|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
EVALUATING THE PRESERVATION OF HURRICANE DEPOSITS IN 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
This thesis s dedicated to my grandmother (Mimi) who has provided support and
encouragement to three generations of my family while studying for their advanced
degrees. Her educational values inspired me to continue my education beyond my
I wish to thank my advisor, Dr. John Jaeger, for his patience and guidance
throughout my thesis work. I would also like to thank my committee members, Dr. Mark
Brenner and Dr. Paul Ciesielski, for their countless advice; and Dr. Jason Curtis for his
assistance with lab work. Also, I wish to thank William Kenney for his lab assistance
and for always having time to answer my endless questions. I also thank Lisa Marie
Mertz for her help with lab work; and Donald Hardison and Jango Bhadha for hours of
discussion and advice.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
LIST OF TABLES ...................... ........ .................. ....... ............ .. vii
LIST OF FIGURES ............. .. ..... ...... ........ ....... .......................... viii
A B ST R A C T ................. .......................................................................................... x
1 INTRODUCTION .............. ................. ........... ....................... ..1..
2 BACKGROUND INFORM ATION ........................................ .......................... 7
S ig n a l .................................................................................. ................................ . 7
Preservation of H hurricane D eposits .................................................................. ...... 10
C coastal Ponds ...................................................................... ......... 12
Study Area ........................................ ..... ............... ........... 13
Known Hurricanes of St. Vincent Island Region.............................................. 14
3 M E T H O D S .............................................................................28
4 R E SU L T S ..............................................................................33
G general Lithology ......................................................................... .. 33
A analyses of 210Pb ....................................................................................... 35
A naly ses of 137C s ............................................................................................. 36
Magnetic Susceptibility .......................................... 36
G ra in S iz e ............... ................................................................................ ........... ...... 3 7
M icro p aleo n to lo g y ................................................................. ............................... 3 8
S a lin ity ............................................................. 3 9
W eight % C and % N .................................................... 39
5 D ISC U S SIO N ............................................................... 59
Geochronology........................... .............. 59
S ig n al ............... ...................................... ........ ...... 6 6
Preservation Potential ....................................................... .. 75
6 C O N CLU SIO N ............... ............................ ............ ............. ........ 87
R E F E R E N C E S C IT E D ............................................................................ .....................89
B IO G R A PH IC A L SK E TCH ...................................................................... ..................94
LIST OF TABLES
2-1. Data on hurricanes to affect St. Vincent Island from 1880 to 2001......................23
2-2. Predicted recurrence interval based on the model developed by Overland .............25
5-1. Results for the calculation of the Peclet number for a Kd value of 102 and an L
v alu e o f 10 cm ................................................... ................ 6 2
5-2. Results for the calculation of the Peclet number for a Kd value of 105 and an L
value of 10 cm ................................................................ .........63
5-3. Results for the calculation of the Peclet number for a Kd value of 102 and an L
v alu e o f 5 cm .................................................... ................ 6 3
5-4. Results for the calculation of the Peclet number for a Kd value of 105 and an L
v alu e o f 5 cm .................................................... ................ 6 3
5-5. Sedimentation and mixing rates for several coastal ponds....................................65
5-6. Synopsis of detection of hurricanes by each of the proxies tested.........................75
LIST OF FIGURES
1-1 Location of St. V incent Island ........................................ ......................... 6
2-1 Location map drafted of the two coastal bays studied along the west-central
2-2 Diagram drafted of the three storm facies ......... ................................ ......... 19
2-3 Location map drafted for cores taken at (A) Lake Shelby and Middle Lake,
Alabama and (B) W western Lake, Florida .................................................... 20
2-4 Location map drafted of a study of coastal Louisiana hurricane deposits in salt
m arshes. .......................................... ............................ 2 1
2-5 Location maps drafted for cores taken in (A) New Jersey and (B) Rhode
Island ..................................... .................................. ......... 22
2-6 Figure of the hurricanes to strike the Florida panhandle from 1885 to 1994 .....26
2-7 Path of hurricane eyewalls passing near St. Vincent Island. ............................27
3-1 C oring locations .......... ...... ........................ ........ ....... .. .. .. ............ 32
4-1 Photographs of cores ......... ................ ................... .................. ............... 40
4-2 Gamma bulk density and x-radiograph data for core OP1. .............................41
4-3 Gamma bulk density and x-radiograph data for core OP2 ..............................42
4-4 R ed-green-blue data for cores....................................... .......................... 43
4-5 Pixel density and x-radiograph data for cores................... .......................... 44
4-6 Plots of gray scale pixel density versus gamma bulk density for cores..............45
4-7 Gamma bulk density and x-radiograph data for cores A) FP1 and B) FP2........46
4-8 Total and excess 210Pb activity for cores.................................. ............... 47
4-9 Measurements of total and excess 210Pb for core FP2 .....................................48
4-10 137Cs activity for cores A) OP1 and B) FP1...................................................49
4-11 Measurements of 137Cs activity for core FP2.................... ........................... 50
4-12 Magnetic susceptibility measurements for cores..............................................51
4-13 Percent sand data for cores. ........................................ .......................... 52
4-14 Mean and median grain sizes for cores.......................................................53
4-15 Sorting and mode measurements for cores. .............................. ......... ...... .54
4-16 Plots of percent sand versus gamma bulk density (gm/cc) for cores.................. 55
4-17 Foraminifera abundances per 0.3 g of sample for cores. ...................................56
4-18 Salinity profiles for cores .......................................................... ............... 57
4-19 Percent organic carbon and nitrogen for cores. .............................................58
5-1 210Pb inventory for cores ......... ................................................. ... ............ 81
5-2 Sediment accumulation rates for cores. ................................... ............... 82
5-3 Plot of percent sand versus percent organic carbon for cores ..........................83
5-4 Diagram of the preservation of an event layer after its deposition .....................84
5-5 Diagram of the destruction of an event layer after its deposition ....................85
5-6 Diagram of using 210Pb concentrations to determine the depth of the mixing
layer ......... ......... ......... ..................................... ........................... 86
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
EVALUATING THE PRESERVATION OF HURRICANE DEPOSITS IN FLORIDA
Chair: John M. Jaeger
Department: Geological Sciences
Cyclones are one of the most energetic geomorphic agents in coastal environments
of tropical, subtropical, and temperate latitudes, causing rapid changes in sediment
deposition and erosion through intense wind and wave energy and coastal flooding.
Since historical records of cyclone activity in the Atlantic Basin only extend back 370
years, a longer record is needed of past cyclone occurrences in order to better evaluate
recurrence intervals (e.g., paleoclimate) and associated geomorphic change caused by
cyclones. Coastal ponds offer an ideal location to study paleocyclone records because
they offer an environment that is near the shoreline, that experiences little disturbance
from waves and tides, and that has the potential for rapid sedimentation rates. Sand beds
in muddy coastal ponds and marshes have frequently been associated with cyclone
overwash deposition, although other transport agents (such as Aeolian) can result in
similar type deposits. The purpose of this study was to evaluate a number of coastal pond
sedimentary proxy records for their utility as paleocyclone indicators. Two sets of piston
cores were taken -30 m from the beach in coastal ponds on St. Vincent Island, Florida, a
relatively undisturbed island off the panhandle that has been frequently disturbed by
hurricane activity. A variety of different proxy records (grain size, magnetic
susceptibility, gamma bulk density, sediment reflectance, micropaleontology, salinity,
%C and %N) were analyzed in these cores to detect three major hurricanes known to
have severely impacted the island (in 1894, 1974-1975, and 1985), as well as additional
minor hurricanes. Measurements of bulk density and magnetic susceptibility were
obtained with a multi-sensor core logger, and cores were split and examined visually and
x-radiographically for lithology. The cores were sampled at 1 cm intervals for
measurements of radioisotopes (210Pb, 226Ra, and 137Cs) and the aforementioned proxy
records. Although it was difficult to establish a geochronology for Flag Pond due to the
dynamics of sedimentary processes, results indicate an average sedimentation rate of 1.8-
3 mm/yr. The sediment accumulation rate for Oyster Pond was calculated as 1.8 mm/yr
based on 210Pb activity. Grain size sorting of the sand fraction, percent sand, x-
radiograph pixel density, and gamma bulk density data display some evidence of
hurricane deposits corresponding to 1894, 1974-1975, and 1985; but the limitations of
establishing a robust age-depth correlation prevent certainties in the correlation of the
event layers to known hurricanes. Aeolian transport of sand makes it difficult to decipher
extreme storms from less extreme storms. Modeling of the preservation potential of the
these ponds indicates that a minimum storm bed thickness of one centimeter is needed in
order for some portion of the bed to remain intact after passing through the surface-mixed
layer. Dissipation time (the time required to completely destroy an event layer) for
Oyster and Flag ponds is estimated to be -5-10 years.
Tropical cyclones are one of nature's most destructive forces. The waves, winds,
and rainfall associated with these storms can cause injury to humans, damage to property,
flooding, and extensive landscape modification. With population increasing in coastal
areas (Pielke and Landsea 1998) where cyclone strikes are most damaging, it would be
helpful to insurance companies and landowners in coastal areas to have a record of the
recurrence intervals of cyclones and the regions most likely to be affected by them in
order to mitigate damage.
To understand the periodicity and intensity of cyclone activity for a region, a record
of cyclone landfalls is needed that spans several thousands of years (Donnelly et al.
2001a). Although written historical accounts of tropical cyclone conditions from North
America extend back 370 years, records of cyclone tracks maintained by the National
Oceanic and Atmospheric Administration only extend back to the late nineteenth century.
To extend this historical record, paleocyclone studies have been initiated that may
provide important scientific information in two respects: (1) the frequency of cyclone
strikes in a region can be calculated based on past occurrences of landfall; and (2)
changes in cyclone patterns (e.g., intensity and frequency) in a region may provide
information about the paleoclimate of that region.
A valuable method for studying paleocyclone activity is to examine the sediment
record of coastal environments in tropical, subtropical, and temperate latitudes. The
chemistry, mineralogy, and stratigraphy of sediments reflect depositional processes
associated with coastal agents, including cyclone activity. Abrupt changes in stratal
composition may indicate an episodic erosional/depositional event associated with the
impact of a cyclone. Coastal sedimentary strata have the potential to preserve a high-
temperal resolution decadall) record of cyclone activity if the particular depositional
environment experiences high sedimentation and low biologic mixing rates.
Cyclone deposits in supratidal environments are typically the result of overwash
deposition from storm surge flooding of an area. Hayes (1967) determined that storm
surge is the dominant characteristic of cyclones, resulting in their importance as
geomorphic agents. Cyclones also move sediment from offshore to onshore through
wind and large waves. In past studies of cyclone deposition, several different proxy
records of storm activity have been measured in cores from the continental shelf, salt
marshes, coastal ponds, and coastal bays. Grain size and sorting are the most common
proxies, but micropaleontology and organic C and N concentrations have also been used
(Parsons 1998, Collins et al. 1999, and Donnely et al. 2001a). Cyclone beds are typically
recognized as being more coarse-grained and poorly sorted than surrounding strata
(Parsons 1998). In addition, during an intense storm, benthic marine foraminifera can be
transported onshore from the continental shelf (Collinset al. 1999). In estuaries, coastal
lakes, and salt marshes where the bottom sediments are fine-grained and organic-rich,
lower organic C and N concentrations correlate with sand-rich layers that could be
indicative of a storm deposit (Parsons 1998).
Recent studies indicate that it is necessary to study a combination of different proxy
records to best identify paleocyclone deposition. Collins et al. (1999) report that using
the presence of sand layers only to mark cyclones can underestimate the number of
storms. Since sand layers are usually only deposited near the location (within -75 km)
where the eyewall comes ashore, other parameters are needed to detect cyclone
deposition. To complicate the matter, biologic and physical mixing can alter the
appearance and preservation of distinct sand layers (Davis et al. 1989). For example, due
to physical and biologic mixing, deposits left by Hurricane Andrew (Category 4) were
only detectable through diatom assemblages in a Louisiana marsh 2 years after landfall,
whereas initial deposits contained distinct sand beds (Parsons 1998). By increasing the
number of proxy records that are used, storm layers can be better distinguished.
Although paleocyclone studies have been performed on a number of different
coastal depositional environments (e.g., subtidal, intertidal, and supertidal) (Davis et al.
1989, Donnelly et al. 2001a, Liu and Fearn 2000), there is a lack of consensus as to
which depositional environment consistently preserves the best record of cyclone
activity. However, an ideal setting would fit the following criteria: (1) have high
sedimentation rates to quickly bury event beds and prevent physical and chemical
mixing; (2) experience regular cyclone activity resulting in likely production of
paleocyclone indicators; and (3) are not influenced by frequent tidal fluctuations that may
subsequently erode/mix event beds after deposition (Wheatcroft and Drake, 2002).
