NCAS at 45-60 cm along transect 14
Soil Chemical Analysis
The acidity of the NCNS ranged from a pH 5.2 to 8.5 (Appendix C-1). The NCNS along
transect 17 had a pH range of 5.2 to 6.1, with lower pH values located in the E horizon and
higher pH associated with A and Bt horizons. The Bt horizon had 10% regolith, indicating
underlying limestone. Soil acidity was contingent upon depth to limestone. Soil samples
collected near the limestone had pH values higher than any other horizon.
The NCAS had a wide pH range of 5.3 to 8.4 across the site area (Figure 4-2). Cemetery
samples had a discernible pattern of pH change based on sample depth. Soil pH increased along
transect 14, from 7.5 at the Ap horizon (15 cm) to 8.4 at a depth of 115 cm. The pH range for
soils sampled from suspected burials is higher due to the mixing of soil horizons, including
limerock and possible diagenesis of bone.
6 -o pH Anthropogenic
5 -1 soil
4 -( -m- pH Natual soil
15 45 85 115 145
Figure 4-2. Depth distribution of Transect 14 pH values and Transect 17 pH values at the
The OC content of NCNS, determined by Walkley-Black digestion, was < 5 g/kgl in all
horizons (Table 4-1). The OC content decreased with depth until the Bt horizon. The NCNS Ap
horizon contained 2.61 % OC. This decreased in the E horizons to approximately 0.35 %
average. The OC content increased in the NCNS Bt to 1.25%. Anthropogenic soil OC content
was similar with the natural soils. The Ap horizon contained 2.8 % OC. A decrease in the E and
Bt horizon occurred to approximately 0.48% and 0.42 % of OC respectively.
The EC (EC) of the NCNS had low values (Table 4-1), ranging from 0.74 to .04 ds/ml
The EC values decreased with depth in the NCNS E horizon, .04 to .06 ds/m- The Bt horizon's
highly conductive clay media had the highest values of .41 ds/m l. The EC of the anthropogenic
soils had a higher range of values. The values for the Ap horizon were .74 ds/m- The EC values
for the E/Bt mixed horizons had a smaller decrease to .35 ds/ml on average.
Table 4-1. Selected soil chemical analysis of the Newberry Cemetery natural and anthropogenic
soils along transects 14 and 17. The EC (EC), extractable (Ext.), Organic carbon
(OC), of the Newberry cemetery natural soils (NCNS), and Newberry cemetery
anthropogenic soils (NCAS).
Soil Depth Horizon EC Ext. P Ext.Ca Ext. Mg OC
(cm) ds m-1 mg/kg mg/kg mg/kg %
NCNS 0-15 Ap 0.28 12.72 188.72 28.06 2.61
Transect 45-60 El 0.06 9.88 60.13 3.94 0.48
17 85-100 E21 0.06 11.96 12.69 3.5 0.23
116-131 E22 0.04 7.1 8.56 1.51 0.23
145-160 Bt 0.41 40.12 1209.4 21.22 1.25
NCAS 0-15 Ap 0.74 32.1 2693.1 53.2 2.8
Transect 45-60 El 0.4 189.8 2141.4 17.2 0.48
14 83-98 E2 0.28 36.6 1167.7 13.1 0.48
115-130 Bt 0.38 80.5 978.6 27.7 0.42
Ext P contents in the NCNS (Table 4-1) ranged from 7.1 to 40.12 mg/kg-l and increased in
depth except between the Ap and El horizons. The Bt horizon had a four-fold increase of Ext P
than from the overlying El and E2 horizon. Overall, the NCAS contained greater values of
extractable P than the NCNS (Table 4-1). Ext P content increased markedly in the El horizon
and reduced to lower values in both the overlying Ap and underlying E2 horizons. The greatest
values of Ext P occurred where the mixing of soil horizons was most pronounced. Ext P content
increased again in the Bt horizon, more than twice as much than the overlying E2 horizon.
Concentrations of Ext Ca in the NCNS ranged from 8.5 to 1209.4 mg/kg with the lowest
value occurring in the E2 horizon and the highest amount in the underlying Bt horizon (Table 4-
1). Proximity of the underlying limerock to the Bt horizon facilitates increased values of Ca in
NCNS. The Ap horizon contained higher values of Ca than that of the underlying El horizon.
Overall values of Ca in the anthropogenic soils were significantly higher in all recorded depths.
Ca content for the NCAS samples were at the high end of the range of all soils in the site area.
Unlike the NCNS, the overall Ca values decreased with depth; both the natural and
anthropogenic soil samples had similar Ca values in the Bt horizon.
Both the natural and anthropogenic soils contained comparable amounts of Mg throughout
(Table 4-1), with the highest values occurring in the Ap horizon and lowest in the El and E2
horizons. The content of Mg in both soils increased in the Bt horizon from overlying El and E2
Soils chemical analyses were conducted to determine the differences between natural and
anthropogenic soils. The NCNS have low values of P, and Mg; while increased Ca values are
associated with depth to limestone. The NCAS have increased values of P, Mg, and Ca from
suspected buried remains. The means of these three elements from natural soils were compared
to those of anthropogenic soils by calculating t-tests (Appendix E)
Samples were collected from both natural and anthropogenic soils at the Newberry site
area. The NCNS from the site were used as the control for determination of potential natural
soils. The NCNS from the site area were lower in extractable P concentrations than from the
anthropogenic soils (t = -3.04, p = 0.01019) (Appendix E-1). The mean average for native soils
was 20.5 mg/kg, while the NCAS had a mean average of 68.4 mg/kg. The overall Ca values of
both the NCNS and NCAS were significantly different. Mixing of the anthropogenic soil
horizons fundamentally altered the concentrations of Ca (t = -3.92, p = 0.00153) (Appendix E-1).
Variability of Ca values throughout the soil horizons is reflected in the standard deviation that is
more than double the NCNS. Concentrations of Mg did change between the NCNS and the
NCAS (t = -2.53, p = 0.09874) (Appendix E-1), but not to the extent of the P and Ca
GPR-Soil Results for Newberry Cemetery
The 500-MHz antenna
Data acquisition of both NCNS and the NCAS soils with the GPR using the 500-MHz
antenna created approximate profiles with distinctive reflections caused by contrasting soil
properties. The NCNS has a planar stratification with occasional voids from the karst subsurface
(Figure 4-3). Multiple, continuous contrasting layers in the NCNS soil represent the presence of
an argillic (Bt) horizon. The depth to argillic horizon in Figure 4-3 is approximately 0.9 to 1.1 m.
Soils associated with Typic and Arenic Hapludalfs are often not stratigraphically homogenous.
Bioturbation, such as animal burrows and tree roots, will frequently alter soil horizons.
Processing steps for the horizontal and vertical correlation of the NCNS and NCAS
profiles include marker interpolation for distance (meters) and static correction time cuts for the
removal of antenna ringing and velocity adaptation (Figure 4-3a). The lines superimposed on the
GPR profiles in Figure 4-4 are manual velocity adaptation picks displayed to highlight the
argillic horizon. After the soil surface and horizontal alignment was determined, a background
filter was applied to the transect profile. In order to accentuate and highlight a horizon interface,
spectral whitening was applied to the GPR profiles when necessary (Figure 4-3b).
A discernable difference in interpreting GPR profiles is from the color scale; this is often
the preference of the GPR operator. However by viewing multiple color scales, subtle images
may become enhanced. In the anthropogenic soils, differences in color scale assisted in the
overall interpretation. The profiles viewed in Rainbow 2 color scale from Figure 4-3a is shown in
Grey 1 color scale (Figure 4-3c).
Energy decay was a final processing step chosen for select GPR profiles to augment data
interpretation. The process extracts an energy decay curve of the line and applies the inverse of
this function on the data. By activating this function a gain curve in the time direction (y-axis) is
applied on the complete profile based on the mean amplitude decay curve. This step was
primarily for highlighting the presence or absence of altered, heterogeneous anthropogenic soils
(Figure 4-4). A pick was inserted into selected profiles to highlight subsurface amplitude value
changes. Such changes included the interface of overlying horizons and the argillic horizon
(Figure 4-3a, c).
The GPR profiles of the NCAS produced distinct non-point source anomalies in both the
E2 and Bt horizons. Due to the soil properties associated with the anomaly and stratigraphic
break of the horizons, radar scatter occurred in the GPR profile (Figure 4- 4a). A low-amplitude
reflection with little or no energy returned to the 500-1VHz antenna. This is demonstrated with
the absence of pronounced horizontal banding within the anomaly. Application of specific
processing procedures may either highlight or mask the selected feature. The processing step of
energy decay enables the operator to both highlight the grave and the backfill overburden (Figure
4-4b). With the application of spectral whitening, used for the NCNS profiles, the subtle image
of overburden is less pronounced (Figure 4-4c). GPR profiles with Grey 1 color scale augment
energy decay and spectral whitening processing features while reducing background noise
8 9 10 11 12 13 14 15 16 17 15 19
0~ ~~ 131 2 6 6 7 10 11 12 3 14 1 16 1 18 1
20~~~~~~ ** "-t gil c H o
10 1 2 3 1 1 1 3 1 1 6 1 8 1
'.''C~-~ ~.ArgCHli2ll Horzo 1c~ ,.
Fiue43 Gz-roudpntaigrdrpoie fteAei aldl sn 0-~ nen
aln rnet1.A akron itri ple n atnaniei e oe.I
this~~ prfie bohteatnaniean h oiotlbnin r eoe ihn
loso ntrrttvedt.Veoiyaapainpiki nere ohihih heagli
horizon B) The depth to the a olrgillcE hoEIriznhsbe etfe n smr
prnune whe spcta whtnn is aple wit a bakron fitr Argic14 15 16 17 18 1
hoio is hihihe wit maua pik C) Bakron nos an arili horzoi
moeprmnntwt fle an Gra color, scal appied
2 3 4 567
0 10 20
0 00 OOO
Fiue44 rudpntrtn aa rflsofnhooei lee Aei aldl sn
The~~ ~ ~ 90-9 nen
subsurfae disGontinuitiesalsoincreasda from tres ofaropotstump antrd Arocks. Mauany ofthse
discntiuites or 'noise were rnot etcted in tern 50-Hz fantuenna. Any usute o pchrangsi P
profleswihdisteinctive rdefletion caued by rcosntrasting soil properies weemreapreti
the90Mzatna h nrae eouinfo he 900-MHz antennawaofstbth
lessened depth of penetration. Radar signals were affected by both the type of anomaly and
depth. Coherent signals were attenuated with specific anthropogenic features.
The 900-MHz antenna was also affected by the depth to the argillic horizon. The depth of
the argillic horizon influenced the higher frequency antenna. On many transects the argillic
horizon was barely visible, or not detected using the 900-MHz antenna (Figure 4-6), but was
strongly evident with the 500-MHz antenna (Figure 4-3).
Figure 4-5. Ground-penetrating radar profile of anthropogenic altered Arenic Hapludalf using
the 500-MHz antenna along transect 14. A) Filtering of horizontal banding and
background clutter is enhanced in Gray 1 color scale. B) Overburden is more
prominent when the energy decay processing step is applied in the Gray 1 color scale.
C) Note the very low amplitude reflection at the 9.5 to 12 m transect range is more
discernable when the Gray 1 color scale is used.
Figure 4-6. Ground-penetrating radar profile of natural Arenic Hapludalf using the 900-MHz
antenna along transect 17. Velocity adaptation pick is inserted to highlight the argillic
horizon. A) Filtered profile reduces horizontal banding. The argillic horizon is less
discernable than from the 500- MHz antenna. B) Spectral whitening step is added to
filtered profile, highlighting the Argillic horizon. C) Filtering with spectral whitening
process combined with gray scale reduces horizontal banding significantly.
Anomalies detected by the 900-MHz antenna from the NCAS along transect 14 (Figure 4-7) are
discernable with post-processing procedures. Filtered profiles with energy decay processing
applied are more discernable than with just the filter alone; both with Gray 1 and Rainbow 1
color scales (Figure 4-6 b, c). Even filtered, numerous micro-reflections are evident throughout
the GPR profile. This background 'noise' may draw focus from subtle sub-surface features.
B t..qi:a-r -.
A I =
f"iv~ r- ~c;
Figure 4-7. Ground-penetrating radar profile of anthropogenic Arenic Hapludalf using a 900-
IVHz antenna along transect 14. A) Background filter reduces subsurface
discontinuities throughout the profile. Low amplitude reflection from a feature is
discernable between the 9.5 to 12 m range along transect from 0.5 to 1.6 m in depth.