One environment that fits these criteria is a coastal pond. Coastal ponds are
typically located very near the beach. Although subject to overwash deposits during
cyclone activity, these ponds are separated from the beach environment by a sand barrier
(e.g., barrier beach or dune ridge) and, thus, not regularly affected by tidal influences
(Liu and Fearn 2000). Bottom sediments in coastal ponds are, in general, composed of
mud-sized particles and organic matter (Liu and Fearn 2000). Sand and associated
marine microfossils from dunes, beaches, and offshore that are transported landward
during a cyclone should be easy to differentiate from organic-rich material commonly
found in these ponds. However, modem sea-level rise and associated landward migration
of the surf zone over Holocene dune sediments and into ponds has the potential to
influence the frequency of deposition of sandy, coarse-grained, poorly sorted intervals
that may be interpreted as being cyclone deposits (Otvos 2002). Therefore, one must use
caution when examining coarse-grained, poorly-sorted sediments in coastal ponds that
date to times of sea-level rise, thus mistakenly interpreting these as cyclone deposits.
Another complicating factor is that sand bed deposition in ponds often times is the result
of extratropical storms (Otvos 2002). Although these storms are not characteristically as
strong as cyclones, they do have the potential to deposit overwash and windblown
material (Donnelly et al. 2001a). For example, Donnelly et al. (2001a) reported New
England winter storms producing extreme storm surges similar in elevation but longer in
duration than most cyclones. Also, vegetation, such as sea oats, is often located in the
dune area between the ponds and the beachface, which could prevent the transfer of
aeolian and overwash sediment by trapping and baffling flow. This would affect the type
of deposit preserved in the sediment record.
Coastal ponds chosen for study of paloecyclone deposition also should be located
in a region that is relatively undisturbed by human activity. Urban development often
results in increased runoff of coarse-grained material from inland areas into low-lying
ponds (Applelboom et al. 2002) could be misinterpreted as cyclone deposits. St. Vincent
Island located along the western panhandle of Florida, offers an ideal environment for
studying paleocyclone activity (Figure 1-1). The island is geographically located in an
area regularly impacted by cyclones (i.e., hurricanes), there are a number of fresh and
saltwater ponds to study, and there has been minimal human disturbance.
The goal of this project was to examine the recent sedimentological record (past
100 years) of St. Vincent Island for evidence of historical hurricanes. Only recent
sediments were studied because of the limitations of the chronological techniques (210Pb)
and historical data of cyclone activity with which to compare the deposits found in the
sediment record. The data set used to address two scientific questions: which proxy
records are best used to identify hurricane deposits along the northern Gulf of Mexico
coastline, and what sedimentary conditions in coastal ponds provide the best preservation
The few studies done on recent (post 1968) hurricane deposits from coastal
environments such as salt marshes, coastal ponds, and coastal bays (Risi 1998, Collins et
al. 1999) suggest general characterizations that allow for the following basic hypotheses
to be tested by this research:
* Because coastal ponds often contain fine-grained, organic mud-rich sediments, the
coarse-grained, poorly-sorted sands typically associated with hurricane overwash
should be easily distinguishable.
* Because of the higher sand content associated with cyclone deposits, bulk density
should show similar increases.
* Particular species of foraminifera are specific to the marine environment. Their
presence in coastal ponds could be indicative of overwash (Collins et al. 1999).
* Overwash deposits in muddy ponds could be observed by a decrease in organic
carbon and nitrogen concentrations that correlates with increased accumulation of
poorly sorted sand.
* Because of the presence of magnetic minerals in offshore sediments (Wheeler et al.
1999), increases in magnetic susceptibility could correlate with overwash.
St. Vinc St ge Island
Figure 1-1. Location of St. Vincent Island drafted from a figure Davis, J.H. and Mokray,
M.F. (2000) Assessment of the effect of road construction and other
modifications on surface-water flow at St. Vincent National Wildlife Refuge,
Franklin County, Florida. USGS Water-Resources Investigations Report 00-
The geologic record contains abundant strata that are associated with the
occurrence of natural disasters (e.g., volcanic ash beds, river flood deposits). Some
layers are well preserved and easily identified through visual examination. Other layers
are more difficult to visually detect and require the use of biostratigraphic, petrologic,
and geochemical markers (i.e. forams, diatoms, % C, % N) (Collins et al. 1999; Parsons
1998). Also, multiple depositional processes can produce the same signal in just one
tracer. It is, therefore, necessary to study a number of different proxy records in order to
clearly define geologic deposits generated by particular natural disasters.
Hurricanes can produce recognizable deposits in coastal regions due to flooding
from intense precipitation or storm surges (Ball et at. 1967, Davis et al. 1989, Risi 1998,
Liu and Fearn 2000). Hurricane stratal characteristics vary depending on storm intensity,
landform shape, distance of sampling site relative to where the eyewall comes ashore,
forward speed and duration of storm, amount of rainfall, and local sedimentary
environment (i.e., subtidal, intertidal, supratidal) (Davis et al. 1989, Risi 1998). In
addition, storm deposits can be altered within months through natural sediment mixing by
physical and biological processes, yet still be detectable through biostratigraphic
evidence (Parsons 1998 and Collins et al. 1999).
Davis et al. (1989) reported the deposition of hurricane beds in cores from two
coastal bays along the west-central coast of Florida (Figure 2-1). Hurricanes are credited
as being a major contributor to the Holocene stratigraphy of Sarasota Bay and Little
Sarasota Bay. Cores were analyzed for textural properties and macrofaunal content.
Three types of storm facies were identified within the cores (Figure 2-2). The graded
storm facies is characterized by the transport of shelly sediment into bays. This facies is
the result of an intense storm, most likely a hurricane. The homogeneous facies
represents the reworking of bay sediments during strong frontal passages or weak
hurricanes. The fluvial storm facies is produced by runoff of terrigenous material into the
bays as the result of extreme rainfall. Although there is some geographic control on the
location of the different facies due to available parent material, they can generally be
found in the same region within Sarasota Bay.
Liu and Fearn (1993 and 2000) studied seven cores from Lake Shelby and Middle
Lake, Alabama and sixteen cores from Western Lake, Florida (Figure 2-3). Lake Shelby
and Middle Lake are freshwater lakes that are separated from the Gulf of Mexico by 250
m of sandy, pine-dominated beach ridges and sand dunes that are 2-4 m high. A canal
dug in 1960 connects the two lakes. A 150-200 m wide barrier beach separates Western
Lake from the Gulf of Mexico. The lake maintains salinities between 2.5 and 5.5 ppt due
to a restricted connection with the Gulf 1 km to the north of the lake. Sand layers were
identified using visual observations and water and organic matter content. The cores
taken from these lakes showed a noticeable absence of thick sand layers below -3 m
attributed to minimal cyclone activity prior to 3.2 ka (14C). Liu and Fearn (1993 and
2000) suggest that the absence coincides with an abrupt regional climate change
documented by Hodell et al. (1991) based on oxygen isotope data from a core from Lake
Miragoane, Haiti. The isotope data indicates that there was a sudden onset of drier
conditions around this time that caused higher evaporation rates and lower lake levels in
Haiti. Liu and Fearn suggest that there may have been a regional shift in circulation
patterns that caused cyclones to change their paths to a more eastern track. Between 4.5
and 3.2 ka, the Pecos River Basin in SW Texas experienced severe flooding, followed by
a period of infrequent flooding, suggesting a change in weather patterns (Patten and
Dibble 1982). Based on their data, combined with the data from Haiti and SW Texas,
Liu and Fearn believe that before 3.2 kya cyclones followed a western track when they
entered the Gulf of Mexico and struck the coasts of Texas and Louisiana; but an abrupt
change in circulation patterns occurred around 3.2 kya that caused cyclones to switch to a
more easterly track, hitting Florida and Alabama more frequently. However, given late
Holocene sea level rise (-1.6 mm/yr) (Bard et al. 1996), sea-level would have been 4-5
meters lower at this time than today. Given a nearshore slope of 1/60 (NOAA
Bathymetric Charts), the shoreline would have been -300 m further seaward. This
increased distance may have contributed to the lack of sandy bedding prior to 3.2 kya.
Parsons (1997) examined cores from a salt marsh pond on the Louisiana coast to
determine if a hurricane layer associated with Hurricane Andrew (1994) could be
distinguished and, if so, the origin of sediments in the hurricane layer (Figure 2-4).
Deposits were identified using grain size, sediment pigments, organic content, and diatom
analysis. The results showed that the sediments were imported and reworked from a
variety of environments and that the layer could be distinguished one year after
deposition. Two years after deposition, the layer was only distinguishable through
diatom assemblages. The only sedimentological evidence of Hurricane Hugo in fresh
water ponds 50 to 75 km apart and parallel to the coast from the landfall location, was the
presence of marine forams at the depth corresponding to the time of Hugo's landfall
Donnelly et al. (2001a; 2001b) describe evidence of hurricane overwash deposits in
cores from salt marshes in New Jersey and Rhode Island (Figure 2-5). The purpose of
these studies was to reconstruct the overwash history of a back-barrier salt marsh in order
to provide a landfall frequency of intense storms. Nine cores were taken in a grid 50 m
apart from Whales Beach Marsh in New Jersey. The marsh is approximately 250 m from
the shoreline. Fourteen cores were taken in the Succotash Marsh, Rhode Island along
transects that went inland from the coast. Succotash Marsh is located 275 m from the
shoreline. Deposits were identified visually for changes in lithology. The cores were
dated with 14C and 137Cs. Pollen stratigraphy provided additional age control. Four of
the six identified overwash deposits were matched to historical photographs from periods
of known overwash. The remaining two overwash fans were dated to between 592 and
570 years B.P. and the other dated to roughly six hundred years B.P.
Preservation of Hurricane Deposits
There have been numerous studies done on storm bed deposition directly after a
hurricane has impacted an area, but few studies have been completed on the preservation
of these deposits through time (Ball et al. 1967, Perkins and Enos 1968, Davis et al. 1989,
Liu and Fearn 2000). Storm intensity and deposit location within the coastal zone (e.g.,
intertidal, or supratidal) determine the preservation potential for a deposit (Davis et al.
1989). Supratidal sediments are likely to preserve a better record of cyclone overwash, as
only powerful hurricanes (Category 4 or 5) can transfer material from offshore into the
supratidal environment (Collins at al. 1999). Also, sediments deposited above the normal
high tide mark are more likely to be preserved because this environment is not constantly
subject to wave resuspension and mixing. Collins et al. (1999) reported that subtidal
deposits of hurricanes were not visually or geochemically decipherable three years after
Wheatcroft and Drake (2002) report that there are four main factors controlling the
preservation potential of sedimentary event beds in the marine environment: 1) sediment
accumulation rate, 2) mixing layer thickness, 3) bioturbation intensity, and 4) event layer
thickness. The sediment accumulation rate determines the time required for an event
layer to be buried and, therefore, better preserved within the sediment record. Low
sedimentation rates allow the event layer to be exposed to physical and biologic mixing
for a greater amount of time allowing the preservation potential to be influenced by the
other three factors to a greater degree (Ravichandran et al. 1995).
The surface mixed layer thickness refers to the depth below the sediment-water
interface at which mixing typically occurs through physical or biological processes.
Bioturbation, in the form of deposit feeding, burrowing, and tube building, can cause
mixing of an event signal within sediments. Bioturbation is more intense at the sediment-
water interface, where infaunal organisms are more numerous and decreases with depth.
Mixing intensity can be measured by naturally occurring tracers (234Th, 7Be, and 210Pb)
and synthesized tracers (glass beads, luminophores, and radio-labeled particles). A
deeper mixing layer (>5 cm) would greatly inhibit preservation of event layers. If an
event layer is thicker than the surface mixed layer is deep, some will be preserved
(Wheatcroft and Drake 2002). Shallower mixing layer depths coupled with high
sediment accumulation rates would favor preservation of signal layers (Ravichandran et
al. 1995). Environments with low accumulation rates, such as lakes and estuaries, are at
a greater risk of signal loss through bioturbation because the animals have an increased
amount of time (5-10 years) to mix the event layer (Ruiz-Femandez et al. 2001).
The preservation potential of an event layer is proportional to its original thickness
and the speed at which it is advected through the surface mixed layer by burial (i.e.
sedimentation rate). For example, in a marine environment Wheatcroft and Drake (in
press) found a thin layer (<1 cm) would be difficult to preserve because other factors,
such as bioturbation, would not require much time (months to years) to destroy the signal,
unless in an area of high sedimentation (1-5 cm/yr) or low mixing (0-0.1 cm2/yr).
Thicker deposits would require more time and energy to be dispersed. Wheatcroft and
Drake (in press) also found that an event such as a flood or hurricane, followed
subsequently by another event, increases the sedimentation rate and allows for quick
burial of the signal left by the first event.
There is no clear definition of the geomorphology or depositional facies of coastal
ponds and lakes. Those mentioned in available scientific literature are generally 0-15 km
from the mean high water mark and vary in area and in depth (e.g. Liu and Fearn 1993
and 2000, Collins 1999, Norton et al. 1997, and O'Sullivan et al. 1991). Some ponds and
lakes are tidally connected and are saline to brackish, while others remain fresh.
Sediment accumulation rate measurements in these ponds are sparse, but range
from -1 to 50 mm/yr, with the average being -5 mm/yr (O'Sullivan et al. 1991, Norton et
al. 1997, Scott and Steenkamp 1996, Ravichandran et al. 1995, Williams 1995, and Hyatt
and Gilbert 2000). In comparison, the sediment accumulation rates for inland Florida
lakes range from 0.2-2 mm/yr (Schelske et al. 2001). Very little research is available on
the sediment mixing rates in coastal pond sediments. A study by Ravichandran et al.