B) Energy decay processing step is added to background filtering, highlighting the
scattering effect of low amplitude reflections. Overburden is discernable overlying
grave. C) Energy decay with background filtering using Gray 1 color scale further
reduces subsurface discontinuities and isolates the overburden and underlying grave.
The Oakland cemetery is located south of Lake Apopka in western Orange County,
Florida. This site was selected for the native soils (Entisols) and the anthropogenic context
(cemetery). The soils from the site area are mapped as both Entisols and Ultisols (Doolittle and
Schellentrager 1989). These soils are classified as a Candler-Apopka complex; the mapped areas
consist of generally 66% Candler soil (Entisols) and similar soils, with about 31% Apopka soil
(Ultisols) and similar soils. This soil complex is too mixed to be mapped separately (Doolittle
and Schellentrager 1989). The marked graves are from the late nineteenth to early twentieth
century; several unmarked graves are located within proximity of the existing cemetery. This
site, similar to the Newberry cemetery, addressed several objectives of the study: the application
of GPR to verify the existence of a clandestine grave while meeting the criteria of a specific
Florida soil Order.
Soil Physical Analysis
Particle-size distribution of Oakland cemetery natural soils (OCNS) was dominated in all
samples by the medium grain to very fine grain sand (Appendix A). The soil data were compared
to the information from the Orange County Soil Survey (Doolittle and Schellentrager 1989). The
site area has been mapped as Candler and Apopka-Candler complex. Candler soils are a
hyperthermic, uncoated Lamellic Quaritzipsamments. The Candler soils series consists of
excessively drained soils, an ochric epipedon, approximately 90% medium to very fine sand in
the A and C horizons, and less than 3% silt and 2.5% clay. The ochric epipedon had an average
silt content of 1.8%. The Candler series at the Oakland cemetery has clear, wavy lamellae at
approximately 1.0 to 1.2 m in depth. The amount of clay content increased from < 1.2% in the
Ap horizon, to 2.5% in the C3 horizons in association with the lamellae. At the Oakland site
area, the Candler soil series is associated with the Apopka series. Apopka soils are loamy,
siliceous, hyperthermic Grossarenic Paleudults. These two soils are morphologically very similar
in the upper 1.5 m. The Apopka series has an argillic horizon within 2.0 m of the surface.
Physical soil analyses from the site determined the absence of a Bt horizon, with the default
classification being a Candler soil series. Further discussion of taxonomic and morphological
classification of the Oakland cemetery soil will be addressed in Chapter 5 and Appendix A.
Oakland cemetery anthropogenic soils (OCAS) were dominated in all samples by medium
to very fine sand. Associated with OCNS, the mixed horizons of the altered soils did not change
noticeably in particle-size distribution. Particle-size distributions of the OCAS horizons were
dominated in all samples by the sand fraction, with a range of 95.1 to 97.9%, with low values of
silt (< 2.7%) and clay (< 2.6%). Potential surveyed graves displayed no significant change in
physical properties from the OCAS.
Soil Chemical Analysis
The acidity of the Oakland cemetery soils ranged from a pH 5.9 to 7.1 (Figure 4-8). The
OCNS had a pH range of 4.9 to 5.3, with lower pH values located in the Ap horizon and higher
pH associated with Cl horizons. Soil acidity for the OCNS was not significantly variable, with a
pH range of 4.9to 5.3 of all horizons. Orange groves are located within 75 meters upslope on the
east and north sides of the site area (Figure 2-3). This tertiary influence on pH and other
chemical analyses will be discussed in the next chapter.
The OCAS had a pH range of 6.0 to 7.2 across the site area. Cemetery samples had a
pattern of pH change based on sample depth. Soil pH increased varied from 6.0 in the lower C
horizon (85cm) to 7.2 (50 cm) at two grave samples (Figure 4-8). The pH range for soils sampled
from suspected burials is higher due to the mixing of soil horizons, including possible diagenesis
I- pH Antropogenic soil
-m- pH natural soil
15 50 85 120
Figure 4-8. Depth distribution of pH values for Transect 2 at the Oakland cemetery.
The OC content of OCNS, determined by Walkley-Black digestion, was < 2 % in all
horizons along transect 2 (Table 4-2). The OC from the OCNS samples decreased with depth
until the C3 horizon. The Ap horizon contained of 1.32 % OC. This decreased in the Cl horizon
to 0.1 %, C2 horizon to 0.16 % and to a 0.29 % in the C3 horizon. The OCAS OC content was
slightly higher than native soils in the Cl horizon and C2, with lower values in the C3 horizon.
The Ap horizon in the OCAS had lower values of OC, with 0.55 %. A decrease in the
anthropogenic Cl horizons occurred to approximately 0.42 % of OC, with 0.48 % in the C2
horizon. The OCAS C3 horizon had lower values than the OCNS C1, C2 horizons with
approximately 0.03 % of OC.
The EC (EC) of the OCNS had low values from transect 2 (Table 4-2), ranging from 0.34
to 0.06 ds/m- The EC values in the Ap horizons of the OCNS were approximately 0.34ds/m l,
and decreased with depth in the Cl horizons, C2 horizons and C3 horizons from .08 to .06 ds/m
1. The EC of the OCAS had a lower range of values. The EC mean average for the Ap horizons
was .14 ds/m- The EC values decreased to .14 ds/m-l in the Cl horizons, .06 ds/m-l in the C2
horizons and increased slightly to.26 ds/m-l in the C3 horizons.
Ext P content in the OCNS (Table 4-2) from transect 2 ranged from 12.8 to 4.29 mg/kg-l
and was variable in depth between the C1, C2 and C3 horizons. The Ap horizon had a value of
12.8 mg/kg- with the C3 horizons having the next highest overall Ext P concentration at 10.41
mg/kg- The OCAS contained greater values of extractable P than the OCNS (Table 4-2). Ext P
content of anthropogenic soils increased markedly in the Ap, with 58.84 mg/kg- with 70.48
mg/kg-l and in the Cl horizon. Lower values of Ext P occurred in C2 with 6.87 mg/kg- and 4.69
mg/kg-l in the C3 horizon. The greatest values of Ext P occurred at a depth of 0-50 cm, where
the mixing of soil horizons was most pronounced.
Concentrations of Ca in OCNS from transect 2 ranged from 26. 14 to 65.45 mg/kg-l with
the lowest value occurring in the lower C2 horizon and the highest amount in the Ap (Table 4-2).
Ext Ca decreased with depth until the C3 horizon (13 5 to 150 cm) with a value of 46.3 8 mg/kg- .
Overall values of Ext Ca in the OCAS were significantly higher in the upper 0.5 m. Ext Ca for
the OCAS ranged from 148.29 mg/kg-l in the Cl horizon to 10.92 mg/kg-l in the C3 horizon.
The OCNS contained lower overall values of Mg than the anthropogenic soils. (Table 4-2).
Ext Mg in OCNS ranged from 25.21 to 9.49 mg/kg- with higher values in the upper 0.5 m and
the lowest values occurring between 0.85 to 1.00 m. Ext Mg in the anthropogenic soils decreased
with values of 14.91 mg/kg-l in the C2 horizon and to 6.01 mg/kg-l in the C3 horizon.
Soils chemical analyses were conducted to determine the differences between natural and
anthropogenic soils. The OGSN have low values of P, Ca, and Mg. Anthropogenic soils have
increased values of P, Mg, and Ca from suspect buried remains. These means of these three
elements from NCNS were compared to those of the NCAS by calculating t-tests
Table 4-2. Selected values of chemical properties of the Oakland Cemetery natural and
anthropogenic soils along transect 2. The EC (EC), extractable (Ext.), Organic carbon
(OC), Oakland cemetery natural soils (OCNS), Oakland cemetery anthropogenic soils
Soil Depth Horizon EC Ext. P Ext.Ca Ext. Mg Org. C
(cm) ds ml mg/kg mg/kg mg/kg %
OCNS 0-15 Ap 0.34 12.8 65.45 25.21 1.32
Transect 35-50 Cl .08 8.02 47.49 12.46 0.1
2 85-100 C2 .06 4.29 26.14 9.49 0.16
135-150 C3 .06 10.41 46.38 13.05 0.29
OCAS 0-15 Ap 0.14 58.84 127 25.21 0.55
Transect 35-50 Cl 0.14 70.48 148.29 20.22 0.42
2 85-100 C2 0.06 6.87 29.91 14.91 0.48
135-150 C3 0.26 4.69 10.92 6.01 0.03
Samples were collected from both natural and anthropogenic soils at the Oakland cemetery
area. The OCNS from the site were used as the control for determination of suspect
anthropogenic soils. The OCNS from the site area were lower in extractable P concentrations and
variability than from the anthropogenic soils (t = -2.62, p = 0.1853) (Appendix E-2). The mean
for OCNS was 8.26 mg/kg, while the impacted soils had a mean of 34.4 mg/kg. The overall Ca
values of both the OCNS and OCAS were markedly different. Mixing of the anthropogenic soil
horizons altered the concentrations of Ca (t = -1.89, p = 0.2329) in the upper 0.5 m. Variability
of Ca values is limited to the upper soil horizons; this is reflected in the standard deviation that is
more than double that of the natural soils. Concentrations of Mg did change significantly
between the OCNS and OCAS (t = -1.69, p = 0.1625) (Appendix E-2).
GPR-Soil Results for Oakland Cemetery
The 500-MHz antenna
The OCNS has a homogenous matrix with little variation in physical properties (Figure 4-
9). The Candler series at this site has clear, wavy lamellae at approximately 1.0 to 1.2 m in depth
that is evident in the GPR profile with the 500-MHz antenna. Bioturbation, such as animal
burrows and tree roots, will be evident at depths less than 2 meters in the OCNS. Anomalies at
depths greater than two meters often entail targets larger than most bioturbation. Both OCNS and
OCAS GPR profiles are along the same transect.
Identical processing steps for the horizontal and vertical correlation of the GPR profiles at
the Newberry cemetery were used for the Oakland cemetery. This included marker interpolation
for distance (meters) and static correction time cuts for the removal of antenna ringing (Figure 4-
9a). After the soil surface and horizontal alignment was determined, a background filter was
applied to the transect profile. In order to accentuate and highlight a horizon interface, spectral
whitening was applied to profiles when necessary (Figure 4-9c). This step was primarily for
highlighting the presence or absence of altered, heterogeneous anthropogenic soils, but was also
used for the homogenous nature of the soils at the site area. Spectral whitening decreased the
hyperbolic tail associated with the anomalies in transect 2
The GPR profiles of anthropogenic soils at the Oakland cemetery produced distinct point
source anomalies in both the Cl and C2 horizons. Due to the soil properties associated with the
elongated and suspect graves, noticeable reflection hyperbolas occurred in the GPR profile
(Figure 4-9a). Medium to high-amplitude reflections with significant energy returned to the 500-
MHz antenna, enabling a distinct signal return separate from the surrounding matrix. This is
demonstrated with the presence of pronounced horizontal banding within the anomaly.
Background filters allow for a clearer image of an already distinct anomaly. With the application
of spectral whitening, the image becomes more pronounced (Figure 4-9c). A pick was inserted
into selected profiles to highlight subsurface amplitude value changes. Such changes included
the start of the lamellae from 1.0 to 1.5 m in depth.
10 20 30
Fiur 4-9. Grudpntaigrdrprfl sn h 0-2 atnawt w upc
bacgrondfilerignote hencrese conras ofsspectgrae and\ te reucio
of- hprboli tal )Bcgon itri pled nen os srmvdadGa
t colo scl sue.Cnrs i nacdadhprolcti sicesd
MFz anenna9 (Figurdee 4-9; wile thea inroieaused rsltion fromz athen 900-Hz atenn wasoffst
by shorter hyperbolic tails in processed profiles. The use of higher frequency antennas may limit
the operator to locating and interpreting anomalies at varied depths along a transect profile.
Anomalies detected by the 900-1VHz antenna from the anthropogenic soils along transect 2
maybe more discernable with post-processing procedures, but not necessary. Filtered profiles
with spectral whitening processing applied are no more discernable than with just the filter alone
(Figure 4-10 b); both with Gray 1 and Rainbow 2 color scales (Figure 4-10 a, c). Filtered, with
spectral whitening allows for less numerous micro-reflections throughout the profile. As with the
500-1VHz antenna profiles, velocity adaptation picks were inserted to the processed profiles to
highlight the lamellae from 1.0 to 1.5 m in depth (Figure 4-10).