(1995) estimated the mixing coefficient for Sabine Lake, located along the Texas coast,
to be -0.04-0.4 cm2/yr based on measurements of 239'240Pu. Mixing rates for other coastal
areas are higher and include 0.3-2.5 cm2/yr for New York Bight, 4-32 cm2/yr for
Narragansett Bay, Rhode Island, and 0.1 to 100 cm2/yr for lacustrine and marine
environments in general (Santschi et al. 1980 and Boudreau, 1994).
The Gulf Coast of the United States is regularly impacted by tropical storms and
hurricanes. Due to prevailing wind currents, hurricanes over the past 200 years have
moved north to northwest when entering the Gulf of Mexico (Williams and Duedell
1997). While it is uncommon for hurricanes to hit the west coast of the Florida
peninsula, the panhandle is frequently affected by such storms. Fifty-six percent of the
hurricanes to hit the Florida panhandle from 1885 to 1984 occurred in the Apalachicola
Bay area (Davis et al. 1989) (Figure 2-6).
One region in the Apalachicola Bay area that has been severely impacted by
hurricane landfall is St. Vincent Island, a Holocene barrier island located on the Gulf side
of the Apalachicola Bay. The island regularly experiences hurricanes moving onshore
from the Gulf of Mexico (National Hurricane Center) (Figure 2-7). The island was
managed as a hunting preserve until 1908, when Dr. R. J. Pierce purchased the island for
use as a hunting area. Over 125 kilometers of dirt roads were built throughout the island
to allow access for timber companies to log the island in the 1940s and 1960s (Doyle and
Krauss 1999). Three permanent structures were built on the island and a dock is located
on the western tip. Some of the ponds were managed for freshwater fishing purposes.
When the Wildlife Refuge took over in 1968, culverts, dams, and other water control
structures were put into place. The ponds have been managed as both salt and freshwater
over the last 30 years. The Wildlife Refuge is currently attempting to return the ponds to
their natural, brackish state. Vegetation in some areas of the island is maintained by fire,
both prescribed and natural. Despite past development, the Wildlife Refuge considers the
island to be in a natural condition (St. Vincent National Wildlife Refuge, Apalachicola,
Fl., 2000, Wildlife and Habitat Management Review, May 31-June 2, 2000.).
St. Vincent Island is 15 km long by 7 km wide and covers -50 km2. The island is
separated from the mainland by the Apalachicola Bay and the St. Vincent Sound. The
elevation of St. Vincent ranges from 0 to -4 meters above mean high water level. The
island is primarily composed of quartz sand with the exception of muddy marshes, which
contain sand-clay, silt, and organic-rich matter (St. Vincent National Wildlife Refuge,
Apalachicola, Fl. ,1968, 1968 Narrative Report. US Department of the Interior, Fish and
Wildlife Service, Bureau of Sport Fisheries and Wildlife, 23p.).
The morphology of the island is characterized by a system of twelve beach ridge
sets, formed between 6000 to 800 years B.P., possibly through cyclicity in late Holocene
sea level, that runs northwest to southeast, a large marsh, and approximately fourteen
enclosed fresh and saltwater lakes (Campbell 1986) (Figure 2-4). Flag Pond is located on
the south side of the island with its south shore approximately -275 m from the island's
south beach face. It has a surface area of approximately -25 m2 and a depth of -0.5 to 1
m. Oyster Pond is located to the east of Flag Pond and is approximately -400 m from the
south beach. The pond has surface area of approximately -350,000 m2 and a depth of
-0.5 to 1.5 m (USGS 7 /2 Minute Quadrangle Map Indian Pass, Fl.).
Known Hurricanes of St. Vincent Island Region
There have been 60 tropical cyclones to follow a course within a 150 km radius of
St. Vincent Island in the last 100 years. Of these, 19 were hurricanes with winds greater
than 110 km/hr and two hurricanes had winds greater than 160 km/ St. Vincent National
Wildlife Refuge, Apalachicola, Fl., 2000, Final Report of the Vegetation Survey and Map
Project, A USFWS-USGS Research Partnership Program Project.) (Table 2-1).
Documentation of damage done by hurricanes and other major storms to hit St. Vincent
Island does not begin until the Wildlife Refuge took over the island in 1968. The refuge
puts out an annual report that documents any major weather events that happen during the
year, and all of the accounts in this section are from those reports unless otherwise
indicated. However, these reports only contain information on specific damage to the
island and it is assumed that other forces, such as wind, rain, and high tides, also affected
the island during these storms.
Williams and Duedell (1997) report six hurricanes in the vicinity of St. Vincent
Island in the years 1885 (2), 1886 (2), 1894, and 1898. One of the 1886 hurricanes
crossed directly over the island and was classified as a Great Hurricane (Category 5). A
tree ring study performed on St. Vincent Island by Doyle and Krauss (1999) showed
evidence of the 1894 hurricane, which had wind speeds sufficient to thin forests on the
island, as indicated by suppressed and released growth patterns, as hurricanes cause
crown and root damage on the windward side of trees resulting in less radial growth.
Williams and Duedell (1997) report another hurricane struck Carrabelle, FL. (-45 km
northeast of St. Vincent Island) on August 1, 1899. The storm remained in a stationary
position for 10 hours, leaving the town severely damaged.
The first hurricane documented in the Wildlife Refuge reports, Hurricane Agnes,
occurred on June 18 and 19, 1972. Although wind gusts in Carabelle would not even
qualify this storm as Category 1 on the Saffir-Simpson Hurricane Scale, storm surges of
-2 meters above astronomical predictions qualify it in this category. The magnitude of
the storm surge put this storm in the 50-year storm category (Table 2-2). The 1972 report
by the Wildlife Refuge states that all of the fresh water ponds on the island were filled
with salt water.
Hurricane Carmen, a Category 3 storm, came ashore along the Louisiana coast near
New Orleans on September 8, 1974. Although the storm hit 550 km away, tides on the
island were 1 to 1.5 m above normal and wind gusts reached 65 km/hr.
In September 1975, Hurricane Eloise came ashore in Northwest Florida. The 1975
Annual Report for the island is missing from the files, so the effects the island
experienced are unknown. The hurricane was classified as Category 3 and came ashore
between Ft. Walton Beach and Panama City, approximately 80 kilometers west of St.
Vincent Island. Tides in the area of impact were 3-5 m above normal and wind gusts
were estimated to be 200-250 km/hr (National Hurricane Center). This storm most likely
had a significant impact on St. Vincent Island.
The next hurricane to impact St. Vincent occurred on September 13, 1979.
Hurricane Frederic, a Category 3 storm, came ashore just west of the Florida-Alabama
border (National Hurricane Center). The main impact on the island came in the form of
rainfall. Approximately 35 cm of rain fell following the hurricane. The culvert that
connects Oyster Pond to the ocean was knocked out. The hurricane contributed to the
wettest September to date for the island.
The year with the most hurricanes on record to hit St. Vincent Island was 1985.
The first to strike was Hurricane Elena on September 1. The eye of the storm passed just
15-25 kilometers south of St. Vincent Island. Winds were recorded up to 200 km/hr,
making this a Category 3 hurricane. In addition to heavy rains, a funnel cloud passed
through the area. The pattern of fallen trees indicated that winds came in from the
northeast. Tropical Storm Juan traveled through the area in October. The main effect to
the island was the 28.44 cm of rainfall. Hurricane Kate hit the island on November 21.
This Category 2 storm moved directly across the island from the south. Storm surges
were recorded to be 2-3 m above normal. The combination of these three storms caused
some of the worst recorded damage the island has ever experienced.
Another year of numerous hurricanes was 1995. Hurricane Allison made landfall
on June 5 at Alligator Point, approximately 30 kilometers to the east. Although the storm
registered as Category 1, it did not have any physical impact on the island. Hurricanes
Erin (Category 2) and Opal (Category 3) both made landfall at Pensacola Beach,
approximately 125 miles to the west. The combination of these three hurricanes caused
beach erosion and maintenance problems on St. Vincent Island. Opal, specifically,
washed out several water control structures at unknown locations put in by the Wildlife
Ho and Myers (1975) performed a study on the storm-tide height frequency for
Franklin County, FL., which includes St. Vincent Island. Using the model developed by
Overland (1975), they predicted the recurrence frequency for different storm surge
heights in Apalachicola (Table 2-2), which can be compared to those experienced on St.
Vincent Island. St. Vincent Island has experienced storm surge heights greater than 3 m
twice since 1885, which is greater than the predicted recurrence interval. Due to
unavailable data, it is not possible to determine if the island has experienced other intense
storms at higher recurrence intervals than those predicted.
Figure 2-1. Location map of the two coastal bays studied by Davis et al. (1989) along the
west-central coast. Contours are the pre-Pleistocene surface. Figure drafted
from an article by Davis, R.A., Knowles, S.C. and Bland, M.J. (1989) Role of
Hurricanes in the Holocene Stratigraphy of Estuaries Examples From the
Gulf Coast of Florida. Journal of Sedimentary Petrology, 59 (6), pg. 1053.
DLithiclast ] Sand
EShells Ei Mud
Figure 2-2. Diagram of the three storm facies. The graded facies is the result of intense
storm activity and is characterized by abundant shells fining upward. The
homogeneous facies is stratigraphically homogeneous and is the result of
strong frontal passages or weak hurricanes. The fluvial facies contains
abundant mud and gravel due to extreme rainfall. Figure drafted from an
article by Davis, R.A., Knowles, S.C. and Bland, M.J. (1989) Role of
Hurricanes in the Holocene Stratigraphy of Estuaries Examples From the
Gulf Coast of Florida. Journal of Sedimentary Petrology, 59 (6), pg. 1057.
-= J/L. *S L -- "-' -' '
,v' rayt I2
(ZS.4 f^W t eastern Lak
1 12, '13
Meico ,ah h. Fld fro lk -
Figure 2-3. Location map for cores taken at A) Lake Shelby and Middle Lake, Alabama
and B) Western Lake, Florida. Figures drafted from articles by A) Liu, K.B.
and Fearn, M.L. (1993) Lake-sediment record of late Holocene hurricane
activities from coastal Alabama. Geology, 21, pg. 793 and B) Liu, K.B. and
Fearn, M.L. (2000) Reconstruction of prehistoric landfall frequencies of
catastrophic hurricanes in northwestern Florida from lake sediment records.
Quaternary Research, 54 (2), pg. 238.
Figure 2-4. Location map of a study of coastal Louisiana hurricane deposits in salt
marshes. Star indicates area where cores were taken. Figure drafted from an
article by Parsons, M.L. (1998) Salt marsh sedimentary record of the landfall
of Hurricane Andrew on the Louisiana coast; diatoms and other
paleoindicators. Journal of Coastal Research, 14 (3), pg. 939.
F e f Done an .
AeB )e Beatn pg,1
Bustlver, KLed R., and Webb III, T. (2001 yb) sedimentary eviec of
e- hu r ine s f ane J orele o .., .
,te/JLd r R. a IIIT -201d Se
Westover, h n., stdrieb frI, TN e Js. Go00 yr 2s100edimentary meted of
intense hurricane landfalls in southern New England. Geological Society of
America Bulletin, 113 (6), pg. 714 and Donnelly, J.P., Roll, S., Wengren, M.,
Butler, J., Lederer, R., and Webb III, T. (2001b) Sedimentary evidence of
intense hurricane strikes from New Jersey. Geology, 29 (7), pg. 615.
Table 2-1. Data on hurricanes to affect St. Vincent Island from 1880 to 2001
Duedell 1997, and Weather Underground 2003).
YEAR Name Category Storm Surge (m) (km/hr)
1885 N/A TS N/A 80 N/A
1885 N/A 1 N/A 130 N/A
1886 N/A 5 N/A N/A N/A
1886 N/A 2 N/A 135 N/A
1894 N/A 3 N/A 170 N/A
1898 N/A 1 N/A 115 N/A
1899 N/A 1 0.91-1.22 130 N/A
1903 N/A 3 N/A 160 N/A
1915 N/A 1 N/A 145 N/A
1924 N/A 1 N/A 130 N/A
1926 N/A 3 N/A 200 N/A
1929 N/A 2 N/A 170 N/A
1939 N/A 1 N/A 130 N/A
1941 N/A 2 N/A 180 N/A
1966 Alma N/A N/A N/A N/A
1968 N/A TS N/A 100 N/A
(St. Vincent National Wildlife Refuge, Williams and
eshwater ponds filled with salt water
34.2 cm of rain, Culvert between Oyster Pond and
the ocean knocked out.
Table 2-1. Continued
Storm Surge (m)
28.44 cm rain
Washed over several water control structures
Table 2-2. The predicted recurrence interval based on the model developed by Overland
(1975) for storm surge heights of 4, 3, 2.5, and 1 m for comparison to the
storm surges observed on St. Vincent Island.