Prehistoric Archeological Site
The prehistoric archaeological site is located in eastern Lake County, Florida on the
southwest shore of Lake Apopka. This site was selected for the study based on the both the soils,
(Inceptisol) and on the anthropogenic context (prehistoric). There are two soils are mapped with
the site area; a well-drained Orlando series (Inceptisol) and a very well-drained Entisols
classified as the Lake soil series (Furman et al. 1975). The Orlando series is the predominate soil
for this study. Two prehistoric archaeological components have been previously excavated in
this study area. Associated features within the prehistoric midden date as early as 1600 B.C.
(Austin 2006). This site addressed several objectives of the study: the application of GPR to
verify the extent of prehistoric anthropogenic influences while meeting the criteria of a specific
Florida soil Order.
10 00a:~ o
10l!c' 05;7 ,5 1
Figure 4-10. Ground-penetrating radar profile using the 900-MHz antenna of two suspect graves
in a Lamellic Quartzipsamment. A) Background filter is applied and antenna noise is
removed. Horizontal banding is still prevalent. B) Spectral whitening processing step
is added to background filtering, note the increased contrast of anomalies and the
reduction of hyperbolic tail. C) Background filter is applied, antenna noise is removed
and Gray 1 color scale is used. Contrast is enhanced and hyperbolic tail is increased.
Soil Physical Analysis
All horizon designations for this site were described morphologically and subdivided for
sampling purposes. The primary numerical suffixes are denote subdivisions within a master
horizon; e.g., Al, A2, C1, C2, and C3. Secondary numerical suffixes denote the sampling
subdivision within the horizon modifier. Particle-size distributions of the prehistoric
archaeological natural soils (PANS) at this site were dominated in all samples by medium to very
fine sand (Appendix A). The soil data was compared to the information in the Lake County Soil
Survey (Furman et al. 1975). The soil survey area has been mapped as a well-drained sandy,
siliceous, hyperthermic Psammentic Dystrudepts, with an umbric epipedon, approximately
93.5% coarse to fine sand throughout the A and C horizons, and less than 1% silt and 5.7% clay.
The umbric epipedon had an average silt content of 0.8% with a decrease to 0.3% in the C
horizon. The amount of clay content increased slightly from < 5.9% in the A horizon to 6.4% in
the C horizon.
Physical properties of the prehistoric archaeological anthropogenic soils (PAAS) at the
prehistoric site were very similar to the PANS. Both the PANS and PAAS had no morphological
differences in any of the horizons, except an A/C horizon at 0.9 to 1.05 m. The PAAS were
dominated with approximately 94.2% coarse to fine grain sands, less than 1% silt and 6.6% clay
in the A and C horizons. The silt content varied less than 0.7% among the A and C horizons.
Less than a 1% difference in clay content occurred among the A and C horizons in all
The acidity of the all prehistoric site soils ranged from a pH 4.3 to 6.9 (Figure 4-11). The
PANS had a pH range of 4.3 to 5.5, with lower pH values located in the A horizon and higher pH
associated at a depth of 0.35 to 0.75 m. Soil acidity for the PANS was variable, with a pH range
of 4.4 to 4.6 below 0.75 m. Orange groves are located within 5 to 10 meters upslope on the west
side of the site area (Figure 2-6). This tertiary influence on pH and other chemical analyses will
be discussed in the next chapter.
The PAAS had a pH range of 6. 1 to 6.9 along selected GPR transects (Figure 4-1 1).
Samples had a pattern of pH change, increasing with sample depth. Soil pH increased from 6. 1 at
the surface (0-15 cm) to 6.9 at the base of the C horizon (1.8 to 1.95 m). The OC content of the
PANS, determined by Walkley-Black digestion, was < 2.5 % in all horizons (Table 4-3). In
general OC amounts decreased with depth. The Al horizon contained 2.4 g/kgl OC. This
decreased in the A2 horizon to approximately 1.32 %, with 0.34 % in the C horizons. The PAAS
OC content was slightly higher than native soils in the C horizon. The A horizon in the PAAS
had slightly lower values of OC (1.96%) than the PANS (Table 4-3). A decrease in the PAAS A2
horizons occurred to approximately 1.55 % of OC, with 0.44 % in the C horizons. The PAAS C
horizon had slightly higher values than the PANS C horizons.
5 -o- IpH (natural soils)
-m- pH (anthropogenic
15 45 75 105 135 165 195
Figure 4-11. Depth distribution of pH values for transects at the prehistoric archeological site.
The EC (EC) of the PANS had values (Table 4-3), ranging from 1.36 to 0.04 ds/m-1. The
EC values in the Al horizons were approximately 1.36 and 0.62 ds/m-1, and decreased with
depth in the A2 horizons. The values in the C horizons were from 0.04 to .48 ds/m-1. EC of the
anthropogenic soils had slightly lower range of values. The EC values for the Al horizon was 0.4
to 0.28 ds/m-1. The EC values decreased to 0.26 to 0.22 ds/m-1 in the A2 horizons. The values
in the C horizons ranged from 0. 18 to 0. 14 ds/m-1. Overall values decreased with depth in both
the PANS and PAAS horizons. Ext P content in the PANS (Table 4-3) ranged from 343.2 to 70.9
mg/kg-1 and decreased in depth between the Al2 and C13 horizons. The Al2 horizon had
highest value of 353.5 mg/kg-1, with the C13 horizons having the lowest overall extractable
phosphorus concentration at 70.9 mg/kg-1. Anthropogenic soils contained much greater values
of Ext P than the PANS. Ext P content of anthropogenic soils increased markedly in the Al2 and
A21 horizons with lower values in C horizons.
Table 4-3. Selected values of chemical properties from natural and anthropogenic soils at
prehistoric archaeological site. The EC (EC), extractable (Ext.), Organic carbon (OC),
prehistoric archaeological natural soils (PANS), and prehistoric archaeological
anthropogenic soils (PAAS). Soil horizons subdivided for sampling depths.
Soil Depth Horizon EC Ext. P Ext. Ca Ext. Mg Org. C
PANS cm ds ml mg/kg mg/kg mg/kg %
0-15 All 1.36 278 310.4 28.4 2.4
30-45 Al2 0.62 343.2 353.3 36 1.62
60-75 A21 0.32 278 231.2 21.6 1.49
90-105 A21 0.38 203.6 112.6 13.3 1.1
120-135 C11 0.04 167.2 80.7 13.2 0.78
150-165 C12 0.6 78.4 39.9 10 0.13
180-195 C13 0.48 70.9 38.6 9.9 0.13
0-15 All 0.4 1400.2 2085 81.4 2.07
30-45 Al2 0.28 1617 2278.1 80.9 2.27
60-75 A21 0.26 1633.5 2347.4 64.7 2.07
90-105 A22 0.22 759.2 801.4 37.4 0.65
120-135 C11 0.18 710 648.6 34.3 0.58
150-165 C12 0.18 449.2 379.5 28.8 0.39
180-195 C13 0.14 388.4 303.4 24.3 0.19
The greatest concentrations of Ext P occurred at a depth of 0.30 to 0.75 m with values of 1617
and 1633.5 mg/kg- Concentrations of Ext P in the anthropogenic C horizon were lower than the
overlying the soil horizons.
Concentrations of Ca in the PANS ranged from 353.3 to 39.9 mg/kg-l with the lowest value
occurring in the lower C horizon and the highest amount in the Al2 (Table 4-3). Ext Ca
decreased with depth. Overall values of Ext Ca in the PAAS were significantly higher in all
horizons. Ext Ca contents for the PAAS ranged from 2347.4 mg/kg-l in the A21 horizon to
303.48 mg/kg lin the C13 horizon. Similar to the PANS the overall Ca values decreased with
depth throughout all horizons.
The PANS contained lower overall values of Mg than the PAAS (Table 4-3). Ext Mg in
the PANS ranged from 36.04 to 9.9 mg/kg-1, with higher values in the Al and the lowest values
occurring in the C horizon. The concentration of Mg in the PAAS decreased with depth with
values from 81.4 mg/kg-1 in the Al horizons to 24.3 mg/kg-1 in the lower C13 horizons.
Native Quartzipsammentic Haplumbrepts from the site were used as the control for
determination of potential anthropogenic soils. Elevated overall concentrations of extractables P,
Ca, and Mg were present throughout the site area; the potential reason for such high values will
be covered in the following chapter. The PANS from the site area were much lower in Ext P
concentrations and variability than from the PAAS (t = -4.71, p = 0.00597) (Appendix E-3). The
mean for the PANS was 224.73 mg/kg, while the impacted soils had a mean of 994 mg/kg. These
values are quite high in both the PANS and the PAAS. The overall Ca values of both the PANS
and the PAAS were noticeably different. Concentrations of Ext Ca in all horizons (t = -1.72, p =
0.01706) (Appendix E-3) were distinct. Variability of Ca values is not limited to any soil
horizons; this is reflected in the standard deviation that is more than five fold that of the PANS.
Concentrations of Mg did change between the PAAS and the PANS (t = -0.541, p = 0. 109), but
not with the significance of both the Ext P and Mg.
GPR-Soil Results for the Prehistoric Archaeological Site
The 500-MHz antenna
The PANS has little variation in physical properties. The Orlando series at this site has a
clear, morphological horizon separation; this depth is evident in the 500-MHz antenna GPR
profile (Figure 4-12). Even in the PANS, punctuated human disturbances are clear throughout
this site area. The GPR profiles of the PANS are along different transects than the GPR profiles
of the PAAS, but run through active orange groves. The GPR transects of the PANS were
conducted upslope and away from the known archaeological site areas, which were situated
along the lake shore. The differences between GPR profiles of the PANS and PAAS are in
context to archaeological significance for this study. Both soils have been altered through human
activities; the difference is temporal instead of spatial and will be covered in the discussion
Identical processing steps for the horizontal and vertical correlation of the GPR profies of the
previous sites were used for the prehistoric site. This includes marker interpolation for distance
(meters) and static correction time cuts for the removal of antenna ringing. After the soil surface
and horizontal alignment is determined, a background fi1ter was applied to each transect profie.
In order to accentuate and highlight a horizon interface, spectral whitening was applied to GPR
profiles when necessary. This step was primarily for highlighting the presence or absence of
altered, heterogeneous anthropogenic soils, but was also used for the homogenous nature of the
epipedon at this site.
The GPR profiles of both the PANS and PAAS produced distinct point source anomalies
and planar subsurface interfaces in both the A2 and C horizons. Due to the soil properties
associated with the elongated and perpendicular anomalies of subsurface irrigation lines,
noticeable reflection hyperbolas occurred in the GPR profile (Figure 4-13a). Medium to high-
amplitude reflections with significant energy returned to the 500-MHz antenna, enabling a
distinct signal return separate from the surrounding matrix.
With planar stratification in the soil horizons, removal of horizontal banding greatly
enhances any anomalies. Background filters also allow for a clearer image of an already distinct
anomaly. With the application of spectral whitening, images become more pronounced and non-
point source anomalies are highlighted (Figure 4-13c).
0 10 20
D i"06 I
A~ Horizon on -
Figure 4-12. Ground-penetrating radar profies of a natural Psammentic Dsystrudepts using the
500-1VHz antenna. A) Background fi1ter is applied and antenna noise is removed.
Velocity adaptation pick is inserted to highlight the C horizon. B) Spectral whitening
processing step is added to background filtering, note the increased contrast between
the A and C horizon. C) Background fi1ter is applied, antenna noise is removed and
Gray 1 color scale is used. Contrast is enhanced between the A and C horizon with
velocity adaptation pick is inserted.
0 10 20
Irrigation Dine -
Figure 4-13. Ground-penetrating radar profiles of an anthropogenic Psammentic Dsystrudepts
using the 500-MHz antenna. A) Background filter is applied and antenna noise is
removed. B) Spectral whitening processing step is added to background filtering, note
the increased contrast between the A and C horizon. C) Background filter is applied,
antenna noise is removed and Gray 1 color scale is used. Hyperbolic tail and contrast
between the A and C horizon is enhanced. Feature banding is highlighted with change
in color scale.
Varying color scales accentuate subtle changes within the GPR profile, allowing the
operator multiple views for data interpretation. A difference in interpreting GPR profiles when
subsurface variation is nominal, offers an advantage with specific sites. In the anthropogenic
soils at this site differences in color scale assisted in the overall interpretation of subsurface
The 900-MHz antenna
The GPR profie of the PANS using the 900-MHz antenna exhibited no distinct reflections
from all horizons. Data acquisition of both PANS and the PAAS with the GPR using the 900-
MHz antenna created shorter wavelengths, thereby increasing resolution of radar images (Figure
0 10 20
I~~l _1 4 Horznnn ji
CL. -Or+Or- 1 er
10~~~- A oizn
-II:L I i E06
f ~SC Horizon 1
30 .l1 5
noise is furhe reovd C) Bakron 1ter is apied anea H znos isreovd
Figuran spectra whin-pntening randa Gray color sal iAs usmed.Futher contrat ips usomewhat
apparent. Contrast is enhanced between the A and C horizon with velocity adaptation
pick is inserted.