Storm Surge Height (m) Recurrence Interval (y)
Data from an article by Ho, F.P. and Myers, V.A. (1975) Joint probability method of tide
frequency analysis applied to Apalachicola Bay and St. George Sound, Florida. NOAA
Technical Report NWS 18, pg. 22.
Figure 2-6. Figure showing that 56% of the hurricanes to strike the Florida panhandle
from 1885 to 1994 hit in the Apalachicola Bay area. Figure drafted from an
article by Davis, R.A. (1995) Geologic impact of Hurricane Andrew on
Everglades coast of Southwest Florida. Environmental Geology, 25 (3), pg.
40 0 40 80 Kilometers
Figure 2-7. Path of hurricane eyewalls passing near St. Vincent Island (1885-1995).
Numbers indicate the category of the storm according to the Saffir Simpson
The purpose of the sampling techniques used in this project is to provide
information on the thickness of storm deposits, their sedimentological, geochemical, and
micropaleontological properties, and the mechanisms of preservation of these deposits in
the sedimenty record. Five piston cores were taken from St. Vincent Island using the
coring method developed by Fisher et al. (1992). Cores approximately seven centimeters
in diameter were taken from Flag Pond (2) and Oyster Pond (2) (Figure 3-1). The length
of these cores ranges from 57 to 75 cm. Sampling locations on the island were chosen
based on the proximity of the pond to the coast as shown in Landsat TM imagery and
accessibility by car and boat. Since both Oyster and Flag Ponds are only separated from
the southern coastline of the island by dunes, they were assumed to have received the
greatest amount of washover during a tropical storm or cyclone.
The cores were kept vertical during transport and stored at 40C. Once in the lab,
the cores were analyzed for bulk density and magnetic susceptibility at 0.5 cm intervals
using a Geotek Multi-sensor Core Logger. The accuracy of the bulk density
measurements was determined by plotting the gamma counts/second (determined using a
standard aluminum density calibration piece) versus density*thickness of the aluminum
(Weber et al. 1997). Cores were then split and processed through the core logger for
detailed digital imaging at a resolution of 40 pixels/cm. Next the cores were x-
radiographed at 50 KeV/450 mAs to reveal internal structures and changes in sediment
density. The x-rays were scanned and then processed by SCION software to generate a
relationship between the sediment density and the gray scale pixel density.
Following the non-destructive core analyses, both halves of each core were
sampled at 1 cm intervals, which was the smallest interval that produced enough material
for grain size and radioisotope measurements. Samples from one half of the core were
divided and prepared for radioisotopic dating, stable isotope analyses, and pore water
analyses. Samples from the other half of the core were reserved for grain size and
Particle-reactive radioisotopes are used in sedimentological studies to examine
sediment mixing and accumulation rates (Appleby and Oldfield 1992) and to distinguish
event beds preserved in the geologic record (Jaeger and Nittrouer 1999). Samples for
radiometric dating were freeze dried, powdered, packed into plastic tubes with up to 3 cm
of dry sediment, and then sealed with a mix of epoxy resin. The activity of naturally
occurring radioisotopes was measured with well-type intrinsic germanium detectors
(Schelske et al. 1994). A large range of y-energies were counted for 24-48 hours
depending on sample height for a minimum of 380 counts to reduce the counting error to
<5%. The activity of 210Pb, a decay product of 226Ra in the 238U decay series, was
measured in each sample. The samples were set aside for a minimum of three weeks to
establish secular equilibrium between radon (222Rn) and radium (226Ra). The excess
activity of 210Pb was determined by subtracting the supported 210Pb (210Pb in secular
equilibrium with 226Ra) from total 210Pb activities. The activity of 137Cs, which may be
used as an additional age marker, was also determined. Sample mass and height and
counting efficiency (98-99%) were factored into calculations of activity of each
radionuclide (Schelske et al. 1994). Blanks were counted before and after measurements
of each core in order to determine background radiation. Amersham International
standards of Americium-241 and Cesium-137 were run within one year of the
measurements. Standards do not have to run before or after each set of samples due to
the 98-99% efficiency of the well detector. The counting error is measured as the square
root of the number of counts for each sample.
The weight percent carbon and nitrogen concentrations were determined with a
Carlo-Erba Elemental Analyzer on one-centimeter subsamples. Approximately 3-5 mg
of sample were placed in tin cups and dried under a heat lamp. Combustion of samples at
1000C determined the total carbon and nitrogen. Four Atropine standards with a
specific percent of carbon and nitrogen were run before each set of samples to determine
a regression line to relate the size of the samples to the percent carbon and nitrogen in
each one. The precision of the analyses was determined by analyzing duplicates of every
tenth sample. The relative percent difference (RPD) was determined using the formula:
(IX1-X2l)/Xmean. The RPD for %C was calculated to be 6.4% and the RPD for %N was
Samples for pore water analyses were centrifuged to separate sediment and water.
Salinity analyses on the porewater were performed using a portable refractometer.
Standard precision and accuracy of portable refractometers is -1 ppt.
Samples for grain size and microfossils were wet-sieved at 63 |tm to separate the
sand and mud-sized fraction. Sand-sized particles were dried in an oven at 600C.
Approximately 0.5 to 1.5 grams of each sample were analyzed on an automated settling
column for grain size measurements (Syvitski et al. 1991). The settling column method
measures grain size as a function of settling velocity, assuming a spherical particle shape.
The results are reported numerically as the percent of the total mass at each phi size class
(0.10), mean, median, mode, and sorting. The accuracy of the automated settling column
was determined with NIST Standard Reference Material glass beads. The precision of
the instrument was determined by running duplicate samples. The RPD for median was
calculated to be 0.5%, 0.7% for mean, 4.3% for mode, and 4.7% for sorting.
The percent sand for each sample was estimated based on the difference between
the total weight of the sample before it was split into sand and mud fractions and the
weight of the sand fraction after the sample was split by Equation 3-1:
(Total Sample Weight Wet (g) Water Weight (g) Dry Sand Weight (g)) = % Sand
Water Weight (g) (3-1)
The water weight for each sample was determined by the following equation:
Total Sample Weight Wet (g) % Water = Water Weight (g)
The percent water was calculated based on the weight of the samples from the archive
core before and after they were freeze dried by Equation 3-2:
Total Sample Weight Wet (g) Total Sample Weight Dry (g) Water Density = % Water
Total Sample Weight Wet (g) (3-2)
Water density was presumed to be 1.03.
Because marine macrofossils and microfossils in terrestrial sediments are useful
indicators of hurricane deposition (Collins et al. 1999, Davis et al. 1989, Parsons 1998),
biostratigraphic measurements were done on each 1-cm sample. A portion of each sand-
sized sample was split to 0.15-0.30 g and microscopically analyzed for foraminiferal
classification and abundance. All of the forams in each sample were picked, mounted,
and identified by comparison to known fresh and saltwater groups at the genus level.
25 0 25 METERS
Figure 3-1. Coring locations
The digital photographs show the Oyster Pond cores to be relatively homogeneous
in color and texture and dominated by muddy sediment with greenish color (Figure 4-1).
Core OP1 has shifts in gamma bulk density at -16 and 34 cm (Figure 4-2). A steady
increase in bulk density down to 16 cm is followed by an interval of scattered values
ranging from 1.2 to 1.4 g/cm3. The values remain at a lower density from 34 to 54 cm
with a gradual increase in bulk density below 54 cm. The red-green-blue color reflective
data for this core shows a transition at -14 cm from lower (lighter) to higher (darker)
values (Figure 4-4). Small peaks occur at 24, 28, 34, and 50 cm. The blue data plots
slightly lower than the red and green data. The gamma bulk density and red-green-blue
data for core OP2 show an increase at 8 cm followed by a drop at 13 cm. There is also a
peak in gamma bulk density at 58 cm (to 1.25 g/cm3).
The Oyster Pond cores show several small bedding changes through visual and x-
radiographic examination (Figures 4-1, 4-2, and 4-3), and the gamma bulk density
measurements for OP1 reveal several small layers of higher density material at 22 cm and
34 cm. The large increase in bulk density at 34 cm is related to a large mollusk shell
found at that interval.
The gray scale pixel intensity data from digitized x-radiographs for core OP1
indicate layers with darker (pixel density of >30 than surrounding data) gray scale pixel
values at 13-15 (a slanted layer), 19, 23, and 36-37 cm (Figure 4-6). The gray scale pixel
intensity data for core OP2 is highly variable, with layers of darker pixel values at 8,17,
and 23-24 cm. Values gradually increase from 21 cm to the bottom of the core. Lighter
pixel values appear to correspond with higher density material (Figure 4-7).
The two cores from Flag Pond have one major change in color and texture
downcore (Figure 4-1). The digital photographs show an abrupt switch from darker
sediment to lighter sediment at 12 cm (FP1) and 20 cm (FP2). The Flag Pond cores are
very sandy (-60-90% sand). The darker sediment is concentrated near the surface
(Figure 4-1). At -14 cm, core FP1 transitions from mud-rich to a sandier texture, with
corresponding increases in gamma bulk density (from -1.1 to 1.6 gm/cc) and red-green-
blue data (Figures 4-4 and 4-5). The gamma bulk density for core FP2 does not show a
change in sediment character until -38 cm. The values for bulk density increase from
-1.1 to 1.6 gm/cc, but they do not remain constant after the increase (i.e., there are some
sand and mud layers mixed together). Other minor changes in density occur at 20, 29,
39, and 48 cm. The red-green-blue data shows a shift from darker to lighter values at -20
cm and small peaks at 9, 14, 29, and 40 cm. Both cores FP1 and FP2 plot the blue data at
lighter values, followed by green, and then red at darker values.
The photographs of core FP2 show changes from dark to light at 13 cm (Figure 4-
1). The x-radiographs from both Flag Pond cores do reveal several gradual bedding
changes (Figure 4-6). Core FP1 has a contact -11 cm below the surface, with a
corresponding increase in the bulk density at this same depth. The x-radiographs reveal
pronounced contacts in core FP2 at 26 cm, 38 cm, and 48 cm, with corresponding
increases in gamma bulk density at 38 cm and 48 cm. All of these are sharp contacts
with definite associated color and lithology changes.
The gray scale pixel data for core FP1 shows a diffuse contact between light and
dark material from 14 to 16 cm (Figure 4-6). Other variations in pixel data occur at 6, 21,
24.5, and 27 cm, and all of these intervals represent slight variations in density and would
only be detected through detailed analysis at <1 cm sampling intervals. The gray scale
pixel data for core FP2 show the same change in lithology at -24-25 cm that is seen in
core FP1 that reflects a transition from light to darker sediment. The gray scale pixel data
also reveals laminations at 11, 14, 19, and 33 cm. The increase in pixel density at 33 cm
appears to be a small bed of lower density material.
Measurements of 210Pb for core OP1 reveal supported 210Pb at 18-19 cm (Figure 4-
8). Total 210Pb activities varied from 2.0 dpm/g to 5.5 dpm/g. Excess 210Pb activities
range from 0.5 dpm/g to 5.5 dpm/g. Excess activity decrease fairly steadily downcore
with the exception of small increases at the 10-11 cm and 13-15 cm intervals.
Core FP1 contains variations in the 210Pb activities downcore with no discernible
trend (Figure 4-8). The measurements of total and excess 210Pb activity range from 0.0
dpm/g to 5.0 dpm/g. Samples at 0.5, 2.5, 4.5, and 5.5 cm have higher activities (5.0
dpm/g decreasing downcore to 1 dpm/g), while samples at 1.5 and 3.5 cm have activities
of <1 dpm/g.
Core FP2 also shows down core variations in 210Pb activity (Figure 4-9). Total
210Pb is greatest at the top of the core (9 dpm/g at 0.5 cm). Excess 210Pb activity is 7.5
dpm/g at the top of the core and decreases to 0.5 dpm/g at approximately 30 cm.
Measurements are missing for several depths because analyses were done on this core
only for the purpose of estimating the depth of supported 210Pb.
Measurements of 137Cs activity are used as a complementary dating tool to 210Pb
because it has a unique source (atomic weapons tests) that overlaps the period of 210Pb
dating (-100 years). Unless mixing has occurred, the depth in the core that first shows
137Cs activity should correspond to 1954.
Core OP1 has a downcore decreasing 137Cs activity profile that reaches 0.0 dpm/g
at a depth of 12 to 13 cm (Figure 4-10). There is a minor decrease in activity at the 2-3
cm depth interval. The 137Cs activity profile for core FP1 does not reveal the same
decrease with depth as core OP1. There is a decrease in activity from 1-2 cm that
corresponds to a 210Pb decrease. The profile for 137Cs for core FP2 (Figure 4-11) also
shows variation in activity with depth, with higher activity of 2.0 dpm/g at 10.5 cm, 0.6
dpm/g at 16.5 and 17.5 cm. Measurements of 137Cs activity remain at 0.0 dpm/g for both
20.5 and 29.5 cm.
The magnetic susceptibility sensor records raw data and does not account for
density changes downcore. Data may reflect changes in density rather than magnetic
susceptibility, therefore, the data have been mass corrected by the following equation:
X = K/p (4-1)
where X is the mass specific susceptibility, K is the uncorrected susceptibility, and p is
the sediment density. Units for susceptibility are 10-6 cgs.