0 10 20
Irrigation Dine Fealture r 1
30 1 5
s: :,;5; '.. 1,6,
Figure ~ ~ ~ ~ ~ ~ 3 4-15. Grudpnertn raa rflsoa nhooei smetcDytuet
using~~~~~~~~~~~~~~ th 0-Mzaten.A) Bckgon fite s plidan ntnaosei
reovd B) Spcrlwieigpoesigse saddt acgon trnnt
the~~~~~~~~~ inrese cotatbtenteAadChrzn.C akrudfle sapid
antnn nos srmvdadGa oo cl i sd yeblcti srdcdi
Thgue -5 GPRon-etrig a profiles of th ASpoie nl difuethratigrapic resoluntion Anystubetle
changes ing GPrflswthe 0~ ditnctiena refectirounsasd byte i cotastlieng soiln proerie s wr
aparntinthe anthrops ognic soil usieng the 900-M z antenona C (Figuren 4-1).The increased,
apresoltion from atheooi 900-M s antenna wa ofet by~ shotern hypeurboi tails ofthe irrigation ip
in processed GPR profiles of the PAAS. A pick was inserted into selected profiles to highlight
subsurface amplitude value changes. Such changes included the A and C horizon interface. A
suspected feature is located near the irrigation pipe; this might be associated with the excavation
of the pipe (Figure 4-15). This suspected feature is less prominent in the GPR profile using the
900-MHz antenna, then with the 500-MHz antenna.
Historic Archeological Site
The British Period occupation site is located in northeastern St. Johns County, Florida,
west Tolomato River. This site was selected for the study based on the both the soils,
(Spodosols) and on the anthropogenic context (historic). The two soil Orders mapped in the site
area are weakly to moderately developed Spodosols with associated Entisols; and are classified
as either Myakka/Immokalee (Spodosols) with Paola (Entisols) soils (Readle 1983). The
Immokalee series is the predominate soil for this study. The archaeological context of this site is
primarily the British Period (1763-1784 A.D.), with Spanish and prehistoric components
associated with the site area. This site addressed several objectives of the study: the application
of GPR to verify the extent of historic anthropogenic influences while meeting the criteria of a
specific Florida soil order.
Soil Physical Analysis
The particle-size distribution of the historic archaeological natural soils (HANS) at this site
was dominated in all samples by the medium to very fine sand fraction throughout (Appendix
A). The soil data was compared to the information from the St. Johns County Soil Survey
(Readle 1983). The sampled site area has been mapped as the Immokalee, sandy, siliceous,
hyperthermic Arenic Alaquods series with approximately 95% medium to fine sand in the Ap
and El, E2, and E3 horizons, with less than 2.5% silt and 3% clay. The underlying Bh horizon is
from 0.74-0.93 m, with 2.4% silt and 2.8% clay. The ochric epipedon had an average silt content
of 1.7% with a linear decrease to 0.9% in the E3 horizon. The amount of clay and silt content
increased slightly in the Bh horizon.
The historic site area has another associated soil, a hyperthermic, uncoated Spodic
Quartzipsamment. This soil is the St. Johns series and has very similar morphological
characteristics to the Immokalee. Due to the depth of water table, the St. Johns series may be
more closely relevant to the site. However, from a morphological interpretation, the Immokalee
series will be referred in this chapter as the dominant soil series.
The human impacted soils at the historic archaeological site were dissimilar to the native
Arenic Alaquods. The historic archaeological anthropogenic soils (HAAS) had soil color
differences, lower values and chroma (Appendix A-4) in the horizons El and E2 horizons from
0.2 to 0.45 m. The HAAS were dominated with approximately 93.7% medium to fine sands, with
values of less than 6.0% silt and 3.5% clay in the A, El, and E2 horizons. The Ap horizon had a
mean average of 3.4% silt and 0.9% clay. The El horizon, from 0.20 to 0.25 m, had a mean of
1.9% silt and 1.6% clay; while the underlying E2 horizon had 1.5% silt and 1.7% clay at 0.40 to
0.45 m below surface. Overall silt and clay values are elevated in the El and E2 horizons when
compared to the HANS.
Soil Chemical Analysis
The acidity of the all prehistoric site soils ranged from a pH 4.7 to 7.8 (Appendix C-4).
Representative samples were taken from a HANS transect and a HAAS transect (Figure 4-16).
The HANS had a pH range of 4.7 to 5.4, with lower pH values located in the E horizons and the
higher pH values associated with the Ap horizon. The HAAS had a higher pH range of 6. 1 to 7.8
across the site area (Appendix C-4). Samples had a pattern of pH change, increasing with
sample depth. Selected soil pH from transect 1 increased with 6.1 at the surface (0-5 cm) to 7.5
and 7.6 in the E horizons (20 to 45 cm) (Figure 4-16).
5 1 -o ~pH (natural soils)
-m- pH (anthropogenic
5 25 45 65 85
Figure 4-16. Depth distribution of selected pH values for transects at historic archaeological site.
Historic site data did not exceed 0.50 m.
The OC content of the HANS, determined by Walkley-Black digestion, was < 6 % in all
horizons (Table 4-4). The OC content decreased with depth in the El and E2 horizons, and
increased in the underlying E3 and Bh horizons. The Ap horizon contained 5.07 % OC. This
decreased to approximately 0.32 %, with 0.26 % in the El and E2 horizons. The Bh horizons
increased to 1.8 %. Both the PANS and anthropogenic soils did not differ significantly in the Ap
horizon. Anthropogenic soil OC content was higher than native soils in the El horizon.
Decreases in the anthropogenic E2 horizon occurred to approximately 0.64 % of OC, yet was
double that of the PANS E2 horizon with 0.25%.
The HANS had values of EC (Table 4-4), ranging from .42 to .02 ds/m- The EC values
in the Ap horizons had the highest PANS values with approximately .42 ds/m l, and decreased
with depth in the E2 horizons, and E3 horizons from 0.04 to 0.02 ds/m- A slight change
occurred in the Bh horizon to a value of 0.06 ds/m- The EC of the anthropogenic soils had
slightly higher range of values. The EC mean average for the Ap
Table 4-4. Selected mean values of chemical properties from PANS and anthropogenic soils at
prehistoric archaeological site. Electrical conductivity (EC), extractable (Ext.),
Organic carbon (OC), historic archaeological natural soils (HANS), historic
archaeological anthropogenic soils (HAAS).
Soil Depth Horizon EC Ext. P Ext. Ca Ext. Mg Org. C
(cm) ds m-1 mg/kg mg/kg mg/kg %
HANS 0-4 Ap 0.42 14.3 7.4 47.6 5.07
4-16 El 0.02 1.1 5.1 3.4 0.32
16-46 E2 0.04 1.2 3.2 2.1 0.26
46-74 E3 0.02 1 3.5 1.4 0.26
74-93 Bh 0.06 27.8 23.4 8.7 1.8
HAAS 0-5 A 0.80 64.3 759.8 130.1 4.49
20-25 El 0.42 383.2 1390.8 50.9 1.28
40-45 E2 0.14 130.8 339.7 9.8 0.64
horizons was 0.80 ds/m- The EC mean values decreased to 0.42 ds/m-l in the El horizons, and
0.14 ds/m-l in the E2 horizons. Overall values decreased with depth in both the PANS and
anthropogenic soil horizons in the upper 0.5 m.
The HANS had Ext P contents that ranged from 1.1 to 27.8 mg/kg-l and decreased in depth
from the Ap to E3 horizons. In the Bh horizon the concentration of Ext P increased. The Ap
horizon had a value of 14.3 mg/kg- with the values stabilizing in the El, E2, and E3 horizons at
1.1 to 1.2 mg/kg- The highest concentrations occurred in the Bh horizon with mean values of
27.8 mg/kg- The HAAS contained more Ext P than the HANS (Table 4-4). Extractable P
content of anthropogenic soils increased distinctly in all horizons. The greatest value of
extractable P occurred in the E2 horizon with 383.2 mg/kg- where morphological alteration was
most prominent. Concentrations of Ext P in the HAAS for the A horizon was 64.3 mg/kg-l and
130.8 mg/kg-l for the E2 horizon.
Concentrations of Ext Ca in the HANS ranged from 3.2 to 23.4 mg/kg-l with the lowest
value occurring in the E2 and the highest amount in the Bh horizon (Table 4-4). Concentrations
of Ext Ca decreased with depth from the Ap horizon (7.4 mg/kg- ), until the Bh. Overall values
of Ext Ca in the HAAS were significantly higher in all horizons. Concentrations of Ext Ca
content for the anthropogenic soil ranged from a low of 339.7 mg/kg-l in the E2 horizon to
1390.8 mg/kg- soils in the El horizon. The Ap horizon also had an elevated concentration of
extractable Ca in the HAAS.
The HANS contained lower overall values of Mg than the HAAS. (Appendix C-4).
Concentrations of Ext Mg in the HANS ranged from 1.4 to 47.6 mg/kg- with higher values in
the Ap and the lowest values occurring in the E3 horizon. The concentration of Ext Mg in the
HAAS also decreased with depth with values from 130.1 mg/kg-l in the Ap horizons to 9.8
mg/kg-l in the E2 horizons (Table 4-4).
The HANS from the site were used as the control for determination of suspected
anthropogenic soils. The HANS from the site area were much lower in extractable P, Ca and Mg
concentrations and variability than from the anthropogenic soils (Appendix E-4). Extractable P
values differed in the HANS than from the HAAS (t = -2.97, p = 0.0934) (Appendix E-4). The
mean for the HANS was 9.65 mg/kg, while the HAAS had a mean of 85.6 mg/kg. The overall Ca
values of both the HANS and the HAAS were significantly different. Concentrations of Ca in all
horizons (t = -2.87, p = 0.0601) varied between soils. Variability of Ca values is reflected in the
standard deviation that is more than ten fold that of the HANS (Appendix E-4). The overall
concentrations of Mg did change between the HAAS and the HANS (t = -1.71, p = 0. 1751i), but
not with the significance of Ext P and Ca.
GPR-Soil Results for the Historic Archaeological Site
The 500-MHz antenna
The natural soil of the HANS has a diverse matrix with variation in physical properties
between the E and Bh horizon (Figure 4-17). The Immokalee series at this site has clear, abrupt
spodic horizon at approximately 0.74 to 0.93 m. This Bh horizon is evident in the GPR profile
from the 500-MHz antenna. Bioturbation, such as tree roots, are evident at depths less than 1
meter (Figure 4-17). The GPR profiles of the HANS and the HAAS are along different transects,
but with 25 m of each other. The length of the GPR transect for the HANS is 3 meters; while the
GPR transects of the HAAS are 2 to 3 meters in length. The length of the anthropogenic GPR
transects were based on the size of excavation unit coverage. In the GPR profile, wall fall debris,
wall footer and a floor from a structure are highlighted in Figures 18 and 20.
Identical processing steps for the horizontal and vertical correlation of the GPR profiles for
the natural soils and the anthropogenic soil at the Newberry cemetery, Oakland cemetery and the
prehistoric site were used for the historic site. This includes marker interpolation for distance
(meters) and static correction time cuts for the removal of antenna ringing (Figure 4-17a-c). In
order to accentuate and highlight the Bh horizon interface, spectral whitening was applied to
profiles when necessary (Figure 4-17b). This step was primarily for highlighting the presence or
absence of the altered, heterogeneous HAAS, but was also used for the homogenous nature of
the soils at the site area. A pick was inserted into selected profiles to highlight subsurface
amplitude value changes. Such changes included a Bh horizon and wall footer. The GPR profiles
of anthropogenic soils at the Historic site produced distinct planar anomalies in both the El and
E2 horizons. Medium to high-amplitude reflections with significant energy returned to the 500-
MHz antenna, enabling a distinct signal return separate from the surrounding matrix. This is
demonstrated with the presence of pronounced horizontal banding within the anomaly
Background filters allow for a clearer image of an already distinct anomaly. With the application
of spectral whitening, any contrast between archaeological features and the overlying soil is
enhanced (Figure 4-18c). The use of Gray 1 color scale enables the operator alternate data