Core OP1 does not have any high amplitude peaks in magnetic susceptibility
(Figure 4-12). It does, however, have a steady increase from 5 cm to 10 cm and then
declines gradually below this point with minor increases at 35 and 41 cm. Core OP2 has
one large peak at 16 cm, declines until -25 cm, and remains steady downcore.
Both Flag Pond cores display two large magnetic susceptibility peaks (Figure 4-
12). For core FP1, the peak at 17 cm has a maximum value of 1.3x10-6 cgs and the peak
at 27 cm has a maximum value of 1.9 "10-6 cgs. The peaks for core FP2 are at 24 and 37
cm. The peaks are separated downcore by 10-15 cm in both cores, but are offset 8-10 cm
All grain size measurements were performed on 1 cm intervals (Figures 4-13, 4-14,
and 4-15). The mean, median, sorting, and mode were calculated for the sand fraction
(>63 tm) based on measurements from an automated settling column (Syvitski et al.
1991) (Figures 4-14 and 4-15). Measurements were only performed on the sand as it
contains the most relevant information about overwash processes. Because there are no
measurements for samples that were too small in mass to be run on the settling column
(< 0.5 g), core OP1 has missing data for the intervals of 1-2, 2-3, 7-8, 8-9, 9-10 and 16-17
cm and core FP1 has missing data for the intervals of 1-2, 3-4, 6-7, and 14-15 cm.
Peaks in mass percent sand for core OP1 occur at 2.5, 6.5, 19.5, and 27.5 cm
(Figure 4-13). The mass percent sand for core FP1 shows a transition from a variable
profile to a constant profile at 12.5 cm. Peaks occur at 2.5, 4.5 and 8.5 cm.
All samples in core OP1 contained less than 50% sand, with only three samples
having greater than 40% sand (Figure 4-13). The data formed a trendline when percent
sand was plotted against gamma bulk density (Figure 4-16). There were only five
samples that did not fit the trendline (at 0.5, 1.5, 2.5, 3.5, and 8.5 cm). Core FP1 was the
opposite of core OP1 and contained increased amounts of sand (40-60% greater) (Figure
4-13). Only four samples contained less than 50% sand, with the majority of the samples
containing greater than 75% sand. The plot of percent sand versus gamma bulk density
showed two trends (Figure 4-16). Samples above 12 cm followed one trend while
samples below 12 cm followed another.
The median and mean sizes of the sand fraction for core OP1 were finer than the
median and mean for core FP1 (Figure 4-14). The measurements for core OP1 cluster
around the average values for the south-facing beach (2.2 phi), while the measurements
for core FP1 cluster around the average values for the ridge sands (1.9 phi). The mean
values for core OP1 are slightly more scattered than the median values. Core FP1 shows
very similar values for both median and mean.
Sand in core OP1 is moderately to moderately well sorted, with sorting values of
ranging from 0.3 to 0.9 (Figure 4-15). There is not a consistent pattern downcore for
sorting. Sand in core FP1 is also moderately to moderately well sorted and does not
show any downcore variation. Neither core matches the sorting values calculated for the
ridge sands, south-facing beach, and east-facing beach.
While the values for median, mean, and sorting are constant downcore for core
FP1, the modal values are quite varied (Figure 4-15). Values range from 1.7 to 2.3 and
there is no consistent downcore pattern. Modal values for core OP1 are uniform
downcore with the exception of the sample from 13-14 cm. This sample has a very low
value at 0.75. The modal values for this core cluster just above the values for the south-
Both ponds show an absence of forams above 17 cm for core OP1 and 25 cm for
core FP1 (Figure 4-17). Below these depths, the number of forams sharply increases to
abundances greater than 100 forams per cm3. Decreases to below 60 forams per cm3
occur at 16, 22, 24, and 26 cm for core OP1 and at 27 and 29 cm for core FP1. Greater
than 90% of the forams were identified as belonging to the Ammonia genus.
Samples from every other centimeter from both ponds were put into a centrifuge
and spun down to separate pore water. Samples from core FP1 were not sufficient to
produce enough pore water to measure, so samples from core FP2 were used instead.
The salinity values for core OP1 range from 10 to 15 ppt with fluctuations near the
surface (until 7cm depth), decreasing in amplitude with depth (Figure 4-18). These
measurements are slightly higher than the overlying water (7 ppt). Core FP2 salinity
values decrease steadily downcore from 17 to 26 ppt. The overlying water in Flag Pond
has a salinity reading of 20 ppt.
Weight %C and %N
Concentrations of organic carbon and nitrogen track one another in cores OP1 and
FP1 (Figure 4-19), but with respect to depth, the trends were different between the cores.
In core OP1 the values decrease from 6.25 to 3 for %C until 5 cm and then leveled off at
-2%. The /oN values decreased from 0.67 to 0.33% and then reached a constant level of
0.2%. Core FP1 had higher values for both elements (14-20% for carbon and 0.9-1.2%
for nitrogen) than core OP1 towards the top of the core. Although there were minor
variations with depth, the values remained constant and elevated until 10 cm. Below this
depth, both the carbon and nitrogen values remained at 0% within the analytical precision
of the instrument.
5 cm .o
Figure 4-1. Photographs of cores A) OP1, B) OP2, C) FP1, and D) FP2.
Gamma Bulk Density (g/cm2)
1.2 1.4 1.6
Figure 4-2. Gamma bulk density and x-radiograph data for core OP1. Yellow boxes represent depth matched to date of known
hurricanes for OP1.
Gamma Bulk Density (g/cm2)
1 1.2 1.4 1.6
10 ------------- ---------------
Figure 4-3. Gamma bulk density and x-radiograph data for core OP2. Yellow boxes represent depth matched to date of known
hurricanes for OP1.
OP2 Red-Green-Blue FP1 Red-Green-Blue
70 1,, 111 111"- 1 70-7 --1 '. ..-70-1 -- 17. 1.. ------- 70. ... 1.
Figure 4-4. Red-green-blue data for cores A) OP1, B) OP2, C) FP1, and D) FP2. Increased values indicate darker sediment. The
spikes in the blue values are an artifact of sampling and not relevant data.
A) OP1 Red-Gree-Blue
Pixel Density Pixel Density
*c- - - :' \^ - --- - -
S20- 20- -20 20
30- 30 30
OP1 FP1 FP2
40 40 40 40
Figure 4-5. Pixel density and x-radiograph data for cores A) OP1, B) OP2, C) FP1, and D) FP2. Yellow boxes represent depth
matched to date of known hurricanes for OP1.
D) Pixel Density
, 1.05 0.8
E 1.15-- ------------------------------------------0.8 E1------ ---------------------------------------
o 0 *
E 1.1 --
Ip Op ^ ^ fL 1.6 -- -- ---- ---- -- -- -- ---- --- --- -- -
E- 1 .*' 3 - - - - i t^ ^ ^ ^ - E\
1.4- .... I....-....-,,,m....2--------------------------------------.
0 50 100 150 200 250 0 50 100 150 200 250
OP1 Pixel Density FP1 Pixel Density
Figure 4-6. Plots of gray scale pixel density versus gamma bulk density for cores A) OP1 and B) FP1. Black lines represent trends in
Bulk Density (g/cm2)
1.5 2 2.5
B) Gamma Bulk Density (g/cm2)
0.5 1 1.5 2 2.5
Figure 4-7. Gamma bulk density and x-radiograph data for cores A) FP1 and B) FP2.
A) OP1 21Pb (dpm/g)
0 12 3 4 5
a18 + -
B) FP1 21Pb (dpm/g)
1993 +/- 5
1986 +/- 5
1974 +/- 6
1968 +/- 6
1961 +/- 7
1946 +/- 8
1925 +/- 12
1880 +/- 47
Figure 4-8. Total and excess 210Pb activity for cores A) OP1 and B) FP1. Dates for core OP1 are calculated by the method developed
by Binford (1990). Core FP1 dates are based on an approximated sediment accumulation rate of 1.5 mm/yr.
FP2 21Pb (dpm/g)
2 4 6 8
Figure 4-9. Measurements of total and excess 210Pb for core FP2. Age dates are based on an approximated sedimentation rate of 3.0
A) OP1 37Cs Activity (dpm/g)
B) FP1 137Cs Activity (dpm/g)
I I I I L
1993 +/- 5
1993 +/- 5
1974 +/- 6
1968 +/- 6
1946 +/- 8
-1925 +/- 12
0.4 0.6 0.8 1
Figure 4-10. 137Cs activity for cores A) OP1 and B) FP1. Age dates for core FP1 are based on an estimated sedimentation rate of 1.5
mm/yr. Activity goes to 0 dpm/g at 12 cm for both core OP1 and FP1.
1.5 2 2.5 3
, , , . .
137Cs Activity (dpm/g)
0 0.5 1 1.5 2
m 0 -1934
Figure 4-11. Measurements of 137Cs activity for core FP2. Age dates are based on an estimated sedimentation rate from 210Pb data of
3.0 mm/yr. Activity falls to 0 dpm/g at a depth of 20 cm.
A) Magnetic Susceptibility (cgs*10 6)
-1 -0.5 0 0.5 1 1.5 2
B) Magnetic Susceptibility (cgs*106)
-1 -0.5 0 0.5 1 1.5 2 2.5
0 I I
Figure 4-12. Magnetic susceptibility measurements for cores A) OP1 and OP2 and B) FP1 and FP2.
10 20 30 40 50 60
20 40 60 80 100 120
l l l l l l l l l l lS
30 30- -T
Figure 4-13. Percent sand data for cores A) OP1 and B) FP1. The light blue line represents the average for all samples for core OP1
and for samples from 0 to 10.5 cm and then 11.5 to 29.5 for core FP1. The green lines represent values that are one
standard deviation from the mean and the pink lines represent values that are two standard deviations away from the mean.
Yellow boxes represent depth matched to date of known hurricanes for OP1.
A) Vlean Grain Size (phi, B) Mean Grain Size (ph.,C)
n n r 1 1 i 9 9 0 0.5 1 1.5 2 2.5
.edian Grain Size (phi, D)
1.4 1.6 1.8 2 2.2 2.4 2.6
.Mledian Grain Size (phi)
30 -301 -- 7-- 30 w----301-
Figure 4-14. Mean grain sizes for cores A) OP1 and B) FP1 and median grain sizes for cores C) OP1 and D) FP1. Lines represent
values for samples taken from the ridge sand, east beach, and south beach. Yellow boxes represent depth matched to date
of known hurricanes for OP1.
0 0.2 0.4 0.6 0.8 1
0 0.2 0.4 0.6 0.8 1
Figure 4-15. Sorting measurements for cores A) OP1 and B) FP1 and mode measurements for cores C) OP1 and D) FP1. Samples
with sorting values less than 0.350 are very well sorted, 0.35-0.5 D are well sorted, 0.5-0.710 are moderately well sorted,
0.71-1 P are moderately sorted, 1-2 O are poorly sorted, and greater than 20 are very poorly sorted. Yellow boxes
represent depth matched to date of known hurricanes for OP1.
1.35------------ m ----------------------2-------------------------------
1.3 u -
E E 1. -
0 111 0
M 1.05 --
0 10 20 30 40 50 60 70 80 0 20 40 60 80 100
% Sand % Sand
Figure 4-16. Plots of percent sand versus gamma bulk density (gm/cc) for cores A) OP1 and B) FP1. Black lines represent trends in
Figure 4-17. Foraminifera abundances per 0.3 g of sample for cores A) OP1 and B) FP1. Yellow boxes represent depth matched to
date of known hurricanes for OP1.
, 10o 11 1 13 ,, 114 15 ,16
18 20 22 24 26 28
Figure 4-18. Salinity profiles for cores A) OP1 and B) FP2.
0 5 10 15 20 25
Figure 4-19. Percent organic carbon and nitrogen for cores A) OP1 and B) FP1.
The constant rate of supply model (CRS) was used to calculate 210Pb sediment
accumulation rates (Appleby and Oldfield 1992). The model assumes that excess 210Pb is
delivered to the sediment at a constant rate. Consequently, as the bulk sedimentation
rates increase, 210Pb content in sediments is diluted and, conversely, as bulk
sedimentation decreases, 210Pb content is enriched (unless sediments are heavily
bioturbated). The cumulative residual unsupported 210Pb activity is calculated by
Equation 5-1 (developed by Appleby and Oldfield 1992).
At Aoekt (5-1)
where Ao is the cumulative residual unsupported 210Pb (dpm/cm2) below sediments of age
t and k is the 210Pb radioactive decay constant (0.03114 yr-) (Figure 5-1). The age of
sediments (t) at depth x is calculated:
tx = k-lln(AoAx-1) (5-2)
where Ao is equal to the total integrated unsupported 210Pb in the core and Ax is equal to
the integrated activity of 210Pb below depth x. The sediment accumulation rate is then
calculated by dividing the dry weight of the sediment in the interval by the time
represented by the interval (Figure 5-2).
Figures 4-8 and 4-9 show the unsupported, i.e. excess, and total 210Pb activity
profiles for cores OP1, FP1, and FP2. Core OP1 shows a decrease in activity with depth.
In addition, measurements of 210Pb activity for core OP1 reveal supported 210Pb activity
at 18-19 cm. Therefore, this depth is approximated (within associated error) to be
approximately 100 years old and gives an average sedimentation rate for the core of 1.8
mm/yr. Sediment accumulation rates calculated by the CRS model (using the dry bulk
density of samples calculated as grams dry per cubic centimeters wet from freeze drying
the samples) are between 0.1 and 7.8 mm/yr with an average of 3 mm/yr. These rates are
much higher than the average estimated from excess 210Pb first appearance and vary
throughout the core.
Cores FP1 and FP2, on the other hand, contain alternating activity with depth,
which complicates the use of the constant rate of supply model for calculating sediment
accumulation rates because the depths with zero activity give accumulation rates of
infinity (Appleby and Oldfield 1992). The average sedimentation rates for these cores
can be estimated by the depth of excess 210Pb first appearance. For core FP1 this depth is
at -15 cm, giving an average sedimentation rate of 1.5 mm/yr. Core FP2 appears to
reach a 210Pb activity of 0 dpm/g at 30 cm for an average sedimentation rate of 3.0
mm/yr. Samples were not measured for 210Pb below this depth, so it is not possible to say
for sure that this is the depth of first appearance.
Activity for 210Pb is commonly reported as inventory (dpm/cm2). The plot of 210Pb
inventory for core OP1 (Figure 5-1) shows similar trends to the plots of total and excess
210Pb activity (dpm/g) (Figure 4-8) with the exception of core FP1 at depths of 11, 13,
and 15 cm. Data for these samples plot higher than when reported as dpm/cm2. The
difference is due to the higher sand content and, therefore, greater mass of samples
All dates given for core FP1 in Figure 4-8 and FP2 in Figure 4-9 use the
sedimentation rates estimated from the first appearance of 210Pb activity. The higher sand
content of Flag Pond is the proposed reason for the sporadic 210Pb activity profile, as
sands have less surface area for radioisotope adsorption. Dating using 210Pb works best if
used in fine grained, highly organic sediments (Appleby and Oldfield 1983, Collins et al.
1999, and Schelske et al. 1994).
Concentrations of 137Cs were used as another means of dating the cores. The
method assumes first appearance of 137Cs occurred -1954. The radioisotope was
introduced into the atmosphere beginning in 1954, from atmospheric atomic weapons
testing (Smith and Comans 1996). 137Cs profiles are also influenced by depositional
diffusion because it is more mobile in sediments than 210Pb. The first appearance for
137Cs in core OP1 occurs at a depth of 12 cm (Figure 4-10), which agrees well with the
age calculated by the 210Pb method. The 137Cs profiles of cores FP1 and FP2 (Figures 4-
10 and 4-11), like the 210Pb profiles, were irregular, but depths for the first appearance of
137Cs activity could be estimated. Core FP1 first showed 137Cs activity at 11.5 cm, while
core FP2 showed activity at 19.5 cm. Although these depths do not correspond with the
depths calculated to be 1954 based on the sedimentation rates approximated from 210Pb
activity (which are -7 cm for core FP1 and -14 cm for core FP2), a chronology can be
estimated for core FP2. According to the 137Cs activity data, sediments at 20 cm depth
for core FP2 are estimated to correspond to -1950. All sediment above this depth is
younger than 1950. Sediments at 30 cm depth correspond to -1880-1900 according to
the first appearance of excess 210Pb activity. The dates calculated by 137Cs and 210Pb
activity give sediment accumulation rates ranging from -3.8 mm/yr for sediments above
20 cm to -3.0 mm/yr for sediments below 30 cm. A rather limited range of accumulation
rates for core FP2 is possibly due to uneven sedimentation. Storm events could have
deposited one or more thick beds of sand at some point after 1950 and increased the
sediment accumulation rate for a brief period. Storm activity would also account for the
variability in the sedimentation rate seen in core FP1 (Figure 5-2). Increases in wind-
blown material into the ponds may also temporarily raise sedimentation rates for both of
the Flag Pond cores.
In order to determine if there is migration of 137Cs within the cores, the Peclet (Pe)
number, a scaling argument between advection and diffusion, was calculated for cores
OP1, FP1, and FP2 using the following equation (Boudreau 1997):
Pe = [(1+Kd)*S*L]/Ds
where Kd is the solid-liquid distribution coefficient (102-103), S is the sedimentation rate
(cm/yr), L is the length of the scale of interest (5-10 cm), and Ds is the sediment diffusion
coefficient for 137Cs (-500 cm2/yr) (Sugai et al. 1994). The Kd value is based on values
reported for lakes by Sugai et al. (1994). If the Pe>>l, then 137Cs diffusion is negligible.
Tables 5-1 through 5-4 shows the results for the above calculations.
Table 5-1. Results for the calculation of the Peclet number for cores OP1, FP1, and FP2
for a Kd value of 102 and an L value of 10 cm For Pe>>l, diffusion of 137Cs is
Core OP1 FP1 FP2
Kd 102 102 102
S (cm/y) 0.18 0.15 0.30
L (cm) 10 10 10
Ds (cm2/y) 500 500 500
Pe 0.36 0.3 0.6
Table 5-2. Results for the calculation of the Peclet number for cores OP1, FP1, and FP2
for a Kd value of 105 and an L value of 10 cm. For Pe>>l, diffusion of 137Cs
Core OP1 FP1 FP2
Kd 103 103 103
S (cm/y) 0.18 0.15 0.30
L (cm) 10 10 10
Ds (cm2/y) 500 500 500
Pe 3.6 3 6
Table 5-3. Results for the calculation of the Peclet number for cores OP1, FP1, and FP2
for a Kd value of 102 and an L value of 5 cm. For Pe>>l, diffusion of 137Cs is
Core OP1 FP1 FP2
Kd 102 102 102
S (cm/y) 0.18 0.15 0.30
L (cm) 5 5 5
Ds (cm2/y) 500 500 500
Pe 0.18 0.15 0.3
Table 5-4. Results for the calculation of the Peclet number for cores OP1, FP1, and FP2
for a Kd value of 105 and an L value of 5 cm. For Pe>>l, diffusion of 137Cs is
Core OP1 FP1 FP2
Kd 103 103 103
S (cm/y) 0.18 0.15 0.30
L (cm) 5 5 5
Ds (cm2/y) 500 500 500
Pe 1.8 1.5 3
The calculated Peclet numbers show that diffusion of 137Cs is influencing the
activity profile, and thus the first appearance, when the solid-liquid distribution
coefficient is 102. When the coefficient is slightly larger (103), there is a balance between
diffusion and advection over 5-10 cm. The length of scale does not appear to affect the
results of the calculation.
According to 210Pb data, the predicted depth that corresponds to 1954 (the
estimated date for the first appearance of 137Cs) is 13 cm for core OP1 and 7 cm for core
FP1. The 137Cs activity data shows a first appearance at 12 cm for both of these cores,
indicating that the diffusion of 137Cs is as much as 9 and 69%. The difference in the
sediment accumulation rates calculated for the two radioisotopes is 0.02 cm/yr for core
OP1 and 0.9 cm/yr for core FP1.
Mass sediment accumulation rates vary with depth in core OP1 (Figure 5-2). The
highest sediment accumulation rates (150-200 mg/cm2/yr) are seen at depths of 8.5, 9.5,
and 11.5 cm, but the increases are only 30-50 mg/cm2/yr (30-35%) greater than other
depths. It is hypothesized that a hurricane would bring in increased amount of material
and, therefore, increase sedimentation rates (as mentioned above). The depths of the
increased sediment accumulation in core OP1 do not match with any known hurricanes.
Error associated with 210Pb age dating is one possible cause for the offsets.
The sediment accumulation rates for core FP1 (Figure 5-2) are much higher than
for core OP1. Three spurious samples (at 1.5, 3.5, and 13.5 cm) reach accumulation rates
greater than 4000 mg/cm2/yr and another sample (at 8.5 cm) has an accumulation rate of
2622 mg/cm2/yr. All other samples for this core vary between 36 and 795 mg/cm2/yr.
The sediment accumulation rates were calculated by the CRS model and may not be
accurate due to the irregularities of the 210Pb activity profile. The model calculates
accumulation rates of infinity for depths with zero or near zero 210Pb activity.
The sedimentation rates for the coastal ponds on St. Vincent Island (1.5-3.0
mm/yr) are on the low end of the rates reported for other coastal ponds (Table 5-5). The
sedimentation rate is comparable to the rates reported by Liu and Fearn (1993) (0.3-4.5
mm/yr) for Lake Shelby in Alabama and Donnelly et al. (2001a) (2-2.5 mm/yr) for
Succotash Marsh in Rhode Island. Accumulation rates are likely low in general for
Oyster and Flag Ponds due to the low elevation and the small drainage basin of the
island, which prevent increased amounts of sediment from entering the ponds (St.
Vincent National Wildlife Refuge, Apalachicola, Fl., 2000, Final Report of the
Vegetation Survey and Map Project, A USFWS-USGS Research Partnership Program
Table 5-5. Sedimentation and mixing rates for several coastal ponds.
Location, Reference Rates (mm/yr) Used Mixing Rates (cm2/yr) Method
Texas Estuary 4--5 239,240 Pu 0.04-0.4 239,240 Pu
(Ravichandran et al. 1995)
Maine Coastal Pond 0.15 210Pb N/A N/A
(Norton et al. 1997)
Texas Tidal Lake 1--45 210Pb, 137Cs Mentioned, N/A
(Williams 1995) not quantified
KwaZulu-Natal Coastal Lake 1.5-5.5 Radiocarbon N/A N/A
(Scott and Steenkamp 1996)
England Coastal Lake 9 210Pb Mentioned, N/A
(O'Sullivan et al. 1991) not quantified
St. Vincent Island,
Florida 1.5-1.8 210Pb,137Cs
There are limitations on establishing a robust and high resolution chronology for
the cores from St. Vincent Island. The high sand content of the Flag Pond cores prevents
a reliable 210Pb activity profile because the lead particles do not adsorb to sand as readily
as organic matter. The measured activity is equal to the activity adsorbed times the mass
flux. If all particles do not adsorb to the sand or if there is variable adsorption, then
measurements of activity do not accurately reflect the initial concentration of 210Pb and,
therefore, cause discrepancies in decay calculations. Consequently, sedimentation rates
for cores FP1 and FP2 can only be estimated based on the first appearance of 210Pb and
137Cs activity. Core OP1 fits logarithmic isotope profiles for 210Pb and 137Cs activity, but
the errors associated with excess 210Pb activity are quite high (up to 1.2 dpm/g). These
limitations make it difficult to correlate the age of known hurricanes to the approximate
depth at which they occur. Due to the error associated with the age-depth relationships, it
is difficult to relate the proxy records to individual hurricanes.
Coastal environments are very dynamic and it is often difficult to observe the
paleocyclone signal within such a dynamic depositional environment. It is best to study
preservation of paleostorm bedding where the signal is strongest. Subtidal environments
are constantly altered due to biological, wave, current and tide activity, resulting in rapid
post-depositional mixing after event bed deposition (Wheatcroft and Drake in press).
Supratidal environments offer the best paleostorm record because there is little physical
activity within the environment from waves and currents and inundation from tides only
occurs during intense storms. Tropical cyclones making landfall, therefore, should
generate enough energy to transport both water and sediment from offshore and deposit
them in the supratidal environment as overwash and aeolian deposits (Collins et al. 1999,
Liu and Fearn 2000).
Based on previous studies, storm event layers are hypothesized to have a coarser
mean grain size and be more poorly sorted than insitu sediments (Davis et al. 1989, Liu
and Fearn 2000, Donnelly et al. 2001). Although the mean and median size of the sand
fraction for core OP1 shows some variations (Figure 4-14), they are on a very small scale
(<1 phi difference) and there are no pronounced intervals (+1 C) that stick out from the
rest of the data as representing an event (Wheatcroft and Drake in press). The interval
from 10-15 cm shows increased variations that may be associated with a storm deposit,
but this interval does not correlate to any of the depths associated with known hurricanes.
Although there is abundant vegetation separating Oyster Pond from the south beach, the
mean size of the sand fraction for this core plots very closely with the mean grain size of
the south beach. Almost all of the samples from Oyster Pond are greater than 2D in size.
Currents of 25 cm/s are required to move grains this large (Prothero and Schwab 1996).
Because velocity is inversely related to shear stress and equal to the volume times the
cross sectional area, sediment brought in as overwash would drop out of suspension very
quickly after crossing over the dunes and into the ponds and deposit very near the
shoreline closest to the beachface. Thus, the sand is most likely brought into Oyster pond
on a semi-regular basis as aeolian deposits during strong storms with winds greater than
Core FP1 (Figure 4-14) shows very little variation (<0.2 phi) in mean and median
size of the sand fraction. The data plot very close to that of the ridge sands, which
implies that the sand is derived from inland or Flag Pond is a submerged part of the ridge
system making up the island. The grain size for this pond appears to be influenced more
by its surrounding environment than by aeolian material from the south beach despite its
closer proximity to the beach than Oyster Pond. Also, the vegetation surrounding Flag
Pond is dense and may prevent transfer of some sediment by the wind.
The sorting of the sand fraction for samples from core OP1 (Figure 4-15) does
show some layers to be more poorly sorted (10-15 cm and 20-29 cm). Within the error
associated with dating, it is possible that the sorting profile shows some evidence of
hurricane deposits. Core FP1 shows a fairly continuous sorting profile downcore for the
sand fraction with no indication of event layers. No other previous paleocyclone studies
document specific data regarding grain size and sorting, other than to say that storm
deposits had higher sand content and were poorly sorted. Parsons (1998) reported that
the deposit left by Hurricane Andrew showed a coarser grain size, but did not have the
graded profile discussed by Davis et al. (1989). He suggested that the sampling interval
(0.5 cm) may have been too large to show any grading associated with the deposit.
Because this sampling interval is smaller than that used for the samples from St. Vincent
Island (1 cm), it could also be the reason for the lack of distinct layers of coarse grain size
and poor sorting. Liu and Fearn (1993) also report storm deposits on the millimeter scale
that were identified visually.
The sorting profiles for both ponds plot higher than all three environments sampled
on the island, implying that the ponds are more poorly sorted than any single
environment and are receiving a mixture of material from different environments. Modal
values, similar to mean and median values, for core OP1, plot very closely with the south
beach and have an outlier point at 13.5 cm. The modal profile for core FP1 plots closely
to the ridge sands below 7.5 cm and then trends toward the south beach upcore indicating
a shift in source material. The shift in modal values corresponds to a shift in values for
percent sand, implying that increased amounts of sand began entering Flag Pond from a
The percent sand is expected to increase for a storm bed as more sand is brought in
as wind-blown material into muddier ponds (Donnelly et al. 2001). There are some very
noticeable increases in percent sand (at 2 cm to 53%, 6 cm to 28%, and 20 cm to 40%)
for core OP1 that are slightly offset from the depths that match known hurricanes (Figure
4-13). However, the offsets are within 2 cm and could be related to sampling and/or
dating error. The samples that did not fit the trendline when percent sand was plotted
against gamma bulk density (Figure 4-16) are most likely related to consolidation effects
after sampling. The increases in percent sand for core FP1 (Figure 4-13) could correlate
with the 1974-75 (60% sand at 2 cm) and 1985 (65% sand at 4 cm) hurricanes. Because
of dating problems with the Flag Pond cores, there is no way to establish an accurate age-
depth relationships relating proxy records of hurricanes.
Decreases in percent organic carbon and percent nitrogen are hypothesized to
correspond to coarse sediment layers that may be associated with hurricane deposits
(Parsons 1998). Figure 5-3 shows that although trends do exist when percent sand is
plotted against percent organic carbon, there is no evidence that increases in percent sand
correspond to decreases in percent organic carbon. Core OP1 shows similar trends with
respect to percent organic carbon and nitrogen (Figure 4-19). The exponential decrease
in core OP1 in both values argues for strong diagenetic decomposition that would mask
any episodic increases. Core OP1 does not have any increases or decreases in percent
organic carbon or nitrogen that could be hurricane related. Although core FP1 shows
trends very different from core OP1, there are similarities in trends between percent
organic carbon and percent nitrogen (Figure 4-19). Core FP1 has small (3-4 %)
decreases in both proxies that could correlate to the 1974-75 and 1985 hurricanes. In
general, the decreases in percent carbon seen in core FP1 at 4.5 and 8.5 cm are minor
compared to the decreases documented by Collins et al. (1999) for the core taken at the
location directly impacted by Hurricane Hugo (from >20% to 0.6-3%) and show a large
decrease in organic carbon due to the low content of organic carbon in beach sand.
However, in the core taken 50-75 km away from the location of landfall of Hurricane
Hugo, there were no noticeable decreases in percent carbon implying a lack of deposition
of a storm bed at this distal site. Parsons (1998) also reported that percent carbon showed
decreases throughout cores taken in a Louisiana marsh, but that this proxy was not a
useful indicator of storm layers due to the diagenetic control of carbon decomposition.
The rapid increase in organic carbon accumulation could be related to a change in
the environment of the pond. Prior to the mid-1900's the area that is now submerged
may have been a dry, sandy low point between beach ridges. A large storm could have
hit the island with winds strong enough to move sand around and create an enclosed area
within the low point of the beach ridges. Water level increases in the enclosed area could
have allowed for the establishment of aquatic vegetation, thereby increasing the percent
of organic carbon in the sediments. Due to difficulties associated with 210Pb dating, it is
not possible to correlate the formation of the pond with any specific storm. Although the
transition from sand to organic rich sediments does not occur at the same depth for both
of the Flag Pond cores (at -10 cm for FP1 and -20 cm for FP2) (Figure 4-1), the age
dates calculated for both cores match within the associated dating error (Figures 4-8 and
The foraminifera abundances (Figure 4-17) were measured because it was
hypothesized that marine forams would be transported into the predominantly freshwater
and brackish ponds by storm surge during cyclones. Greater than 90% of the
foraminifera belonged to the genus Ammonia, which is characteristic of detrital-rich
environments such as continental shelves and lagoons. This species is known to exist in a
wide variety of environments from brackish and hypersaline waters to freshwater, as
such, Ammonia are able to exist under very stressful conditions. In addition, there is a
noticeable absence of Miliolidae and Elphidium, which are common to shallow,
nearshore and marsh sediments in Florida (Rose and Lidz 1977). This is a contrast to the
findings of Collins et al. (1999), where they observed that even in cores taken at 50-75
km from where Hurricane Hugo came ashore, offshore species of foraminifera were
abundant at depths dated to the time of the hurricane landfall. The cores from South
Carolina showed no other sedimentological evidence for hurricane landfall other than the
presence of offshore foraminifera.
While foraminifera have been useful in other paleohurricane studies (Collins et al.
1999), there are no foraminifera in samples above 15 cm for core OP1 and 25 cm for core
FP1. Therefore, the foraminifera data cannot be used as a proxy for hurricane activity at
this site. The conditions in the pond may no longer be suitable for the foraminifera to
survive, although this is unlikely given the ability of Ammonia to tolerate very stressful
conditions. Rather, an increase in organic matter to the sediments could have caused an
increase in organic carbon decomposition, which would increase CO2 production in
sediment porewaters. Increases in CO2 lower pH, leading to abundant calcite dissolution
and loss of fossil record (Green et al. 2001).
Magnetic susceptibility (Figure 4-12) was used as a proxy for storm deposits
because increases in magnetic susceptibility may reflect changes in sediment provenance.
Magnetic susceptibility measures whether minerals are diamagnetic (biogenic carbonate
and silica) or paramagnetic (Fe-rich silicates including clays). Diamagnetic minerals
have a negative magnetic susceptibility, while paramagnetic minerals are positive
(Frederichs et al. 1999). There appear to be clear intervals of increased (positive)
magnetic susceptibility, which indicate intervals with more Fe-rich silicates. Such
increases in magnetic susceptible minerals may reflect periodic input of heavy mineral
sands to the ponds, although the increases are small and may reflect very minor additions
of such minerals. Both cores from each respective pond correlate well to each other, with
common patterns in layers of positive magnetic susceptibility reflecting a common
source. The increases in magnetic susceptibility at 14-16 cm for core FP2 and at 38 cm
for core OP2 match with increases in gamma bulk density. The offset of the peaks for
cores FP1 and FP2 is most likely related to the difference in the estimated sediment
accumulation rates. However, these layers do not seem to correlate to the depths of
known hurricanes. The loop used to test the cores integrates over a range of 10 cm
(Weber et al. 1997), which is far too coarse a sample interval (-20 years) to detect
individual storm beds that may be only one mm thick. Also, inflection points may be due
to dilution from diamagnetic silica. No other studies have used magnetic susceptibility as
a proxy for hurricane deposits.
The gamma bulk density (Figures 4-2 and 4-3) shows some correlation to known
hurricanes for the Oyster Pond cores, but because the sampling interval is 0.5 cm, it is
difficult to correlate with other variables sampled at coarser resolution. Core OP1 shows
a peak in the 17-22 cm range that may correlate with the 1886 hurricane, whereas core
OP2 has peaks in bulk density at 8-9 cm and 16-19 cm that may relate to the 1974-1975
and 1886 hurricanes using a chronology that is approximated from that of core OP1. The
Flag Pond (Figure 4-5) cores do not show any peaks that relate to known hurricanes. The
large increases in bulk density (at 12 cm for core FP1 and 36 cm for core FP2) are likely
related to the environmental shift mentioned above in relation to organic carbon.
Difficulties establishing age-depth relationships make correlations difficult. Bulk density
could correlate well with percent sand if there were no compaction or bioturbation.
Figure 4-16 shows a potential relationship between bulk density and percent sand for core
OP1. The two separate trends for core FP1 in Figure 4-16 indicate a consolidation due to
the change in lithology. Because percent sand and bulk density correlate well, it may be
necessary to measure one of these parameters in paleocyclone studies. Bulk density was
not used as a parameter for identifying paleostorm deposits in previous studies.
The x-radiographs (Figure 4-6) show greater detail of bedding, changes in
lithology, and bioturbation than the photographs. Bedding changes, changes in density,
and worm tubes that are beneath the surface are visible in the x-radiographs. They reveal
that some of the bedding is slanted across the core, while other beds are parallel across
the core. The subsampling of the cores was done perpendicular to the core wall and
would have, thus, cut across these slanted beds and, therefore, prevented samples for
other proxy records from fully representing event layers. The x-radiographs also reveal
information about the depth of biologic activity. Worm tubes (<0.5 cm in diameter)
extend down to a depth of 3 cm in core OP 1 and 2 cm in core FP 1.
In addition to visual examination of sedimentary structures, the x-radiograph pixel
density data are the proxy record that resolves a strong signal for paleocyclones (Figure
4-6). Pixel density (0-255 gray scale) is controlled by the absorption of x-rays by the
film and variability in pixel intensity roughly corresponds to the bulk density of the
sediments. The x-radiographs were sampled at high resolution (-1 mm) and can detect
layers that are thinner than the 0.5 cm layers measured for gamma bulk density. Both of
the Flag Pond cores show increases in pixel density that may correlate to the 1985 and
1974-75 hurricanes (40 pixel density increases at 3 and 5 cm), but difficulties with the
Flag Pond chronology prevent an exact correlation. Core FP1 may also show evidence of
the 1886 hurricane, but the increase is masked by the increase related to the
environmental shift mentioned above. The Oyster Pond cores show several increases in
pixel density (at 14, 31, and 36 for core OP1 and at 8, 24, and 29 for core OP2), but they
are offset from the depths corresponding to the dates of hurricanes, possibly due to dating
errors. Both cores have increases (at 19 cm for core OP1 and 17 cm for core OP2) that
correspond to the depth of the 1886 hurricane. The chronology for core OP2 is estimated
by correlation with OP1. Due to its high-resolution sampling, the pixel density shows a
detailed record of event layers that supports the percent sand data. Increases in both
proxies are seen at 2.5, 6.5, and 19.5 cm for core OP1 and at 2.5, 4.5, and 23.5 for core
The salinity data for core OP1 (Figure 4-18) have an irregular profile above 4 cm
and then remain between 13 and 15 ppt downcore. Core FP2 trends from saline to
fresher water with increasing depth. The trend suggests that the pond was previously
either a fresh marsh or influenced by fresh groundwater and that the saltwater has not
fully diffused through the sediments. Also, the pond may still experience freshwater
intrusion from the groundwater and saltwater spray from the ocean and the trend reflects
a mixture of the two sources.
Table 5-6 is a summary of all of the proxies tested and whether they showed any
evidence of the 1886, 1974-75, and 1985 hurricanes. Due to the uncertainties associated
with the dating of the Flag Pond cores, it is impossible to match proxy records to specific
hurricanes. All data in Table 5-6 only relate to the Oyster Pond cores. There are large
errors associated with the dates calculated for core OP1, which also make it difficult to
match specific hurricanes. The 1886 hurricane appeared to be detected the most
frequently. The 1974-75 hurricanes appeared to be recorded by two of the proxies tested.
The more recent 1985 hurricane appeared to be evident in three of the proxy records.
Because of the increased error associated with the 1886 date, the range of possible depths
that could correspond with the date of this hurricane is quite large allowing for a greater
number of proxy records to fit within this range.
Table 5-6. Synopsis of detection of hurricanes by each of the proxies tested. The 1886
hurricane was a category five. Hurricane Carmen (Category 3) occurred in
1974 with Hurricane Eloise (Category 3) followed in 1975. Hurricanes Elena
(Category 3) and Kate (Category 2) and Tropical Storm Juan all occurred in
Proxies 1886 1974-75 1985
Visual Examination No No No
Gamma Bulk Density Yes Yes Yes
Pixel Density Yes Yes Yes
Sediment Accumulation Rates No No No
Magnetic Susceptibility No No No
Percent Sand Yes No Yes
Mean Grain Size of Sand No No No
Median Grain Size of Sand No No No
Sorting of Sand Fraction Yes No No
Mode Grain Size of Sand No No No
Foram Abundance No No No
Salinity No No No
% Carbon No No No
% Nitrogen No No No
It is important to determine preservation potential of coastal depositional
environments when studying paleocyclone deposits because a combination of strong
sediment mixing and low sedimentation rates may make it difficult to preserve deposited
cyclone deposits. If coastal sedimentary strata do not show any evidence of past
hurricane activity, then a low preservation potential for that area could be one
explanation. If that same environment has a high preservation potential, but lacks a
strong sedimentological signal, then it is likely that that no bedding from hurricanes has
been deposited during the time period that the strata represent.
The preservation potential of an event layer (Figure 5-4 and 5-5) can be estimated
from comparing transit time (the time required for an event layer to travel through the
surface mixed layer) to dissipation time (the time required for an event deposit to be
completely destroyed). Transit time is calculated by the equation developed by
Wheatcroft and Drake (in press):
[(Lb-Ls)/2]/Burial Rate = Transit time of event layer
where Lb is the thickness of the surface mixed layer and Ls is the thickness of the event
layer. Lb incorporates both physical and biologic mixing. The units for both Lb and Ls are
centimeters, while burial rate is recorded as cm/yr and preservation potential is calculated
as a percentage of the original signal. The thickness of the event bed represents the
sediment transport potential of the storm and is a function of available sediment for
transport, shoreline vegetation, dune morphology, distance of coastal pond from shore,
and storm surge. Storm surge is related to the forward speed of the storm, amount of
rainfall, wind speed, and duration of storm (Davis et al. 1989 and Risi 1998). Other
hurricane studies report event bed thickness ranging from 0.1 to 30 cm (Liu and Fearn
1993 and 2000, Donnelly et al. 2001a,b, Davis et al. 1989, and Collins et al. 1999).
However, Liu and Fearn (1993) do not detail how they were able to detect event layers
that were 0.1 cm thick, nor why these beds were preserved.
In order for an event layer to be well preserved, transit time must be greater than
dissipation time. Therefore, Ls needs to be much greater than Lb. Figure 5-4
demonstrates that when Lb is thicker than Ls, the event layer is mixed by biologic and
physical mixing but remains detectable. When Ls is greater than Lb, the upper part of the
event layer is mixed, but preserved further down as it is below the depth of mixing
For St. Vincent Island, the thickness of the surface mixed layer was calculated
based on 210Pb profiles and x-radiographs. If surface sediments have been rapidly mixed,
the 210Pb profile may have a near-surface interval of constant activity before
exponentially decreasing (Figure 5-6) (Sugai et al. 1994). Core OP1 did not have such an
interval, but did show evidence for mixing in its x-radiographs, as worm tubes extend
down 1-2 cm from the surface, depending on the core. 210Pb data for core FP1 were not
reliable for determining mixing depth. The x-radiographs for this core also show worm
tubes down 1-2 cm from the surface. Taking all of these factors into consideration, the
rapidly mixed Lb for St. Vincent Island was estimated to be -1 cm.
The methods used to detect the signal of hurricane deposition in coastal ponds on
St. Vincent Island examined cores for evidence of washover sand, marine microfossils,
and geochemical data (C and N). Because the proxy records showed marginal evidence
of event beds that matched depths with 210Pb ages of known hurricanes, Ls was estimated
to be 0-3 cm. The complicating fact is that the thickness of a storm layer would need to
be measured directly following a hurricane for an accurate estimate of Ls.
Using the above data and a sedimenation rate of 2 mm/yr, the equation developed
by Wheatcroft and Drake (in press) yields a transit time of 0-5 years (Lb > -1 cm) for a
deposit within the surface mixed zone. The fast transit time suggests that an event layer
should be preserved, but does not account for dissipation of the event layer due to
physical and biologic mixing. Dissipation time (the time required for an event layer to be
completely destroyed) is best calculated by time-series cores (Wheatcroft and Drake in
press). Since time series cores are not available for the ponds on St. Vincent Island, the
dissipation time for the cores can be estimated based on estimates of biodiffusion
coefficients (Db), as increased biodiffusion (i.e. bioturbation) should lead to a
concomitant increase in dispersion. Db can be calculated from (a) 210Pb activity data or
(b) data from previous studies. 210Pb data can yield measurements of Db of over decades
of mixing if sediments are completely homogenized and there is no sedimentation.
Under these conditions, the 210Pb data can be modeled as such:
6A/6t = Db (62A/6z2) XA (i.e., no sedimentation)
where A is 210Pb activity (dpm/g), t is time (years), and z is depth (cm). Since the
sediments in the coastal ponds on St. Vincent Island show evidence of bedding (in the x-
radiographs) and 137Cs data indicate sedimentation, Db values can be estimated based on
the shape of the 210Pb profile but are most likely very large overestimates. Fitting the
data yields a maximum mixing coefficient of 2 cm2/yr. If contributions from sediment
accumulation are also accounted for, the mixing coefficient becomes much smaller.
Fitting a mixing profile to the 210Pb activity data could only be done for core OP1
because the data for cores FP1 and FP2 were too variable. Wheatcroft and Drake (in
press) report Db ranging from 10 to 100 cm2/yr for continental margin sediments where
sediments are more biologically active and correspond to dissipation times of 3-5 years.
Ravichandran et al. (1995) measured dispersion rates of 0.04 to 0.4 cm2/yr for a Texas
estuary. Dispersion rates for the ponds on St. Vincent Island are estimated to be low,
between 0.1 and 2 cm2/yr, due to the 210Pb estimates and because the x-radiographs
showed little evidence of biologically mixed sediments. Laminae preserved within the
sediments would have been destroyed if mixing coefficients were >5 cm2/yr (Jaeger and
Nittrouer in press).
There are not enough observations or data presented by Wheatcroft and Drake (in
press) to establish a quantitative relationship between Db and dissipation time. A semi-
quantitative relation can be estimated based on the two relationships that are reported in
the paper (Db=10 cm2/yr, dissipation time=3 years and Db=100cm2/yr, dissipation time=1
year). The estimated Db values for St. Vincent Island (0.1 to 2 cm2/yr) yield estimated
dissipation times ranging from 5 to 10 years. If dispersion rates on St. Vincent Island are
very low due to little biologic activity, and therefore dissipation time is slow, then event
layers are more likely to pass through the surface mixed layer partially preserved. If an
event layer is thin (<1 cm), then a short dissipation time (<5 years) is likely to result in
the destruction of the layer. Since dissipation times of less than five years are not likely
to exist on St. Vincent Island (would require Db>1Ocm2/yr), most storm beds should be
partially preserved. Obviously, if a storm bed is greater than 1 cm thick, some portion of
it will be preserved.
Consequently, the ponds on St. Vincent offer a good environment for studying
paleocyclones because they offer an environment with low biologic and physical mixing
and thus long dissipation times. The sedimentation rates are low compared to other
coastal ponds, but moderate compared to other coastal areas studied for hurricane
deposits and lead to fairly fast transit through the surface mixed layer. Because these
ponds apparently offer an ideal environment for studying paleocyclone deposits, they
should preserve a record of the many large historical hurricanes. However, they show
only marginal evidence for storm deposits actually existing within the sediments when
using a variety of proxy records. Diagenetic processes mask evidence of storm deposits
and prevent recognition of distinct layers. Thus, one should use caution when examining
other published data related to paleocyclone activity.
A) OP1 210pb (dpm/cm2)
0 0.5 1 1.5
. 10 ,
B) "P1 21Pb Activity (dpm/cm2)
0 0.2 0.4 0.6 0.8 1
Figure 5-1. 21Pb inventory for cores A) OP1 and B) FP1 calculated by the equation developed by Appleby and Oldfield (1992). Age
dates for core OP1 are calculated by the CRS model. Age dates for core FP1 are estimated based on an approximated
sediment accumulation rate of 1.5 mm/yr.
Sediment Accumulation Rates (mg/cm2/yr) Sediment Accumulation Rate (mg/cm2/yr)
0 50 100 150 200 0 900 1800 2700 3600
4 -- -- -- -- -
8 ------- ---- --- -- ----------
--E I6 ------
1 6 -- - - - - --- - -- 1 - - - - -
12-----------------14 --- ----------------------
Figure 5-2. Sediment accumulation rates for core A) OP1 and B) FP1 based the CRS model.
1. I U
30 40 50
0 20 40 60
80 100 120
Figure 5-3. Plot of percent sand versus percent organic carbon for cores A) OP1 and B) FP1. Although trends exist in the data, there is
no correlation between increased sand and decreased organic carbon. Some core FP1 samples plot above 100% because
samples for organic carbon included only a small mass and may not have been representative of the entire sample.
10 20 ;
- / Lb
% Sand (z) >2 IMean %
Figure 5-4. Diagram of the preservation of an event layer after its deposition.
% Sand "
1- 1 -1 -
% Sand (z) > 2 a 1% Sand
Figure 5-5. Diagram of the destruction of an event layer after its deposition.
Sediments Not Mixed
SDepth of Mixing
Figure 5-6. Diagram of using 210Pb concentrations to determine the depth of the mixing layer.
Establishing a robust age-depth relationship in coastal ponds is difficult because of
non-steady, heterogeneous sedimentary conditions. The cores from Flag Pond showed
irregular patterns of 210Pb activity due to the high sand content in certain layers of
sediment. Ideal coring locations have organic rich sediments that are rich in 210Pb
throughout the core.
The results of the preservation potential modeling show that detection of storm
bedding is sensitive to sampling interval (1mm vs. 1cm) and coring artifacts (tilted
bedding). Depending on storm intensity, event beds may be on the millimeter scale and,
therefore, homogenized with non-event bedding during sampling. Also, the cores are
placed on their side for splitting which would cause the beds to artificially tilt in the x-
The signal left in sediments by hurricane activity is best observed in the x-
radiographs, the pixel density data, and percent sand. The x-radiographs and the gray
scale pixel intensity data give the highest resolution and can detect sub-centimeter scale
event beds. All other proxies require larger sampling intervals in order to have the
amount of sample required by the different detection devices. Percent sand showed a
possible correlation to the known hurricanes in core OP1, although it was offset from the
depth associated with the age of these hurricanes. This proxy is commonly used in
paleohurricane studies. A smaller sampling interval would provide more precise data on
these deposits because it is likely that storm deposits were partially mixed within the 1
cm sampling interval. Also, constant aeolian transport of sand makes it difficult to
distinguish extreme storm event.
The data show very little evidence of storm bedding associated with the three
largest storms of the past century (1886, 1974-75, 1985) for the Oyster Pond cores (Table
5-6). Because there is possible evidence of these hurricanes in the percent sand and gray
scale intensity data, the Ls for these storms must have been >1 cm. Proxy records that
can be recorded while the core is still intact give a better record. It is too difficult to
sample a core at intervals less than one centimeter. For proxy records that require the
core to be cut into sections, only storms that had an initial Ls of>1 cm will be detected.
Data from Flag Pond are overshadowed by the dynamics in sedimentary processes.
Detecting paleocyclone "signal" from natural "noise" of dynamic coastal
sedimentary processes is difficult in this location unless signal is very strong. Examples
of "noise" on the island are the higher natural sand content and the change in
environmental conditions marked by a gradual shift from predominantly organic rich
sediments to sand rich sediments.
This research is part of a larger Florida coastal depositional study examining the
preservation potential of storm event layers. Additional work will be performed on St.
Vincent to better quantify Lb in these ponds and to examine additional sedimentary
environments (e.g., salt marshes).
Appelboom, T.W., Chescheir, G.M., Skaggs, R.W., and Hesterberg, H.L. (2002)
Management practices for sediment reduction from forest roads in the coastal
plain. Transactions of the American Society ofAgricultural Engineers, 45 (5),
Appleby, P.G. and Oldfield, F. (1992) Application oflead-210 to sedimentation studies.
In: Uranium-series Disequalibrium: Applications to Earth, Marine, and
Environmental Sciences, Vol. 0 (Ed. By I.M. and R.S. Harmon), pp. 731-778.
Clarendon Press, Oxford.
Ball, M.M., Shinn, E.A. and Stockman, K.W. (1967) Geologic Effects of Hurricane
Donna in South Florida. Journal of Geology, 75 (5), 583 -587.
Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G., Faure, G., and
Rougerie, F.(1996) Deglacial sea-level record from Tahiti corals and the timing of
global meltwater discharge. Nature, 382, 241-244.
Binford, M.W. (1990) Calculations and uncertainty of 210Pb dates for PIRLA project
sediment cores. Journal ofPaleolimnology, 3, 253-267.
Boudreau, B.P. (1994) Is burial velocity a master parameter for bioturbation?
Geochimica et Cosmochimica Acta, 58, 1243-1249.
Brooks, H.K. (1981) Map of Florida. Cooperative Extension Service, University of
Florida, Institute for Food and Agricultural Science.
Campbell, K.M. (1986) St. Vincent Island (St. Vincent National Wildlife Refuge),
Florida. Geological Society of America Centennial Field Guide-.SoNwu 't\e l n,
Collins, E.S., Scott, D.B. and Gayes, P.T. (1999) Hurricane records on the South Carolina
coast: Can they be detected in the sediment record? Quaternary International, 56,
Committee on Engineering Implications of Changes in Relative Mean Sea Level (1987)
Responding to Changes in Sea Level. National Academy Press, Washington, D.C.,